# Instructor's Solutions Manual For Modern Control Systems, 12th Ed Systems Edition

User Manual: Pdf

Open the PDF directly: View PDF .

Page Count: 754

Download | |

Open PDF In Browser | View PDF |

© 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. MODERN CONTROL SYSTEMS SOLUTION MANUAL Richard C. Dorf Robert H. Bishop University of California, Davis Marquette University A companion to MODERN CONTROL SYSTEMS TWELFTH EDITION Richard C. Dorf Robert H. Bishop Prentice Hall Upper Saddle River Boston Columbus San Francisco New York Indianapolis London Toronto Sydney Singapore Tokyo Montreal Dubai Madrid Hong Kong Mexico City Munich Paris Amsterdam Cape Town Educator Home | eLearning & Assessment | Support/Contact Us | Find your rep | Exam copy bookbag Instructor's Solutions Manual for Modern Control Systems, 12/E Richard C. Dorf, University of California, Davis Robert H. Bishop, University of Texas at Austin ISBN-10: 013602498X ISBN-13: 9780136024989 Publisher: Prentice Hall Copyright: 2011 Format: On-line Supplement Published: 08/16/2010 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. P R E F A C E In each chapter, there are five problem types: Exercises Problems Advanced Problems Design Problems/Continuous Design Problem Computer Problems In total, there are over 1000 problems. The abundance of problems of increasing complexity gives students confidence in their problem-solving ability as they work their way from the exercises to the design and computer-based problems. It is assumed that instructors (and students) have access to MATLAB and the Control System Toolbox or to LabVIEW and the MathScript RT Module. All of the computer solutions in this Solution Manual were developed and tested on an Apple MacBook Pro platform using MATLAB 7.6 Release 2008a and the Control System Toolbox Version 8.1 and LabVIEW 2009. It is not possible to verify each solution on all the available computer platforms that are compatible with MATLAB and LabVIEW MathScript RT Module. Please forward any incompatibilities you encounter with the scripts to Prof. Bishop at the email address given below. The authors and the staff at Prentice Hall would like to establish an open line of communication with the instructors using Modern Control Systems. We encourage you to contact Prentice Hall with comments and suggestions for this and future editions. Robert H. Bishop rhbishop@marquette.edu iii © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. T A B L E - O F - C O N T E N T S 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. iv Introduction to Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Mathematical Models of Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 State Variable Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Feedback Control System Characteristics . . . . . . . . . . . . . . . . . . . . . . . 133 The Performance of Feedback Control Systems . . . . . . . . . . . . . . . . . 177 The Stability of Linear Feedback Systems . . . . . . . . . . . . . . . . . . . . . . 234 The Root Locus Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 Frequency Response Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382 Stability in the Frequency Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445 The Design of Feedback Control Systems . . . . . . . . . . . . . . . . . . . . . . . 519 The Design of State Variable Feedback Systems . . . . . . . . . . . . . . . . 600 Robust Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 659 Digital Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 714 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 1 Introduction to Control Systems There are, in general, no unique solutions to the following exercises and problems. Other equally valid block diagrams may be submitted by the student. Exercises E1.1 A microprocessor controlled laser system: Controller Desired power output Error - Microprocessor Current i(t) Laser Power Sensor power A driver controlled cruise control system: Controller Process Foot pedal Desired speed Power out Measurement Measured E1.2 Process - Driver Car and Engine Actual auto speed Measurement Visual indication of speed E1.3 Speedometer Although the principle of conservation of momentum explains much of the process of fly-casting, there does not exist a comprehensive scientific explanation of how a fly-fisher uses the small backward and forward motion of the fly rod to cast an almost weightless fly lure long distances (the 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 2 CHAPTER 1 Introduction to Control Systems current world-record is 236 ft). The fly lure is attached to a short invisible leader about 15-ft long, which is in turn attached to a longer and thicker Dacron line. The objective is cast the fly lure to a distant spot with deadeye accuracy so that the thicker part of the line touches the water first and then the fly gently settles on the water just as an insect might. Fly-fisher Desired position of the fly Controller - Wind disturbance Mind and body of the fly-fisher Process Rod, line, and cast Actual position of the fly Measurement Visual indication of the position of the fly E1.4 Vision of the fly-fisher An autofocus camera control system: One-way trip time for the beam Conversion factor (speed of light or sound) K1 Beam Emitter/ Receiver Beam return Distance to subject Subject Lens focusing motor Lens © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 3 Exercises E1.5 Tacking a sailboat as the wind shifts: Error Desired sailboat direction - Controller Actuators Sailor Rudder and sail adjustment Wind Process Sailboat Actual sailboat direction Measurement Measured sailboat direction Gyro compass E1.6 An automated highway control system merging two lanes of traffic: Controller Error Desired gap - Embedded computer Actuators Brakes, gas or steering Process Active vehicle Actual gap Measurement Measured gap Radar E1.7 Using the speedometer, the driver calculates the difference between the measured speed and the desired speed. The driver throotle knob or the brakes as necessary to adjust the speed. If the current speed is not too much over the desired speed, the driver may let friction and gravity slow the motorcycle down. Controller Desired speed Error - Driver Actuators Throttle or brakes Measurement Visual indication of speed Speedometer Process Motorcycle Actual motorcycle speed © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 4 CHAPTER 1 E1.8 Introduction to Control Systems Human biofeedback control system: Controller Desired body temp Process Hypothalumus - Message to blood vessels Actual body temp Human body Measurement Visual indication of body temperature E1.9 TV display Body sensor E-enabled aircraft with ground-based flight path control: Corrections to the flight path Desired Flight Path - Controller Aircraft Gc(s) G(s) Flight Path Health Parameters Meteorological data Location and speed Optimal flight path Ground-Based Computer Network Optimal flight path Meteorological data Desired Flight Path E1.10 Specified Flight Trajectory Health Parameters Corrections to the flight path Gc(s) G(s) Controller Aircraft Location and speed Flight Path Unmanned aerial vehicle used for crop monitoring in an autonomous mode: Trajectory error - Controller UAV Gc(s) G(s) Flight Trajectory Sensor Location with respect to the ground Map Correlation Algorithm Ground photo Camera © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 5 Exercises E1.11 An inverted pendulum control system using an optical encoder to measure the angle of the pendulum and a motor producing a control torque: Actuator Voltage Error Desired angle - Controller Process Torque Motor Pendulum Angle Measurement Measured angle E1.12 In the video game, the player can serve as both the controller and the sensor. The objective of the game might be to drive a car along a prescribed path. The player controls the car trajectory using the joystick using the visual queues from the game displayed on the computer monitor. Controller Desired game objective Optical encoder Error - Player Actuator Joystick Measurement Player (eyesight, tactile, etc.) Process Video game Game objective © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 6 CHAPTER 1 Introduction to Control Systems Problems P1.1 Desired temperature set by the driver An automobile interior cabin temperature control system block diagram: Error - Controller Process Thermostat and air conditioning unit Automobile cabin Automobile cabin temperature Measurement Measured temperature P1.2 Temperature sensor A human operator controlled valve system: Controller Process Error * Desired fluid output * - Tank Valve Fluid output Measurement Visual indication of fluid output * Meter * = operator functions P1.3 A chemical composition control block diagram: Controller Process Error Desired chemical composition - Mixer tube Valve Measurement Measured chemical composition Infrared analyzer Chemical composition © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 7 Problems P1.4 A nuclear reactor control block diagram: Controller Process Error Desired power level Reactor and rods Motor and amplifier - Output power level Measurement Measured chemical composition P1.5 A light seeking control system to track the sun: Measurement Light source Dual Photocells P1.6 Ionization chamber Controller Ligh intensity Trajectory Planner Desired carriage position Controller - Motor, carriage, and gears K Photocell carriage position If you assume that increasing worker’s wages results in increased prices, then by delaying or falsifying cost-of-living data you could reduce or eliminate the pressure to increase worker’s wages, thus stabilizing prices. This would work only if there were no other factors forcing the cost-of-living up. Government price and wage economic guidelines would take the place of additional “controllers” in the block diagram, as shown in the block diagram. Controller Process Market-based prices Initial wages Process Motor inputs Error - Industry Government price guidelines Controller Wage increases Government wage guidelines Cost-of-living K1 Prices © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 8 CHAPTER 1 P1.7 Introduction to Control Systems Assume that the cannon fires initially at exactly 5:00 p.m.. We have a positive feedback system. Denote by ∆t the time lost per day, and the net time error by ET . Then the follwoing relationships hold: ∆t = 4/3 min. + 3 min. = 13/3 min. and ET = 12 days × 13/3 min./day . Therefore, the net time error after 15 days is ET = 52 min. P1.8 The student-teacher learning process: Process Controller Lectures Error Desired knowledge - Teacher Knowledge Student Measurement Exams Measured knowledge P1.9 A human arm control system: Process Controller u Desired arm location e y s Brain Nerve signals z Measurement Visual indication of arm location Pressure Eyes and pressure receptors Arm & muscles d Arm location © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 9 Problems P1.10 An aircraft flight path control system using GPS: Controller Desired flight path from air traffic controllers Actuators Computer Auto-pilot Error - Process Ailerons, elevators, rudder, and engine power Flight path Aircraft Measurement Measured flight path P1.11 The accuracy of the clock is dependent upon a constant flow from the orifice; the flow is dependent upon the height of the water in the float tank. The height of the water is controlled by the float. The control system controls only the height of the water. Any errors due to enlargement of the orifice or evaporation of the water in the lower tank is not accounted for. The control system can be seen as: Desired height of the water in float tank P1.12 Global Positioning System - Controller Process Float level Flow from upper tank to float tank Actual height Assume that the turret and fantail are at 90◦ , if θw 6= θF -90◦ . The fantail operates on the error signal θw - θT , and as the fantail turns, it drives the turret to turn. y Wind qW = Wind angle qF = Fantail angle qT = Turret angle Controller * qW qF qT qW * Turret x - Process Torque Error Fantail Fantail Gears & turret qT © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 10 CHAPTER 1 P1.13 Introduction to Control Systems This scheme assumes the person adjusts the hot water for temperature control, and then adjusts the cold water for flow rate control. Controller Error Desired water temperature Process Hot water system Valve adjust - Hot water Actual water temperature and flow rate Desired water flow rate Cold water system Valve adjust - Cold water Measurement Measured water flow Measured water temperature P1.14 Human: visual and touch If the rewards in a specific trade is greater than the average reward, there is a positive influx of workers, since q(t) = f1 (c(t) − r(t)). If an influx of workers occurs, then reward in specific trade decreases, since c(t) = −f2 (q(t)). Controller Average rewards r(t) P1.15 Desired Fuel Pressure Error - f1(c(t)-r(t)) Process q(t) - f2(q(t)) Total of rewards c(t) A computer controlled fuel injection system: - Controller Process Electronic Control Unit High Pressure Fuel Supply Pump and Electronic Fuel Injectors Measurement Measured fuel pressure Fuel Pressure Sensor Fuel Pressure © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 11 Problems P1.16 With the onset of a fever, the body thermostat is turned up. The body adjusts by shivering and less blood flows to the skin surface. Aspirin acts to lowers the thermal set-point in the brain. Controller Desired temperature or set-point from body thermostat in the brain Process Adjustments within the body - Body temperature Body Measurement Measured body temperature Internal sensor P1.17 Hitting a baseball is arguably one of the most difficult feats in all of sports. Given that pitchers may throw the ball at speeds of 90 mph (or higher!), batters have only about 0.1 second to make the decision to swing—with bat speeds aproaching 90 mph. The key to hitting a baseball a long distance is to make contact with the ball with a high bat velocity. This is more important than the bat’s weight, which is usually around 33 ounces (compared to Ty Cobb’s bat which was 41 ounces!). Since the pitcher can throw a variety of pitches (fast ball, curve ball, slider, etc.), a batter must decide if the ball is going to enter the strike zone and if possible, decide the type of pitch. The batter uses his/her vision as the sensor in the feedback loop. A high degree of eye-hand coordination is key to success—that is, an accurate feedback control system. P1.18 Define the following variables: p = output pressure, fs = spring force = Kx, fd = diaphragm force = Ap, and fv = valve force = fs - fd . The motion of the valve is described by ÿ = fv /m where m is the valve mass. The output pressure is proportional to the valve displacement, thus p = cy , where c is the constant of proportionality. Constant of proportionality Spring Screw displacement x(t) K fs - Valve position fv Valve c y Diaphragm area fd A Output pressure p(t) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 12 CHAPTER 1 P1.19 Introduction to Control Systems A control system to keep a car at a given relative position offset from a lead car: Throttle Position of follower Follower car Actuator u - Controller Relative position - Position of lead Lead car Fuel throttle (fuel) Video camera & processing algorithms Reference photo Desired relative position P1.20 A control system for a high-performance car with an adjustable wing: Desired road adhesion - Process Actuator Controller Computer Adjustable wing Road conditions Race Car Road adhesion Measurement Measured road adhesion P1.21 K Tire internal strain gauges A control system for a twin-lift helicopter system: Measurement Measured separation distance Desired separation distance - Controller Process Separation distance Pilot Desired altitude Radar Helicopter Altitude Measurement Measured altitude Altimeter © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 13 Problems P1.22 The desired building deflection would not necessarily be zero. Rather it would be prescribed so that the building is allowed moderate movement up to a point, and then active control is applied if the movement is larger than some predetermined amount. Process Controller Desired deflection Hydraulic stiffeners - Building Deflection Measurement Measured deflection P1.23 Strain gauges on truss structure K The human-like face of the robot might have micro-actuators placed at strategic points on the interior of the malleable facial structure. Cooperative control of the micro-actuators would then enable the robot to achieve various facial expressions. Controller Process Error Desired actuator position - Voltage Electromechanical actuator Amplifier Actuator position Measurement Position sensor Measured position P1.24 We might envision a sensor embedded in a “gutter” at the base of the windshield which measures water levels—higher water levels corresponds to higher intensity rain. This information would be used to modulate the wiper blade speed. Process Controller Desired wiper speed Wiper blade and motor Electronic Control Unit - Measurement K Measured water level Water depth sensor Wiper blade speed © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 14 CHAPTER 1 Introduction to Control Systems A feedback control system for the space traffic control: P1.25 Controller Error Desired orbit position Control law - Actuator Jet commands Process Applied forces Reaction control jets Satellite Actual orbit position Measurement Measured orbit position Radar or GPS Earth-based control of a microrover to point the camera: P1.26 Microrover Camera position command Receiver/ Transmitter Controller G(s) Gc(s) Rover position Camera Camera Position m Ca ap er Sensor ea iti os M Measured camera position on d re su d an m m co ap er m ca on iti os P1.27 Desired Charge Level Control of a methanol fuel cell: - Controller Recharging System Gc(s) GR(s) Methanol water solution G(s) Sensor Measured charge level Fuel Cell H(s) Charge Level © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 15 Advanced Problems Advanced Problems AP1.1 Control of a robotic microsurgical device: Microsurgical robotic manipulator Controller Desired End-effector Position - G(s) Gc(s) End-effector Position Sensor H(s) AP1.2 An advanced wind energy system viewed as a mechatronic system: AERODYNAMIC DESIGN STRUCTURAL DESIGN OF THE TOWER ELECTRICAL AND POWER SYSTEMS SENSORS Rotor rotational sensor Wind speed and direction sensor ACTUATORS Motors for manipulatiing the propeller pitch Physical System Modeling CONTROL SYSTEM DESIGN AND ANALYSIS ELECTRICAL SYSTEM DESIGN AND ANALYSIS POWER GENERATION AND STORAGE Sensors and Actuators WIND ENERGY SYSTEM Software and Data Acquisition CONTROLLER ALGORITHMS DATA ACQUISTION: WIND SPEED AND DIRECTION ROTOR ANGULAR SPEED PROPELLOR PITCH ANGLE AP1.3 Signals and Systems Computers and Logic Systems COMPUTER EQUIPMENT FOR CONTROLLING THE SYSTEM SAFETY MONITORING SYSTEMS The automatic parallel parking system might use multiple ultrasound sensors to measure distances to the parked automobiles and the curb. The sensor measurements would be processed by an on-board computer to determine the steering wheel, accelerator, and brake inputs to avoid collision and to properly align the vehicle in the desired space. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 16 CHAPTER 1 Introduction to Control Systems Even though the sensors may accurately measure the distance between the two parked vehicles, there will be a problem if the available space is not big enough to accommodate the parking car. Controller Desired automobile position Error Actuators On-board computer - Steering wheel, accelerator, and brake Process Actual automobile position Automobile Measurement Position of automobile relative to parked cars and curb Ultrasound There are various control methods that can be considered, including placing the controller in the feedforward loop (as in Figure 1.3). The adaptive optics block diagram below shows the controller in the feedback loop, as an alternative control system architecture. AP1.4 Process Astronomical object Uncompensated image Astronomical telescope mirror Compensated image Measurement Wavefront reconstructor Wavefront corrector Wavefront sensor Actuator & controller AP1.5 Desired floor Error - The control system might have an inner loop for controlling the acceleration and an outer loop to reach the desired floor level precisely. Controller #2 Outer Loop Desired acceleration Error - Controller #1 Elevator motor, cables, etc. Inner Loop Measured acceleration Acceleration Measurement Elevator Floor © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 17 Advanced Problems An obstacle avoidance control system would keep the robotic vacuum cleaner from colliding with furniture but it would not necessarily put the vacuum cleaner on an optimal path to reach the entire floor. This would require another sensor to measure position in the room, a digital map of the room layout, and a control system in the outer loop. AP1.6 Process Desired distance from obstacles Error - Controller Measured distance from obstacle Motors, wheels, etc. Infrared sensors Robotic vacuum cleaner Distance from obstacles © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 18 CHAPTER 1 Introduction to Control Systems Design Problems CDP1.1 The machine tool with the movable table in a feedback control configuration: Controller Error Desired position x Amplifier - Actuator Process Machine tool with table Positioning motor Actual position x Measurement Position sensor Measured position DP1.1 Use the stereo system and amplifiers to cancel out the noise by emitting signals 180◦ out of phase with the noise. Process Controller Noise signal Desired noise = 0 Shift phase by 180 deg - Machine tool with table Positioning motor Noise in cabin Measurement Microphone DP1.2 Desired speed of auto set by driver 1/K An automobile cruise control system: Controller Desired shaft speed - Electric motor Process Automobile and engine Valve Measurement Measured shaft speed Shaft speed sensor Drive shaf t speed K Actual speed of auto © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 19 Design Problems DP1.3 An automoted cow milking system: Measurement Cow location Vision system Motor and gears - Desired cup location Process Actuator Controller Location of cup Robot arm and cup gripper Cow and milker Milk Measurement Vision system Measured cup location DP1.4 A feedback control system for a robot welder: Controller Desired position Process Computer and amplifier Error - Voltage Motor and arm Weld top position Measurement Vision camera Measured position DP1.5 A control system for one wheel of a traction control system: Antislip controller Engine torque + - Wheel dynamics + - Wheel speed Sensor + Actual slip 1/Rw Vehicle dynamics Brake torque + Vehicle speed Antiskid controller Rw = Radius of wheel Sensor Measured slip © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 20 CHAPTER 1 Introduction to Control Systems A vibration damping system for the Hubble Space Telescope: DP1.6 Controller Desired jitter = 0 Error Computer - Actuators Gyro and reaction wheels Process Signal to cancel the jitter Spacecraft dynamics Jitter of vibration Measurement Measurement of 0.05 Hz jitter DP1.7 A control system for a nanorobot: Controller Desired nanorobot position Rate gyro sensor Error - Biocomputer Actuators Plane surfaces and propellers Process Nanorobot Actual nanorobot position Measurement External beacons Many concepts from underwater robotics can be applied to nanorobotics within the bloodstream. For example, plane surfaces and propellers can provide the required actuation with screw drives providing the propulsion. The nanorobots can use signals from beacons located outside the skin as sensors to determine their position. The nanorobots use energy from the chemical reaction of oxygen and glucose available in the human body. The control system requires a bio-computer–an innovation that is not yet available. For further reading, see A. Cavalcanti, L. Rosen, L. C. Kretly, M. Rosenfeld, and S. Einav, “Nanorobotic Challenges n Biomedical Application, Design, and Control,” IEEE ICECS Intl Conf. on Electronics, Circuits and Systems, Tel-Aviv, Israel, December 2004. DP1.8 The feedback control system might use gyros and/or accelerometers to measure angle change and assuming the HTV was originally in the vertical position, the feedback would retain the vertical position using commands to motors and other actuators that produced torques and could move the HTV forward and backward. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 21 Design Problems Process Desired angle from vertical (0o) Error - Controller Measured angle from vertical Motors, wheels, etc. Gyros & accelerometers HTV Angle from vertical © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 2 Mathematical Models of Systems Exercises E2.1 We have for the open-loop y = r2 and for the closed-loop e = r − y and y = e2 . So, e = r − e2 and e2 + e − r = 0 . 16 14 12 y 10 8 open-loop 6 4 closed-loop 2 0 0 0.5 1 1.5 2 r 2.5 3 3.5 4 FIGURE E2.1 Plot of open-loop versus closed-loop. For example, if r = 1, then e2 + e − 1 = 0 implies that e = 0.618. Thus, y = 0.382. A plot y versus r is shown in Figure E2.1. 22 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 23 Exercises E2.2 Define f (T ) = R = R0 e−0.1T and ∆R = f (T ) − f (T0 ) , ∆T = T − T0 . Then, ∆R = f (T ) − f (T0 ) = ∂f ∂T T =T0 =20◦ ∆T + · · · where ∂f ∂T T =T0 =20◦ = −0.1R0 e−0.1T0 = −135, when R0 = 10, 000Ω. Thus, the linear approximation is computed by considering only the first-order terms in the Taylor series expansion, and is given by ∆R = −135∆T . The spring constant for the equilibrium point is found graphically by estimating the slope of a line tangent to the force versus displacement curve at the point y = 0.5cm, see Figure E2.3. The slope of the line is K ≈ 1. 2 1.5 Spring breaks 1 0.5 0 Force (n) E2.3 -0.5 -1 -1.5 -2 -2.5 -3 -2 Spring compresses -1.5 -1 -0.5 0 0.5 1 y=Displacement (cm) FIGURE E2.3 Spring force as a function of displacement. 1.5 2 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 24 CHAPTER 2 E2.4 Mathematical Models of Systems Since R(s) = 1 s we have Y (s) = 4(s + 50) . s(s + 20)(s + 10) The partial fraction expansion of Y (s) is given by Y (s) = A1 A2 A3 + + s s + 20 s + 10 where A1 = 1 , A2 = 0.6 and A3 = −1.6 . Using the Laplace transform table, we find that y(t) = 1 + 0.6e−20t − 1.6e−10t . The final value is computed using the final value theorem: 4(s + 50) lim y(t) = lim s =1. 2 t→∞ s→0 s(s + 30s + 200) E2.5 The circuit diagram is shown in Figure E2.5. R2 v+ A + vin - FIGURE E2.5 Noninverting op-amp circuit. With an ideal op-amp, we have vo = A(vin − v − ), + v0 - R1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 25 Exercises where A is very large. We have the relationship R1 vo . R1 + R2 v− = Therefore, vo = A(vin − R1 vo ), R1 + R2 and solving for vo yields vo = A 1+ AR1 R1 +R2 1 Since A ≫ 1, it follows that 1 + RAR ≈ 1 +R2 vo simplifies to vo = E2.6 vin . AR1 R1 +R2 . Then the expression for R1 + R2 vin . R1 Given y = f (x) = ex and the operating point xo = 1, we have the linear approximation y = f (x) = f (xo ) + ∂f ∂x x=xo (x − xo ) + · · · where df dx f (xo ) = e, = e, x=xo =1 and x − xo = x − 1. Therefore, we obtain the linear approximation y = ex. E2.7 The block diagram is shown in Figure E2.7. R(s) Ea(s) + G1(s) G2(s) - H(s) FIGURE E2.7 Block diagram model. I(s) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 26 CHAPTER 2 Mathematical Models of Systems Starting at the output we obtain I(s) = G1 (s)G2 (s)E(s). But E(s) = R(s) − H(s)I(s), so I(s) = G1 (s)G2 (s) [R(s) − H(s)I(s)] . Solving for I(s) yields the closed-loop transfer function G1 (s)G2 (s) I(s) = . R(s) 1 + G1 (s)G2 (s)H(s) E2.8 The block diagram is shown in Figure E2.8. H2(s) - R(s) K - E(s) - G1(s) W(s) - A(s) G2(s) Z(s) 1 s Y(s) H3(s) H1(s) FIGURE E2.8 Block diagram model. Starting at the output we obtain Y (s) = 1 1 Z(s) = G2 (s)A(s). s s But A(s) = G1 (s) [−H2 (s)Z(s) − H3 (s)A(s) + W (s)] and Z(s) = sY (s), so 1 Y (s) = −G1 (s)G2 (s)H2 (s)Y (s) − G1 (s)H3 (s)Y (s) + G1 (s)G2 (s)W (s). s Substituting W (s) = KE(s) − H1 (s)Z(s) into the above equation yields Y (s) = −G1 (s)G2 (s)H2 (s)Y (s) − G1 (s)H3 (s)Y (s) 1 + G1 (s)G2 (s) [KE(s) − H1 (s)Z(s)] s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 27 Exercises and with E(s) = R(s) − Y (s) and Z(s) = sY (s) this reduces to Y (s) = [−G1 (s)G2 (s) (H2 (s) + H1 (s)) − G1 (s)H3 (s) 1 1 − G1 (s)G2 (s)K]Y (s) + G1 (s)G2 (s)KR(s). s s Solving for Y (s) yields the transfer function Y (s) = T (s)R(s), where T (s) = E2.9 KG1 (s)G2 (s)/s . 1 + G1 (s)G2 (s) [(H2 (s) + H1 (s)] + G1 (s)H3 (s) + KG1 (s)G2 (s)/s From Figure E2.9, we observe that Ff (s) = G2 (s)U (s) and FR (s) = G3 (s)U (s) . Then, solving for U (s) yields U (s) = 1 Ff (s) G2 (s) FR (s) = G3 (s) U (s) . G2 (s) and it follows that Again, considering the block diagram in Figure E2.9 we determine Ff (s) = G1 (s)G2 (s)[R(s) − H2 (s)Ff (s) − H2 (s)FR (s)] . But, from the previous result, we substitute for FR (s) resulting in Ff (s) = G1 (s)G2 (s)R(s)−G1 (s)G2 (s)H2 (s)Ff (s)−G1 (s)H2 (s)G3 (s)Ff (s) . Solving for Ff (s) yields G1 (s)G2 (s) Ff (s) = R(s) . 1 + G1 (s)G2 (s)H2 (s) + G1 (s)G3 (s)H2 (s) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 28 CHAPTER 2 Mathematical Models of Systems H2(s) + - R(s) U(s) G2(s) Ff (s) U(s) G3(s) FR(s) G1(s) - H2(s) FIGURE E2.9 Block diagram model. E2.10 The shock absorber block diagram is shown in Figure E2.10. The closedloop transfer function model is T (s) = Gc (s)Gp (s)G(s) . 1 + H(s)Gc (s)Gp (s)G(s) Controller Gear Motor Plunger and Piston System Gc(s) Gp(s) G(s) + R(s) Desired piston travel - Y(s) Piston travel Sensor H(s) Piston travel measurement FIGURE E2.10 Shock absorber block diagram. E2.11 Let f denote the spring force (n) and x denote the deflection (m). Then K= ∆f . ∆x Computing the slope from the graph yields: (a) xo = −0.14m → K = ∆f /∆x = 10 n / 0.04 m = 250 n/m (b) xo = 0m → K = ∆f /∆x = 10 n / 0.05 m = 200 n/m (c) xo = 0.35m → K = ∆f /∆x = 3n / 0.05 m = 60 n/m © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 29 Exercises E2.12 The signal flow graph is shown in Fig. E2.12. Find Y (s) when R(s) = 0. -K Td(s) 1 1 K2 G(s) Y (s) -1 FIGURE E2.12 Signal flow graph. The transfer function from Td (s) to Y (s) is Y (s) = G(s)(1 − K1 K2 )Td (s) G(s)Td (s) − K1 K2 G(s)Td (s) = . 1 − (−K2 G(s)) 1 + K2 G(s) If we set K1 K2 = 1 , then Y (s) = 0 for any Td (s). E2.13 The transfer function from R(s), Td (s), and N (s) to Y (s) is K K 1 R(s)+ 2 Td (s)− 2 N (s) Y (s) = 2 s + 10s + K s + 10s + K s + 10s + K Therefore, we find that Y (s)/Td (s) = E2.14 s2 1 + 10s + K and Y (s)/N (s) = − s2 K + 10s + K Since we want to compute the transfer function from R2 (s) to Y1 (s), we can assume that R1 = 0 (application of the principle of superposition). Then, starting at the output Y1 (s) we obtain Y1 (s) = G3 (s) [−H1 (s)Y1 (s) + G2 (s)G8 (s)W (s) + G9 (s)W (s)] , or [1 + G3 (s)H1 (s)] Y1 (s) = [G3 (s)G2 (s)G8 (s)W (s) + G3 (s)G9 (s)] W (s). Considering the signal W (s) (see Figure E2.14), we determine that W (s) = G5 (s) [G4 (s)R2 (s) − H2 (s)W (s)] , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 30 CHAPTER 2 Mathematical Models of Systems H1(s) + G1(s) R1(s) + G7(s) R2(s) G4(s) - + G2(s) G3(s) + Y1(s) G9(s) G8(s) + + G6(s) G5(s) Y2(s) W(s) - H2(s) FIGURE E2.14 Block diagram model. or [1 + G5 (s)H2 (s)] W (s) = G5 (s)G4 (s)R2 (s). Substituting the expression for W (s) into the above equation for Y1 (s) yields Y1 (s) G2 (s)G3 (s)G4 (s)G5 (s)G8 (s) + G3 (s)G4 (s)G5 (s)G9 (s) = . R2 (s) 1 + G3 (s)H1 (s) + G5 (s)H2 (s) + G3 (s)G5 (s)H1 (s)H2 (s) E2.15 For loop 1, we have di1 1 R1 i1 + L1 + dt C1 Z (i1 − i2 )dt + R2 (i1 − i2 ) = v(t) . And for loop 2, we have 1 C2 E2.16 Z di2 1 i2 dt + L2 + R2 (i2 − i1 ) + dt C1 Z (i2 − i1 )dt = 0 . The transfer function from R(s) to P (s) is P (s) 4.2 = 3 . 2 R(s) s + 2s + 4s + 4.2 The block diagram is shown in Figure E2.16a. The corresponding signal flow graph is shown in Figure E2.16b for P (s)/R(s) = s3 + 4.2 . + 4s + 4.2 2s2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 31 Exercises v1(s) R(s) v2(s) 7 - q(s) 0.6 s 1 s2+2s+4 P(s) (a) R(s ) 1 V1 7 1 s2 + 2 s + 4 0.6 s V2 P (s) -1 (b) FIGURE E2.16 (a) Block diagram, (b) Signal flow graph. E2.17 A linear approximation for f is given by ∆f = ∂f ∂x ∆x = 2kxo ∆x = k∆x x=xo where xo = 1/2, ∆f = f (x) − f (xo ), and ∆x = x − xo . E2.18 The linear approximation is given by ∆y = m∆x where m= ∂y ∂x . x=xo (a) When xo = 1, we find that yo = 2.4, and yo = 13.2 when xo = 2. (b) The slope m is computed as follows: m= ∂y ∂x = 1 + 4.2x2o . x=xo Therefore, m = 5.2 at xo = 1, and m = 18.8 at xo = 2. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 32 CHAPTER 2 E2.19 Mathematical Models of Systems The output (with a step input) is Y (s) = 15(s + 1) . s(s + 7)(s + 2) The partial fraction expansion is 18 1 3 1 15 − + . 14s 7 s+7 2s+2 Y (s) = Taking the inverse Laplace transform yields y(t) = E2.20 15 18 −7t 3 −2t − e + e . 14 7 2 The input-output relationship is A(K − 1) Vo = V 1 + AK where K= Z1 . Z1 + Z2 Assume A ≫ 1. Then, Vo K−1 Z2 = =− V K Z1 where Z1 = R1 R1 C 1 s + 1 and Z2 = R2 . R2 C 2 s + 1 Therefore, Vo (s) R2 (R1 C1 s + 1) 2(s + 1) =− =− . V (s) R1 (R2 C2 s + 1) s+2 E2.21 The equation of motion of the mass mc is mc ẍp + (bd + bs )ẋp + kd xp = bd ẋin + kd xin . Taking the Laplace transform with zero initial conditions yields [mc s2 + (bd + bs )s + kd ]Xp (s) = [bd s + kd ]Xin (s) . So, the transfer function is bd s + kd 0.7s + 2 Xp (s) = = 2 . Xin (s) mc s2 + (bd + bs )s + kd s + 2.8s + 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 33 Exercises E2.22 The rotational velocity is ω(s) = 2(s + 4) 1 . 2 (s + 5)(s + 1) s Expanding in a partial fraction expansion yields ω(s) = 81 1 1 3 13 1 1 + − − . 2 5 s 40 s + 5 2 (s + 1) 8 s+1 Taking the inverse Laplace transform yields ω(t) = E2.23 8 1 3 13 + e−5t − te−t − e−t . 5 40 2 8 The closed-loop transfer function is Y (s) K1 K2 = T (s) = 2 . R(s) s + (K1 + K2 K3 + K1 K2 )s + K1 K2 K3 E2.24 The closed-loop tranfser function is Y (s) 10 = T (s) = 2 . R(s) s + 21s + 10 E2.25 Let x = 0.6 and y = 0.8. Then, with y = ax3 , we have 0.8 = a(0.6)3 . Solving for a yields a = 3.704. A linear approximation is y − yo = 3ax2o (x − xo ) or y = 4x − 1.6, where yo = 0.8 and xo = 0.6. E2.26 The equations of motion are m1 ẍ1 + k(x1 − x2 ) = F m2 ẍ2 + k(x2 − x1 ) = 0 . Taking the Laplace transform (with zero initial conditions) and solving for X2 (s) yields X2 (s) = (m2 s2 k F (s) . + k)(m1 s2 + k) − k 2 Then, with m1 = m2 = k = 1, we have X2 (s)/F (s) = 1 . s2 (s2 + 2) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 34 CHAPTER 2 E2.27 Mathematical Models of Systems The transfer function from Td (s) to Y (s) is Y (s)/Td (s) = E2.28 G2 (s) . 1 + G1 G2 H(s) The transfer function is R2 R4 C R2 R4 Vo (s) = s+ = 24s + 144 . V (s) R3 R1 R3 E2.29 (a) If G(s) = s2 1 + 15s + 50 and H(s) = 2s + 15 , then the closed-loop transfer function of Figure E2.28(a) and (b) (in Dorf & Bishop) are equivalent. (b) The closed-loop transfer function is T (s) = (a) The closed-loop transfer function is T (s) = G(s) 1 10 = 2 1 + G(s) s s(s + 2s + 20) where G(s) = 0.8 0.7 0.6 Amplitude E2.30 1 . s2 + 17s + 65 0.5 0.4 0.3 0.2 0.1 0 0 1 2 3 Time sec FIGURE E2.30 Step response. 4 5 6 s2 10 . + 2s + 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 35 Exercises (b) The output Y (s) (when R(s) = 1/s) is Y (s) = 0.5 −0.25 + 0.0573j −0.25 − 0.0573j − + . s s + 1 − 4.3589j s + 1 + 4.3589j (c) The plot of y(t) is shown in Figure E2.30. The output is given by √ √ 1 1 y(t) = 1 − e−t cos 19t − √ sin 19t 2 19 E2.31 The partial fraction expansion is V (s) = a b + s + p1 s + p2 where p1 = 4 − 19.6j and p2 = 4 + 19.6j. Then, the residues are a = −10.2j b = 10.2j . The inverse Laplace transform is v(t) = −10.2je(−4+19.6j)t + 10.2je(−4−19.6j)t = 20.4e−4t sin 19.6t . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 36 CHAPTER 2 Mathematical Models of Systems Problems P2.1 The integrodifferential equations, obtained by Kirchoff’s voltage law to each loop, are as follows: R1 i1 + 1 C1 R3 i2 + 1 C2 Z i1 dt + L1 d(i1 − i2 ) + R2 (i1 − i2 ) = v(t) dt (loop 1) and P2.2 Z i2 dt + R2 (i2 − i1 ) + L1 d(i2 − i1 ) =0 dt (loop 2) . The differential equations describing the system can be obtained by using a free-body diagram analysis of each mass. For mass 1 and 2 we have M1 ÿ1 + k12 (y1 − y2 ) + bẏ1 + k1 y1 = F (t) M2 ÿ2 + k12 (y2 − y1 ) = 0 . Using a force-current analogy, the analagous electric circuit is shown in Figure P2.2, where Ci → Mi , L1 → 1/k1 , L12 → 1/k12 , and R → 1/b . FIGURE P2.2 Analagous electric circuit. P2.3 The differential equations describing the system can be obtained by using a free-body diagram analysis of each mass. For mass 1 and 2 we have M ẍ1 + kx1 + k(x1 − x2 ) = F (t) M ẍ2 + k(x2 − x1 ) + bẋ2 = 0 . Using a force-current analogy, the analagous electric circuit is shown in Figure P2.3, where C→M L → 1/k R → 1/b . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 37 Problems FIGURE P2.3 Analagous electric circuit. (a) The linear approximation around vin = 0 is vo = 0vin , see Figure P2.4(a). (b) The linear approximation around vin = 1 is vo = 2vin − 1, see Figure P2.4(b). (a) (b) 0.4 4 3.5 0.3 3 0.2 2.5 0.1 2 vo vo P2.4 0 1.5 linear approximation 1 -0.1 0.5 -0.2 0 -0.3 -0.4 -1 linear approximation -0.5 -0.5 0 vin 0.5 FIGURE P2.4 Nonlinear functions and approximations. 1 -1 -1 0 1 vin 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 38 CHAPTER 2 P2.5 Mathematical Models of Systems Given Q = K(P1 − P2 )1/2 . Let δP = P1 − P2 and δPo = operating point. Using a Taylor series expansion of Q, we have Q = Qo + ∂Q ∂δP (δP − δPo ) + · · · δP =δPo where Qo = KδPo1/2 ∂Q ∂δP and = δP =δPo K −1/2 . δP 2 o Define ∆Q = Q − Qo and ∆P = δP − δPo . Then, dropping higher-order terms in the Taylor series expansion yields ∆Q = m∆P where m= P2.6 K 1/2 2δPo . From P2.1 we have R1 i1 + 1 C1 R3 i2 + 1 C2 Z i1 dt + L1 d(i1 − i2 ) + R2 (i1 − i2 ) = v(t) dt and Z i2 dt + R2 (i2 − i1 ) + L1 d(i2 − i1 ) =0. dt Taking the Laplace transform and using the fact that the initial voltage across C2 is 10v yields [R1 + 1 + L1 s + R2 ]I1 (s) + [−R2 − L1 s]I2 (s) = 0 C1 s and [−R2 − L1 s]I1 (s) + [L1 s + R3 + 1 10 + R2 ]I2 (s) = − . C2 s s Rewriting in matrix form we have R1 + 1 C1 s + L 1 s + R2 −R2 − L1 s −R2 − L1 s L 1 s + R3 + 1 C2 s + R2 I1 (s) I2 (s) = 0 −10/s . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 39 Problems Solving for I2 yields 1 0 R2 + L 1 s 1 L1 s + R3 + C2 s + R2 = . 1 ∆ −10/s I2 (s) R2 + L 1 s R1 + C1 s + L1 s + R2 I1 (s) or I2 (s) = −10(R1 + 1/C1 s + L1 s + R2 ) s∆ where ∆ = (R1 + P2.7 1 1 + L1 s + R2 )(L1 s + R3 + + R2 ) − (R2 + L1 s)2 . C1 s C2 s Consider the differentiating op-amp circuit in Figure P2.7. For an ideal op-amp, the voltage gain (as a function of frequency) is V2 (s) = − Z2 (s) V1 (s), Z1 (s) where Z1 = R1 1 + R1 Cs and Z2 = R2 are the respective circuit impedances. Therefore, we obtain V2 (s) = − Z R2 (1 + R1 Cs) V1 (s). R1 Z 1 C + R1 2 R2 + + V1(s) V2(s) - - FIGURE P2.7 Differentiating op-amp circuit. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 40 CHAPTER 2 Let ∆= G2 + Cs −Cs −G2 −Cs G1 + 2Cs −Cs −G2 −Cs Cs + G2 . Then, Vj = ∆ij I1 ∆ or or V3 ∆13 I1 /∆ = . V1 ∆11 I1 /∆ Therefore, the transfer function is −Cs 2Cs + G1 T (s) = ∆13 V3 = = V1 ∆11 −G2 −Cs 2Cs + G1 −Cs −Cs Cs + G2 Pole-zero map (x:poles and o:zeros) 3 2 o 1 Imag Axis P2.8 Mathematical Models of Systems 0 x x -1 -2 -3 -8 o -7 -6 -5 -4 Real Axis FIGURE P2.8 Pole-zero map. -3 -2 -1 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 41 Problems = C 2 R1 R2 s2 + 2CR1 s + 1 . C 2 R1 R2 s2 + (2R1 + R2 )Cs + 1 Using R1 = 0.5, R2 = 1, and C = 0.5, we have T (s) = s2 + 4s + 8 (s + 2 + 2j)(s + 2 − 2j) √ √ . = 2 s + 8s + 8 (s + 4 + 8)(s + 4 − 8) The pole-zero map is shown in Figure P2.8. From P2.3 we have M ẍ1 + kx1 + k(x1 − x2 ) = F (t) M ẍ2 + k(x2 − x1 ) + bẋ2 = 0 . Taking the Laplace transform of both equations and writing the result in matrix form, it follows that M s2 + 2k −k M s2 + bs + k −k X1 (s) X2 (s) = F (s) 0 , Pole zero map 0.4 0.3 0.2 0.1 Imag Axis P2.9 0 - 0.1 -0.2 -0.3 -0.4 -0.03 FIGURE P2.9 Pole-zero map. -0.025 -0.02 -0.015 Real Axis -0.01 -0.005 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 42 CHAPTER 2 Mathematical Models of Systems or k F (s) 1 M s2 + bs + k = 2 ∆ X2 (s) k M s + 2k 0 X1 (s) where ∆ = (M s2 + bs + k)(M s2 + 2k) − k 2 . So, G(s) = M s2 + bs + k X1 (s) = . F (s) ∆ When b/k = 1, M = 1 , b2 /M k = 0.04, we have G(s) = s2 + 0.04s + 0.04 . s4 + 0.04s3 + 0.12s2 + 0.0032s + 0.0016 The pole-zero map is shown in Figure P2.9. P2.10 From P2.2 we have M1 ÿ1 + k12 (y1 − y2 ) + bẏ1 + k1 y1 = F (t) M2 ÿ2 + k12 (y2 − y1 ) = 0 . Taking the Laplace transform of both equations and writing the result in matrix form, it follows that or M1 s2 + bs + k1 + k12 M2 s2 + k12 −k12 −k12 Y1 (s) Y2 (s) = F (s) 0 k12 F (s) 1 M2 s2 + k12 = ∆ Y2 (s) k12 M1 s2 + bs + k1 + k12 0 Y1 (s) where 2 ∆ = (M2 s2 + k12 )(M1 s2 + bs + k1 + k12 ) − k12 . So, when f (t) = a sin ωo t, we have that Y1 (s) is given by Y1 (s) = aM2 ωo (s2 + k12 /M2 ) . (s2 + ωo2 )∆(s) For motionless response (in the steady-state), set the zero of the transfer function so that (s2 + k12 ) = s2 + ωo2 M2 or ωo2 = k12 . M2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 43 Problems P2.11 The transfer functions from Vc (s) to Vd (s) and from Vd (s) to θ(s) are: K1 K2 , and (Lq s + Rq )(Lc s + Rc ) Km . θ(s)/Vd (s) = 2 (Js + f s)((Ld + La )s + Rd + Ra ) + K3 Km s Vd (s)/Vc (s) = The block diagram for θ(s)/Vc (s) is shown in Figure P2.11, where θ(s)/Vc (s) = K1 K2 Km θ(s) Vd (s) = , Vd (s) Vc (s) ∆(s) where ∆(s) = s(Lc s + Rc )(Lq s + Rq )((Js + b)((Ld + La )s + Rd + Ra ) + Km K3 ) . Vc 1 L cs+R c Ic K1 Vq 1 L qs+R q Iq K2 Vd + 1 (L d+L a)s+R d+R a Id Tm Km - 1 Js+f w 1 s q Vb K3 FIGURE P2.11 Block diagram. P2.12 The open-loop transfer function is Y (s) K = . R(s) s + 20 With R(s) = 1/s, we have Y (s) = K . s(s + 20) The partial fraction expansion is K Y (s) = 20 1 1 − , s s + 20 and the inverse Laplace transform is y(t) = K 1 − e−20t , 20 As t → ∞, it follows that y(t) → K/20. So we choose K = 20 so that y(t) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 44 CHAPTER 2 Mathematical Models of Systems approaches 1. Alternatively we can use the final value theorem to obtain y(t)t→∞ = lim sY (s) = s→0 K =1. 20 It follows that choosing K = 20 leads to y(t) → 1 as t → ∞. P2.13 The motor torque is given by Tm (s) = (Jm s2 + bm s)θm (s) + (JL s2 + bL s)nθL (s) = n((Jm s2 + bm s)/n2 + JL s2 + bL s)θL (s) where n = θL (s)/θm (s) = gear ratio . But Tm (s) = Km Ig (s) and Ig (s) = 1 Vg (s) , (Lg + Lf )s + Rg + Rf and Vg (s) = Kg If (s) = Kg Vf (s) . Rf + L f s Combining the above expressions yields θL (s) Kg Km = . Vf (s) n∆1 (s)∆2 (s) where ∆1 (s) = JL s2 + bL s + Jm s2 + bm s n2 and ∆2 (s) = (Lg s + Lf s + Rg + Rf )(Rf + Lf s) . P2.14 For a field-controlled dc electric motor we have ω(s)/Vf (s) = Km /Rf . Js + b © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 45 Problems With a step input of Vf (s) = 80/s, the final value of ω(t) is 80Km = 2.4 Rf b ω(t)t→∞ = lim sω(s) = s→0 or Km = 0.03 . Rf b Solving for ω(t) yields 80Km −1 1 ω(t) = L Rf J s(s + b/J) = 80Km (1−e−(b/J)t ) = 2.4(1−e−(b/J)t ) . Rf b At t = 1/2, ω(t) = 1, so ω(1/2) = 2.4(1 − e−(b/J)t ) = 1 implies b/J = 1.08 sec . Therefore, ω(s)/Vf (s) = P2.15 0.0324 . s + 1.08 Summing the forces in the vertical direction and using Newton’s Second Law we obtain ẍ + k x=0. m The system has no damping and no external inputs. Taking the Laplace transform yields X(s) = s2 x0 s , + k/m where we used the fact that x(0) = x0 and ẋ(0) = 0. Then taking the inverse Laplace transform yields x(t) = x0 cos P2.16 s k t. m Using Cramer’s rule, we have 1 1.5 x1 or 2 4 x1 x2 = 6 11 1 4 −1.5 6 = ∆ −2 x2 1 11 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 46 CHAPTER 2 Mathematical Models of Systems where ∆ = 4(1) − 2(1.5) = 1 . Therefore, x1 = 4(6) − 1.5(11) = 7.5 1 and x2 = −2(6) + 1(11) = −1 . 1 The signal flow graph is shown in Figure P2.16. 11 1/4 6 1 -1/2 X2 X1 -1.5 FIGURE P2.16 Signal flow graph. So, x1 = P2.17 6(1) − 1.5( 11 4 ) = 7.5 3 1− 4 and x2 = 11( 41 ) + 1− −1 2 (6) 3 4 = −1 . (a) For mass 1 and 2, we have M1 ẍ1 + K1 (x1 − x2 ) + b1 (ẋ3 − ẋ1 ) = 0 M2 ẍ2 + K2 (x2 − x3 ) + b2 (ẋ3 − ẋ2 ) + K1 (x2 − x1 ) = 0 . (b) Taking the Laplace transform yields (M1 s2 + b1 s + K1 )X1 (s) − K1 X2 (s) = b1 sX3 (s) −K1 X1 (s) + (M2 s2 + b2 s + K1 + K2 )X2 (s) = (b2 s + K2 )X3 (s) . (c) Let G1 (s) = K2 + b2 s G2 (s) = 1/p(s) G3 (s) = 1/q(s) G4 (s) = sb1 , where p(s) = s2 M2 + sf2 + K1 + K2 and q(s) = s2 M1 + sf1 + K1 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 47 Problems The signal flow graph is shown in Figure P2.17. G4 G3 X3 G1 G2 X1 K1 K1 FIGURE P2.17 Signal flow graph. (d) The transfer function from X3 (s) to X1 (s) is X1 (s) K1 G1 (s)G2 (s)G3 (s) + G4 (s)G3 (s) = . X3 (s) 1 − K12 G2 (s)G3 (s) P2.18 The signal flow graph is shown in Figure P2.18. I1 V1 Va Z2 Y3 Ia Z4 V2 Y1 -Z 2 -Y 1 -Y 3 FIGURE P2.18 Signal flow graph. The transfer function is V2 (s) Y 1 Z2 Y 3 Z4 = . V1 (s) 1 + Y 1 Z2 + Y 3 Z2 + Y 3 Z4 + Y 1 Z2 Z4 Y 3 P2.19 For a noninerting op-amp circuit, depicted in Figure P2.19a, the voltage gain (as a function of frequency) is Vo (s) = Z1 (s) + Z2 (s) Vin (s), Z1 (s) where Z1 (s) and Z2 (s) are the impedances of the respective circuits. In the case of the voltage follower circuit, shown in Figure P2.19b, we have © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 48 CHAPTER 2 Mathematical Models of Systems Z2 Z1 + vin v0 + vin (a) v0 (b) FIGURE P2.19 (a) Noninverting op-amp circuit. (b) Voltage follower circuit. Z1 = ∞ (open circuit) and Z2 = 0. Therefore, the transfer function is Vo (s) Z1 = = 1. Vin (s) Z1 P2.20 (a) Assume Rg ≫ Rs and Rs ≫ R1 . Then Rs = R1 + R2 ≈ R2 , and vgs = vin − vo , where we neglect iin , since Rg ≫ Rs . At node S, we have vo = gm vgs = gm (vin − vo ) or Rs vo gm Rs = . vin 1 + gm Rs (b) With gm Rs = 20, we have vo 20 = = 0.95 . vin 21 (c) The block diagram is shown in Figure P2.20. vin(s) gmRs - FIGURE P2.20 Block diagram model. P2.21 From the geometry we find that ∆z = k l1 − l2 l2 (x − y) − y . l1 l1 vo(s) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 49 Problems The flow rate balance yields A dy = p∆z dt which implies Y (s) = p∆Z(s) . As By combining the above results it follows that l2 p l1 − l2 Y (s) = k (X(s) − Y (s)) − Y (s) . As l1 l1 Therefore, the signal flow graph is shown in Figure P2.21. Using Mason’s -1 (l 1 - l 2)/l 1 k X DZ p/As Y 1 -l 2 / l 1 FIGURE P2.21 Signal flow graph. gain formula we find that the transfer function is given by Y (s) = X(s) 1+ k(l1 −l2 )p l1 As k(l1 −l2 )p l2 p l1 As + l1 As = K1 , s + K2 + K1 where K1 = P2.22 k(l1 − l2 )p p l1 A and K2 = l2 p . l1 A (a) The equations of motion for the two masses are L 2 L M L θ¨1 + M gLθ1 + k (θ1 − θ2 ) = f (t) 2 2 2 L M L2 θ¨2 + M gLθ2 + k (θ2 − θ1 ) = 0 . 2 2 With θ˙1 = ω1 and θ˙2 = ω2 , we have g k k f (t) ω˙1 = − + θ1 + θ2 + L 4M 4M 2M L k g k ω˙2 = θ1 − + θ2 . 4M L 4M © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 50 CHAPTER 2 Mathematical Models of Systems a - F (t) w1 1/s 1/s 1/2ML (a) q1 b w2 1/s 1/s q2 a Imag(s) + j + j (b) g k L + 4M g k L + 2M X O X + j g L Re(s) FIGURE P2.22 (a) Block diagram. (b) Pole-zero map. (b) Define a = g/L + k/4M and b = k/4M . Then θ1 (s) 1 s2 + a = . F (s) 2M L (s2 + a)2 − b2 (c) The block diagram and pole-zero map are shown in Figure P2.22. P2.23 The input-output ratio, Vce /Vin , is found to be β(R − 1) + hie Rf Vce = . Vin −βhre + hie (−hoe + Rf ) P2.24 (a) The voltage gain is given by vo RL β1 β2 (R1 + R2 ) . = vin (R1 + R2 )(Rg + hie1 ) + R1 (R1 + R2 )(1 + β1 ) + R1 RL β1 β2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 51 Problems (b) The current gain is found to be ic2 = β1 β2 . ib1 (c) The input impedance is vin (R1 + R2 )(Rg + hie1 ) + R1 (R1 + R2 )(1 + β1 ) + R1 RL β1 β2 = , ib1 R1 + R2 and when β1 β2 is very large, we have the approximation vin RL R1 β1 β2 ≈ . ib1 R1 + R2 P2.25 The transfer function from R(s) and Td (s) to Y (s) is given by Y (s) = G(s) R(s) − 1 (G(s)R(s) + Td (s)) + Td (s) + G(s)R(s) G(s) = G(s)R(s) . Thus, Y (s)/R(s) = G(s) . Also, we have that Y (s) = 0 . when R(s) = 0. Therefore, the effect of the disturbance, Td (s), is eliminated. P2.26 The equations of motion for the two mass model of the robot are M ẍ + b(ẋ − ẏ) + k(x − y) = F (t) mÿ + b(ẏ − ẋ) + k(y − x) = 0 . Taking the Laplace transform and writing the result in matrix form yields M s2 + bs + k −(bs + k) −(bs + k) ms2 + bs + k X(s) Y (s) k m . = Solving for Y (s) we find that 1 Y (s) mM (bs +k) = m b F (s) s2 [s2 + 1 + M ms + ] F (s) 0 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 52 CHAPTER 2 P2.27 Mathematical Models of Systems The describing equation of motion is mz̈ = mg − k i2 . z2 Defining f (z, i) = g − ki2 mz 2 leads to z̈ = f (z, i) . The equilibrium condition for io and zo , found by solving the equation of motion when ż = z̈ = 0 , is ki2o = zo2 . mg We linearize the equation of motion using a Taylor series approximation. With the definitions ∆z = z − zo and ∆i = i − io , ˙ = ż and ∆z ¨ = z̈. Therefore, we have ∆z ¨ = f (z, i) = f (zo , io ) + ∂f ∆z ∂z z=zo i=io ∆z + ∂f ∂i z=zo i=io ∆i + · · · But f (zo , io ) = 0, and neglecting higher-order terms in the expansion yields 2 ¨ = 2kio ∆z − 2kio ∆i . ∆z mzo3 mzo2 Using the equilibrium condition which relates zo to io , we determine that ¨ = 2g ∆z − g ∆i . ∆z zo io Taking the Laplace transform yields the transfer function (valid around the equilibrium point) ∆Z(s) −g/io = 2 . ∆I(s) s − 2g/zo © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 57 Problems R(s ) + K1 s (s+1) Y (s) 1 +K 2s 1 0.9 0.8 0.7 <---- time to 90% = 0.39 sec y(t) 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 time(sec) FIGURE P2.35 The equivalent block diagram and the system step response. P2.36 (a) Given R(s) = 1/s2 , the partial fraction expansion is Y (s) = s2 (s 24 3 8/3 3/4 1 13/12 = − + + 2− . + 2)(s + 3)(s + 4) s+2 s+3 s+4 s s Therefore, using the Laplace transform table, we determine that the ramp response is 8 3 13 y(t) = 3e−2t − e−3t + e−4t + t − , 3 4 12 t≥0. (b) For the ramp input, y(t) ≈ 0.21 at t = 1. second (see Figure P2.36a). (c) Given R(s) = 1, the partial fraction expansion is Y (s) = 24 12 24 12 = − + . (s + 2)(s + 3)(s + 4) s+2 s+3 s+4 Therefore, using the Laplace transform table, we determine that the impulse response is y(t) = 12e−2t − 24e−3t + 412e−4t , t≥0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 53 Problems P2.28 The signal flow graph is shown in Figure P2.28. -d G B +b P +c D +a -m +e -k M +g +f +h S C FIGURE P2.28 Signal flow graph. (a) The PGBDP loop gain is equal to -abcd. This is a negative transmission since the population produces garbage which increases bacteria and leads to diseases, thus reducing the population. (b) The PMCP loop gain is equal to +efg. This is a positive transmission since the population leads to modernization which encourages immigration, thus increasing the population. (c) The PMSDP loop gain is equal to +ehkd. This is a positive transmission since the population leads to modernization and an increase in sanitation facilities which reduces diseases, thus reducing the rate of decreasing population. (d) The PMSBDP loop gain is equal to +ehmcd. This is a positive transmission by similar argument as in (3). P2.29 Assume the motor torque is proportional to the input current Tm = ki . Then, the equation of motion of the beam is J φ̈ = ki , where J is the moment of inertia of the beam and shaft (neglecting the inertia of the ball). We assume that forces acting on the ball are due to gravity and friction. Hence, the motion of the ball is described by mẍ = mgφ − bẋ © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 54 CHAPTER 2 Mathematical Models of Systems where m is the mass of the ball, b is the coefficient of friction, and we have assumed small angles, so that sin φ ≈ φ. Taking the Laplace transfor of both equations of motion and solving for X(s) yields X(s)/I(s) = P2.30 gk/J . + b/m) s2 (s2 Given H(s) = k τs + 1 where τ = 4µs = 4 × 10−6 seconds and 0.999 ≤ k < 1.001. The step response is Y (s) = k 1 k k · = − . τs + 1 s s s + 1/τ Taking the inverse Laplace transform yields y(t) = k − ke−t/τ = k(1 − e−t/τ ) . The final value is k. The time it takes to reach 98% of the final value is t = 15.6µs independent of k. P2.31 From the block diagram we have Y1 (s) = G2 (s)[G1 (s)E1 (s) + G3 (s)E2 (s)] = G2 (s)G1 (s)[R1 (s) − H1 (s)Y1 (s)] + G2 (s)G3 (s)E2 (s) . Therefore, Y1 (s) = G1 (s)G2 (s) G2 (s)G3 (s) R1 (s) + E2 (s) . 1 + G1 (s)G2 (s)H1 (s) 1 + G1 (s)G2 (s)H1 (s) And, computing E2 (s) (with R2 (s) = 0) we find G4 (s) E2 (s) = H2 (s)Y2 (s) = H2 (s)G6 (s) Y1 (s) + G5 (s)E2 (s) G2 (s) or E2 (s) = G4 (s)G6 (s)H2 (s) Y1 (s) . G2 (s)(1 − G5 (s)G6 (s)H2 (s)) Substituting E2 (s) into equation for Y1 (s) yields Y1 (s) = G1 (s)G2 (s) R1 (s) 1 + G1 (s)G2 (s)H1 (s) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 55 Problems + G3 (s)G4 (s)G6 (s)H2 (s) Y1 (s) . (1 + G1 (s)G2 (s)H1 (s))(1 − G5 (s)G6 (s)H2 (s)) Finally, solving for Y1 (s) yields Y1 (s) = T1 (s)R1 (s) where T1 (s) = G1 (s)G2 (s)(1 − G5 (s)G6 (s)H2 (s)) (1 + G1 (s)G2 (s)H1 (s))(1 − G5 (s)G6 (s)H2 (s)) − G3 (s)G4 (s)G6 (s)H2 (s) . . Similarly, for Y2 (s) we obtain Y2 (s) = T2 (s)R1 (s) . where T2 (s) = P2.32 G1 (s)G4 (s)G6 (s) (1 + G1 (s)G2 (s)H1 (s))(1 − G5 (s)G6 (s)H2 (s)) − G3 (s)G4 (s)G6 (s)H2 (s) The signal flow graph shows three loops: L1 = −G1 G3 G4 H2 L2 = −G2 G5 G6 H1 L3 = −H1 G8 G6 G2 G7 G4 H2 G1 . The transfer function Y2 /R1 is found to be Y2 (s) G1 G8 G6 ∆1 − G2 G5 G6 ∆2 = , R1 (s) 1 − (L1 + L2 + L3 ) + (L1 L2 ) where for path 1 ∆1 = 1 and for path 2 ∆ 2 = 1 − L1 . Since we want Y2 to be independent of R1 , we need Y2 /R1 = 0. Therefore, we require G1 G8 G6 − G2 G5 G6 (1 + G1 G3 G4 H2 ) = 0 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 56 CHAPTER 2 P2.33 Mathematical Models of Systems The closed-loop transfer function is G3 (s)G1 (s)(G2 (s) + K5 K6 ) Y (s) = . R(s) 1 − G3 (s)(H1 (s) + K6 ) + G3 (s)G1 (s)(G2 (s) + K5 K6 )(H2 (s) + K4 ) P2.34 The equations of motion are m1 ÿ1 + b(ẏ1 − ẏ2 ) + k1 (y1 − y2 ) = 0 m2 ÿ2 + b(ẏ2 − ẏ1 ) + k1 (y2 − y1 ) + k2 y2 = k2 x Taking the Laplace transform yields (m1 s2 + bs + k1 )Y1 (s) − (bs + k1 )Y2 (s) = 0 (m2 s2 + bs + k1 + k2 )Y2 (s) − (bs + k1 )Y1 (s) = k2 X(s) Therefore, after solving for Y1 (s)/X(s), we have Y2 (s) k2 (bs + k1 ) = . 2 X(s) (m1 s + bs + k1 )(m2 s2 + bs + k1 + k2 ) − (bs + k1 )2 P2.35 (a) We can redraw the block diagram as shown in Figure P2.35. Then, T (s) = K1 /s(s + 1) K1 = 2 . 1 + K1 (1 + K2 s)/s(s + 1) s + (1 + K2 K1 )s + K2 (b) The signal flow graph reveals two loops (both touching): L1 = −K1 s(s + 1) and L2 = −K1 K2 . s+1 Therefore, T (s) = K1 /s(s + 1) K1 = 2 . 1 + K1 /s(s + 1) + K1 K2 /(s + 1) s + (1 + K2 K1 )s + K1 (c) We want to choose K1 and K2 such that s2 + (1 + K2 K1 )s + K1 = s2 + 20s + 100 = (s + 10)2 . Therefore, K1 = 100 and 1 + K2 K1 = 20 or K2 = 0.19. (d) The step response is shown in Figure P2.35. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 58 CHAPTER 2 Mathematical Models of Systems (d) For the impulse input, y(t) ≈ 0.65 at t = 1 seconds (see Figure P2.36b). (a) Ramp input (b) Impulse input 0.8 2 1.8 0.7 1.6 0.6 1.4 0.5 y(t) y(t) 1.2 1 0.8 0.4 0.3 0.6 0.2 0.4 0.1 0.2 0 0 1 2 0 3 0 1 Time (sec) 2 3 Time (sec) FIGURE P2.36 (a) Ramp input response. (b) Impulse input response. P2.37 The equations of motion are m1 d2 x = −(k1 + k2 )x + k2 y dt2 and m2 d2 y = k2 (x − y) + u . dt2 When m1 = m2 = 1 and k1 = k2 = 1, we have d2 x = −2x + y dt2 P2.38 and d2 y =x−y+u . dt2 The equation of motion for the system is J d2 θ dθ + b + kθ = 0 , dt2 dt where k is the rotational spring constant and b is the viscous friction coefficient. The initial conditions are θ(0) = θo and θ̇(0) = 0. Taking the © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 59 Problems Laplace transform yields J(s2 θ(s) − sθo ) + b(sθ(s) − θo ) + kθ(s) = 0 . Therefore, θ(s) = (s + Jb θo ) (s2 + b Js + K J) = (s + 2ζωn )θo . + 2ζωn s + ωn2 s2 Neglecting the mass of the rod, the moment of inertia is detemined to be J = 2M r 2 = 0.5 kg · m2 . Also, s ωn = k = 0.02 rad/s J and ζ = b = 0.01 . 2Jωn Solving for θ(t), we find that q θo θ(t) = p e−ζωn t sin(ωn 1 − ζ 2 t + φ) , 1 − ζ2 where tan φ = p 1 − ζ 2 /ζ). Therefore, the envelope decay is θo θe = p e−ζωn t . 1 − ζ2 So, with ζωn = 2 × 10−4 , θo = 4000o and θf = 10o , the elapsed time is computed as t= P2.39 1 θo ln p = 8.32 hours . ζωn 1 − ζ 2 θf When t < 0, we have the steady-state conditions i1 (0) = 1A , va (0) = 2V and vc (0) = 5V , where vc (0) is associated with the 1F capacitor. After t ≥ 0, we have 2 di1 + 2i1 + 4(i1 − i2 ) = 10e−2t dt and Z i2 dt + 10i2 + 4(i2 − i1 ) − i1 = 0 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 60 CHAPTER 2 Mathematical Models of Systems Taking the Laplace transform (using the initial conditions) yields 2(sI1 − i1 (0)) + 2I1 + 4I1 − 4I2 = 10 s+2 or (s + 3)I1 (s) − 2I2 (s) = s+7 s+2 and 1 [ I2 −vc (0)]+10I2 +4(I2 −I1 ) = I1 (s) or s −5sI1 (s)+(14s+1)I2 (s) = 5s . Solving for I2 (s) yields I2 = 5s(s2 + 6s + 13) , 14(s + 2)∆(s) where ∆(s) = s+3 −2 −5s 14s + 1 = 14s2 + 33s + 3 . Then, Vo (s) = 10I2 (s) . P2.40 The equations of motion are J1 θ̈1 = K(θ2 − θ1 ) − b(θ̇1 − θ̇2 ) + T and J2 θ̈2 = b(θ̇1 − θ̇2 ) . Taking the Laplace transform yields (J1 s2 + bs + K)θ1 (s) − bsθ2 (s) = Kθ2 (s) + T (s) and (J2 s2 + bs)θ2 (s) − bsθ1 (s) = 0 . Solving for θ1 (s) and θ2 (s), we find that θ1 (s) = (Kθ2 (s) + T (s))(J2 s + b) ∆(s) and θ2 (s) = b(Kθ2 (s) + T (s)) , ∆(s) where ∆(s) = J1 J2 s3 + b(J1 + J2 )s2 + J2 Ks + bK . P2.41 Assume that the only external torques acting on the rocket are control torques, Tc and disturbance torques, Td , and assume small angles, θ(t). Using the small angle approximation, we have ḣ = V θ © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 61 Problems J θ̈ = Tc + Td , where J is the moment of inertia of the rocket and V is the rocket velocity (assumed constant). Now, suppose that the control torque is proportional to the lateral displacement, as Tc (s) = −KH(s) , where the negative sign denotes a negative feedback system. The corresponding block diagram is shown in Figure P2.41. Td H desired=0 K + Tc + + - 1 Js 2 V s H( s) FIGURE P2.41 Block diagram. P2.42 (a) The equation of motion of the motor is J dω = Tm − bω , dt where J = 0.1, b = 0.06, and Tm is the motor input torque. (b) Given Tm (s) = 1/s, and ω(0) = 0.7, we take the Laplace transform of the equation of motion yielding sω(s) − ω(0) + 0.6ω(s) = 10Tm or ω(s) = 0.7s + 10 . s(s + 0.6) Then, computing the partial fraction expansion, we find that ω(s) = A B 16.67 15.97 + = − . s s + 0.6 s s + 0.6 The step response, determined by taking the inverse Laplace transform, is ω(t) = 16.67 − 15.97e−0.6t , t≥0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 62 CHAPTER 2 P2.43 Mathematical Models of Systems The work done by each gear is equal to that of the other, therefore Tm θm = TL θL . Also, the travel distance is the same for each gear, so r1 θ m = r2 θ L . The number of teeth on each gear is proportional to the radius, or r1 N 2 = r2 N 1 . So, θm r2 N2 = = , θL r1 N1 and N1 θ m = N2 θ L N1 θL = θm = nθm , N2 where n = N1 /N2 . Finally, Tm θL N1 = = =n. TL θm N2 P2.44 The inertia of the load is JL = πρLr 4 . 2 Also, from the dynamics we have T2 = JL ω̇2 + bL ω2 and T1 = nT2 = n(JL ω̇2 + bL ω2 ) . So, T1 = n2 (JL ω̇1 + bL ω1 ) , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 63 Problems since ω2 = nω1 . Therefore, the torque at the motor shaft is T = T1 + Tm = n2 (JL ω̇1 + bL ω1 ) + Jm ω̇1 + bm ω1 . P2.45 Let U (s) denote the human input and F (s) the load input. The transfer function is P (s) = G(s) + KG1 (s) Gc (s) + KG1 (s) U (s) + F (s) , ∆(s) ∆(s) where ∆ = 1 + GH(s) + G1 KBH(s) + Gc E(s) + G1 KE(s) . P2.46 Consider the application of Newton’s law ( mv we obtain P F = mẍ). From the mass mv ẍ1 = F − k1 (x1 − x2 ) − b1 (ẋ1 − ẋ2 ). Taking the Laplace transform, and solving for X1 (s) yields X1 (s) = b1 s + k1 1 F (s) + X2 (s), ∆1 (s) ∆1 (s) where ∆1 := mv s2 + b1 s + k1 . From the mass mt we obtain mt ẍ2 = −k2 x2 − b2 ẋ2 + k1 (x1 − x2 ) + b1 (ẋ1 − ẋ2 ). Taking the Laplace transform, and solving for X2 (s) yields X2 (s) = b1 s + k1 X1 (s), ∆2 (s) where ∆2 := mt s2 + (b1 + b2 )s + k1 + k2 . Substituting X2 (s) above into the relationship fpr X1 (s) yields the transfer function ∆2 (s) X1 (s) = . F (s) ∆1 (s)∆2 (s) − (b1 s + k1 )2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 64 CHAPTER 2 P2.47 Mathematical Models of Systems Using the following relationships h(t) = Z (1.6θ(t) − h(t))dt ω(t) = θ̇(t) J ω̇(t) = Km ia (t) va (t) = 50vi (t) = 10ia (t) + vb (t) θ̇ = Kvb we find the differential equation is d3 h Km + 1+ 3 dt 10JK P2.48 Km dh 8Km d2 h + = vi . 2 dt 10JK dt J (a) The transfer function is V2 (s) (1 + sR1 C1 )(1 + sR2 C2 ) = . V1 (s) R1 C 2 s (b) When R1 = 100 kΩ, R2 = 200 kΩ, C1 = 1 µF and C2 = 0.1 µF , we have V2 (s) 0.2(s + 10)(s + 50) = . V1 (s) s P2.49 (a) The closed-loop transfer function is T (s) = G(s) 6205 = 3 . 1 + G(s) s + 13s2 + 1281s + 6205 (b) The poles of T (s) are s1 = −5 and s2,3 = −4 ± j35. (c) The partial fraction expansion (with a step input) is Y (s) = 1 − 1.0122 0.0061 + 0.0716j 0.0061 − 0.0716j + + . s+5 s + 4 + j35 s + 4 − j35 (d) The step response is shown in Figure P2.49. The real and complex roots are close together and by looking at the poles in the s-plane we have difficulty deciding which is dominant. However, the residue at the real pole is much larger and thus dominates the response. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 65 Problems 1 0.9 0.8 Amplitude 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Time (secs) FIGURE P2.49 Step response. P2.50 (a) The closed-loop transfer function is T (s) = s3 + 45s2 14000 . + 3100s + 14500 (b) The poles of T (s) are s1 = −5 and s2,3 = −20 ± j50. (c) The partial fraction expansion (with a step input) is Y (s) = 0.9655 1.0275 0.0310 − 0.0390j 0.0310 + 0.0390j − + + . s s+5 s + 20 + j50 s + 20 − j50 (d) The step response is shown in Figure P2.50. The real root dominates the response. (e) The final value of y(t) is yss = lim sY (s) = 0.9655 . s→0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 2 Mathematical Models of Systems 1 0.9 0.8 0.7 Amplitude 66 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Time (secs) FIGURE P2.50 Step response. P2.51 Consider the free body diagram in Figure P2.51. Using Newton’s Law and summing the forces on the two masses yields M1 ẍ(t) + b1 ẋ(t) + k1 x(t) = b1 ẏ(t) M2 ÿ(t) + b1 ẏ(t) + k2 y(t) = b1 ẋ(t) + u(t) k1x M1 k1 x . . b1(x - y) k2 M1 x . . b1(y - x) k2 y b1 M2 M2 y y u(t) FIGURE P2.51 Free body diagram. u(t) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 67 Advanced Problems Advanced Problems AP2.1 The transfer function from V (s) to ω(s) has the form ω(s) Km = . V (s) τm s + 1 In the steady-state, ωss = lim s s→0 Km 5 = 5Km . τm s + 1 s So, Km = 70/5 = 14 . Also, ω(t) = Vm Km (1 − e−t/τm ) where V (s) = Vm /s. Solving for τm yields τm = −t . ln(1 − ω(t)/ωss ) When t = 2, we have τm = −2 = 3.57 . ln(1 − 30/70) Therefore, the transfer function is ω(s) 14 = . V (s) 3.57s + 1 AP2.2 The closed-loop transfer function form R1 (s) to Y2 (s) is Y2 (s) G1 G4 G5 (s) + G1 G2 G3 G4 G6 (s) = R1 (s) ∆ where ∆ = [1 + G3 G4 H2 (s)][1 + G1 G2 H3 (s)] . If we select G5 (s) = −G2 G3 G6 (s) then the numerator is zero, and Y2 (s)/R1 (s) = 0. The system is now decoupled. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 68 CHAPTER 2 AP2.3 Mathematical Models of Systems (a) Computing the closed-loop transfer function: G(s)Gc (s) R(s) . Y (s) = 1 + Gc (s)G(s)H(s) Then, with E(s) = R(s) − Y (s) we obtain E(s) = 1 + Gc (s)G(s)(H(s) − 1) R(s) . 1 + Gc (s)G(s)H(s) If we require that E(s) ≡ 0 for any input, we need 1 + Gc (s)G(s)(H(s) − 1) = 0 or H(s) = Gc (s)G(s) − 1 n(s) = . Gc (s)G(s) d(s) Since we require H(s) to be a causal system, the order of the numerator polynomial, n(s), must be less than or equal to the order of the denominator polynomial, d(s). This will be true, in general, only if both Gc (s) and G(s) are proper rational functions (that is, the numerator and denominator polynomials have the same order). Therefore, making E ≡ 0 for any input R(s) is possible only in certain circumstances. (b) The transfer function from Td (s) to Y (s) is Gd (s)G(s) Y (s) = Td (s) . 1 + Gc (s)G(s)H(s) With H(s) as in part (a) we have Gd (s) Y (s) = Td (s) . Gc (s) (c) No. Since Y (s) = Gd (s)G(s) Td (s) = T (s)Td (s) , 1 + Gc (s)G(s)H(s) the only way to have Y (s) ≡ 0 for any Td (s) is for the transfer function T (s) ≡ 0 which is not possible in general (since G(s) 6= 0). AP2.4 (a) With q(s) = 1/s we obtain τ (s) = 1/Ct s+ QS+1/R Ct · 1 . s Define α := QS + 1/R Ct and β := 1/Ct . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 69 Advanced Problems Then, it follows that τ (s) = β 1 −β/α β/α · = + . s+α s s+α s Taking the inverse Laplace transform yields τ (t) = β −β −αt β e + = [1 − e−αt ] . α α α 1 (b) As t → ∞, τ (t) → αβ = Qs+1/R . (c) To increase the speed of response, you want to choose Ct , Q, S and R such that α := Qs + 1/R Ct is ”large.” AP2.5 Considering the motion of each mass, we have M3 ẍ3 + b3 ẋ3 + k3 x3 = u3 + b3 ẋ2 + k3 x2 M2 ẍ2 + (b2 + b3 )ẋ2 + (k2 + k3 )x2 = u2 + b3 ẋ3 + k3 x3 + b2 ẋ1 + k2 x1 M1 ẍ1 + (b1 + b2 )ẋ1 + (k1 + k2 )x1 = u1 + b2 ẋ2 + k2 x2 In matrix form the three equations can be written as 0 M1 0 M2 AP2.6 0 0 0 0 0 ẍ1 b1 + b2 −b2 0 0 b2 + b3 −b3 ẍ2 + −b2 M3 ẍ3 −b3 b3 −k2 0 k1 + k2 + k2 + k3 −k3 −k2 −k3 k3 ẋ1 ẋ 2 ẋ3 x 1 u1 x = u . 2 2 x3 u3 Considering the cart mass and using Newton’s Law we obtain M ẍ = u − bẋ − F sin ϕ where F is the reaction force between the cart and the pendulum. Considering the pendulum we obtain m d2 (x + L sin ϕ) = F sin ϕ dt2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 70 CHAPTER 2 Mathematical Models of Systems m d2 (L cos ϕ) = F cos ϕ + mg dt2 Eliminating the reaction force F yields the two equations (m + M )ẍ + bẋ + mLϕ̈ cos ϕ − mLϕ̇2 sin ϕ = u mL2 ϕ̈ + mgL sin ϕ + mLẍ cos ϕ = 0 If we assume that the angle ϕ ≈ 0, then we have the linear model (m + M )ẍ + bẋ + mLϕ̈ = u mL2 ϕ̈ + mgLϕ = −mLẍ AP2.7 The transfer function from the disturbance input to the output is Y (s) = 1 Td (s) . s + 20 + K When Td (s) = 1, we obtain y(t) = e−(20+K)t . Solving for t when y(t) < 0.1 yields t> 2.3 . 20 + K When t = 0.05 and y(0.05) = 0.1, we find K = 26.05. AP2.8 The closed-loop transfer function is T (s) = 200K(0.25s + 1) (0.25s + 1)(s + 1)(s + 8) + 200K The final value due to a step input of R(s) = A/s is v(t) → A 200K . 200K + 8 We need to select K so that v(t) → 50. However, to keep the percent overshoot to less than 10%, we need to limit the magnitude of K. Figure AP2.8a shows the percent overshoot as a function of K. Let K = 0.06 and select the magnitude of the input to be A = 83.3. The inverse Laplace transform of the closed-loop response with R(s) = 83.3/s is v(t) = 50 + 9.85e−9.15t − e−1.93t (59.85 cos(2.24t) + 11.27 sin(2.24t)) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 71 Advanced Problems The result is P.O. = 9.74% and the steady-state value of the output is approximately 50 m/s, as shown in Figure AP2.8b. 25 Percent Overshoot (%) 20 15 10 5 0 0 0.01 0.02 0.03 0.04 0.05 K 0.06 0.07 0.08 0.09 0.1 Step Response 60 System: untitled1 Peak amplitude: 54.9 Overshoot (%): 9.74 At time (sec): 1.15 50 Amplitude 40 30 20 10 0 0 0.5 1 1.5 Time (sec) FIGURE AP2.8 (a) Percent overshoot versus the gain K. (b) Step response. AP2.9 The transfer function is Vo (s) Z2 (s) =− , Vi (s) Z1 (s) 2 2.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 72 CHAPTER 2 Mathematical Models of Systems where Z1 (s) = R1 R1 C 1 s + 1 and Z2 (s) = R2 C 2 s + 1 . C2 s Then we can write KI Vo (s) = Kp + + KD s Vi (s) s where KP = − R1 C 1 +1 , R2 C 2 KI = − 1 , R1 C 2 KD = −R2 C1 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 73 Design Problems Design Problems CDP2.1 The model of the traction drive, capstan roller, and linear slide follows closely the armature-controlled dc motor model depicted in Figure 2.18 in Dorf and Bishop. The transfer function is T (s) = rKm , s [(Lm s + Rm )(JT s + bm ) + Kb Km ] where JT = Jm + r 2 (Ms + Mb ) . Va(s) - 1 JTs+bm Km Lms+Rm 1 s q r X(s) Kb Back EMF DP2.1 w The closed-loop transfer function is Y (s) G1 (s)G2 (s) = . R(s) 1 + G1 (s)H1 (s) − G2 (s)H2 (s) When G1 H1 = G2 H2 and G1 G2 = 1, then Y (s)/R(s) = 1. Therefore, select G1 (s) = DP2.2 1 G2 (s) and H1 (s) = G2 (s)H2 (s) = G22 (s)H2 (s) . G1 (s) At the lower node we have 1 1 v + + G + 2i2 − 20 = 0 . 4 3 Also, we have v = 24 and i2 = Gv . So 1 1 v + + G + 2Gv − 20 = 0 4 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 74 CHAPTER 2 Mathematical Models of Systems and G= DP2.3 20 − v 1 4 + 1 3 3v = 1 S. 12 Taking the Laplace transform of 3 1 1 y(t) = e−t − e−2t − + t 4 4 2 yields Y (s) = 1 1 3 1 − − + 2 . s + 1 4(s + 2) 4s 2s Similarly, taking the Laplace transform of the ramp input yields R(s) = 1 . s2 Therefore G(s) = DP2.4 Y (s) 1 = . R(s) (s + 1)(s + 2) For an ideal op-amp, at node a we have vin − va vo − va + =0, R1 R1 and at node b vin − vb = C v̇b , R2 from it follows that 1 1 + Cs Vb = Vin . R2 R2 Also, for an ideal op-amp, Vb − Va = 0. Then solving for Vb in the above equation and substituting the result into the node a equation for Va yields Vo = Vin 1 R2 " 2 1 − R + Cs 2 1 R2 + Cs 2 or Vo (s) R2 Cs − 1 =− . Vin (s) R2 Cs + 1 # © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 75 Design Problems For vin (t) = At, we have Vin (s) = A/s2 , therefore 2 2 vo (t) = A e−βt + t − β β where β = 1/R2 C. DP2.5 The equation of motion describing the motion of the inverted pendulum (assuming small angles) is ϕ̈ + g ϕ=0. L Assuming a solution of the form ϕ = k cos ϕ, taking the appropriate derivatives and substituting the result into the equation of motion yields the relationship ϕ̇ = r g . L If the period is T = 2 seconds, we compute ϕ̇ = 2π/T . Then solving for L yields L = 0.99 meters when g = 9.81 m/s2 . So, to fit the pendulum into the grandfather clock, the dimensions are generally about 1.5 meters or more. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 76 CHAPTER 2 Mathematical Models of Systems Computer Problems CP2.1 The m-file script is shown in Figure CP2.1. pq = 1 9 P= -5 -2 Z= -2 value = 4 p=[1 7 10]; q=[1 2]; % Part (a) pq=conv(p,q) % Part (b) P=roots(p), Z=roots(q) % Part (c) value=polyval(p,-1) 24 20 FIGURE CP2.1 Script for various polynomial evaluations. The m-file script and step response is shown in Figure CP2.2. numc = [1]; denc = [1 1]; sysc = tf(numc,denc) numg = [1 2]; deng = [1 3]; sysg = tf(numg,deng) % part (a) sys_s = series(sysc,sysg); sys_cl = feedback(sys_s,[1]) % part (b) step(sys_cl); grid on Transfer function: s+2 ------------s^2 + 5 s + 5 Step Response From: U(1) 0.4 0.35 0.3 To: Y(1) 0.25 Amplitude CP2.2 0.2 0.15 0.1 0.05 0 0 0.5 1 1.5 2 Time (sec.) FIGURE CP2.2 Step response. 2.5 3 3.5 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 77 Computer Problems Given ÿ + 4ẏ + 3y = u with y(0) = ẏ = 0 and U (s) = 1/s, we obtain (via Laplace transform) Y (s) = s(s2 1 1 = . + 4s + 3) s(s + 3)(s + 1) Expanding in a partial fraction expansion yields Y (s) = 1 1 1 − − . 3s 6(s + 3) 2(s + 1) Taking the inverse Laplace transform we obtain the solution y(t) = 0.3333 + 0.1667e−3t − 0.5e−t . The m-file script and step response is shown in Figure CP2.3. Step Response 0.35 0.3 0.25 Amplitude CP2.3 n=[1]; d=[1 4 3]; sys = tf(n,d); t=[0:0.1:5]; y = step(sys,t); ya=0.3333+0.1667*exp(-3*t)-0.5*exp(-t); plot(t,y,t,ya); grid; title('Step Response'); xlabel('Time (sec)'); ylabel('Amplitude'); 0.2 0.15 0.1 0.05 0 0 FIGURE CP2.3 Step response. 0.5 1 1.5 2 2.5 3 Time (sec) 3.5 4 4.5 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 78 CHAPTER 2 CP2.4 Mathematical Models of Systems The mass-spring-damper system is represented by mẍ + bẋ + kx = f . Taking the Laplace transform (with zero initial conditions) yields the transfer function X(s)/F (s) = s2 1/m . + bs/m + k/m The m-file script and step response is shown in Figure CP2.4. m=10; k=1; b=0.5; num=[1/m]; den=[1 b/m k/m]; sys = tf(num,den); t=[0:0.1:150]; step(sys,t) Step Response From: U(1) 1.8 1.6 1.4 1 To: Y(1) Amplitude 1.2 0.8 0.6 0.4 0.2 0 0 50 100 150 Time (sec.) FIGURE CP2.4 Step response. CP2.5 The spacecraft simulations are shown in Figure CP2.5. We see that as J is decreased, the time to settle down decreases. Also, the overhoot from 10o decreases as J decreases. Thus, the performance seems to get better (in some sense) as J decreases. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 79 Computer Problems Nominal (solid); Off-nominal 80% (dashed); Off-nominal 50% (dotted) 18 16 Spacecraft attitude (deg) 14 12 10 8 6 4 2 0 0 10 20 30 40 50 60 70 80 90 100 Time (sec) %Part (a) a=1; b=8; k=10.8e+08; J=10.8e+08; num=k*[1 a]; den=J*[1 b 0 0]; sys=tf(num,den); sys_cl=feedback(sys,[1]); % % Part (b) and (c) t=[0:0.1:100]; % % Nominal case f=10*pi/180; sysf=sys_cl*f ; y=step(sysf,t); % % Off-nominal case 80% J=10.8e+08*0.8; den=J*[1 b 0 0]; sys=tf(num,den); sys_cl=feedback(sys,[1]); sysf=sys_cl*f ; y1=step(sysf,t); % % Off-nominal case 50% J=10.8e+08*0.5; den=J*[1 b 0 0]; sys=tf(num,den); sys_cl=feedback(sys,[1]); sysf=sys_cl*f ; y2=step(sysf,t); % plot(t,y*180/pi,t,y1*180/pi,'--',t,y2*180/pi,':'),grid xlabel('Time (sec)') ylabel('Spacecraft attitude (deg)') title('Nominal (solid); Off-nominal 80% (dashed); Off-nominal 50% (dotted)') FIGURE CP2.5 Step responses for the nominal and off-nominal spacecraft parameters. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 80 CHAPTER 2 CP2.6 Mathematical Models of Systems The closed-loop transfer function is T (s) = 4s6 + 8s5 + 4s4 + 56s3 + 112s2 + 56s , ∆(s) p= 7.0709 -7.0713 1.2051 + 2.0863i 1.2051 - 2.0863i 0.1219 + 1.8374i 0.1219 - 1.8374i -2.3933 -2.3333 -0.4635 + 0.1997i -0.4635 - 0.1997i num1=[4]; den1=[1]; sys1 = tf(num1,den1); num2=[1]; den2=[1 1]; sys2 = tf(num2,den2); num3=[1 0]; den3=[1 0 2]; sys3 = tf(num3,den3); num4=[1]; den4=[1 0 0]; sys4 = tf(num4,den4); num5=[4 2]; den5=[1 2 1]; sys5 = tf(num5,den5); num6=[50]; den6=[1]; sys6 = tf(num6,den6); num7=[1 0 2]; den7=[1 0 0 14]; sys7 = tf(num7,den7); sysa = feedback(sys4,sys6,+1); sysb = series(sys2,sys3); sysc = feedback(sysb,sys5); sysd = series(sysc,sysa); syse = feedback(sysd,sys7); sys = series(sys1,syse) poles % pzmap(sys) % p=pole(sys) z=zero(sys) z= 0 1.2051 + 2.0872i 1.2051 - 2.0872i -2.4101 -1.0000 + 0.0000i -1.0000 - 0.0000i Polezero map 2.5 2 1.5 1 Imag Axis 0.5 0 -0.5 -1 -1.5 -2 -2.5 -8 -6 -4 -2 0 Real Axis FIGURE CP2.6 Pole-zero map. 2 4 6 8 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 81 Computer Problems where ∆(s) = s10 + 3s9 − 45s8 − 125s7 − 200s6 − 1177s5 − 2344s4 − 3485s3 − 7668s2 − 5598s − 1400 . CP2.7 The m-file script and plot of the pendulum angle is shown in Figure CP2.7. With the initial conditions, the Laplace transform of the linear system is θ(s) = s2 θ0 s . + g/L To use the step function with the m-file, we can multiply the transfer function as follows: θ(s) = s2 θ0 , 2 s + g/L s which is equivalent to the original transfer function except that we can use the step function input with magnitude θ0 . The nonlinear response is shown as the solid line and the linear response is shown as the dashed line. The difference between the two responses is not great since the initial condition of θ0 = 30◦ is not that large. 30 L=0.5; m=1; g=9.8; theta0=30; % Linear simulation sys=tf([1 0 0],[1 0 g/L]); [y,t]=step(theta0*sys,[0:0.01:10]); % Nonlinear simulation [t,ynl]=ode45(@pend,t,[theta0*pi/180 0]); plot(t,ynl(:,1)*180/pi,t,y,'--'); xlabel('Time (s)') ylabel('\theta (deg)') 20 θ (deg) 10 0 -10 function [yd]=pend(t,y) L=0.5; g=9.8; yd(1)=y(2); yd(2)=-(g/L)*sin(y(1)); yd=yd'; -20 -30 0 2 4 6 Time (s) FIGURE CP2.7 Plot of θ versus xt when θ0 = 30◦ . 8 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 82 CHAPTER 2 CP2.8 Mathematical Models of Systems The system step responses for z = 5, 10, and 15 are shown in Figure CP2.8. z=5 (solid), z=10 (dashed), z=15 dotted) 1.5 x(t) 1 0.5 0 0 0.5 1 1.5 2 2.5 3 Time (sec) 3.5 4 4.5 5 FIGURE CP2.8 The system response. CP2.9 (a,b) Computing the closed-loop transfer function yields T (s) = G(s) s2 + 2s + 1 = 2 . 1 + G(s)H(s) s + 4s + 3 The poles are s = −3, −1 and the zeros are s = −1, −1. (c) Yes, there is one pole-zero cancellation. The transfer function (after pole-zero cancellation) is T (s) = s+1 . s+3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 83 Computer Problems Pole?Zero Map 1 0.8 0.6 0.4 Imaginary Axi s 0.2 0 ?-0.2 ?-0.4 ?-0.6 ?-0.8 ?-1 ?-3 ?-2.5 ?-2 ?-1.5 ?-1 ?-0.5 0 Real Axi s ng=[1 1]; dg=[1 2]; sysg = tf(ng,dg); nh=[1]; dh=[1 1]; sysh = tf(nh,dh); sys=feedback(sysg,sysh) % pzmap(sys) % pole(sys) zero(sys) >> Transfer function: s^2 + 2 s + 1 ------------s^2 + 4 s + 3 poles p= -3 -1 zeros z= -1 -1 FIGURE CP2.9 Pole-zero map. CP2.10 Figure CP2.10 shows the steady-state response to a unit step input and a unit step disturbance. We see that K = 1 leads to the same steady-state response. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 2 Mathematical Models of Systems 0.35 K=[0.1:0.1:10]; sysg=tf([1],[1 20 20]); for i=1:length(K) nc=K(i); dc=[1];sysc=tf(nc,dc); syscl=feedback(sysc*sysg,1); systd=feedback(sysg,sysc); y1=step(syscl); Tf1(i)=y1(end); y2=step(systd); Tf2(i)=y2(end); end plot(K,Tf1,K,Tf2,'--') xlabel('K') ylabel('Steady-state response') 0.3 0.25 Steady−state response 84 0.2 0.15 0.1 Disturbance Response Steady-State 0.05 K=1 0 FIGURE CP2.10 Gain K versus steady-state value. Input Response Steady-State 0 1 2 3 4 5 K 6 7 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 3 State Variable Models Exercises E3.1 One possible set of state variables is (a) the current iL2 through L2 , (b) the voltage vC2 across C2 , and (c) the current iL1 through L1 . We can also choose vC1 , the voltage across C1 as the third state variable, in place of the current through L1 . E3.2 We know that the velocity is the derivative of the position, therefore we have dy =v , dt and from the problem statement dv = −k1 v(t) − k2 y(t) + k3 i(t) . dt This can be written in matrix form as 0 1 y 0 d y + i . = dt v −k2 −k1 v k3 Define u = i, and let k1 = k2 = 1. Then, ẋ = Ax + Bu where A= 0 1 −1 −1 , B= 0 k3 , and x= y v . 85 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 86 CHAPTER 3 E3.3 State Variable Models The charactersitic roots, denoted by λ, are the solutions of det(λI − A) = 0. For this problem we have λ det(λI − A) = det −1 1 λ+2 = λ(λ + 2) + 1 = λ2 + 2λ + 1 = 0 . Therefore, the characteristic roots are λ1 = −1 and λ2 = −1 . E3.4 The system in phase variable form is ẋ = Ax + Bu y = Cx where E3.5 A= 0 0 1 0 1 , 0 −8 −6 −4 0 B= 0 , C= 20 h 1 0 0 i . From the block diagram we determine that the state equations are ẋ2 = −(f k + d)x2 + ax1 + f u ẋ1 = −kx2 + u and the output equation is y = bx2 . Therefore, ẋ = Ax + Bu y = Cx + Du , where A= E3.6 0 −k a −(f k + d) , B= 1 f , C= h 0 b i (a) The state transition matrix is Φ(t) = eAt = I + At + 1 2 2 A t + ··· 2! and D = [0] . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 87 Exercises But A2 = 0, thus A3 = A4 = · · · = 0. So, Φ(t) = eAt = I + At = 1 0 0 1 + 0 1 0 0 t = 1 t 0 1 . (b) The state at any time t ≥ 0 is given by x(t) = Φ(t)x(0) and since x1 (0) = x2 (0) = 1, we determine that x1 (t) = x1 (0) + tx2 (0) = 1 + t x2 (t) = x2 (0) = 1 . E3.7 The state equations are ẋ1 = x2 x˙2 = −100x1 − 20x2 + u or, in matrix form ẋ = 0 1 −100 −20 x + 0 1 u . So, the characteristic equation is determined to be det(λI − A) = det λ −1 100 λ + 20 = λ2 + 20λ + 100 = (λ + 10)2 = 0 . Thus, the roots of the characteristic equation are λ1 = λ2 = −10 . E3.8 The characteristic equation is λ −1 det(λI − A) = det 0 λ 0 6 0 −1 λ+3 = λ(λ2 + 3λ + 6) = 0 . Thus, the roots of the characteristic equation are λ1 = 0 , λ2 = −1.5 + j1.9365 and λ3 = −1.5 − j1.9365 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 88 CHAPTER 3 E3.9 State Variable Models Analyzing the block diagram yields 1 ẋ1 = −x1 + x2 + r 2 3 ẋ2 = x1 − x2 − r 2 3 y = x1 − x2 − r. 2 In state-variable form we have ẋ = −1 1 1 2 − 32 x+ 1 r , −1 The characteristic equation is h i 3 x + −1 r . y= 1 − 2 5 1 s2 + s + 1 = (s + 2)(s + ) = 0 . 2 2 E3.10 (a) The characteristic equation is det[λI − A] = det λ −6 = λ(λ+ 5)+ 6 = (λ+ 2)(λ+ 3) = 0 . 1 (λ + 5) So, the roots are λ1 = −2 and λ2 = −3. (b) We note that −1 Φ(s) = [sI − A] = s −6 1 s+5 −1 s+5 6 1 . = (s + 2)(s + 3) −1 s Taking the inverse Laplace transform yields the transition matrix E3.11 Φ(t) = 3e−2t − 2e−3t −e−2t + 6e−2t − 6e−3t e−3t −2e−2t + 3e−3t A state variable representation is . ẋ = Ax + Br y = Cx where A= 0 1 −12 −8 , B= 0 1 , C= h 12 4 i . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 89 Exercises The equation of motion is L di + Ri + vc = vin dt where vc = 1 C Z i dt . Unit step response 1.8 1.6 1.4 1.2 State response E3.12 x1: capacitor voltage 1 0.8 0.6 0.4 0.2 x2: inductor current 0 −0.2 0 0.05 0.1 0.15 0.2 Time(sec) 0.25 0.3 FIGURE E3.12 State variable time history for a unit step input. Selecting the state variables x1 = vc and x2 = i, we have 1 x2 C R 1 1 ẋ2 = − x2 − x1 + vin . L L L ẋ1 = This can be written in matrix form as ẋ = 0 1/C −1/L −R/L x + 0 1/L vin . 0.35 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 90 CHAPTER 3 State Variable Models When C = 0.001F , R = 4Ω, and L = 0.1H, we have ẋ = 0 1000 −10 −40 x+ 0 10 vin . The step response is shown in Figure E3.12. E3.13 (a) Select the state variables as x1 = y and x2 = ω. (b) The corresponding state equation is ẋ1 = −x1 − ax2 + 2u ẋ2 = bx1 − 4u or, in matrix form ẋ = −1 −a b 0 x + 2 −4 u and x = x1 x2 . (c) The characteristic equation is det[λI − A] = det λ+1 a −b λ = λ2 + λ + ab = 0 . So, the roots are 1 1√ 1 − 4ab . λ=− ± 2 2 E3.14 Assume that the mass decay is proportional to the mass present, so that Ṁ = −qM + Ku where q is the constant of proportionality. Select the state variable, x, to be the mass, M . Then, the state equation is ẋ = −qx + Ku . E3.15 The equations of motion are mẍ + kx + k1 (x − q) + bẋ = 0 mq̈ + kq + bq̇ + k1 (q − x) = 0 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 91 Exercises In state variable form we have 0 (k+k1 ) − m ẋ = 0 k1 m 1 0 0 b −m k1 m 0 0 0 1 0 1) − (k+k m b −m where x1 = x, x2 = ẋ, x3 = q and x4 = q̇. E3.16 x The governing equations of motion are m1 ẍ + k1 (x − q) + b1 (ẋ − q̇) = u(t) m2 q̈ + k2 q + b2 q̇ + b1 (q̇ − ẋ) + k1 (q − x) = 0 . Let x1 = x, x2 = ẋ, x3 = q and x4 = q̇. Then, 0 − k1 ẋ = m1 0 k1 m2 1 0 0 b1 −m 1 k1 m1 b1 m1 0 0 1 b1 m2 2) − (k1m+k 2 − (b1m+b2 2 ) E3.17 h 0 0 1 0 i x. At node 1 we have C1 v̇1 = va − v1 v2 − v1 + R1 R2 C2 v˙2 = vb − v2 v1 − v2 + . R3 R2 and at node 2 we have Let x 1 = v1 and x 2 = v2 . 0 1 m x + 1 u(t) . 0 Since the output is y(t) = q(t), then y= 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 92 CHAPTER 3 State Variable Models Then, in matrix form we have − E3.18 ẋ = 1 R1 C1 + 1 R2 C1 − R21C2 1 R2 C1 − 1 R3 C2 + 1 R2 C2 The governing equations of motion are x+ 1 R1 C1 0 0 1 R3 C2 va vb . di1 + v = va dt di2 L2 + v = vb dt dv . iL = i1 + i2 = C dt Ri1 + L1 Let x1 = i1 , x2 = i2 , x3 = v, u1 = va and u2 = vb . Then, ẋ = y= E3.19 h − LR1 0 0 0 1 C 1 C 0 0 1 First, compute the matrix − L11 − L12 1 L1 x+ 0 0 i 0 1 L2 0 0 x + [0] u . sI − A = s −1 3 s+4 u . Then, Φ(s) is Φ(s) = (sI − A)−1 where ∆(s) = s2 + 4s + 3, and G(s) = E3.20 h 10 0 i s+4 ∆(s) 3 − ∆(s) 1 s+4 1 = ∆(s) −3 s 1 ∆(s) s ∆(s) 0 1 = s2 10 . + 4s + 3 The linearized equation can be derived from the observation that sin θ ≈ θ when θ ≈ 0. In this case, the linearized equations are θ̈ + g k θ + θ̇ = 0 . L m © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 93 Exercises Let x1 = θ and x2 = θ̇. Then in state variable form we have ẋ = Ax y = Cx where A= E3.21 0 1 −g/L −k/m The transfer function is , C= h 1 0 i , G(s) = C [sI − A]−1 B + D = and s2 x(0) = θ(0) θ̇(0) . −1 . + 2s + 1 The unit step response is y(t) = −1 + e−t + te−t . E3.22 The transfer function is G(s) = s2 s−6 . − 7s + 6 The poles are at s1 = 1 and s2 = 6. The zero is at s = 6. So, we see that there is a pole-zero cancellation. We can write the system in state variable form as √ ẋ = x − 2u √ 2 y=− x 2 and the transfer function is G(s) = E3.23 1 . s−1 The system in state variable form can be represented by ẋ = Ax + Bu y = Cx + Du where A= 0 1 0 0 0 1 , −1 −3 −3 0 B= 0 , 1 C= h 0 1 −1 i , D= h 1 i . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 94 CHAPTER 3 State Variable Models + - U(s) + - -- 1 s x3 3 3 FIGURE E3.23 Block diagram. 1 s x2 1 s x1 + X(s) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 95 Problems Problems P3.1 The loop equation, derived from Kirchoff’s voltage law, is di 1 R 1 = v− i − vc dt L L L where vc = 1 C Z i dt . (a) Select the state variables as x1 = i and x2 = vc . (b) The corresponding state equations are R 1 1 v− x1 − x2 L L L 1 x1 . ẋ2 = C ẋ1 = (c) Let the input u = v. Then, in matrix form, we have ẋ = −R/L −1/L 1/C 0 x+ 1/L 0 u . -R/L 1/C v 1/s 1/L x1 1/s x2 -1/L FIGURE P3.1 Signal flow graph. P3.2 Let −2 −2R1 R2 , a22 = , (R1 + R2 )C (R1 + R2 )L 1 R2 = b12 = , b21 = −b22 = . (R1 + R2 )C (R1 + R2 )L a11 = b11 The corresponding block diagram is shown in Figure P3.2. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 96 CHAPTER 3 State Variable Models 2/(R1+R2)C v1 1/s 1/(R1+R2)C x1 R2 (a) 1/(R1+R2)C v2 1/s x2 - 2R1R2/(R1+R2)C a 11 v1 b11 x1 1/s b12 (b) b21 v2 1/s x2 b22 a 22 FIGURE P3.2 (a) Block diagram. (b) Signal flow graph. P3.3 Using Kirchoff’s voltage law around the outer loop, we have L diL − vc + v2 − v1 = 0 . dt Then, using Kirchoff’s current law at the node, we determine that C dvc = −iL + iR , dt where iR is the current through the resistor R. Considering the right loop we have iR R − v2 + vc = 0 or iR = − vc v2 + . R R © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 97 Problems Thus, dvc vc iL v2 =− − + dt RC C RC vc v1 v2 diL = + − . dt dt L L and In matrix form, the state equations are ẋ1 ẋ2 = 0 1/L −1/C −1/RC x1 x2 + 1/L −1/L 0 1/RC v1 v2 , where x1 = iL and x2 = vc . The signal flow graph is shown in Figure P3.3. v1 1/L -1/L -1/C 1/s v2 1/L x2 1/s x1 -1/RC 1/RC FIGURE P3.3 Signal flow graph. P3.4 (a) The block diagram model for phase variable form is shown in Figure P3.4a. The phase variable form is given by ẋ = y= h 0 1 0 0 −10 −6 −4 10 2 1 i 0 0 1 x + 0 r x. 1 (b) The block diagram in input feedforward form is shown in Figure P3.4b. The input feedforward form is given by ẋ = y= h −4 1 0 1 −6 0 1 x + 2 r(t) −10 0 0 1 0 0 i x. 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 98 CHAPTER 3 State Variable Models 1 2 x2 x3 R(s) 1 s + - -- 1 s x1 1 s + 10 + + Y(s) 4 6 10 (a) 1 2 R(s) 10 + . . x3 1 + + x2 s - 1 + s - + . x1 1 s Y(s) 4 6 10 (b) FIGURE P3.4 (a)Block diagram model for phase variable form. (b) Block diagram model for input feedforward form. P3.5 (a) The closed-loop transfer function is T (s) = s3 s+1 . + 4s2 − 11s + 1 (b) A matrix differential equation is ẋ = Ax + Bu y = Cx where A= 0 1 0 0 0 1 , −1 11 −4 0 B= 0 , 1 The block diagram is shown in Figure P3.5. C= h 1 1 0 i . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 99 Problems 1 R(s) - x3 1 s + -- x2 1 s x1 1 s + 1 + Y(s) 4 -11 1 FIGURE P3.5 Block diagram model. P3.6 The node equations are dv1 vi − v1 + iL − =0 dt 4000 dv2 v2 0.0005 − iL + − i3 = 0 dt 1000 diL 0.002 + v2 − v1 = 0 . dt 0.00025 Define the state variables x 1 = v1 x 2 = v2 x3 = iL . Then, ẋ = Ax + Bu where −1 A= 0 P3.7 0 −4000 −2 2000 , 500 −500 Given K = 1, we have KG(s) · 0 0 1 B= 0 2000 0 0 (s + 1)2 1 = . s s(s2 + 1) We then compute the closed-loop transfer function as T (s) = s2 + 2s + 1 s−1 + 2s−2 + s−3 = . 3s3 + 5s2 + 5s + 1 3 + 5s−1 + 5s−2 + s−3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 100 CHAPTER 3 State Variable Models A state variable model is ẋ = y= P3.8 h 0 1 0 0 −1/3 −5/3 −5/3 1 2 1 The state-space equations are i 0 0 1 x + 0 r x. 1/3 ẋ1 = x2 ku ẋ2 = −g x3 ẋ3 = u . This is a set of nonlinear equations. P3.9 (a) The closed-loop transfer function is T (s) = 10s−3 10 = , Js3 + (b + 10J)s2 + 10bs + 10K1 1 + 10.1s−1 + s−2 + 5s−3 where K1 = 0.5, J = 1, and b = 0.1. (b) A state-space model is ẋ = ω= h 0 1 0 0 1 0 0 0 0 1 x + 0 r −5 −1 −10.1 i x. 10 (c) The characteristic equation is s −1 det[sI − A] = det 0 s 5 1 0 −1 s + 10.1 = s3 + 10.1s2 + s + 5 = 0 . The roots of the characteristic equation are s1 = −10.05 and s2,3 = −0.0250 ± 0.7049j . All roots lie in the left hand-plane, therefore, the system is stable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 101 Problems P3.10 (a) From the signal flow diagram, we determine that a state-space model is given by −K1 ẋ = y= K2 x + −K1 −K2 y1 y2 = K1 −K2 K1 K2 1 0 r1 r2 x . 0 1 (b) The characteristic equation is det[sI − A] = s2 + (K2 + K1 )s + 2K1 K2 = 0 . (c) When K1 = K2 = 1, then A= −1 1 −1 −1 . The state transition matrix associated with A is o n Φ = L−1 [sI − A]−1 = e−t P3.11 cos t sin t − sin t cos t The state transition matrix is Φ(t) = (2t − 1)e−t 2te−t −2te−t (−2t + 1)e−t . So, when x1 (0) = x2 (0) = 10, we have x(t) = Φ(t)x(0) or x1 (t) = 10e−t x2 (t) = 10e−t P3.12 (a) A state variable representation is given by ẋ = 0 1 0 0 0 0 1 x + 0 r −48 −44 −12 1 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 102 CHAPTER 3 State Variable Models y = [40 8 0]x . (b) The state transition matrix is Φ(t) = . . Φ1 (t)..Φ2 (t)..Φ3 (t) , where e−6t − 3e−4t + 3e−2t −6t + 12e−4t − 6e−2t Φ1 (t) = −6e 36e−6t − 48e−4t + 12e−2t P3.13 Φ3 (t) = Φ2 (t) = 1 −6t − 41 e−4t + 18 e−2t 8e − 43 e−6t + e−4t − 14 e−2t 9 −6t − 4e−4t + 12 e−2t 2e (a) The RLC circuit state variable representation is ẋ = −10 −4 6 0 x+ 4 0 3 −6t − 2e−4t + 54 e−2t 4e − 29 e−6t + 8e−4t − 25 e−2t 27e−6t − 32e−4t + 5e−2t . u . The characteristic equation is s2 + 10s + 24 = 0 . All roots of the characteristic equation (that is, s1 = −4 and s2 = −6) are in the left half-plane; therefore the system is stable. (b) The state transition matrix is Φ(t) = 3e−6t − 2e−4t −3e−6t + 3e−4t 2e−6t + 2e−4t −2e−6t + 3e−4t . (c) Given x1 (0) = 0.1 , x2 (0) = 0 and e(t) = 0 , we have i(t) = x1 (t) = 0.3e−6t − 0.2e−4t vc (t) = x2 (t) = −0.3e−6t + 0.3e−4t . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 103 Problems (d) When x(0) = 0 and u(t) = E, we have x(t) = Z t Φ(t − τ )Bu(τ )dτ , 0 where Bu(t) = Integrating yields 4E 0 . x1 (t) = (−2e−6t + 2e−4t )E x2 (t) = (1 + 2e−6t − 3e−4t )E . P3.14 A state space representation is ẋ = Ax + Br , y = Cx where P3.15 A= 0 1 0 0 0 1 0 0 0 0 0 , 1 −50 −34 −10 −12 , 0 B= A state variable representation is ẋ = 0 1 0 0 0 0 C = [50 1 0 0] . 1 0 0 1 x+ 0 r −16 −31 −10 y = [56 14 0]x . The block diagram is shown in Figure P3.15. 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 104 CHAPTER 3 State Variable Models 14 R(s) 1 s + - -- x3 x2 1 s 1 s x1 56 + + Y(s) 10 31 16 FIGURE P3.15 Block diagram model. (a) The characteristic equation is x1 - solid; x2 - dotted; x3 - dashed 0.5 0 -0.5 Step response) P3.16 -1 -1.5 -2 -2.5 -3 -3.5 0 20 40 60 80 100 Time (s) FIGURE P3.16 Step response of magnitude 0.285◦ . s det(sI − A) = det 0.0071 0 −1 s + 0.111 −0.07 0 −0.12 s + 0.3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 105 Problems = s3 + 0.411s2 + 0.032s + 0.00213 = 0 . The roots are s1 = −0.3343 and s2,3 = −0.0383 ± 0.0700j . All the poles lie in the left half-plane, therefore, the system is stable. (b) The solution of the system to a step of magnitude 0.285◦ is given by x1 (t) = −2.66 − 0.11e−0.33t + e−0.038t (2.77 cos 0.07t + 0.99 sin 0.07t) x2 (t) = 0.037e−0.33t − e−0.038t (0.037 cos 0.07t + 0.23 sin 0.07t) x3 (t) = 0.069 − 0.075e−0.33t + e−0.038t (0.006 cos 0.07t − 0.06 sin 0.07t) P3.17 The transfer function is G(s) = C(sI − A)−1 B = P3.18 −4s + 12 . s3 − 14s2 + 37s + 20 Define the state variables as x1 = φ1 − φ2 ω1 x2 = ωo ω2 x3 = . ωo Then, the state equations of the robot are ẋ1 = ωo x2 − ωo x3 −J2 ωo x1 − ẋ2 = J1 + J2 J1 ωo ẋ3 = x2 + J1 + J2 b x2 + J1 b x2 − J2 b Km x3 + i J1 J1 ωo b x3 J2 or, in matrix form 0 1 ẋ = ωo a−1 a where a= J1 , (J1 + J2 ) b1 = −b1 b2 −b2 b , J1 ωo −1 0 b1 x + d i b2 = 0 b and J2 ωo d= Km . J1 ωo © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 106 CHAPTER 3 P3.19 State Variable Models The state equation is given by 0 ẋ = 1 −2 −3 x where x1 (0) = 1 and x2 (0) = −1. The state transition matrix is Φ(t) = −e−2t + 2e−t −e−2t + e−t 2e−2t − 2e−t 2e−2t − e−t . The system response is x1 (t) = −e−2t + 2e−t x1 (0) + −e−2t + e−t x2 (0) x2 (t) = 2e−2t − 2e−t x1 (0) + 2e−2t − e−t x2 (0) . The state response is shown in Figure P3.19. 1 0.8 x1 0.6 System response 0.4 0.2 0 -0.2 -0.4 x2 -0.6 -0.8 -1 0 1 2 3 Time (s) 4 5 6 FIGURE P3.19 Response with x1 (0) = 1 and x2 (0) = −1. P3.20 The state equation is given by ẋ = − 0.693 6.7 −1 0 − 0.693 9.2 x where x(0) = 0.3 × 1016 7× 1016 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 107 Problems The state transition matrix is Φ(t) = e−0.103433t 35.5786(e−0.103433t − 0 e−0.0753261t ) The system response is e−0.075326t . x1 (t) = e−0.103433t x1 (0) h i x2 (t) = 35.5786 e−0.103433t − e−0.0753261t x1 (0) + e−0.075326t x2 (0) . The state response is shown in Figure P3.20. 7 Nucleide densities in atoms per unit volume 6 X=Xenon 135 I=Iodine 135 5 4 3 2 1 0 -1 0 10 20 30 Time (hours) 40 FIGURE P3.20 Nuclear reactor state response to initial conditions. 50 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 108 CHAPTER 3 P3.21 State Variable Models Referring to Figure P3.21 we have 1 1 1 Y (s) = W (s) = h1 U (s) + Q(s) s s s 1 h1 U (s) + 2 [h0 U (s) − a0 Y (s) − a1 sY (s) + a1 h1 U (s)] . = s s Gathering like terms and re-arranging yields 1+ a1 a0 + 2 Y (s) = s s h1 h0 a1 h1 U (s) + 2 + 2 s s s or Y (s) = h1 s + h0 + a1 h1 U (s) . s 2 + a1 s + a0 Computing the transfer function from the state variable representation yields G(s) = C (sI − A)−1 B = h 1 0 i s+a1 s2 +a1 s+a0 −a0 s2 +a1 s+a0 1 s2 +a1 s+a0 s s2 +a1 s+a0 h1 h0 = h1 s + h0 + a1 h1 . s 2 + a1 s + a0 h1 U(s) Q(s) h0 + -- 1 s + + 1 s W(s) a1 a0 FIGURE P3.21 Block diagram with labeled signals. P3.22 The governing equations are L C1 di = v2 dt dv1 1 1 + (v1 − v) + (v1 − v2 ) = 0 dt R1 R2 Y(s) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 109 Problems C2 dv2 1 v2 + (v2 − v1 ) + i + =0. dt R2 R3 Let u = v, x1 = i, x2 = v1 and x3 = v2 . Then, ẋ = 0 0 − C12 − a1 1 R1 + 1 R2 1 R2 C2 1 L 1 C1 R2 0 − 1 R2 C2 + 1 R3 C2 x+ 0 1 R1 C1 0 u y = [0 0 1]x . P3.23 A state variable representation is given by ẋ = 0 1 0 0 0 0 1 x + 0 r −30 −31 −10 y = [1 0 0]x . 1 Other representations include the input feedforward representation −10 1 0 0 ẋ = −31 0 1 x + 0 r −30 0 0 y = [1 0 0]x , 1 the physical variable representation 1 −3 ẋ = 0 −2 0 0 0 1 x+ 0 r 0 −5 y = [1 0 0]x , 1 and the decoupled representation 0 −3 ẋ = 0 −2 0 0 1 0 x + 1 r 0 −5 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 110 CHAPTER 3 State Variable Models y= P3.24 1 1 1 x. − 6 3 2 The matrix representation of the state equations is ẋ = 3 0 0 2 x + 1 1 0 1 u1 u2 + 0 1 d . When u1 = 0 and u2 = d = 1, we have ẋ1 = 3x1 + u2 ẋ2 = 2x2 + 2u2 So we see that we have two independent equations for x1 and x2 . With U2 (s) = 1/s and zero initial conditions, the solution for x1 is found to be x1 (t) = L −1 1 {X1 (s)} = L s(s − 3) 1 1 1 1 −1 =L − + = − 1 − e3t 3s 3 s − 3 3 −1 and the solution for x2 is x2 (t) = L−1 {X2 (s)} = L−1 P3.25 2 s(s − 2) 1 1 = L−1 − + s s−2 = −1+e2t . Since Φ(s) = (sI − A)−1 , we have Φ(s) = s+1 −2 0 s+3 −1 = s+3 2 0 s+1 1 ∆(s) where ∆(s) = (s + 1)(s + 3). The state transition matrix is Φ(t) = L−1 {Φ(s)} = P3.26 e−t 0 e−t − e−3t e−3t The state variable differential equation is ẋ = 0 1 −25 −6 y = [1 0]x . x + 0 25 r . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 111 Problems and s+6 1 Φ(s) = (sI − A)−1 = −25 s 1 ∆(s) where ∆(s) = s2 + 6s + 25. P3.27 Equating the change in angular momentum to the sum of the external torques yields J θ̈ − Hω cos θ = −bθ̇ − kθ where b is the damping coefficient, k is the spring constant, and J is the wheel moment of inertia. Defining the state variables x1 = θ and x2 = ẋ and the input u = ω, we can write the equations of motion as ẋ1 = x2 k b H ẋ2 = − x1 − x2 + u cos x1 J J J With a small angle assumption (that is, cos x1 ≈ 1) we have ẋ = 0 1 −k/J −b/J y=θ= P3.28 h 1 0 i x. x + 0 u H/J The governing equations of motion are m1 y¨1 + k(y1 − y2 ) + by˙1 = u m2 y¨2 + k(y2 − y1 ) + by˙2 = 0 y = y2 . Let x1 = y1 , x2 = ẏ1 , x3 = y2 and x4 = ẏ2 . Then 0 − k ẋ = m1 0 k m2 y= h 1 0 − mb1 k m1 0 0 0 1 − mk2 − mb2 0 0 0 1 0 i x. 0 0 1 m x + 1 u 0 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 112 CHAPTER 3 P3.29 State Variable Models The equations of motion are I q¨1 + M gL sin q1 + k(q1 − q2 ) = 0 J q¨2 − k(q1 − q2 ) = u . Let x1 = q1 , x2 = q̇1 , x3 = q2 , and x4 = q̇2 and linearize the equations using small angle assumptions (i.e. sin q1 ≈ q1 ). Then, we have x˙1 = x2 M gL k x˙2 = − x1 − (x1 − x3 ) I I x˙3 = x4 k 1 x˙4 = (x1 − x3 ) + u . J J P3.30 Using Kirchoff’s current law, we find that C dv c = i2 + i3 dt where i3 = current in R3 . Let i1 = current in R1 . Using Kirchoff’s voltage law, we have L diL = v1 − R1 i1 dt and R1 i1 + R2 i2 + vc = v1 . But i2 = i1 − iL , so (R1 + R2 )i1 = v1 − vc + R2 iL . Using Kirchoff’s voltage law once again, we calculate i3 as i3 = v2 − vc . R3 Utilizing the above equations, we can solve for diL /dt and dvc /dt, as follows: diL R2 R1 R1 R2 = v1 + vc − iL dt L(R1 + R2 ) L(R1 + R2 ) L(R1 + R2 ) vc v1 vc vc R1 iL v2 = − − − + dt C(R1 + R2 ) C(R1 + R2 ) CR3 C(R1 + R2 ) CR3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 113 Problems Define the state variables x1 = vc and x2 = iL . Then, in matrix form we have ẋ = 1 +R2 +R3 ) − (R CR3 (R1 +R2 ) R1 L(R1 +R2 ) y = i2 = P3.31 h 1 − (R1 +R 2) 1 − C(RR1 +R 2) x+ R1 R2 − L(R 1 +R2 ) 1 − (R1R+R 2) i x+ h 1 CR3 1 C(R1 +R2 ) R2 L(R1 +R2 ) 1 (R1 +R2 ) 0 i 0 v1 v2 v1 v2 A state variable representation is ẋ = 0 1 x + −3 −4 0 30 u . The state transition matrix can be computed as follows: n Φ = L−1 [sI − A]−1 o s 1 s+4 = L−1 ∆(s) −3 = 3 −t − 21 e−3t 2e − 32 e−t + 23 e−3t 1 1 −t − 21 e−3t 2e − 12 e−t + 32 e−3t where ∆(s) = s2 + 4s + 3 = (s + 1)(s + 3) . P3.32 A state variable representation is ṁ1 = −k1 m1 + r ṁ2 = k1 m1 − k2 m2 where k1 and k2 are constants of proportionality. In matrix form, we have ẋ = Ax + Br = −k1 0 k1 −k2 x+ 1 0 r where x1 = m1 and x2 = m2 . Let k1 = k2 = 1 and assume that r(t) = 0 and x1 = 1 and x2 = 0. Then x(t) = Φ(t)x(0) = e−t 0 te−t e−t x(0) = e−t te−t . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 114 CHAPTER 3 State Variable Models The simulation is shown in Figure P3.32. 0.4 1 0.9 0.35 0.8 0.3 0.25 x1 0.6 x2 state history, x(t) 0.7 0.5 0.4 0.2 0.15 0.3 0.1 0.2 0.05 x2 0.1 0 0 5 t=0 0 0 10 0.5 time (sec) 1 x1 FIGURE P3.32 Actual versus approximate state response. P3.33 The system (including the feedback) is described by ẋ = Ax = 0 1 −1/2 −1 x . The charactersitic equation is det[λI − A] = det λ −1 1/2 λ + 1 = λ2 + λ + The roots of the characteristic equation are 1 1 λ1,2 = − ± j . 2 2 1 =0. 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 115 Problems The system response is x(t) = eAt x(0) = e−t/2 cos 2t + e−t/2 sin 2t 2e−t/2 sin 2t −e−t/2 sin 2t e−t/2 cos 2t − e−t/2 sin 2t = e−t/2 2 sin 2t cos 2t − sin 2t x(0) where x1 (0) = 0 and x2 (0) = 1. P3.34 (a) The state space representation is 1 −6 −11 −6 0 ẋ = 0 0 0 1 x + 0 r 0 y = [6 0 0] x . 1 (b) The element φ11 (t) of the state transition matrix is φ11 (t) = e−3t − 3e−2t + 3e−t . P3.35 The state equations are 1 8 [80θ − 50h] = −x1 + x2 50 5 θ̇ = ẋ2 = ω = x3 Km Km Kb Km Ka 353 25000 ω̇ = ẋ3 = ia = − ω+ vi = − x3 + vi . J JRa JRa 30 3 ḣ = ẋ1 = In state variable form, we have 8 −1 5 ẋ = 0 0 P3.36 0 0 1 0 − 353 30 x+ 0 0 25000 3 vi . Using Newton’s Law and summing the forces on the two masses yields M1 ẍ(t) + b1 ẋ(t) + k1 x(t) = b1 ẏ(t) M2 ÿ(t) + b1 ẏ(t) + k2 y(t) = b1 ẋ(t) + u(t) Let z1 = x, z2 = ẋ, z3 = y, and z4 = ẏ . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 116 CHAPTER 3 State Variable Models Then we write the system in state variable form as 0 − k1 ż = M1 0 1 0 0 b1 −M 1 0 b1 M1 0 0 1 b1 M2 k2 −M 2 b1 −M 2 0 y= P3.37 h 1 0 0 0 i z. 0 0 z + u 0 1 M2 From the block diagram in Figure P3.37, we obtain ẋ1 ẋ2 ẋ3 y = x2 = x3 = −10x1 − 4x2 − 3x3 + u = x1 + 12x2 + 5x3 or ẋ = 0 1 0 0 −10 −4 −3 y = [1 12 5] x . 0 0 1 x + 0 u 1 The third-order differential equation model is ... y +3ÿ + 4ẏ + 10y = 5ü + 12u̇ + u . 5 12 U(s) + - -- ∫ x3 ∫ x2 3 4 10 FIGURE P3.37 Block diagram with states labeled. ∫ x1 ++ + Y(s) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 117 Advanced Problems Advanced Problems AP3.1 With the state variables are defined as x z= ẋ , i the nonlinear equations of motion are ż1 ż = g − 2 ż3 z2 K (Io +z3 )2 m (Xo +z1 )2 1 L (v − Rz3 ) , where the control is the voltage v. We assume that z1 = x is measurable. The linearized equations of motion are ż = Az + Bv y = Cz where A= 0 1 0 2K Io2 m Xo3 0 0 0 Io − 2K m Xo2 −R L The transfer function is , 0 h i B= . 0 , and C = 1 0 0 1 L G(s) = C(sI − A)−1 B . With the constants R = 23.2 L = 0.508 m = 1.75 K = 2.9 × 10−4 Io = 1.06 Xo = 4.36 × 10−3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 118 CHAPTER 3 State Variable Models the transfer function is G(s) = AP3.2 s3 + −36.38 . + 4493s + 205195 45.67s2 The differential equation describing the motion of y is mÿ + bẏ + ky = bu̇ + ku . Taking Laplace tranforms (with zero initial conditions) yields the transfer function Y (s) (b/m)s + (k/m) = 2 . U (s) s + (b/m)s + (k/m) In state space form, we have ẋ = y= AP3.3 h 0 1 −k/m −b/m k/m b/m i x + 0 1 x. u The transfer function is Y (s) 2s2 + 6s + 5 = 3 . R(s) s + 4s2 + 5s + 2 In (nearly) diagonal form, we have 1 −1 A= 0 −1 0 0 0 , 0 −2 0 B= 1 , and 1 C= h 1 1 1 i . The matrix A is not exactly diagonal due to the repeated roots in the denominator of the transfer function. AP3.4 The differential equations describing the motion of y and q are mÿ + k2 ẏ + k1 (y − q) = f −bq̇ + k1 (y − q) = f where k1 = 2 and k2 = 1. Assume the mass m = 1. Then with the state © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 119 Advanced Problems variables defined as z = h y ẏ q 0 1 ż = −3 0 iT y= 1 0 0 i 0 −1/b 0 2 z + 2/b 0 −2/b h , we have the state variable model z 1 f If we model a large bump at high speeds as an impulse and a small bump at low speeds as a step, then b = 0.8 provides good performance. In both cases, the ride settles out completely in about 10 seconds. AP3.5 The differential equations describing the motion of x and θ are (M + m)ẍ + M L cos θ θ̈ − M L sin θ θ̇ 2 = −kx g sin θ + cos θẍ + Lθ̈ = 0 Assuming θ and θ̇ are small, it follows that (M + m)ẍ + M Lθ̈ = −kx ẍ + Lθ̈ = −gθ Define the state variables as z = able model is 0 −k/m ż = 0 h x ẋ θ θ̇ 1 0 0 gM/m 0 0 iT . Then, the state vari 0 0 k/(Lm) 0 −g(M + m)/(Lm) 0 AP3.6 AP3.7 z 1 Computing the closed-loop system yields A − BK = −1 1 −K1 −K2 , The characteristic polynomial is B= 0 1 , and C= h 2 1 i . |sI − (A − BK)| = s2 + (K2 + 1)s + K1 + K2 = 0. The roots are in the left-half plane whenever K2 +1 > 0 and K1 +K2 > 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 120 CHAPTER 3 AP3.8 State Variable Models (a) A state variable representation is given by x˙1 x˙2 x˙3 y = x2 = x3 = −Kx1 − 12x2 − 6x3 + Kr = x1 or, in matrix form ẋ = y= h 0 1 0 0 1 0 0 x 0 0 1 x + 0 r −K −12 −6 i K (b) The characteristic roots are found by solving det [λI − A] = 0 or λ3 + 6λ2 + 12λ + K = 0 When K = 8, we have characteristic roots at λ1 = −2, λ2 = −2, and λ3 = −2, as desired. (c) The unit step response is given by y(t) = 1 − e−2t − 2te−2t − 2t2 e−2t . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 121 Design Problems Design Problems CDP3.1 The transfer model of the traction drive, capstan roller, and linear slide was given in CDP2.1 as rKm X(s) = , Va (s) s [(Lm s + Rm )(JT s + bm ) + Kb Km ] where JT = Jm + r 2 (Ms + Mb ) . Define x1 = x, x2 = ẋ, and x3 = ẍ. Then, a state variable representation is ẋ = Ax + Bva y = Cx where 0 A= 0 0 C= DP3.1 h 1 0 0 1 b Km − Rm bLmm+K JT m JT − Lm bLmm+R JT 1 0 0 i . , B= (a) The equation of motion of the spring-mass-damper is mÿ + bẏ + ky = u or ÿ = − b k 1 ẏ − y + u . m m m Select the state variables x1 = y and x2 = ẏ . Then, we have ẋ = Ax + Bu y = Cx 0 0 rKm Lm J T © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 3 State Variable Models where A= 0 1 −20 −9 , 0 B= 1 , C= h 1 0 i . A is the system matrix. The characteristic equation is det[λI − A] = det s −1 20 s + 9 = s2 + 9s + 20 = 0 . The roots of the characteristic equation are s1 = −4 and and the transistion matrix is Φ(t) = 5e−4t − 4e−5t e−4t − e−5t −20e−4t + 20e−5t −4e−4t + 5e−5t s2 = −5 , . (b) Assume the initial conditions are x1 (0) = 1 and x2 (0) = 2. The zeroinput response is shown in Figure DP3.1. (c) Suppose we redesign the system by choosing b and k to quickly damp out x2 and x1 . We can select b and k to achieve critical damping. Critical damping: b/m=20, k/m=100 b/m=9, k/m=20 2 2 1.5 1 x1 1 0 x1 State response, x 0.5 State response, x 122 0 x2 -1 -0.5 -2 x2 -1 -3 -1.5 -2 0 0.5 1 Time(sec) FIGURE DP3.1 Zero input state response. 1.5 2 -4 0 0.5 1 Time(sec) 1.5 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 123 Design Problems If we desire the characteristic polynomial to be pd (s) = (s + 10)2 = s2 + 20s + 100, then we need b = 20 and k = 100. DP3.2 The desired transfer function is 6 Y (s) = 2 . U (s) s + 7s + 10 The transfer function derived from the phase variable representation is Y (s) d = 2 . U (s) s + bs + a Therefore, we select d = 6, a = 10 and b = 7. Assume the aircraft lands precisely on the centerline. The linearized equations of motion are m3 ẍ3 + KD ẋ3 + K2 (x3 − x2 ) = 0 m2 ẍ2 + K2 (x2 − x3 ) + K1 (x2 − x1 ) = 0 2 m1 ẍ1 = − √ K2 (x1 − x2 ) 2 where x1 (0) = x2 (0) = ẋ2 (0) = ẋ3 = 0 and ẋ1 (0) = 60. The system response is shown in Figure DP3.3 where KD = 215. The aircraft settles out at 30 m, although initially it overshoots by about 10 m at 1 second. 45 40 35 30 Amplitude DP3.3 25 20 15 10 5 0 0 1 2 3 4 5 Time (secs) FIGURE DP3.3 Aircraft arresting gear response. 6 7 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 124 CHAPTER 3 DP3.4 State Variable Models We can model the bungi cord system as a mass-spring-damper. This is actually an over-simplification because the bungi cord cannot “push” the jumper down as a spring would—it can only exert a restoring force when the cord is stretched (that is, when the jumper exceeds the length, L, of the cord). The problem is nonlinear! When the distance of the jumper from the platform is less than L we should model the cord spring constant and damping as K = 0 and b = 0, respectively. Only gravity acts on the jumper. Also, when ẋ (the jumper velocity) is negative (where we define positive towards the ground), then we should model b = 0. A reasonable set of equations of motion are ẋ1 = x2 K b ẋ2 = − x1 − x2 + g m m where x1 is the distance measured from the top of the platform and x2 is the jumper velocity. For the initial conditions we have x1 (0) = 10 and x2 (0) = 0. A reasonable set of parameters for the bungi cord are L = 40 m, K = 40 N/m and b = 20 kg/m. The system response is shown in Figure DP3.4 for a person with m = 100 kg. The accelerations experienced by the jumper never exceed 1.5 g. global MASS GRAVITY LENGTH K b MASS=100; HEIGHT=100; GRAVITY=9.806; LENGTH=40; SPRINGCONSTANT=40; SPRINGDAMPING=20; x0=[10;0]; t=0; dt=0.1; n=round(120/dt); for i=1:n; if x0(1)1. For example, we can take K = 8. (b) The transfer function from Td (s) to ω(s) is given by −10s ω(s) = 2 . Td (s) s + 10s + 100 The error plot is shown in Figure AP4.4, where e(s) = −ω(s) (V (s) = 0.) 0.2 0.18 0.16 0.14 e(t) 0.12 0.1 0.08 0.06 0.04 0.02 0 0 1 2 3 4 5 6 7 8 9 10 Time (sec) FIGURE AP4.4 Error plot with a ramp disturbance input. AP4.5 (a) The transfer function from the disturbance Td (s) to the output Y (s) is Y (s) −s = 3 . 2 Td (s) s + 4s + 4s + K The steady-state error (when Td (s) = 1/s) is ess = lim s s→0 s3 + 4s2 1 s =0. + 4s + K s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 158 CHAPTER 4 Feedback Control System Characteristics (b) The closed-loop transfer function is Y (s) K = 3 . R(s) s + 4s2 + 4s + K The steady-state error (when R(s) = 1/s2 ) is ess = lim s(1 − T (s)) s→0 s3 + 4s2 + 4s 4 1 = lim = . 2 3 2 s→0 s s(s + 4s + 4s + K) K (c) Let K = 8. Then, Y (s) −s = 3 . Td (s) s + 4s2 + 4s + 8 The error plot is shown in Figure AP4.5, for r(t) = 0. 0.15 0.1 e(t) 0.05 0 -0.05 -0.1 0 2 4 6 8 10 12 Time (sec) FIGURE AP4.5 Error plot with a step disturbance input and K =8. AP4.6 (a) The transfer function is Vo (s) 1 + RCs = . V (s) 2 + RCs 14 16 18 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 159 Advanced Problems (b) The system sensitivity is defined as SCG = ∂G/G . ∂C/C Therefore, the sensitivity is determined to be SCG = RCs = (2 + RCs)(1 + RCs) 1+ 1 2 RCs (c) Let V (s) = 1/s. Then Vo (s) = 1+ 1 RCs . 1 + RCs 1 0.5 0.5RC = + . 2 + RCs s s RCs + 2 Taking the inverse Laplace transform yields vo (t) = 0.5(1 + e−2t/RC )u(t) where u(t) is the unit step function. A plot of vo (t) versus t/RC is shown in Figure AP4.6. 1 0.95 0.9 0.85 Vo 0.8 0.75 0.7 0.65 0.6 0.55 0.5 0 0.5 1 1.5 2 2.5 t / RC FIGURE AP4.6 Step response. 3 3.5 4 4.5 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 160 CHAPTER 4 AP4.7 Feedback Control System Characteristics (a) The transfer function from Td (s) to Y (s) is s Y (s) =− . Td (s) s(s + 1) + K (b) The transfer function from N (s) to Y (s) is Y (s) K = . N (s) s(s + 1) + K (c) Let Td (s) = A/s and N (s) = B/s. Then, ess = −yss = lim s s→0 A K B s − lim s = −B . s(s + 1) + K s s→0 s(s + 1) + K s So, K has no effect on the steady-state errors. However, choosing K = 100 will minimize the effects of the disturbance Td (s) during the transient period. AP4.8 (a) The closed-loop transfer function is T (s) = Kb . s + Kb + 2 (b) The sensitivity is determined to be SbT = ∂T /T s+2 = . ∂b/b s + Kb + 2 (c) The transfer function from Td (s) to Y (s) is Y (s) b = . Td (s) s + Kb + 2 So, choose K as large as possible, to make Y (s)/Td (s) as “small” as possible. Thus, select K = 50 . This also minimizes SbT at low frequencies. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 161 Design Problems Design Problems The model of the traction drive, capstan roller, and linear slide was developed in CDP2.1: θ(s) Km = . Va (s) s [(Lm s + Rm )(JT s + bm ) + Kb Km ] The step response for the closed-loop system (with the tachometer not in the loop) and various values of the controller gain Ka is shown below. % System parameters Ms=5.693; Mb=6.96; Jm=10.91e-03; r=31.75e-03; bm=0.268; Km=0.8379; Kb=0.838; Rm=1.36; Lm=3.6e-03; Lm=0; % Controller gain Ka=100; % Motor and slide model Jt=Jm+r^2*(Ms+Mb); num=[Km]; den=[Lm*Jt Rm*Jt+Lm*bm Kb*Km+Rm*bm 0]; sys=tf(num,den); %Closed-loop tf and step response sys_cl=feedback(Ka*sys,[1]); step(sys_cl) 1.5 Ka=2 Ka=5 Ka=10 Ka=100 1 Theta step response CDP4.1 0.5 0 0 0.2 0.4 0.6 Time (sec) 0.8 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 162 CHAPTER 4 (a) The transfer function from the load disturbance to the output speed is −G(s) −s ω(s) = = 2 . Td (s) 1 + Gc G(s) s + 4s + K Thus, the effect on ω(s) (of a unit step disturbance) at steady-state is lim ω(t) = lim s t→∞ s→0 −s 2 s + 4s + K 1 =0. s We see that the load disturbance has no effect on the output at steadystate. (b) The system response for 10 ≤ K ≤ 25 is shown in Figure DP4.1. K=10,12,16,18,20,23,25 100.04 100.02 100 K=25 99.98 99.96 w(t) DP4.1 Feedback Control System Characteristics 99.94 K=10 99.92 99.9 99.88 99.86 99.84 0 0.5 1 1.5 2 2.5 3 Time(sec) FIGURE DP4.1 Speed control system response. For example , if we select K = 16, then ωn = 4, ζ = response due to a unit step disturbance is −s ω(s) = 2 s + 4s + 16 1 s = 1 2, −1 . (s + 2)2 + 12 Hence, if we are originally at ω(t) = 100 for t < τ , we have √ 1 ω(t) = 100 − √ e−2t sin 12t 12 t≥τ . and the © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 163 Design Problems DP4.2 With θd = 0, we have θ(s) = s G(s) T (s) = 3 Td (s) . KK1 d 2 s + 4s + 9s + KK1 1 + G(s) s For Td = A/s, we have θ(s) = A . s3 + 4s2 + 9s + KK1 The system response to a unit step disturbance for various values of KK1 are shown in Figure DP4.2. From the plot we see that when KK1 is small the response is slow but not oscillatory. On the other hand, when KK1 is large the response is fast but highly oscillatory. In fact, if KK1 > 35, the system is unstable. Thus, we might select KK1 = 10 as a reasonable trade-off between fast performance and stability. Unit step response for KK1=1,5,10,15,20,25 0.12 0.1 KK1=1 0.08 0.06 KK1=5 q 0.04 0.02 0 -0.02 KK1=25 -0.04 -0.06 0 1 2 3 4 5 6 7 8 time(sec) FIGURE DP4.2 Aircraft roll angle control system response to a disturbance. DP4.3 (a) The closed-loop transfer function is T (s) = ω(s) K = 2 . ωd (s) s + 5s + KK1 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 164 CHAPTER 4 Feedback Control System Characteristics Then, E(s) = (1 − T (s))ωd (s) = s2 + 5s + K(K1 − 1) 1 . s2 + 5s + KK1 s So, if 0.99 < K1 < 1.01 , then |ess | < 0.01 . (b) The transfer function from Td (s) to ω(s) is ω(s) = s2 −s Td (s) . + 5s + KK1 So, with E(s) = −ω(s) and Td (s) = 2/s2 , we have lim sE(s) = s→0 2 . KK1 Therefore, we select KK1 > 20 to obtain ess < 0.1. DP4.4 The steady-state error for a step input command is zero for any K1 . The transfer function from Td (s) to Y (s) is Y (s) G(s) 2 = = 3 . 2 Td (s) 1 + KG(s) s + 5s + 4s + 2K Thus, the output at steady-state due to a step disturbance Td (s) = A/s is lim sY (s) = s→0 A . K We want to maximize K to reduce the effect of the disturbance. As we will see in Chapter 6, we cannot select K too high or the system will become unstable. That is why the problem statement suggests a maximum gain of K = 10. For the design we choose K = 10 . DP4.5 The transfer function from V (s) to Vo (s) is Vo (s)/V (s) = ks s+a © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 165 Design Problems where k= R2 + R3 R2 and a = 1 . R1 C Computing the step response, we find that vo (t) = ke−at = 5e−100t . Solving for R1 , R2 , R3 and C yields R1 C = 0.01 DP4.6 and R2 =4. R3 (a) The closed-loop transfer function is θ(s) = s2 K/J θd (s) . + K/J Since J > 0, the system is unstable when K < 0 and marginally stable when K > 0. (b) Since the system is marginally stable, the system response does not have a steady-state value—it oscillates indefinitely. (c) The closed-loop transfer function is θ(s) = KD s + KP θd (s) . Js2 + KD s + KP The system is stable for all KD > 0 and KP > 0, given that J > 0. (d) The tracking error E(s) = θd (s) − θ(s) is E(s) = Js2 . Js2 + KD s + KP Therefore, using the final value theorem we obtain the steady-state value lim sE(s) = lim s s→0 DP4.7 s→0 Js2 1 · =0. 2 Js + KD s + KP s (a) The closed-loop transfer function is Y (s) = s2 s 1 Td (s) = 2 + Ks + 2K s + Ks + 2K where the disturbance is a unit step Td (s) = 1/s. Considering the poles of the closed-loop system, we find that when K > 8 the system © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 166 CHAPTER 4 Feedback Control System Characteristics has two real poles. In that case the disturbance step response is 1 y(t) = √ e−αt − e−βt , K 2 − 8K where α= K− √ K 2 − 8K 2 and β = K+ √ K 2 − 8K 2 Bounding the maximum y(t) yields the inequality |y(t)| = √ K2 1 e−αt − e−βt ≤ 0.05. − 8K We know that e−αt − e−βt ≤ 1, for any α and β computed as shown above where K > 8. So, if we choose K such that √ 1 ≤ 0.05. K 2 − 8K we will guarantee that the maximum bound of 0.05 is not exceeded. Solving for K yields K > 24.4. For any K > 24.4 we know that the maximum value of the disturbance step response will be less than 0.05. When K = 24.4 the maximum unit step disturbance response is 0.035. Solving explicitly for K so that the maximum is 0.05 we find that K = 16.3 (this was found numerically since it is very difficult to obtain analytically). (b) Since the system is type 2, we know that the steady-state value of the disturbance step response is zero for a unit step disturbance. DP4.8 (a) The sensitivities are " # " # SτT1 ∂T /T −s2 (τ2 s + 1) τ1 = = T (s) ∂τ1 /τ1 K SτT2 ∂T /T −s2 (τ1 s + 1) τ2 = = T (s) ∂τ2 /τ2 K and where we assume that K 6= 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 167 Design Problems (b) Computing the closed-loop transfer function yields Y (s) = s (τ1 s + 1) Td (s) s (τ1 s + 1) (τ2 s + 1) + K When Td (s) = 1/s, using the final value theorem we find that s (τ1 s + 1) =0 s→0 s (τ1 s + 1) (τ2 s + 1) + K lim sY (s) = lim s→0 as long as K 6= 0. We assume here that final value theorem applies (i.e., the system is stable, more on this in Chapter 6). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 168 CHAPTER 4 Feedback Control System Characteristics Computer Problems CP4.1 The step response and an m-file script which generates the step response is shown in Figure CP4.1. The closed-loop transfer function is T (s) = s2 12 . + 2s + 22 The percent overshoot is P.O. = 50.2% and the steady-state error is ess = 0.45. Step Response 0.9 0.8 System: sys_cl Peak amplitude: 0.82 Overshoot (%): 50.2 At time (sec): 0.67 0.7 Amplitude 0.6 num = [12]; den = [1 2 10]; sys = tf(num,den); sys_cl = feedback(sys,[1]) step(sys_cl) 0.5 0.4 0.3 0.2 0.1 0 0 1 2 3 Time (sec) 4 5 6 FIGURE CP4.1 Step response. CP4.2 The transfer function is G(s) = s2 4 . + 2s + 20 An m-file script which generates the step response is shown in Figure CP4.2. The step response is also shown in Figure CP4.2. The step response is generated using the step function. In the script, the transfer function numerator is represented by num and the denominator is represented by den. The steady-state value is yss = 0.2 and the desired value is 1.0. Therefore, the steady-state error is ess = 0.8 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 169 Computer Problems Step Response 0.35 0.3 0.25 Amplitude num=[4]; den=[1 2 20]; sys = tf(num,den); axis([0 6 0 1]); t=[0:0.01:6]; step(sys,t) y = step(sys,t); yss = y(length(t)) 0.2 0.15 0.1 yss = 0.20 0.05 0 0 1 2 3 Time (sec) 4 5 6 FIGURE CP4.2 Step response. CP4.3 The step responses and the m-file script which generates the step responses is shown in Figure CP4.3. 7 K=10 K=200 K=500 6 K=[10,200,500]; t=[0:0.01:7]; for i=1:3 num=5*K(i); den=[1 15 K(i)]; sys = tf(num,den) y(:,i)= step(sys,t); end plot(t,y(:,1),t,y(:,2),'--',t,y(:,3),':') legend('K=10','K=200','K=500',-1) 5 4 3 2 1 0 FIGURE CP4.3 Step responses for K = 10, 100, 500. 0 1 2 3 4 5 6 7 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 170 CHAPTER 4 CP4.4 Feedback Control System Characteristics (a,b) The m-file and plots are shown in Figure CP4.4. Step response 1.5 y(t) 1 ng=1;dg=[1 1.91 0];sysg=tf(ng,dg); K=10; syscl=feedback(K*sysg,1); figure(1) subplot(211) step(syscl) subplot(212) syst=feedback(sysg,K) step(syst) 0.5 0 0 1 2 3 Time (sec) 4 5 6 Disturbance response 0.2 0.15 y(t) 0.1 0.05 0 0 1 2 3 Time (sec) 4 5 6 FIGURE CP4.4 Step response and disturbance response. (c) The estimated steady-state tracking error due to a unit step input is zero, and the estimated steady-state tracking error to a unit disturbance is 0.1. (d) The estimated maximum tracking error due to a unit step input is 0.4, and the estimated maximum tracking error to a unit disturbance is 0.14. The maximum occurs at approximately t = 1 s. CP4.5 The step response and the m-file script which generates the step response is shown in Figure CP4.5. The closed-loop transfer function is determined to be T (s) = s2 10 . + 3.7s + 10 Using the m-file script, a trial-and-error search on k yields k = 3.7 . The percent overshoot P.O. = 10% and the steady-state value is 1, as expected. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 171 Computer Problems 1.4 1.2 1 0.8 y(t) k = 3.7; % Final value of k=3.7 numcg = [10]; dencg = [1 k 0]; sys_o = tf(numcg,dencg); sys_cl = feedback(sys_o,[1]) t = [0:0.1:5]; [y,t] = step(sys_cl,t); plot(t,y,[0 5],[1.1 1.1],'--'); grid xlabel('Time (sec)'); ylabel('y(t)'); 0.6 0.4 Transfer function: 10 ---------------s^2 + 3.7 s + 10 0.2 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 5 FIGURE CP4.5 Step response. CP4.6 The closed-loop transfer function is T (s) = K s−a+K where K = 2. When a = 1 and R(s) = 1/s, the final value is lim sT (s)R(s) = lim T (s) = s→0 s→0 K =2. K −a The output is within 2% of the final value at around t = 4.6 seconds. The plot of the step responses for a = 1, 0.5, 2, 5 is shown in Figure CP4.6. The output is unstable for a > 2. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 172 CHAPTER 4 Feedback Control System Characteristics K=2; t=[0:0.1:5]; num=K*[1]; a=[1 0.5 2 5]; for i=1:4 den=[1 -a(i)]; sys = tf(num,den); sys_cl = feedback(sys,[1]); y(:,i)=step(sys_cl,t); end plot(t,y(:,1),t,y(:,2),':',t,y(:,3),'--',t,y(:,4),'-.') axis([0 5 0 5]); xlabel('Time (sec)'), ylabel('y(t)') title('a=1 (solid); a=0.5 (dotted); a=2 (dashed); a=5 (dashdot)') a=1 (solid); a=0.5 (dotted); a=2 (dashed); a=5 (dashdot) 5 4.5 4 3.5 y(t) 3 2.5 2 1.5 1 0.5 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 5 FIGURE CP4.6 Step response for a=1, 0.5, 2, and 5. CP4.7 The transfer function from the disturbance to the output is T (s) = G(s) 1 = . 2 1 + K0 G(s) Js + bs + k + K0 The disturbance response is shown in Figure CP4.7. The compensated system response is significantly reduced from the uncompensated system response. The compensated system output is about 11 times less than the uncompensated system output. So, closed-loop feedback has the advantage of reducing the effect of unwanted disturbances on the output. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 173 Computer Problems J=1; k=5; c=0.9; num=[1/J]; den=[1 c/J k/J]; sys = tf(num,den); t=[0:0.1:10]; % yu=step(sys,t); % Part (a) K0=50; numk=[K0]; denk=[1]; sysk = tf(numk,denk); sys_cl = feedback(sys,sysk); yc=step(sys_cl,t); % Part (b) plot(t,yu,t,yc,'--') xlabel('Time (sec)'), ylabel('\theta') title('Uncompensated response (solid) & Compensated response (dashed)') Uncompensated response (solid) & Compensated response (dashed) 0.35 0.3 0.25 q 0.2 0.15 0.1 0.05 0 0 1 2 3 4 5 Time (sec) 6 7 8 9 10 FIGURE CP4.7 Disturbance responses for both the uncompensated and compensated systems. CP4.8 The step responses for the proportional and PI controller are shown in Figure CP4.8. The steady-state tracking error for the proportional controller is ess = 0.33 . Increasing the complexity of the controller from a proportional controller to a proportional plus integral (PI) controller allows the closed-loop system to track the unit step response with zero steady-state error. The cost is controller complexity, which translates into higher costs ($). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 174 CHAPTER 4 Feedback Control System Characteristics numg=[10]; deng=[1 10]; sysg = tf(numg,deng); t=[0:0.001:0.5]; % Part (a) numc=[2]; denc=[1]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); yk=step(sys_cl,t); % Part (b) numc=[2 20]; denc=[1 0]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); yp=step(sys_cl,t); % plot(t,yk,t,yp,'--') xlabel('Time (sec)'),ylabel('y(t)') title('Proportional controller (solid) & PI controller (dashed)') Proportional controller (solid) & PI controller (dashed) 1 0.9 0.8 0.7 y(t) 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.05 0.1 0.15 0.2 0.25 Time (sec) 0.3 0.35 0.4 0.45 0.5 FIGURE CP4.8 Step response for proportional controller and PI controller. CP4.9 (a) The closed-loop transfer function is T (s) = G(s) 10s2 + 500s R(s) = 2 R(s) . 1 + G(s)H(s) s + 200s + 5000 The step response is shown in Figure CP4.9a. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 175 Computer Problems (b) The response of the system to the sinusoidal disturbance N (s) = s2 100 + 100 is shown in Figure CP4.9b. (c) In the steady-state, the magnitude of the peak response is 0.095 and the frequency is 10 rad/sec (see Figure CP4.9b). % Part (a) ng=10*[1 0]; dg=[1 100]; sysg=tf(ng,dg); nh=[5]; dh=[1 50]; sysh=tf(nh,dh); sys=feedback(sysg,sysh) figure(1) step(sys) >> Transfer function: 10 s^2 + 500 s -----------------s^2 + 200 s + 5000 Step Response % Part (b) sysn=-feedback(sysg*sysh,1) syss=tf([100],[1 0 100]); % This is the sinusoidal input figure(2) t=[0:0.001:7]; step(syss*sysn,t) 10 Amplitude 8 6 4 2 Step Response 0.1 0 0.08 0 0.02 0.04 0.06 0.08 0.1 Time (sec) 0.06 (a) Amplitude 0.04 0.02 0 −0.02 −0.04 −0.06 −0.08 −0.1 0 1 2 3 4 Time (sec) 5 6 7 (b) FIGURE CP4.9 (a) Unit step response. (b) Response to sinusoidal noise input at ω = 10 rad/sec. CP4.10 (a) The closed-loop transfer function is T (s) = Gc (s)G(s) K(s + 1) R(s) = R(s) . 2 1 + G(s)Gc (s) (s + 15)(s + s + 6.5) + K(s + 1) (b) The step responses are shown in Figure CP4.10a. 0.12 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 176 CHAPTER 4 Feedback Control System Characteristics (c) The unit disturbance response of the system is shown in Figure CP4.10b. The steady-state value is 0.14. Step Response 0.7 K=5 K=10 K=50 0.6 0.2 0.18 0.16 System: syscl Final Value: 0.14 0.5 0.4 Amplitude Step response 0.14 0.3 0.12 0.1 0.08 0.2 0.06 0.04 0.1 0.02 0 0 0 1 2 3 4 Time (s) 5 6 7 0 1 2 3 Time (sec) (a) 4 5 6 (b) FIGURE CP4.10 (a) Unit step responses for K = [5, 10, 50]. (b) Disturbance unit step response. The m-file is shown in Figure CP4.11a and the step responses in Figure CP4.11b. CP4.11 5 K=10 K=12 K=15 4 3 Step response K=[10, 12, 15]; t=[0:0.1:20]; ng=[20]; dg=[1 4.5 64]; sysg=tf(ng,dg); nh=[1]; dh=[1 1]; sysh=tf(nh,dh); for i=1:length(K) sys=K(i)*sysg; syscl=feedback(sys,sysh) y(:,i)= step(syscl,t); end plot(t,y(:,1),t,y(:,2),'--',t,y(:,3),':') xlabel('Time (s)') ylabel('Step response') legend('K=10','K=12','K=15',-1) 2 1 0 −1 −2 −3 0 5 (a) FIGURE CP4.11 (a) M-file script. (b) Unit step responses for K = [10, 12, 15]. 10 Time (s) (b) 15 20 7 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 5 The Performance of Feedback Control Systems Exercises E5.1 For a zero steady-state error, when the input is a step we need one integration, or a type 1 system. A type 2 system is required for ess = 0 for a ramp input. E5.2 (a) The closed-loop transfer function is T (s) = Y (s) G(s) 240 240 = = = 2 . R(s) 1 + G(s) (s + 4)(s + 6) + 240 s + 2ζωn s + ωn2 The steady-state error is given by ess = A , 1 + Kp where R(s) = A/s and Kp = lim G(s) = s→0 240 = 10 . 24 Therefore, ess = A . 11 (b) The closed-loop system is a second-order system with natural frequency √ ωn = 264 , 177 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 178 CHAPTER 5 The Performance of Feedback Control Systems and damping ratio 10 ζ= √ = 0.31 . 2 264 The percent overshoot is thus computed to be √ 2 P.O. = 100e−πζ/ 1−ζ = 36% . E5.3 The closed-loop transfer function is G(s) K K Y (s) = = = 2 . I(s) 1 + G(s) s(s + 14) + K s + 14s + K Utilizing Table 5.6 in Dorf & Bishop, we find that the optimum coefficients are given by s2 + 1.4ωn s + ωn2 . We have s2 + 14s + K , so equating coefficients yields ωn = 10 and K = ωn2 = 100 . We can also compute the damping ratio as ζ= 14 = 0.7 . 2ωn Then, using Figure 5.8 in Dorf & Bishop, we find that P.O. ≈ 5%. E5.4 (a) The closed-loop transfer function is T (s) = G(s) 2(s + 8) = 2 . 1 + G(s) s + 6s + 16 (b) We can expand Y (s) in a partial fraction expansion as 2(s + 8) A 1 s+4 Y (s) = 2 =A − 2 (s + 6s + 16) s s s + 6s + 16 . Taking the inverse Laplace transform (using the Laplace transform tables), we find √ y(t) = A[1 − 1.07e−3t sin( 7t + 1.21)] . (c) Using the closed-loop transfer function, we compute ζ = 0.75 and © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 179 Exercises ωn = 4. Thus, 8 a = = 2.67 , ζωn 3 where a = 8. From Figure 5.13(a) in Dorf & Bishop, we find (approximately) that P.O. = 4% . (d) This is a type 1 system, thus the steady-state error is zero and y(t) → A as t → ∞. E5.5 The closed-loop transfer function is Y (s) K = 2 . R(s) s + 4s + K Utilizing Table 5.6 in Dorf & Bishop, we find that the optimum coefficients are given by s2 + 1.4ωn s + ωn2 . We have s2 + 4s + K , so equating coefficients yields ωn = 2.86 and K = ωn2 = 8.16 . We can also compute the damping ratio as ζ= 4 = 0.7 . 2ωn Then, using Figure 5.8 in Dorf & Bishop, we find that P.O. ≈ 5%. E5.6 (a) The closed-loop transfer function is T (s) = Y (s) G(s) 100 = = 2 , R(s) 1 + GH(s) s + 100Ks + 100 where H(s) = 1 + Ks and G(s) = 100/s2 . The steady-state error is computed as follows: ess = lim s[R(s) − Y (s)] = lim s[1 − T (s)] s→0 s→0 " = lim 1 − s→0 1+ 100 s2 100 (1 + s2 A s2 # A = KA . Ks) s (b) From the closed-loop transfer function, T (s), we determine that ωn = © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 180 CHAPTER 5 The Performance of Feedback Control Systems 10 and ζ= 100K = 5K . 2(10) We want to choose K so that the system is critically damped, or ζ = 1.0. Thus, K= 1 = 0.20 . 5 The closed-loop system has no zeros and the poles are at p s1,2 = −50K ± 10 25K 2 − 1 . The percent overshoot to a step input is P.O. = 100e √−5πK 1−25K 2 for 0 < K < 0.2 and P.O. = 0 for K > 0.2. E5.7 The closed-loop transfer function is Y (s) KG(s) K(s + 2) K(s + 2) = = = 2 . R(s) 1 + KG(s) s(s + 1) + K(s + 2) s + s(K + 1) + 2K √ √ Therefore, ωn = 2K and ζ = 2K+1 . So, 2K T (s) = a 4 = . ζωn K +1 From Figure 5.13a in Dorf & Bishop, we determine that a ≈ 1.5 ζωn when ζ = 0.707. Thus, solving for K yields 4 = 1.5 K +1 or K = 1.67. E5.8 The pole-zero map is shown in Figure E5.8. Since the dominant poles are at s = −2 ± 2.45i we have a damping ratio ζ = 0.63. We estimate the percent overshoot to be √ 2 P.O. = 100e−πζ/ 1−ζ = 7.69% The step response is shown in Figure E5.8b. The actual overshoot is 8%. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 181 Exercises Pole−Zero Map 2.5 2 1.5 Imaginary Axis 1 0.5 0 −0.5 −1 −1.5 −2 −2.5 −25 −20 −15 −10 −5 0 Real Axis Step Response 1.4 System: sys Time (sec): 1.28 Amplitude: 1.08 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 Time (sec) 2 FIGURE E5.8 (a) Pole-zero map. (b) Unit step response. E5.9 (a) The closed-loop transfer function is T (s) = s2 + √ K . 2Ks + K 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 182 CHAPTER 5 The Performance of Feedback Control Systems √ The √ damping ratio is ζ = 2/2 and the natural frequency is ωn = K. Therefore, we compute the percent overshoot to be √ 2 P.O. = 100e−πζ/ 1−ζ = 4.3% for ζ = 0.707. The settling time is estimated via Ts = 8 4 =√ . ζωn 2K (b) The settling time is less than 1 second whenever K > 32. E5.10 The second-order closed-loop transfer function is given by T (s) = ωn2 . s2 + 2ζωn s + ωn2 From the percent overshoot specification, we determine that P.O. ≤ 5% implies ζ ≥ 0.69. From the settling time specification, we find that Ts < 4 implies ωn ζ > 1. p And finally, from the peak time specification we have Tp < 1 implies ωn 1 − ζ 2 > π. The constraints imposed on ζ and ωn by the performance specifications define the permissible area for the poles of T (s), as shown in Figure E5.10. Im(s) wn z = 0.69 1-z 2 = P desired region 46o for poles wn Re(s) 1-z 2 = - P z wn = -1 FIGURE E5.10 Permissible area for poles of T (s). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 183 Exercises E5.11 The system is a type 1. The error constants are Kp = ∞ and Kv = 1.0 . Therefore, the steady-state error to a step input is 0; the steady-state error to a ramp input is 1.0A0 , where A0 is the magnitude (slope) of the ramp input. (a) The tracking error is given by E(s) = (s + 9)(s + 2)(s + 4) R(s) = R(s) . 1 + Gc G(s) (s + 9)(s + 2)(s + 4) + K(s + 6) The steady-state tracking error (with R(s) = 1/s) is lim sE(s) = s→0 72 . 72 + 6K We require ess < 0.05, so solving for K yields K > 228. (b) The tracking error due to the disturbance is E(s) = −G(s) −(s + 9)(s + 6) Td (s) = Td (s) . 1 + Gc G(s) (s + 9)(s + 2)(s + 4) + K(s + 6) The tracking error is shown in Figure E5.12. 0 -0.01 -0.02 -0.03 Amplitude E5.12 -0.04 -0.05 -0.06 -0.07 -0.08 0 0.1 0.2 0.3 0.4 0.5 Time (secs) FIGURE E5.12 Tracking error due a step disturbance. 0.6 0.7 0.8 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 184 CHAPTER 5 E5.13 The Performance of Feedback Control Systems The system is a type 0. The error constants are Kp = 0.4 and Kv = 0. The steady-state error to a ramp input is ∞. The steady-state error to a step input is ess = E5.14 1 = 0.71. 1 + Kp (a) The tracking error is given by E(s) = [1 − T (s)] R(s) . The steady-state tracking error (with R(s) = 1/s) is ess = lim s [1 − T (s)] R(s) = lim [1 − T (s)] = 1 − T (0) . s→0 s→0 The closed-loop transfer function is T (s) = K(s + 0.1) , s(s + 0.1)(s + 2) + K(s + 3) and T (0) = 0.033. Therefore, ess = 1 − T (0) = 0.967. (b) Use Gp (s) = 30. Then, ess = lim s [1 − T (s)Gp (s)] R(s) = 1−lim T (s)Gp (s) = 1−30 T (0) = 0 . s→0 The plot of y(t) is shown in Figure E5.15. 1.4 1.2 Response using only dominate poles 1 0.8 y(t) E5.15 s→0 0.6 Actual response 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 Time (sec) 1.2 1.4 1.6 1.8 FIGURE E5.15 Plot of y(t) with T (s) (solid line) and approximate Ta (s) (dashed line). 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 185 Exercises Using the dominant poles, the approximate closed-loop transfer function is Ta (s) = s2 50 . + 10s + 50 The actual transfer function is T (s) = E5.16 500 . (s + 10)(s2 + 10s + 50) The partial fraction expansion is y(t) = − 10(z − 1) −t 10(z − 8) −8t e + e + 1.25 . 7z 56z The plot of y(t) for z = 2, 4, 6 is shown in Figure E5.16. z=2 (solid) & z=4 (dashed) & z=6 (dotted) 1.4 1.2 1 y(t) 0.8 0.6 0.4 0.2 0 0 1 2 3 Time (sec) 4 5 6 FIGURE E5.16 Plot of y(t) for z=2, 4, 6. E5.17 The desired pole locations for the 5 different cases are shown in Figure E5.17. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 186 CHAPTER 5 The Performance of Feedback Control Systems Im(s) desired region for poles wn = 10 37 o o 53 Re(s) (a) 0.6 < z < 0.8 and wn <10 desired region for poles Im(s) wn = 10 45 o o Re(s) 60 (b) 0.5 < z < 0.707 and wn > 1 0 Im(s) desired region for poles wn = 10 o Re(s) 60 wn = 5 (c) 0.5 < z and 5 < wn <10 FIGURE E5.17 Desired pole locations. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 187 Exercises Im(s) desired region for poles wn = 10 o Re(s) 45 wn = 5 (d) 0.707 > z and 5 < wn <10 Im(s) o wn = 6 Re(s) 53 desired region for poles (e) 0.6 < z and wn < 6 FIGURE E5.17 CONTINUED: Desired pole locations. E5.18 The output is given by Y (s) = T (s)R(s) = K G(s) R(s) . 1 + G(s) When K = 1, the steady-state error is ess = 0.2 which implies that lim sY (s) = 0.8 . s→0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 188 CHAPTER 5 The Performance of Feedback Control Systems Since we want ess = 0, it follows that lim sY (s) = 1 , s→0 or 0.8K = 1 . Therefore, K = 1.25. (a) The characteristic equation is s2 = 2ζωn s + ωn2 = s2 + 3.17s + 7 = 0 , from which it follows that √ ωn = 7 = 2.65, ζ= 3.17 = 0.6 . 2ωn Therefore, we compute the percent overshoot and the estimated settling time to be √ 4 2 P.O. = 100e−πζ/ 1−ζ = 9.53% and Ts = = 2.5 s . ζωn (b) The unit step response is shown in Figure E5.19. Step Response 1.4 System: sys Peak amplitude: 1.1 Overshoot (%): 9.53 At time (sec): 1.47 1.2 System: sys Settling Time (sec): 2.25 1 Amplitude E5.19 0.8 0.6 0.4 0.2 0 0 FIGURE E5.19 Unit step response. 0.5 1 1.5 Time (sec) 2 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 189 Exercises E5.20 (a) The closed-loop transfer function is T (s) = s2 + √ K . 2Ks + K The damping ratio is √ 2 2 √ and the natural frequency is ωn = K. Therefore, we compute the percent overshoot to be √ 2 P.O. = 100e−πζ/ 1−ζ = 4.3% ζ= for ζ = 0.707. The settling time is estimated via Ts = 4 8 =√ . ζωn 2K (b) The settling time is less than 1 second whenever K > 32. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 190 CHAPTER 5 The Performance of Feedback Control Systems Problems P5.1 (a) The system error is E(s) = 1 1+ Ka Km sτm +1 R(s) where R(s) = 25o /sec . s So, lim e(t) = lim sE(s) = t→0 s→0 25 . 1 + Ka Km (b) If we desire ess ≤ 1o /sec, then 25o /s ≤ 1o /sec , 1 + Ka Km and solving for Ka Km yields Ka Km ≥ 24 . (c) The closed-loop transfer function is T (s) = Vb (s) Ka Km = . Vc (s) sτm + 1 + Ka Km The step response of the system (i.e. vc (t) = A) is −(Ka Km +1) AKa Km t τm vb (t) = 1−e 1 + Ka Km . So, at settling time, we have 1−e −(1+Ka Km ) t τm ≥ 0.98 , where τm = 0.4. Setting t = 0.03 and solving for Ka Km yields Ka Km ≥ 52 . P5.2 (a) The settling time specification Ts = 4 < 0.6 ζωn © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 191 Problems is used to determine that ζωn > 6.67. The P.O. < 20% requirement is used to determine ζ < 0.45 which implies θ < 63o and the P.O. > 10% requirement is used to determine ζ > 0.60 which implies θ > 53o , since cos θ = ζ. The desired region for the poles is shown in Figure P5.2. Im(s) desired region for poles 53 o 63 o Re(s) s = -6.67 FIGURE P5.2 Desired region for pole placement. (b) The third root should be at least 10 times farther in the left halfplane, so |r3 | ≥ 10|ζωn | = 66.7 . (c) We select the third pole such that r3 = −66.7. Then, with ζ = 0.45 and ζωn = 6.67, we determine that ωn = 14.8. So, the closed-loop transfer function is T (s) = 66.7(219.7) , (s + 66.7)(s2 + 13.3s + 219.7) where the gain K = (66.7)(219.7) is chosen so that the steady-state © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 192 CHAPTER 5 The Performance of Feedback Control Systems tracking error due to a step input is zero. Then, T (s) = G(s) , 1 + G(s) G(s) = T (s) . 1 − T (s) or P5.3 Given the input R(s) = 1 , s3 we compute the steady-state error as ess 1 = lim s s→0 1 + G(s) 1 1 = lim 2 3 s→0 s s G(s) = lim s→0 Since we require that ess ≤ 0.5 cm, we determine 1 s2 K s2 = 1 . K K≥2. P5.4 (a) The closed-loop transfer function is T (s) = G(s) K ωn2 = 2 = 2 . 1 + G(s) s + 2s + K s + 2ζωn s + ωn2 Thus, ωn = √ K and √ ζ = 1/ωn = 1/ K . √ Our percent overshoot requirement√of 5% implies that ζ = 1/ 2 , which in turn implies that ωn = 2. However, the corresponding time to peak would be 4.4 Tp = √ = 3.15 . 2 Our desired Tp = 1.1—we cannot meet both specification simultaneously. (b) Let Tp = 1.1∆ and P.O. = 0.05∆, where ∆ is the relaxation factor to be determined. We have that K = ωn2 and ζωn = 1, so 1 ζ=√ . K © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 193 Problems Thus, √ √ 2 P.O. = e−πζ/ 1−ζ = e−π/ K−1 . Also, Tp = √ π = 1.1∆ . K −1 Therefore, from the proceeding two equations we determine that P.O. = 0.05∆ = e−1.1∆ . Solving for ∆ yields f (∆) = ln ∆ + ln(0.05) + 1.1∆ = 0 . The plot of f (∆) versus ∆ is shown in Figure P5.4. From the plot we 2 1 0 * D=2.07 -1 f(D) -2 -3 -4 -5 -6 -7 -8 0 0.5 1 1.5 2 D FIGURE P5.4 Solving for the zeros of f. see that ∆ = 2.07 results in f (∆) = 0. Thus, P.O. = 0.05∆ = 10% Tp = 1.1∆ = 2.3 sec. 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 194 CHAPTER 5 The Performance of Feedback Control Systems So, we can meet the specifications if they are relaxed by a factor of about 2 (i.e. ∆ = 2.07). P5.5 (a) The closed-loop transfer function is T (s) = s2 K1 K2 (s + 1) . + K1 K2 s + K1 K2 A percent overshoot less than 5% implies ζ ≥ 0.69. So, choose ζ = 0.69. Then set 2ζωn = K1 K2 and ωn2 = K1 K2 . Then 2(0.69)ωn = ωn2 ; and solving for ωn yields ωn = 1.38 . Therefore K1 K2 = ωn2 = 1.9. When K1 K2 ≥ 1.9 it follows that ζ ≥ 0.69. (b) We have a type 2 system, so the steady-state tracking error to both a step and ramp input is zero. (c) For a step input, the optimum ITAE characteristic equation is s2 + 1.4ωn s + ωn2 = 0 . For a ramp input, the optimum ITAE characteristic equation is s2 + 3.2ωn s + ωn2 = 0 . Thus, K1 K2 = ωn2 = 3.2ωn . So, ωn = 3.2 and K1 K2 = 10.24. P5.6 We have a ramp input, r(t) = t. So Kv = lim sG(s) = lim s s→0 s→0 75 75(s + 1) = = 0.6 , s(s + 5)(s + 25) 125 and ess = P5.7 |R| 1 = = 1.67 . Kv 0.6 (a) The closed-loop transfer function is T (s) = Is2 K1 K2 . + K1 K2 K3 s + K1 K2 The steady-state tracking error for a ramp input is ess = lim sE(s) = lim s(1 − T (s))R(s) = lim s(1 − T (s)) s→0 s→0 s→0 1 s2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 195 Problems = lim s→0 Is + K1 K2 K3 = K3 . Is2 + K1 K2 K3 s + K1 K2 But we desire ess = 0.01 m, so K3 = 0.01. (b) For P.O. = 10%, we have ζ = 0.6. Also, 2ζωn = 0.01K1 K2 25 and ωn2 = K1 K2 . 25 Thus, solving for K1 K2 yields K1 K2 = 36 × 104 . P5.8 (a) The closed-loop transfer function is T (s) = P (s) G(s)/s 20 = = . R(s) 1 + G(s)H(s)/s s(s + 40) Therefore, the closed-loop system time constant is τ = 1/40 sec. (b) The transfer function from Td (s) to the output P (s) is P (s) −G(s) −20 = = . Td (s) 1 + G(s)H(s)/s s + 40 The response to a unit step disturbance is 1 p(t) = − (1 − e−40t ) . 2 At settling time, p(t) = 0.98pss = −0.49. Thus, solving for t(= Ts ) we determine that Ts = 0.098 sec. P5.9 We need to track at the rate ω= v 16000 = = 1.78 × 10−3 radians/sec . r 2500 The desired steady-state tracking error is ess ≤ 1 degree = 0.1754 × 10−2 rad . 10 Therefore, with ess = |ω| , Kv © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 196 CHAPTER 5 The Performance of Feedback Control Systems we compute Kv as Kv = 1.78 × 10−3 = 1.02 . 0.1754 × 10−2 This assumes that the system is type 1. (a) The armature controlled DC motorblock diagram is shown in Figure P5.10. P5.10 amplifier R(s ) + Km K 1 J s+b R a+ L as - w(s) Kb back emf FIGURE P5.10 Armature controlled DC motor block diagram. (b) The closed-loop transfer function is T (s) = ω(s) KG(s) = , R(s) 1 + KKb G(s) where G(s) = Km . (Ra + La s)(Js + b) Thus, T (s) = K , s2 + 2s + 1 + K where Ra = La = J = b = Kb = Km = 1. The steady-state tracking error is ess A = lim s(R(s) − Y (s)) = lim s (1 − T (s)) s→0 s→0 s K A = A(1 − T (0)) = 1 − = . 1+K 1+K (c) For a percent overshoot of 15%, we determine that ζ = 0.5. √ From our characteristic polynomial we have 2ζωn = 2 and ωn = 1 + K. Solving for ωn yields ωn = 2, thus K = 3. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 197 Problems P5.11 (a) The closed-loop transfer function is T (s) = K Y (s) = . R(s) s+K To include the initial condition, we obtain the differential equation: ẏ(t) + Ky(t) = Kr(t) . Taking the Laplace transform yields: sY (s) − y(to ) + KY (s) = K A s , where y(to ) = Q. Computing the inverse Laplace transform, L−1 {Y (s)} yields y(t) = A(1 − e−Kt ) + Qe−Kt . Also, the tracking error is given by e(t) = A − y(t) = e−Kt (A − Q) . Thus, the performance index, I is determined to be (for K > 0) I= Z 0 ∞ 2 −2Kt (A − Q) e 2 dt = (A − Q) = Q)2 (A − 2K 1 −2K e−2Kt ∞ 0 . (b) The minimum I is obtained when K = ∞, which is not practical. (c) Set K at the maximum value allowable such that the process does not saturate. For example, if K = 50, then I= P5.12 (A − Q)2 . 100 The optimum ITAE transfer function for a ramp input is T (s) = 3.25ωn2 s + ωn3 . s3 + 1.75ωn s + 3.25ωn2 s + ωn3 The steady-state tracking error, ess = 0, for a ramp input. The step response is shown in Figure P5.12 for ωn = 10. The percent overshoot is P.O. = 39%, and the settling time is Ts = 0.72 s . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 198 CHAPTER 5 The Performance of Feedback Control Systems 1.4 PO=39% 1.2 Ts=0.72s 1 y(t) 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 time (sec) 1.2 1.4 1.6 1.8 2 FIGURE P5.12 Step input system response. The step responses for the actual system and the approximate system are shown in Figure P5.13. It can be seen that the responses are nearly identical. 1 0.9 0.8 0.7 0.6 y(t) P5.13 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 5 FIGURE P5.13 Closed-loop system step response: Actual T(s) (solid line) and second-order approximation (dashed line). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 199 Problems P5.14 Consider L(s) = 2(c1 s + 1) . (s + 1)(s + 2) After cancellation of like factors, we compute H(s)/L(s), H(s) s3 + 7s2 + 24s + 24 = . L(s) (s + 3)(s + 4)2(c1 s + 1) Therefore, M (s) = s3 + 7s2 + 24s + 24 , and ∆(s) = 2[c1 s3 + (7c1 + 1)s2 + (12c1 + 7)s + 12] . Then, following the procedure outlined in Section 5.10, we have M o (0) = 24, M 1 (0) = 24, M 2 (0) = 14, M 3 (0) = 6, and ∆0 (0) = 24, ∆1 (0) = (12c1 + 7)2, ∆2 (0) = 2(2 · (7c1 + 1)), ∆3 (0) = 12c1 . For q = 1: M2 = 240, and ∆2 = 4[144c21 + 25] . Then, equating ∆2 and M2 , we find c1 , c1 = 0.493 . So, L(s) = P5.15 0.986s + 2 0.986(s + 2.028) 2(0.493s + 1) = 2 = . (s + 1)(s + 2) s + 3s + 2 (s + 1)(s + 2) The closed-loop transfer function is T (s) = (s + 4)(s2 K(s + 1) . + s + 10) + K(s + 1) The percent overshoot as function of the gain, K, is shown in Figure P5.15. It can be seen that the percent overshoot decreases as the gain increases approaching a minimum around 85%. The larger the gain, the smaller the percent overshoot. For a gain K ≈ 250, we have essentially minimized the percent overshoot. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 200 CHAPTER 5 The Performance of Feedback Control Systems 160 150 Percent Overshoot (%) 140 130 120 110 100 90 80 0 50 100 150 200 250 300 350 K FIGURE P5.15 Percent overshoot versus the gain, K. P5.16 The open-loop transfer function is G(s) = 10 . (s + 1)(50Cs + 1) Define τ = 50C. Then, the closed-loop transfer function is Vo (s) 10 10/τ = = τ +1 2 Vin (s) (s + 1)(τ s + 1) + 10 s + τ s+ 11 τ . With ωn2 = 11 τ τ +1 1 and ζ = √ = , 2τ ωn 2 we can solve for τ , yielding τ 2 − 20τ + 1 = 0 . Therefore, τ = 19.95 and 0.05. For each value of τ we determine C as follows: τ = 19.95 = 50C, implies C = 0.399F , and τ = 0.05 = 50C, implies C = 1mF . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 201 Problems P5.17 (a) The closed-loop transfer function is T (s) = Y (s) 12K = 2 . R(s) s + 12s + 12K The percent overshoot specification P.O. ≤ 10% implies ζ ≥ 0.59. From the characteristic equation we find that ωn2 = 12K Solving for K yields √ 2(0.59) 12K = 12 and ζωn = 6 . which implies that K = 8.6 . So, any gain in the interval 0 < K < 8.6 is valid. The settling time is Ts = 4/ζωn = 4/6 seconds and satisfies the requirement. Notice that Ts is not a function of K. (b) The sensitivity is T SK (s) = 1 s(s + 12) = 2 1 + G(s) s + 12s + 120 when K = 10. (c) The sensitivity at DC (s = 0) is T SK (0) = 0 . (d) In this case, s = j2π · 1 beat/sec = j2π. So, the sensitivity at s = 2πj is T |SK (j2π)| = P5.18 85.1084 = 0.77 . 110.31 We select L(s) as L(s) = 1 , αs + 1 then H(s) 6αs + 6 = 3 . L(s) s + 6s2 + 11s + 6 Therefore, M (s) = 6αs + 6 , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 202 CHAPTER 5 The Performance of Feedback Control Systems and M o (0) = 6, M 1 (0) = 6α, M 2 (0) = 0. Also, ∆(s) = s3 + 6s2 + 11s + 6 , and ∆o (0) = 6 , ∆1 (0) = 11 , ∆2 (0) = 12. So, computing M2 and ∆2 yields M2 = 36α2 , and ∆2 = 49 . Finally, equating M2 = ∆2 yields 36α2 = 49 , or α = 1.167 . Thus, L(s) = P5.19 1 0.857 = . 1.167s + 1 s + 0.857 (a) The closed-loop transfer function is T (s) = s3 + 6s2 8 . + 12s + 8 (b) The second-order approximation is L(s) = d2 s2 1 , + d1 s + 1 where d1 and d2 are to be determined using the methods of Section 5.10 in Dorf & Bishop. Given M (s) = 8d2 s2 + 8d1 s + 8 ∆(s) = s3 + 6s2 + 12s + 8 we determine that M2 M4 ∆2 ∆4 = −128d2 + 64d21 = 64d22 = 48 = 18 . Equating M2 = ∆2 and M4 = ∆4 , and solving for d1 and d2 yields d1 = 1.35 and d2 = 0.53 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 203 Problems 1 0.9 0.8 0.7 y(t) 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 5 FIGURE P5.19 Closed-loop system step response: Actual T (s) (solid line) and second-order approximation (dashed line). Thus, the second-order approximation is L(s) = 0.53s2 1 . + 1.35s + 1 (c) The plot of the step response for the actual system and the approximate system is shown in Figure P5.19. P5.20 The steady-state error is (s + 5)(s + 11) + K(1 − K1 ) 55 + K(1 − K1 ) = . s→0 (s + 5)(s + 11) + K 55 + K ess = lim To achieve a zero steady-state tracking error, select K1 as follows: K1 = 1 + P5.21 55 . K The closed-loop transfer function is T (s) = s+a . s2 + (2k + a)s + 2ak + 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 204 CHAPTER 5 The Performance of Feedback Control Systems (a) If R(s) = 1/s, we have the tracking error E(s) = R(s) − Y (s) = [1 − T (s)]R(s) or E(s) = s2 + (2k + a − 1)s + 2ak + 1 − a 1 · . s2 + (2k + a)s + 2ak + 1 s From the final value theorem we obtain ess = lim sE(s) = s→0 2ak + 1 − a . 2ak + 1 Selecting k = (a − 1)/(2a) leads to a zero steady-state error due to a unit step input. (b) To meet the percent overshoot specification we desire ζ ≥ 0.69. From T (s) we find ωn2 = 2ak + 1 and 2ζωn = 2k + a. Therefore, solving for a and k yields a = 1.5978 and k = 0.1871 when we select ζ = 0.78. We select ζ > 0.69 to account for the zero in the closed-loop transfer function which will impact the percent overshoot. With a and k, as chosen, we have T (s) = s2 s + 1.598 + 1.972s + 1.598 and the step response yields P.O. ≈ 4%. P5.22 The closed-loop transfer function is T (s) = 2(2s + τ ) . (s + 0.2K)(2s + τ ) + 4 (a) If R(s) = 1/s, we have the unit step response Y (s) = 2(2s + τ ) 1 . (s + 0.2K)(2s + τ ) + 4 s From the final value theorem we obtain yss = lim sY (s) = s→0 2τ . 0.2Kτ + 4 Selecting K = 10− 20/τ leads to yss = 1 and a zero steady-state error due to a unit step input. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 205 Problems (b) The characteristic equation is (s + 0.2K)(2s + τ ) + 4 = 2s2 + (0.4K + τ )s + 0.2Kτ + 4 = 0 . So, with K = 10 − 20/τ , the natural frequency and damping ratio are: ωn = √ τ and ζ = τ 2 + 4τ − 8 . 4τ 3/2 The settling time and percent overshoot are found using the standard design formulas √ π −ζπ 1−ζ 2 p Tp = and P.O. = 100e ωn 1 − ζ 2 with ωn and ζ given above (as a function of τ ). Since the closed-loop system has a zero at s = −τ /2, the formulas for Tp and P.O. will only be approximate. Also, note that for the closed-loop system poles to be in the left half-plane (that √ is, all the poles have negative real parts), we require that τ > 2 3 − 2 ≈ 1.4642. As seen √ in the next chapter, this is the condition for stability. Having τ > 2 3 − 2 insures that the damping ratio ζ is positive. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 206 CHAPTER 5 The Performance of Feedback Control Systems Advanced Problems (a) The steady-state error is ess = lim s(1 − T (s))R(s) = 1 − T (0) = 1 − s→0 108(3) =0. 9(36) (b) Assume the complex poles are dominant. Then, we compute a = 0.75 , ζωn since a = 3, ζ = 0.67 and ωn = 6. Using Figure 5.13 in Dorf & Bishop, we estimate the settling time and percent overshoot to be P.O. = 45% and Ts = 4 = 1 second . ζωn (c) The step response is shown in Figure AP5.1. The actual settling time and percent overshoot are P.O. = 34.4% and Ts = 1.18 second . Step Response 1.4 1.2 1 Amplitude AP5.1 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 FIGURE AP5.1 Closed-loop system step response. 0.8 1 Time (sec) 1.2 1.4 1.6 1.8 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 207 Advanced Problems The closed-loop transfer function is T (s) = s3 + 28s2 5440(τz s + 1) . + (432 + 5440τz )s + 5440 The closed-loop step responses are shown in Figure AP5.2. The performance results are summarized in Table AP5.2. tau=0 (solid) & tau=0.05 (dashed) & tau=0.1 (dotted) & tau=0.5 (dot-dash) 1.4 1.2 1 0.8 y(t) AP5.2 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 Time (sec) 1 1.2 1.4 1.6 FIGURE AP5.2 Closed-loop system step responses. τz Tr Ts P.O. 0 0.16 0.89 32.7% p = −20, −4 ± 16j 0.05 0.14 0.39 4.5% p = −10.4, −8.77 ± 21.06j 0.1 0.10 0.49 0% p = −6.5, −10.74 ± 26.84j 0.5 0.04 1.05 29.2% p = −1.75, −13.12 ± 54.16j TABLE AP5.2 closed-loop poles Performance summary. As τz increases from 0 to 0.1, the P.O. decreases and the response is faster © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 208 CHAPTER 5 The Performance of Feedback Control Systems and more stable. However, as τz is increased beyond 0.1, the P.O. and Ts increase, although Tr continues to decrease. The closed-loop transfer function is T (s) = τp s3 1 . + (1 + 2τp )s2 + 2s + 1 The closed-loop step responses for τp = 0, 0.5, 2, 5 are shown in Figure AP5.3. The performance results are summarized in Table AP5.3. tau=5 (solid) & tau=2 (dotted) & tau=0.5 (dashed) & tau=0 (dot-dash) 1.5 1 y(t) AP5.3 0.5 0 0 5 10 15 20 25 Time (sec) 30 35 40 45 FIGURE AP5.3 Closed-loop system step responses. τp Tr Ts P.O. 0 4 5.8 0% 0.5 3.6 7.4 4.75% p = −2.84, −0.58 ± 0.6j 2 4.6 22.4 27.7% p = −2.14, −0.18 ± 0.45j 5 6 45.8 46% p = −2.05, −0.07 ± 0.3j TABLE AP5.3 Performance summary. closed-loop poles p = −1, −1 50 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 209 Advanced Problems As τp increases, the P.O., Tr and Ts also increase; adding the pole makes the system less stable with more overshoot. The system transfer function is Y (s) = 15 15K R(s) + Td (s) . (s + 5)(s + 7) + 15K (s + 5)(s + 7) + 15K When considering the input response, we set Td (s) = 0, and similarly, when considering the disturbance response, we set R(s) = 0. The closedloop step input and disturbance responses for K = 1, 10, 100 are shown in Figure AP5.4. The performance results are summarized in Table AP5.4. Unit step input response Unit step distrubance response 1.6 0.35 1.4 0.3 1.2 0.25 1 y(t) 0.2 y(t) AP5.4 0.8 0.15 0.6 0.1 0.4 0.05 0.2 0 0 0.2 0.4 0.6 Time (sec) 0.8 1 0 0 0.2 0.4 0.6 Time (sec) 0.8 1 FIGURE AP5.4 Closed-loop system input and disturbance responses (K =1: solid line, K =10: dotted line, and K =100:dashed line). TABLE AP5.4 K ess Ts P.O. |y/d|max 1 0.7 0.45 0% 0.3 10 0.19 0.6 17.3% 0.1 100 0.023 0.59 60.0% 0.01 Performance summary. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 210 CHAPTER 5 The Performance of Feedback Control Systems The best value of the gain is K = 10, which is compromise between (i) percent overshoot, and (ii) disturbance rejection and tracking error. The system transfer function is 50(s + α)(s + 2) R(s) s(s + 3)(s + 4) + 50(s + α)(s + 2) 50s(s + 2) Td (s) . + s(s + 3)(s + 4) + 50(s + α)(s + 2) Y (s) = Disturbance response: alpha=0 (solid) & alpha=10 (dashed) & alpha=100 (dotted) 10 9 8 7 6 5 y(t) AP5.5 4 3 2 1 0 -1 0 0.05 0.1 0.15 0.2 0.25 Time (sec) FIGURE AP5.5 Closed-loop system disturbance response. When considering the input response, we set Td (s) = 0, and similarly, when considering the disturbance response, we set R(s) = 0. The steadystate tracking error is ess = lim s(1 − T (s))R(s) = lim 1 − s→0 s→0 50(s + α)(s + 2) . s(s + 3)(s + 4) + 50(s + α)(s + 2) When α = 0, we have ess = 1 − 100 = 0.11 , 100 + 12 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 211 Advanced Problems and, for α 6= 0 ess = 0 . The closed-loop step input and disturbance responses for α = 0, 10, 100 are shown in Figure AP5.5. For disturbance rejection and steady-state tracking error the best value of the parameter is α = 100 . However, when considering both the disturbance and input response we would select the parameter α = 10 , since it offers a good compromise between input response overshoot (about 5% for α = 10) and disturbance rejection/tracking error. AP5.6 (a) The closed-loop transfer function is T (s) = KKm . KKm + s(s + Km Kb + 0.01) The steady-state tracking error for a ramp input R(s) = 1/s2 is ess = lim s(1 − T (s))R(s) s→0 s + Km Kb + 0.01 s→0 KKm + s(s + Km Kb + 0.01) Km Kb + 0.01 = . KKm = lim (b) With Km = 10 and Kb = 0.05 , we have Km Kb + 0.01 10(0.05) + 0.01 = =1. KKm 10K Solving for K yields K = 0.051 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 212 CHAPTER 5 The Performance of Feedback Control Systems (c) The plot of the step and ramp responses are shown in Figure AP5.6. The responses are acceptable. Step input response 1.4 1.2 y(t) 1 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 Time (sec) 12 14 16 18 20 14 16 18 20 Ramp input response 20 y(t) 15 10 5 0 0 2 4 6 8 10 Time (sec) 12 FIGURE AP5.6 Closed-loop system step and ramp responses. AP5.7 The performance is summarized in Table AP5.7 and shown in graphical form in Fig. AP5.7. K Estimated Percent Overshoot Actual Percent Overshoot 1000 8.8 % 8.5 % 2000 32.1 % 30.2 % 3000 50.0 % 46.6 % 4000 64.4 % 59.4 % 5000 76.4 % 69.9 % TABLE AP5.7 Performance summary. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 213 Advanced Problems 80 Percent Overshoot (% ) 70 Actual P.O . Estimated P.O. 60 50 40 30 20 10 0 1000 2000 3000 K 4000 5000 FIGURE AP5.7 Percent overshoot versus K. The closed-loop transfer function is T (s) = 100K . s(s + 50)(s + 100) + 100K The impact of the third pole is more evident as K gets larger as the estimated and actual percent overshoot deviate in the range 0.3% at K = 1000 to 6.5% at K = 5000. AP5.8 The closed-loop transfer function is T (s) = K(s + 2) . s2 + ( 23 + K)s + 13 + 2K Comparing T (s) to a second-order system we have ωn = q 1/3 + 2K 2/3 + K ζ= p 2 1/3 + 2K For the closed-loop transfer function to have complex roots, we require K 2 − (20/3)K − (8/9) < 0. This occurs when −0.13 ≤ K ≤ 6.8. When K = 1/3, we have the minimum ζ = 0.5, as shown in Figure AP5.8. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 214 CHAPTER 5 The Performance of Feedback Control Systems 0.58 0.57 0.56 ζ 0.55 0.54 0.53 0.52 0.51 0.5 0 0.05 0.1 0.15 0.2 0.25 K 0.3 0.35 0.4 0.45 0.5 FIGURE AP5.8 Damping ratio, ζ, versus K. AP5.9 The closed-loop characteristic equation is s4 + 40s3 + 375s2 + KP s + KI = 0. The desired characteristic equation is √ (s + a)(s + b)(s2 + 2ωn s + ωn2 ) = 0. Expanding the desired characteristic equation and equating terms to the actual characteristic equation yields √ abωn2 = KI , ωn2 (a + b) + 2abωn = KP √ 2(a + b)ωn + ab = 375, √ 2ωn + a + b = 40 This represents 4 equations with 5 unknowns (a, b, KP , KI , and ωn ). We can choose one variable as part of the controller design. Let KI = 0.1KP . Then, solving the 4 equations for the remaining 4 variables yields a = 29.15, b = 0.1, KP = 1720, KI = 172, and ωn = 7.6. The resulting Ts = 1.1s and P.O. = 6.4%, as shown in Figure AP5.9. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 215 Advanced Problems Step Response 1.4 System: sysa Peak amplitude: 1.06 Overshoot (%): 6.39 At time (sec): 0.618 1.2 Amplitude 1 System: sysa Settling Time (sec): 1.09 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 Time (sec) FIGURE AP5.9 Step response for KP = 1720 and KI = 172. 6 7 8 9 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 216 CHAPTER 5 The Performance of Feedback Control Systems Design Problems The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: 26.035 θ(s) = , Va (s) s(s + 33.142) where we neglect the motor inductance Lm . The closed-loop transfer function from the disturbance to the output is θ(s) 26.035 = 2 . Td (s) s + 33.142s + 26.035Ka For a unit step disturbance input the steady-state response is θss = 1 . Ka Therefore, we want to use the maximum Ka while keeping the percent overshoot less than 5%. The step response for the closed-loop system (with the tachometer not in the loop) and Ka = 22 is shown below. Values of Ka greater than 22 lead to overshoots greater than 5%. Step response 1.4 1.2 q(t)/A 1 0.8 0.6 0.4 0.2 0 0 0.1 0.2 0.3 0.4 0.5 Time (sec) 0.6 0.7 0.8 0.9 1 0.7 0.8 0.9 1 Unit disturbance response 0.05 0.04 0.03 q(t) CDP5.1 0.02 0.01 0 0 0.1 0.2 0.3 0.4 0.5 Time (sec) 0.6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 217 Design Problems DP5.1 (a) The closed-loop transfer function is 12.2K 12.2K φ(s) = = 3 . 2 φd (s) s(s + 2.2)(s + 7) + 12.2K s + 9.2s + 15.4s + 12.2K (b) For K = 0.7, we have the characteristic equation s3 + 9.2s2 + 15.4s + 8.54 = 0 , with roots s1 = −7.23 and s2,3 = −0.98 ± 0.46j. For K = 3, we have the characteristic equation s3 + 9.2s2 + 15.4s + 36.6 = 0 , with roots s1 = −7.83 and s2,3 = −0.68 ± 2.05j. And for K = 6, we have the characteristic equation s3 + 9.2s2 + 15.4s + 73.2 = 0 , with roots s1 = −8.4 and s2,3 = −0.4 ± 2.9j. (c) Assuming the complex conjugate pair are the dominant roots, we expect the following: (i) for K = 0.7: P.O.=0.13% and Tp = 6.8 sec (ii) for K = 3: P.O.=35.0% and Tp = 1.5 sec (iii) for K = 6: P.O.=65.2% and Tp = 1.1 sec (d),(e) We select K = 1.71 to have a P.O. = 16% and Tp = 2.18sec. All four cases (K = 0.7, 3, 6, 1.71) are shown in Figure DP5.1. In each case, the approximate transfer function is derived by neglecting the non-dominant real pole and adjusting the gain for zero steady-state error. The approximate transfer functions are 1.18 0.7908 = + 1.965s + 1.18 (s + 0.98 + 0.46j)(s + 0.98 − 0.46j) 4.67 3.299 TK=3 (s) = 2 = s + 1.37s + 4.67 (s + 0.68 + 2.05j)(s + 0.68 − 2.05j) 8.71 6.399 TK=6 (s) = 2 = s + 0.796s + 8.71 (s + 0.4 + 2.9j)(s + 0.4 − 2.9j) 2.77 1.458 TK=1.71 (s) = 2 = s + 1.679s + 2.77 (s + 0.83 + 1.43j)(s + 0.83 − 1.43j) TK=0.7 (s) = s2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 218 CHAPTER 5 The Performance of Feedback Control Systems K=0.7 K=3 1 1 phi 1.5 phi 1.5 0.5 0 0.5 0 5 time (sec) K=6 0 10 1.5 10 0 5 time (sec) 10 1 phi phi 5 time (sec) K=1.71 1.5 2 1 0.5 0.5 0 0 0 5 time (sec) 10 0 FIGURE DP5.1 Step responses (actual response:solid lines; approximate response: dotted lines). DP5.2 The closed-loop transfer function is T (s) = Kωn2 , s3 + 2ζωn s2 + ωn2 s + Kωn2 where ζ = 0.6. From the second-order system approximation, we have Tp = ωn π . 1 − ζ2 p So, with ζ = 0.6 given, we should select ωn “large” to make Tp “small.” Also, from the problem hint, let 0.2 < K/ωn < 0.4 . As a first attempt, we can select ωn = 10. See Figure DP5.8 for various values of K/ωn . Our final selection is K = 3.33 and ωn = 10 . This results in P.O. = 3.6% and Tp = 0.66 second. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 219 Design Problems 1.4 1.2 K/ωn=0.4 1 K/ωn=0.33 y(t) 0.8 K/ω =0.2 n 0.6 0.4 0.2 0 0 0.5 1 1.5 Time (sec) 2 2.5 3 FIGURE DP5.2 Closed-loop system response. DP5.3 The closed-loop transfer function is T (s) = s2 K . + qs + K From the ITAE specification, we desire T (s) = ωn2 . s2 + 1.4ωn s + ωn2 But 2ζωn = 1.4ωn which implies ζ = 0.7 . Since we want Ts ≤ 0.5, we require ζωn ≥ 8. So, ωn ≥ 8 = 11.4 . 0.7 We can select ωn = 12. Then, T (s) = 144 . s2 + 16.8s + 144 Therefore, K = 144 and q = 16.8. The predicted percent overshoot is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 220 CHAPTER 5 The Performance of Feedback Control Systems P.O. = 4.5%. DP5.4 The loop transfer function is Gc (s)G(s) = 10K 10K/70 = . (s + 70)(s + 3)(s + 7) (s/70 + 1)(s + 3)(s + 7) The second-order approximation is obtained by neglecting the fastest firstorder pole. Thus, K/7 . (s + 3)(s + 7) Gc (s)G(s) ≈ The closed-loop transfer function is T (s) = s2 K/7 . + 10s + 21 + K/7 When ζ ≥ 0.52, we have less than 15% overshoot. So, we have 2ζωn = 10 and ωn = q 21 + K/7. Eliminating ωn and solving for K (with P.O. ≤ 15%) yields K ≤ 500.19 . Also, Kp = lim GGc (s) = s→0 K 7(21) and ess = 1 1 = < 0.12 K 1 + Kp 1 + 147 implies K ≥ 1078 . Therefore, we have an inconsistency. We require 1078 ≤ K to meet the steady-state requirement and K ≤ 500.18 to meet the percent overshoot requirement. It is not possible to meet both specifications. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 221 Design Problems DP5.5 The closed-loop characteristic equation is 1 + K1 G1 (s) + K2 G1 G2 (s) = 1 + 2K2 K1 − =0 s(s + 1) s(s + 1)(s + 2) or s3 + 3s2 + (2 + K1 )s + 2(K1 − K2 ) = 0 . Assuming that K1 > 0 and K2 > 0, the range of the gains for stability is 0 < K2 < K1 . DP5.6 The closed-loop transfer function is T (s) = s2 K1 . + (K1 K2 + 1)s + K1 The percent overshoot specification P.O. ≤ 2% is satisfied when ζ > 0.78. The peak time specification Tp ≤ 0.5 s is satisfied when ωn = 10 and ζ = 0.78. So, given K1 = ωn2 and K1 K2 + 1 = 2ζωn , we determine that the specifications are satisfied when K1 = 100 and K2 = 0.15 . DP5.7 The plant is G(s) = 2 s(s + 1)(s + 4) and the PD controller is Gc (s) = KD s + KP . The characteristic equation is s3 + 6s2 + (8 + 2KD )s + 2KP = 0. The desired characteristic equation is (s + a)(s2 + 2ζωn s + ωn2 ) = s3 + (2ζωn + a)s2 + (ωn2 + 2ζωn a)s + aωn2 = 0. Equating the desired characteristic equation to the actual characteristic © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 222 CHAPTER 5 The Performance of Feedback Control Systems equation yields 2ζωn + a = 6, ωn2 + 2ζωn a = 8 + 2KD , aωn2 = 2KP , where ζ = 0.69 and ωn = 3 to meet the design specifications. This represents 3 equations in 3 unknowns (a, KD , and KP ). Solving yields a = 1.86, KD = 4.35 and KP = 8.37. The step response is shown in Figure DP5.7. Step Response 1.4 System: sys_cl Peak amplitude: 1.04 Overshoot (%): 4.28 At time (sec): 1.48 1.2 Amplitude 1 System: sys_cl Settling Time (sec): 1.96 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 Time (sec) 2 2.5 3 FIGURE DP5.7 Step response withKD = 4.35 and KP = 8.37. DP5.8 The closed-loop transfer function is T (s) = s2 K + 6s + 5 + K The damping ratio and natural frequency is ζ=√ 3 K +5 and ωn = √ K +5 Using the design formulas for second-order systems we have √ 4 2 P O = 100e−ζπ/ 1−ζ and Ts = . ζωn We know that the formula for Ts is approximate and that the formulas apply only to systems with ζ < 1. For K = 1 the closed-loop poles are © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 223 Design Problems both real, so there is no overshoot and the design formula for settling time does not apply. Thus we obtain the results shown in Table DP5.8. We can choose K = 10 as a good trade-off between percent overshoot, settling time, and steady-state tracking error. The disturbance response is shown in Figure DP5.8. TABLE DP5.8 Step response for K=1, 10, and 20. K % P.O. Ts , sec Ts , sec Estimated Actual ess 1 0 - 3.24 0.83 10 2.13 1.33 1.38 0.33 20 9.48 1.33 1.19 0.20 1 0.9 0.8 0.7 y(t) 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.2 0.4 0.6 0.8 1 Time (sec) 1.2 FIGURE DP5.8 Closed-loop system disturbance response for K = 10. 1.4 1.6 1.8 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 224 CHAPTER 5 The Performance of Feedback Control Systems Computer Problems With the impulse input we have R(s) = 1. The transfer function is Y (s) = 15 15 R(s) = . (s + 3)(s + 5) (s + 3)(s + 5) Therefore, taking the inverse Laplace transforms yields the output response: y(t) = 15 −3t 15 −5t e − e . 2 2 The impulse response and the analytic response is shown in Figure CP5.1. n=15; d=[1 8 15]; t=[0:0.1:6]; ya=(15/2)*exp(-3.*t)-(15/2)*exp(-5.*t); sys = tf(n,d) y=impulse(sys,t); plot(t,y,t,ya,'o') xlabel('Time (sec)'), ylabel('y(t)'), legend('Computer','Analytic',-1) 1.4 Computer Analytic 1.2 1 0.8 y(t) CP5.1 0.6 0.4 0.2 0 0 FIGURE CP5.1 Impulse responses. 1 2 3 Time (sec) 4 5 6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 225 Computer Problems CP5.2 The ramp response is shown in Figure CP5.2. The unity feedback system is type 2, so that the steady-state tracking error is lim ess → 0. t→∞ 60 50 Amplitude 40 n=[1 10]; d=[1 15 0 0]; t=[0:0.1:50]; sys= tf(n,d); sys_cl = feedback(sys,[1]); u=t; lsim(sys_cl,u,t); 30 20 10 0 0 5 10 15 20 25 Time (sec) 30 35 40 45 50 FIGURE CP5.2 Ramp responses. CP5.3 The m-file script and the four plots are shown in Figure CP5.3. The plots can be compared to Figure 5.17 in Dorf & Bishop. wn=2, zeta=0 2 1 1 0 0 -1 -1 -2 0 5 10 15 20 wn=1, zeta=0 1 wn=2, zeta=0.1 2 -2 0 5 15 20 wn=1, zeta=0.2 1 0.5 10 0.5 0 0 -0.5 -1 0 FIGURE CP5.3 Impulse responses. 5 10 15 20 -0.5 0 5 10 15 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 226 CHAPTER 5 The Performance of Feedback Control Systems w1=2; z1=0; w2=2; z2=0.1; w3=1; z3=0; w4=1; z4=0.2; t=[0:0.1:20]; % num1=[w1^2]; den1=[1 2*z1*w1 w1^2]; sys1 = tf(num1,den1); [y1,x1]=impulse(sys1,t); % num2=[w2^2]; den2=[1 2*z2*w2 w2^2]; sys2 = tf(num2,den2); [y2,x2]=impulse(sys2,t); % num3=[w3^2]; den3=[1 2*z3*w3 w3^2]; sys3 = tf(num3,den3); [y3,x3]=impulse(sys3,t); % num4=[w4^2]; den4=[1 2*z4*w4 w4^2]; sys4 = tf(num4,den4); [y4,x4]=impulse(sys4,t); % clf subplot(221),plot(t,y1),title('wn=2, zeta=0') subplot(222),plot(t,y2),title('wn=2, zeta=0.1') subplot(223),plot(t,y3),title('wn=1, zeta=0') subplot(224),plot(t,y4),title('wn=1, zeta=0.2') FIGURE CP5.3 CONTINUED: Impulse response m-file script. CP5.4 The closed-loop system is T (s) = s2 21 . + 2s + 21 Therefore, the natural frequency is √ ωm = 21 = 4.58 and the damping ratio is computed as 2ζωn = 2 , which implies ζ = 0.218 . The percent overshoot is estimated to be √ 2 P.O. = 100e−ζπ/ 1−ζ = 50% , since ζ = 0.218. The actual overshoot is shown in Figure CP5.4. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 227 Computer Problems Step Response From: U(1) 1.5 To: Y(1) 1 Amplitude numc=[21]; denc=[1 0]; sysc = tf(numc,denc); numg=[1]; deng=[1 2]; sysg = tf(numg,deng); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]) step(sys_cl) 0.5 0 0 1 2 3 4 5 6 Time (sec.) FIGURE CP5.4 Impulse responses. CP5.5 For a step input, the optimum ITAE characteristic equation is s3 + 1.75ωs2 + 2.15ω 2 s + ω 3 = 0 . Examining Figure 5.30 for n=3 in Dorf & Bishop, we estimate that ωTs = 8. So, once we decide on the desired Ts we can estimate ω. For this design we let Ts =8 so that ω = 1. Computing the desired characteristic equation and the actual characteristic equation and comparing the coefficients leads to the following relationships: p = 1.75ω − 2ζωn K = (2.15ω 2 − ωn2 − 2ζωn p)/ωn2 z = (ω 3 − pωn2 )/(Kωn2 ) where ζ = 0.59 and ωn = 0.45. The controller and prefilter are Gc (s) = 6.42 s + 0.58 s + 1.22 and Gp (s) = The unit step response is shown in Figure CP5.5. 1 . 1.3s + 0.75 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 228 CHAPTER 5 The Performance of Feedback Control Systems Step Response 1.4 wn=0.45; zeta=0.59 ng=wn^2; dg=[1 2*zeta*wn wn^2]; sysg=tf(ng,dg); Ts=8; w=8/Ts; p=1.75*w-2*zeta*wn; K=(2.15*w^2-wn^2-2*zeta*wn*p)/wn^2; z=(w^3-p*wn^2)/(K*wn^2); nc=K*[1 z]; dc=[1 p]; sysc=tf(nc,dc); sys=series(sysc,sysg); syscl=feedback(sys,1); [num,den]=tfdata(syscl,'v'); sysp=tf([den(end)],num); step(syscl*sysp) System: untitled1 Peak amplitude: 1.02 Overshoot (%): 1.98 At time (sec): 4.68 1.2 1 Amplitude System: untitled1 Settling Time (sec): 7.54 0.8 0.6 0.4 0.2 0 0 2 4 6 Time (sec) 8 10 12 FIGURE CP5.5 Closed-loop system step response m-file script. CP5.6 The unit step response is shown in Figure CP5.6. The performance numbers are as follows: Mp = 1.16, Tp = 0.73, and Ts = 1.62. Step Response 1.4 System: sys_cl Peak amplitude: 1.16 Overshoot (%): 16.3 At time (sec): 0.73 1.2 1 System: sys_cl Settling Time (sec): 1.62 Amplitude numg=[25]; deng=[1 5 0]; sys = tf(numg,deng); sys_cl = feedback(sys,[1]); t=[0:0.01:2]; step(sys_cl,t); 0.8 0.6 0.4 0.2 0 0 FIGURE CP5.6 Closed-loop system step response m-file script. 0.2 0.4 0.6 0.8 1 Time (sec) 1.2 1.4 1.6 1.8 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 229 Computer Problems The m-file script and the simulations are shown in Figure CP5.7. % Part (a) numc=[2]; denc=[1]; sys_c = tf(numc,denc); nums=[-10]; dens=[1 10]; sys_s = tf(nums,dens); numg=[-1 -5]; deng=[1 3.5 6 0]; sys_g = tf(numg,deng); sysa = series(sys_c,sys_s); sysb = series(sysa,sys_g); sys = feedback(sysb,[1]); f=0.5*pi/180; % Convert to rad/sec t=[0:0.1:10]; u=f*t; [y,x]=lsim(sys,u,t);(y(length(t),1)-u(1,length(t)))*180/pi subplot(211) plot(t,y*180/pi,t,u*180/pi,'--'), grid xlabel('Time (sec)'),ylabel('theta') title('Constant gain C(s) = 2: theta (solid) & input (dashed)') % Part (b) numc=[2 1]; denc=[1 0]; sys_c = tf(numc,denc); [numa,dena]=series(numc,denc,nums,dens); sysa = series(sys_c,sys_s); sysb = series(sysa,sys_g); sys = feedback(sysb,[1]); [y,x]=lsim(sys,u,t);(y(length(t),1)-u(1,length(t)))*180/pi subplot(212), plot(t,y*180/pi,t,u*180/pi,'--'), grid xlabel('Time (sec)'),ylabel('theta') title('PI controller C(s) = 2 + 1/s: theta (solid) & input (dashed)') Constant gain C(s) = 2: theta (solid) & input (dashed) 5 theta 4 3 2 1 0 0 1 2 3 4 5 Time (sec) 6 7 8 9 10 8 9 10 PI controller C(s) = 2 + 1/s: theta (solid) & input (dashed) 6 5 4 theta CP5.7 3 2 1 0 0 1 2 3 4 5 Time (sec) 6 7 FIGURE CP5.7 Closed-loop system response to a ramp input for two controllers. For the constant gain controller, the attitude error after 10 seconds is ess = −0.3 deg. On the other hand, the PI controller has a zero steadystate error ess = 0 deg. So, we can decrease the steady-state error by © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 230 CHAPTER 5 The Performance of Feedback Control Systems using a more sophisticated controller, in this case a PI controller versus a constant gain controller. CP5.8 The closed-loop characteristic equation is s3 + 12s2 + 610s + 500 = (s + 0.8324)(s2 + 11.1676s + 600.7027) = 0 . The natural frequency and damping ratio of the complex roots are ωn = 24.5 and ζ = 0.23. From this we predict Mp = 1.48, Ts = 0.72, and Tp = 0.13. The actual response is shown in Figure CP5.8. The differences Step Response From: U(1) 1.4 1.2 1 0.8 To: Y(1) theta dot numg=[100 100]; deng=[1 2 100]; sysg = tf(numg,deng); numc=[0.1 5]; denc=[1 0]; sysc = tf(numc,denc); sys_o = series(sysg,sysc); sys_cl = feedback(sys_o,[1]) t=[0:0.01:3]; step(sys_cl,t); ylabel('theta dot') 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 Time (sec.) FIGURE CP5.8 Missile rate loop autopilot simulation. can be explained by realizing that the system is not a second-order system. The closed-loop system actually has two zeros, one real pole, and two complex-conjugate poles: T (s) = (s + 50)(s + 1) . (s + 0.8324)(s2 + 11.1676s + 600.7027) The effect of the pole at s = −0.8324 is diminished by the zero at s = −1. The third pole and the zeros affect the overall response such that the analytic formulas for second-order systems are not exact predictors of the transient response. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 231 Computer Problems CP5.9 Figure CP5.9 shows an m-file to compute the closed-loop transfer function and to simulate and plot the step response. Step Response 1 System: sys Peak amplitude: 0.979 Overshoot (%): 95.7 At time (sec): 0.533 0.9 0.8 numg=[10]; deng=[1 10]; sysg = tf(numg,deng); numh=[0.5]; denh=[10 0.5]; sysh = tf(numh,denh); sys = feedback(sysg,sysh) step(sys); Amplitude 0.7 Transfer function: 100 s + 5 --------------------10 s^2 + 100.5 s + 10 0.6 System: sys Settling Time (sec): 39.1 0.5 0.4 0.3 0.2 0.1 0 0 10 20 30 Time (sec) 40 50 60 FIGURE CP5.9 M-file to compute the transfer function and to simulate the step response. CP5.10 Figure CP5.10 shows an m-file to compute the closed-loop transfer function and to simulate and plot the ramp response. The steady-state error Linear Simulation Results 100 90 80 70 Amplitude numg=[10]; deng=[1 20 75 0]; sysg = tf(numg,deng); sys = feedback(sysg,1) t=[0:0.1:100]; u=t; % Unit ramp input lsim(sys,u,t); 60 50 40 30 20 10 0 0 10 20 30 40 50 Time (sec) FIGURE CP5.10 M-file to compute the transfer function and to simulate the ramp response. is 7.5. 60 70 80 90 100 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 232 CHAPTER 5 CP5.11 The Performance of Feedback Control Systems Figure CP5.11 shows an m-file to compute the closed-loop transfer function and to simulate and plot the impulse, step, and ramp responses. Notice that the closed-loop system is unstable. 0 Amplitude -10 0 2 4 6 0 2 4 6 0 2 4 6 8 10 Time (sec) Step Response 12 14 16 18 20 8 10 12 Time (sec) Linear Simulation Results 14 16 18 20 14 16 18 20 10 0 -10 Amplitude numg=[1]; deng=[1 2 0]; sysg = tf(numg,deng); numc=[0.5 2]; denc=[1 0]; sysc = tf(numc,denc); syss=series(sysg,sysc); sys = feedback(syss,1) t=[0:0.1:20]; subplot(311) impulse(sys,t); subplot(312) step(sys,t); subplot(313) u=t; % Unit ramp input lsim(sys,u,t); Amplitude Impulse Response 10 40 20 0 8 10 Time (sec) 12 FIGURE CP5.11 M-file to compute the transfer function and to simulate the ramp response. CP5.12 Figure CP5.12 shows an m-file to simulate and plot the step response for the original system and the 2nd-order approximation. For the original system, we find Ts = 2.28 and P.O. = 80.6%. For the 2nd-order approximation we find Ts = 2.16 and P.O. = 101% © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 233 Computer Problems 2.5 2nd order approximation 2 3rd order system response Step response num=77*[1 2]; den=conv([1 7],[1 4 22]); sys = tf(num,den) na=(77/7)*[1 2]; da=[1 4 22]; sysa=tf(na,da); t=[0:0.01:5]; y=step(sys,t); ya=step(sysa,t); plot(t,y,t,ya,'--') xlabel('Time (s)'), ylabel('Step response') 1.5 1 0.5 0 0 1 2 3 Time (s) FIGURE CP5.12 Step response. 4 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 6 The Stability of Linear Feedback Systems Exercises E6.1 The Routh array is s3 1 K +1 s2 K 6 s1 b 0 so 6 where b= K(K + 1) − 6 . K For stability, we require K > 0 and b > 0. Therefore, using the condition that b > 0, we obtain K2 + K − 6 > 0 , and solving for K yields K > 2 and K < −3. We select K > 2, since we also have the condition that K > 0. E6.2 The Routh array is s3 1 2 s2 10 30 s1 -1 0 so 30 The system is unstable since the first column shows two sign changes. 234 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 235 Exercises E6.3 The Routh array is s4 1 32 s3 10 37 s2 28.3 20 s1 29.9 s0 20 20 By the Routh-Hurwitz criterion, the system is stable (i.e., all the numbers in the first column are positive). E6.4 The closed-loop transfer function is T (s) = s3 + −K(s − 2) . + (4 − K)s + 2K 5s2 Therefore, the characteristic equation is s3 + 5s2 + (4 − K)s + 2K = 0 . The corresponding Routh array is given by s3 1 s2 4−K 5 2K s1 b 0 so 2K where b= 5(4 − K) − 2K 20 − 7K = . 5 5 For stability we require K > 0 and b > 0 . Thus, the range of K for stability is 0 < K < 20/7. E6.5 The closed-loop transfer function is T (s) = s3 + 10s2 K . + 27s + 18 + K When K = 20, the roots of the characteristic polynomial are s1,2 = −1.56 ± j1.76 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 236 CHAPTER 6 The Stability of Linear Feedback Systems and s3 = −6.88 . E6.6 When K = 252, the roots of the characteristic equation are on the imaginary axis. The roots are s1,2 = ±j5.2 E6.7 and s3 = −10 . (a) The closed-loop system characteristic equation is 1 + GH(s) = 1 + or K(s + 2) =0, s(s − 1) s2 + (K − 1)s + 2K = 0 . √ We have the relationships ωn = 2K and 2ζωn = K − 1, where ζ = 0.707. Thus, 1 √ 2 √ 2K = K − 1 , 2 or 2 √ 2 2 = K −1 √ 2K 2 , and K 2 − 6K + 1 = 0 . Solving for K yields K = 5.83 and K = 0.17. However, for stability we require K > 1 (from the Routh array), so we select K = 5.83. √ (b) The two roots on the imaginary axis when K = 1 are s1,2 = ±j 2. E6.8 The closed-loop system characteristic equation is 3 + 20s2 + (100 + K)s + 20K = 0 . The corresponding Routh array is s3 1 (100 + K) s2 20 20K s1 b 0 so 20K © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 237 Exercises where b= 20(100 + K) − 20K 20(100) = = 100 . 20 20 Therefore, the system is stable for all K > 0. E6.9 The characteristic equation is s3 + 2s2 + (K + 1)s + 8 = 0 , and the Routh array is given by s3 1 K +1 s2 2 8 s1 b 0 so 8 where b= 2(K + 1) − 8 =K −3 . 2 Setting b = 0, yields K − 3 = 0 or K>3. E6.10 Stable with your eyes open and (generally) unstable with your eyes closed. E6.11 The system is unstable. The poles are s1 = −5.66, s2 = −0.90 and s3,4 = 0.28 ± j0.714. E6.12 The characteristic equation is s2 + as + b = 0, so, the Routh array is s2 1 b s1 a 0 so b The system is stable if and only if a > 0 and b > 0. For a second-order system, a necessary and sufficient condition for stability is that all the coefficients have the same sign. E6.13 The characteristic equation is s2 + (KD + 2)s + 4KP = 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 238 CHAPTER 6 The Stability of Linear Feedback Systems The Routh array is s2 1 4KP s1 KD + 2 0 so 4KP The system is stable if and only if KP > 0 and KD > −2. E6.14 The characteristic equation associated with the system matrix is s3 + 3s2 + 5s + 6 = 0 . The roots of the characteristic equation are s1 = −2 and s2,3 = −5±j1.66. The system is stable. E6.15 The roots of q(s) are s1 = −4, s2 = −3, s3,4 = −1 ± j2 and s5,6 = ±j0.5. The system is marginally stable. The Routh array is s6 1 31.25 67.75 s5 9 61.25 14.75 s4 24.44 66.11 15 s3 31.909 9.2273 0 s2 60 15 s1 0 0 15 so The auxillary equation is 60s2 + 15 = 0 . Solving the auxillary equation yields two roots at s1,2 = ±j0.5. After accounting for the row of zeros, the completed Routh array verifies that the system has no poles in the right half-plane. E6.16 The Routh array is s4 1 45 s3 9 87 s2 35.33 50 s1 74.26 0 so 50 50 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 239 Exercises The system is stable. The roots of q(s) are s1,2 = −3 ± j4, s3 = −2 and s4 = −1. E6.17 The characteristic equation is s3 + 7s2 + 36s + 24 = 0 . The system is stable. The roots of the characteristic equation are s1 = −0.77, s2,3 = −3.12 ± 4.64i. E6.18 The roots of q(s) are s1 = −20 and s2,3 = ±j2.24. The system is marginally stable. The Routh array is s3 1 5 s2 20 100 s1 0 0 so The auxillary equation is 20s2 + 100 = 0 . The roots are s = ±j2.24. So, the system has roots at s = ±j2.24. Completing the Routh array (after accounting for the row of zeros) verifies that no poles lie in the right half-plane. E6.19 (a) Unstable. (b) Stable. (c) Stable. E6.20 (a) The roots are s1,2 = −2 and s3 = −1. (b) The roots are s1,2,3 = −3. E6.21 The characteristic equation is (sn − 2)3 + 10(sn − 2)2 + 29(sn − 2) + K = 0 or s3n + 4s2n + sn − 26 + K = 0 . The Routh array is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 240 CHAPTER 6 The Stability of Linear Feedback Systems s3 1 1 s2 4 s1 30−K 4 K − 26 so K − 26 0 If K = 30, then the auxillary equation is 4s2n + 4 = 0 or sn = ±j. Therefore, s = sn − 2 implies s = −2 ± j. E6.22 This system is not stable. The output response to a step input is a ramp y(t) = kt. E6.23 The characteristic polynomial is s3 + 4s2 + ks + 8 = 0 . The Routh array is s3 1 k s2 4 8 s1 4k−8 4 so 8 So, k > 2 for stability. E6.24 The transfer function is G(s) = C(sI − A)−1 B + D k 0 k s+k 2 s + ks + k = [ 1 0 0 ] −k 0 −1 0 s −1 = [ 1 0 0 ] 0 s −ks 1 s+k s2 + ks 1 0 s 0 −ks − k s2 1 1 ∆(s) where ∆(s) = s3 + ks2 + ks + k. Thus, the transfer function is G(s) = The Routh array is s3 + ks2 1 . + ks + k © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 241 Exercises s3 1 k s2 k k s1 k−1 so k For stability k > 1. E6.25 The closed-loop transfer function is T (s) = Ks + 1 . s2 (s + p) + Ks + 1 Therefore, the characteristic equation is s3 + ps2 + Ks + 1 = 0 . The Routh array is s3 1 K s2 p 1 s1 (pK − 1)/p so 1 We see that the system is stable for any value of p > 0 and pK − 1 > 0. E6.26 The closed-loop transfer function is T (s) = 10 . 2s2 + (K − 20)s + 10 − 10K Therefore, the characteristic equation is 2s2 + (K − 20)s + 10 − 10K = 0 . The Routh array is s2 2 s1 K − 20 so 10-10K 10 − 10K We see that the system is stable for any value of K > 20 and any K < 1. Therefore, the system is unstable for all K > 0 since the gain K cannot be simultaneously greater than 20 and less than 1. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 242 CHAPTER 6 The Stability of Linear Feedback Systems Problems P6.1 (a) Given s2 + 5s + 2 , we have the Routh array s2 1 2 s1 5 0 so 2 Each element in the first column is positive, thus the system is stable. (b) Given s3 + 4s2 + 8s + 4 , we have the Routh array s3 1 8 s2 4 4 s1 7 0 so 4 Each element in the first column is positive, thus the system is stable. (c) Given s3 + 2s2 − 6s + 20 , we determine by inspection that the system is unstable, since it is necessary that all coefficients have the same sign. There are two roots in the right half-plane. (d) Given s4 + s3 + 2s2 + 12s + 10 , we have the Routh array s4 1 2 10 s3 1 12 0 s2 -10 10 0 s1 13 0 so 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 243 Problems There are two sign changes in the first column, thus the system is unstable with two roots in the right half-plane. (e) Given s4 + s3 + 3s2 + 2s + K , we have the Routh array s4 1 3 K s3 1 2 0 s2 1 K s1 2−K 0 so K Examining the first column, we determine that the system is stable for 0 < K < 2. (f) Given s5 + s4 + 2s3 + s + 6 , we know the system is unstable since the coefficient of the s2 term is missing. There are two roots in the right half-plane. (g) Given s5 + s4 + 2s3 + s2 + s + K , we have the Routh array s5 1 2 1 s4 1 1 K s3 1 s2 1−K K K s1 −K 0 so K Examining the first column, we determine that for stability we need K > 0 and K < 0. Therefore the system is unstable for all K. P6.2 (a) The closed-loop characteristic polynomial is s4 + 27.88s3 + 366.4s2 + 1500s + 1500ka = 0 . The Routh array is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 244 CHAPTER 6 The Stability of Linear Feedback Systems s4 1 366.4 s3 27.88 1500 s2 312.6 1500ka s1 b so 1500ka 1500ka where b = 1500 − 133.78ka . Examining the first column of the Routh array, we find that b > 0 and 1500ka > 0 for stability. Thus, 0 < ka < 11.21 . (b) With Ts = 1.5 = 4 , ζωn we determine that ζωn = 2.67 . So, shift the axis by s = so − 2.67, and (so − 2.67)4 + 27.88(so − 2.67)3 + 366.4(so − 2.67)2 + 1500(so − 2.67) + 1500ka = s4o + 17.2s3o + 185.85s2o + 63.55so − 1872.8 + 1500ka . The Routh array is s4 1 185.85 s3 17.2 63.55 s2 182.16 1500ka -1872.8 s1 b so 1500ka -1872.8 1500ka -1872.8 where b = 240.38 − 141.63ka . Examining the first column of the Routh array, we find that b > 0 and 1500ka − 1872.8 > 0. Thus, 1.25 < ka < 1.69. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 245 Problems P6.3 (a) Given G(s) = K , (s + 1)(s + 2)(0.5s + 1) and H(s) = 1 , 0.005s + 1 the closed-loop transfer function is T (s) = 0.0025s4 K(0.005s + 1) . + 0.5125s3 + 2.52s2 + 4.01s + 2 + K Therefore, the characteristic equation is 0.0025s4 + 0.5125s3 + 2.52s2 + 4.01s + (2 + K) = 0 . The Routh array is given by s4 0.0025 2.52 2+K s3 0.5125 4.01 0 s2 2.50 2+K s1 3.6 − 0.205K 0 so 2+K Examining the first column, we determine that for stability we require −2 < K < 17.6 . (b) Using K = 9, the roots of the characteristic equation are s1 = −200 , s2,3 = −0.33 ± 2.23j , and s4 = −4.35 . Assuming the complex roots are dominant, we compute the damping ratio ζ = 0.15. Therefore, we estimate the percent overshoot as √ 2 P.O. = 100e−πζ/ 1−ζ = 62% . The actual overshoot is 27%, so we see that assuming that the complex poles are dominant does not lead to accurate predictions of the system response. P6.4 (a) The closed-loop characteristic equation is 1 + GH(s) = 1 + K(s + 40) =0, s(s + 10)(s + 20) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 246 CHAPTER 6 The Stability of Linear Feedback Systems or s3 + 30s2 + 200s + Ks + 40K = 0 . The Routh array is s3 1 200 + K s2 30 40K s1 200 − so K 3 0 40K Therefore, for stability we require 200 − K/3 > 0 and 40K > 0. So, the range of K for stability is 0 < K < 600 . (b) At K = 600, the auxilary equation is 30s2 + 40(600) = 0 or s2 + 800 = 0 . The roots of the auxiliary equation are s = ±j28.3 . (c) Let K = 600/2 = 300. Then, to the shift the axis, first define so = s + 1. Substituting s = so − 1 into the characteristic equation yields (so −1)3 +30(so −1)2 +500(so −1)+12000 = s3o +27s2o +443so +11529 . The Routh array is s3 1 443 s2 27 11529 s1 16 0 so 11529 All the elements of the first column are positive, therefore all the roots lie to left of s = −1. We repeat the procedure for s = so − 2 and obtain s3o + 24s2o + 392so + 10992 = 0 . The Routh array is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 247 Problems s3 1 392 s2 24 10992 s1 -66 0 so 10992 There are two sign changes in the first column indicating two roots to right of s = −2. Combining the results, we determine that there are two roots located between s = −1 and s = −2. The roots of the characteristic equation are s1 = −27.6250 and s2,3 = −1.1875 ± 20.8082j . We see that indeed the two roots s2,3 = −1.1875±20.8082j lie between -1 and -2. P6.5 (a) Given the characteristic equation, s3 + 3s2 + 4s + 2 = 0 , we compute the roots s1 = −1, and s2,3 = −1 ± j. (b) The roots of the characteristic equation s4 + 9s3 + 30s2 + 42s + 20 = 0 are s1 = −1, s2 = −2, and s3,4 = −3 ± j1. (c) The roots of the characteristic equation s3 + 19s2 + 110s + 200 = 0 are s1 = −4, s2 = −5, and s3 = −10. P6.6 (a) The characteristic equation is 1 + G(s) = 0 , or s3 + s2 + 10s + 2 = 0 . The roots are: s1 = −0.2033, and s2,3 = −0.3984 ± j3.1112. (b) The characteristic equation is s4 + 10s3 + 35s2 + 50s + 24 = 0 . The roots are s1 = −1, s2 = −2, s3 = −3, and s4 = −4. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 248 CHAPTER 6 The Stability of Linear Feedback Systems (c) The characteristic equation is s3 + 11s2 + 29s + 6 = 0 . The roots are s1 = −0.2258, s2 = −3.8206 and s3 = −6.9536. P6.7 (a) The closed-loop characteristic equation is s3 + 101s2 + (100 + 10KKa )s + 100KKa = 0 . The Routh array is s3 1 100 + 10KKa s2 101 100KKa s1 b so 100KKa where b = 100 + 910 KKa > 0 . 101 Thus, examing the first column, we determine that KKa > 0 stabilizes the system. (b) The tracking error is e(s) = lim s(1 − T (s)) s→0 100 100 = . 2 s KKa We require E(s) < 1o = 0.01745. So, KKa > 100 = 5729 . 0.01745 When KKa = 5729, the roots of the characteristic polynomial are s1 = −10.15 P6.8 and s2,3 = −45.43 ± j233.25 . (a) The closed-loop characteristic equation is 1+ K =0, (0.5s + 1)(s + 1)( 14 s + 1) or s3 + 7s2 + 14s + 8(1 + K) = 0 . The Routh array is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 249 Problems s3 1 14 s2 7 8(1 + K) s1 b so 8(1 + K) where 7(14) − 8(1 + K) . 7 b= For stability, we require b > 0 and 8(1 + K) > 0. Therefore, the range of K for stability is −1 < K < 11.25 . (b) Let K = 11.25/3 = 3.75. Then, the closed-loop transfer function is T (s) = s3 + 3.37 . + 14s + 38 7s2 The settling time to a step input is Ts ≈ 6 seconds. (c) We want Ts = 4 sec, so Ts = 4 = 4 ζωn implies ζωn = 1 . Our desired characteristic polynomial is (s + b)(s2 + 2ζωn s + ωn2 ) = s3 + (2 + b)s2 + (ωn2 + 2b)s + bωn2 where we have used the fact that ζωn = 1 and ωn and b are to be determined. Our actual characteristic polynomial is s3 + 7s2 + 14s + 8(1 + K) = 0 . Comparing the coefficients of the actual and desired characteristic polynomials, we find the following relationships: ωn2 2+b=7 + 2b = 14 bωn2 = 8(1 + K) . Solving these three equations yields b=5, ωn = 2 and K = 1.5 . The actual settling time is Ts = 4.17 sec. This is not exactly our © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 250 CHAPTER 6 The Stability of Linear Feedback Systems desired Ts since we have the contribution of the additional pole at s = −5. The closed-loop poles are s1 = −5 and s2,3 = −1 ± 1.73j . P6.9 (a) The closed-loop characteristic equation is 1 + GH(s) = 1 + 10K , (s + 100)(s + 20)2 or s3 + 140s2 + 4400s + 40000 + 10K = 0 . The Routh array is s3 1 4400 s2 140 40000 + 10K s1 b so 40000 + 10K where b= 140(4400) − (40000 + 10K) . 140 Examining the first column and requiring all the terms to be positive, we determine that the system is stable if −4000 < K < 57600 . (b) The desired characteristic polynomial is (s+b)(s2 +1.38ωn s+ωn2 ) = s3 +(1.38ωn +b)s2 +(ωn2 +1.38ωn b)s+bωn2 where we have used the fact that ζ = 0.69 to achieve a 5% overshoot, and ωn and b are to be determined. The actual characteristic polynomial is s3 + 140s2 + 4400s + 40000 + 10K = 0 . Equating the coefficients of the actual and desired characteristic polynomials, and solving for K, b, and ωn yields b = 104.2 , ωn = 25.9 So, a suitable gain is K = 3003. and K = 3003 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 251 Problems P6.10 (a) The closed-loop characteristic equation is s4 + 7s3 + 20s2 + (24 + K)s + 10K = 0 . The Routh array is s4 1 20 10K s3 7 24 + K 0 s2 116−K 7 10K s1 b so 10K where b= 116−K 7 (24 + K) − 70K Setting b > 0 yields 116−K 7 . 2784 − 398K − K 2 > 0 , which holds when −404.88 < K < 6.876 . Examining the first column, we also find that K < 116 and K > 0 for stability. Combining all the stability regions, we determine that for stability 0 < K < 6.876 . (b) When K = 6.876, the roots are s1,2 = −3.5 ± 1.63j , P6.11 and s3,4 = ±2.1j . Given s3 + (1 + K)s2 + 10s + (5 + 15K) = 0 , the Routh array is s3 1 10 s2 1+K 5 + 15K s1 b so 5 + 15K © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 252 CHAPTER 6 The Stability of Linear Feedback Systems where b= 5 − 5K (1 + K)10 − (5 + 15K) = . 1+K 1+K Given that K > 0, we determine that the system is stable when 5−5K > 0 or 0 0, ab − c > 0 and c > 0. When a > 0 and c > 0, we know that b > 0. So, a necessary condition for stability is that all coefficients a, b, and c be positive. The necessary and sufficient conditions for stability also require that b > c/a, in addition to a > 0 and c > 0. P6.13 The characteristic equation is s3 + (p + 2ζωn )s2 + (2ζωn p + Kωn2 )s + Kωn2 z = 0. The conditions for stability (see P6.12) are p + 2ζωn > 0, 2ζωn p + Kωn2 > (Kωn2 z)/(p + 2ζωn ), and Kωn2 z > 0. Since we know that K > 0, ζ > 0, and ωn > 0, it follows that for stability z > 0, p > −2ζωn , and 2ζωn p + Kωn2 > P6.14 Kωn2 z . p + 2ζωn The system has the roots s1,2 = ±j , s3,4 = ±j , and s5,6 = −1 ± 3j , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 253 Problems Therefore, the system is not stable since there are repeated roots on the jω-axis. P6.15 (a) Neglecting the zeros and poles, we have the characteristic equation s4 + 30s3 + 325s2 + 2500s + K = 0 . The Routh array is s4 1 325 K s3 30 2500 0 s2 241.67 K s1 b so K where b= 604166.67 − 30K . 241.67 Therefore, the system is stable for 0 < K < 20139. (b) Without neglecting the zeros and poles, the closed-loop characteristic equation is s6 + 90s5 + 5525s4 + 12400s3 + (1255000 + K)s2 + (8500000 + 30K)s + 1125K = 0 . This is stable for 0 < K < 61818 . We see that the additional poles and zero makes the system stable for a much larger gain K. P6.16 (a) The Routh array is s3 1 5 s2 5 6 s1 3.8 so 6 Examining the first column of the Routh array, we see no sign changes. So, the system is stable. (b) The roots of the system are s1 = −0.3246 and s2,3 = −2.3377 ± 3.6080j. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 254 CHAPTER 6 The Stability of Linear Feedback Systems (c) The step response is shown in Figure P6.16. Step Response 0.18 0.16 0.14 Amplitude 0.12 0.1 0.08 0.06 0.04 0.02 0 0 5 10 15 Time (sec ) FIGURE P6.16 Unit step response. P6.17 The closed-loop transfer function is T (s) = s3 + K +1 . + 3s + K + 1 3s2 The Routh array is s3 1 3 s2 3 K+1 s1 8−K 3 so K +1 So, for stability we require −1 < K < 8. P6.18 The system characteristic equation is s2 + (h − k)s + ab − kh = 0 . For stability we require h > k and ab > kh. If k > h, the system is unstable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 255 Problems P6.19 (a) The characteristic equation is s3 + 9s2 + (K − 10)s + 2K = 0 . The Routh array is s3 1 s2 K − 10 9 2K s1 7K−90 9 so 2K For stability K > 90/7 . (b) When K = 90/7, the system is marginally stable. The roots are q s1,2 = ±j 20/7 , at the jω-axis crossing. P6.20 The closed-loop characteristic equation is q(s) = s5 + s4 + 9s3 + Ks2 + 2Ks + K . The range of stability for the vertical-liftoff vehicle is 5.177 < K < 7.823 . Therefore, for K = 6, the system is stable. When K = 6 we have q(s) = s5 + s4 + 9s3 + 6s2 + 12s + 6 The Routh array is s5 1 9 12 s4 1 6 6 s3 3 6 s2 4 6 s1 1.5 so 6 All entries in the first column are positive, so the system is stable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 256 CHAPTER 6 P6.21 The Stability of Linear Feedback Systems The state transition matrix is (k − p1 )e−p1 t − (k2 − p2 )e−p2 t e−p1 t − e−p2 t 1 2 Φ(t, 0) = −p t −p t −p t −p t p2 − p1 1 2 1 2 −k1 e + k1 e −p1 e + p2 e where p1 p2 = k1 and p1 + p2 = k2 . We assume that p1 6= p2 . In the case when p1 = p2 , the state transition matrix will change, but the factors e−p1 t and e−p2 t will remain. The eigenvalues of A are given by the solution to det |λI − A| = λ2 + k2 λ + k1 = 0 . q Therefore, the eigenvalues are λ1,2 = −k2 /2 ± k22 − 4k1 . If k2 > 0 and k1 > 0, then the eigenvalues are in the left half-plane, and the system is stable. The transfer function is given by G(s) = C (sI − A)−1 B = − s−1 . s2 + k2 s + k1 Therefore the characteristic equation is s2 + k2 s + k1 = 0 and the poles q are s1,2 = −k2 /2 ± k22 − 4k1 . If k2 > 0 and k1 > 0, then the poles are in the left half-plane, and the system is stable. Notice that the values of λ1,2 and s1,2 are the same. Also, the eigenvalues are the same as the values of −p1 and −p2 . So, if the eigenvalues are negative, then the elements of the state transition matrix will decay exponentially. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 257 Advanced Problems Advanced Problems The Routh array is s4 1 K1 s3 20 4 s2 20K1 −4 20 K2 s1 b 0 so K2 K2 where b= 20K1 − 4 − 100K2 . 5K1 − 1 For stability, we require K2 > 0, K1 > 0.2, and b > 0. Therefore, using the condition that b > 0, we obtain K2 < 0.2K1 − 0.04 . The stability region is shown in Figure AP6.1. 0. 4 0.35 0. 3 0.25 K2 AP6.1 0. 2 0.15 0. 1 0.05 STABLE REGION 0 0. 2 0. 4 0. 6 0. 8 1 1. 2 K1 FIGURE AP6.1 Stability region. 1. 4 1. 6 1. 8 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 258 CHAPTER 6 AP6.2 The Stability of Linear Feedback Systems The Routh array is s4 1 30 s3 9 s2 310−K 9 K − 40 s1 b so K K K 0 where b= (310 − K)(K − 40) − 81K . 310 − K Therefore, using the condition that b > 0, we obtain the stability range for K: 59.07 < K < 209.94 . AP6.3 (a) The steady-state tracking error to a step input is ess = lim s(1 − T (s))R(s) = 1 − T (0) = 1 − α . s→0 We want |1 − α| < 0.05 . This yields the bounds for α 0.95 < α < 1.05 . (b) The Routh array is s3 1 α s2 1+α 1 s1 b 0 so 1 where b= α2 + α − 1 . 1+α Therefore, using the condition that b > 0, we obtain the stability range for α: α > 0.618 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 259 Advanced Problems (c) Choosing α = 1 satisfies both the steady-state tracking requirement and the stability requirement. The closed-loop transfer function is T (s) = s3 K . + (p + 1)s2 + ps + K The Routh array is s3 1 p s2 1+p K s1 b 0 so K where b= p2 + p − K . 1+p Therefore, using the condition that b > 0, we obtain the the relationship K < p2 + p . The plot of K as a function of p is shown in Figure AP6.4. 120 100 80 K AP6.4 60 40 STABLE REGION 20 0 0 1 2 3 4 5 p FIGURE AP6.4 Stability region. 6 7 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 260 CHAPTER 6 The closed-loop transfer function is T (s) = 30K1 K2 . (s + 1 + K1 K3 )(s − 10)(2s + K2 K3 − 4) + 30K1 K2 K4 The Routh array is s3 2 a s2 b c s1 d 0 so c where a = −9K2 K3 + 16 + K1 K2 K32 − 24K1 K3 , b = 2K1 K3 + K2 K3 − 22, and c = −10K2 K3 + 40 − 10K1 K2 K32 + 40K1 K3 and d = (ab − 2c)/b . The conditions for stability are 2K1 K3 + K2 K3 − 22 > 0 −10K2 K3 + 40 − 10K1 K2 K32 + 40K1 K3 > 0 −2(−10K2 K3 + 40 − 10K1 K2 K32 + 40K1 K3 ) + (9K2 K3 +16 + K1 K2 K32 − 24K1 K3 )(2K1 K3 + K2 K3 − 22) > 0 Valid values for the various gains are: K1 = 50, K2 = 30, K3 = 1, and K4 = 0.3. The step response is shown in Figure AP6.5. Step Response 350 300 250 Amplitude AP6.5 The Stability of Linear Feedback Systems 200 150 100 50 0 FIGURE AP6.5 Stability region. 0 5 10 15 Time (sec) 20 25 30 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 261 Advanced Problems AP6.6 The characteristic equation is s3 + 7s2 + (KD + 14)s + KP + 8 = 0. For stability we require that KP > −8 and KD > KP + 8 − 14. 7 The relationship between KD and KP is shown in Figure AP6.6. 2 0 −2 STABLE REGION K D −4 −6 −8 UNSTABLE REGION −10 −12 −14 0 10 20 30 40 50 60 70 80 90 100 K P FIGURE AP6.6 Stability region. AP6.7 The characteristic equation is 0.1s4 + 2.05s3 + s2 + 8KP s + 8KI = 0. From the Routh array we find the conditions for stability are 0 < KI < 0.3125 1.2812 − p 1.6416 − 5.2531KI < KP < 1.2812 + p 1.6416 − 5.2531KI © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 262 CHAPTER 6 The Stability of Linear Feedback Systems Design Problems CDP6.1 The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: θ(s) 26.035 = . Va (s) s(s + 33.142) In the above transfer function we have neglected the motor inductance Lm . The closed-loop transfer function from the input to the output is 26.035Ka θ(s) = 2 . R(s) s + 33.142s + 26.035Ka The Routh array is s2 1 26.035Ka s1 33.142 0 s0 26.035Ka Stability is achieved for any 0 < Ka < ∞. DP6.1 The closed-loop characteristic polynomial is 1 1 1 s3 + s2 (5 + p + K) + s( Kp + K + 5p) + K = 0 . 5 5 5 (i) When p = 2, we have 1 3 s3 + s2 (7 + K) + s(10 + K) + K = 0 . 5 5 The Routh array is s3 1 s2 7+ s1 b so K 10 + 35 K K 5 K where b= (7 + K/5)(10 + 3K/5) − K . 7 + 51 K When −32.98 < K < −17.69, we find that b > 0. Examining the other terms in the first column of the array, we find that the system © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 263 Design Problems is stable for any K > 0. (ii) When p = 0, we have 1 1 s3 + s2 (5 + K) + s( K) + K = 0 . 5 5 The Routh array is s3 1 1 5K s2 5 + 15 K K s1 b so K where b= (5 + 15 K) 15 K − K K 2 /25 = . (5 + K/5) (5 + K/5) Again, examination of the first column reveals that any K > 0 results in a stable system. So, we just need to select any K > 0; e.g. K = 10. DP6.2 (a) The closed-loop characteristic equation is 1+ 20(Ks + 1) =0, s2 (s + 20) or s3 + 20s2 + 20Ks + 20 = 0 . The Routh array is s3 1 20K s2 20 20 s1 b so 1 where b= 20K − 1 . 1 For stability, we require K > 0.05. (b) The desired characteristic polynomial is (s2 + as + b)(s + 5) = s3 + s2 (a + 5) + s(5a + b) + 5b = 0 . Equating coefficients with the actual characteristic equation we can © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 264 CHAPTER 6 The Stability of Linear Feedback Systems solve for a, b and K, yielding b = 4, a = 15, and K= 5a + b 79 = . 20 20 (c) The remaining two poles are s1 = −14.73 and s2 = −0.27. (d) The step response is shown in Figure DP6.2. 1 0.9 0.8 0.7 y(t) 0.6 0.5 0.4 0.3 0.2 0.1 0 0 2 4 6 8 10 time (sec) 12 14 16 18 FIGURE DP6.2 Mars guided vehicle step response. DP6.3 (a) The closed-loop characteristic equation is 2τ s3 + (τ + 2)s2 + (K + 1)s + 2K = 0 . The Routh array is s3 2τ K+1 s2 τ +2 2K s1 b so 2K where b= (τ + 2)(K + 1) − 4Kτ . (τ + 2) 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 265 Design Problems Examining the first column of the Routh array, we determine that for stability τ > 0, K > 0 and setting b > 0 yields the relationships: (1) K < τ +2 2 when τ > 3τ − 2 3 (2) K > 0 when 0 < τ ≤ 2 . 3 The plot of τ versus K is shown in Figure DP6.3a. 5 4.5 4 3.5 tau 3 2.5 2 1.5 1 0.5 0 STABLE REGION 1 2 3 4 5 6 7 K FIGURE DP6.3 (a) The plot of τ versus K. (b) The steady-state error is ess = A , Kv where Kv = 2K . So, ess 1 = . A 2K We require that ess ≤ 0.25A, therefore K≥2. One solution is to select τ = 0.5, then we remain within the stable region. (c) The step response is shown in Figure DP6.3b. The percent overshoot is P.O. = 57%. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 266 CHAPTER 6 The Stability of Linear Feedback Systems 1.6 P.O. = 56.77 % 1.4 1.2 y(t) 1 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 20 time (sec) FIGURE DP6.3 CONTINUED: (b) Closed-loop system step response. DP6.4 (a) The closed-loop characteristic polynomial is s3 + Ks2 + [(2 + m)K − 1]s + 2mK = 0 . The Routh array is s3 1 s2 2K + mK − 1 K 2mK s1 b so 2mK Examining the first column of the Routh array, we see that for stability we require m > 0, K > 0, and b > 0, where b= (2K + mK − 1)K − 2mK = (2 + m)K − (1 + 2m) > 0 , K or K> 1 + 2m . 2+m The plot of K vs m is shown in Figure DP6.4a. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 267 Design Problems 1.6 1.4 STABLE REGION 1.2 K 1 0.8 0.6 0.4 0 0.5 1 1.5 2 2.5 3 3.5 6 7 4 4.5 5 m FIGURE DP6.4 (a) The plot of K versus m. 1.8 P.O. = 64.3208 % 1.6 1.4 1.2 y(t) 1 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 time (sec) FIGURE DP6.4 CONTINUED: (b) Shuttle attitude control step response. (b) The steady-state error is ess 1 1 = < 0.10 , = A Kv 2mK 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 268 CHAPTER 6 The Stability of Linear Feedback Systems or mK > 5. For example, we can select m = 0.5 and K = 2. (c) See Figure DP6.4b for the step response where P.O. = 64.3%. DP6.5 The closed-loop transfer function is T (s) = s3 + K . + 20s + K 10s2 The range of K for stability is 0 < K < 200. If we let K = Km /N where Km = 200, then N = 6.25 results in a step response with P.O. = 15.7% and Ts = 1.96 seconds. DP6.6 The closed-loop system is given by ẋ = 0 1 2 − K1 −2 − K2 x+ 0 1 r The characteristic polynomial is s2 +(2+K2 )s+K1 −2h = 0. Soithe system is stable for K1 > 2 and K2 > −2. Selecting K = 10 2 results in closed-loop eigenvalues at s = −2 ± 2j. The closed-loop step response has a settling time of 2.11 s and a percent overshoot of 4.32%. Im(s) sin-1 ζ = sin-1 0.69=43.63ο Re(s) desired region for eigenvalues ζωn = -1 FIGURE DP6.6 Performance region. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 269 Design Problems DP6.7 (a) The inner loop closed-loop transfer function is 20s Y (s) = 3 . 2 U (s) s + 10s + 20s + 20K1 The Routh array is s3 1 20 s2 ω 20K1 s1 200−20K1 10 so 20K1 For stability 0 < K1 < 10. (b) The fastest response (that is, the quickest settling time) occurs when K1 = 2.2 (c) With K1 = 2.2, the closed-loop transfer function is Y (s) 20K2 s = 3 . R(s) s + 10s2 + (20 + 20K2 )s + 44 The Routh array is s3 1 20(K2 + 1) s2 10 44 s1 200K2 +156 10 so 44 For stability, we require 200K2 + 156 > 0 . Therefore, K2 > −0.78. DP6.8 The closed-loop characteristic equation is s2 + 4KD s + 4(KP + 1) = 0. So, it is possible to find KP and KD to stabilize the system. For example, any KP > 0 and KD > 0 leads to stability. Choosing KP ≥ 9 results in a steady-state tracking error √ less than 0.1 due to a unit step input. Then, the damping ratio ζ = 2/2 is achieved by selecting √ √ 2 KP + 1 KD = . 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 270 CHAPTER 6 The Stability of Linear Feedback Systems Computer Problems CP6.1 The m-file script is shown in Figure CP6.1. ans = -0.6063 + 2.7322i -0.6063 - 2.7322i -1.7874 pa=[1 3 10 14]; roots(pa) pb=[1 8 24 32 16]; roots(pb) pc=[1 0 2 1]; roots(pc) ans = -2.0004 -2.0000 + 0.0004i -2.0000 - 0.0004i -1.9996 ans = 0.2267 + 1.4677i 0.2267 - 1.4677i -0.4534 FIGURE CP6.1 Computing the polynomial roots with the rootsfunction. CP6.2 The m-file script is shown in Figure CP6.2. K1=1;K2=2;K3=5; den=[1 2 1]; num1=K1*[1 -1 2];num2=K2*[1 -1 2];num3=K3*[1 -1 2]; sys1 = tf(num1,den); sys2 = tf(num2,den); sys3 = tf(num3,den); sys1_cl=feedback(sys1,[1]); sys2_cl=feedback(sys2,[1]); sys3_cl=feedback(sys3,[1]); p1 = pole(sys1_cl), p2 = pole(sys2_cl), p3 = pole(sys3_cl) ans = -2.5000e -01 + 1.1990e+00i -2.5000e -01 - 1.1990e+00i ans = ans = 2.5000e -01 + 1.3307e+00i 2.5000e -01 - 1.3307e+00i 0 + 1.2910e+00i 0 - 1.2910e+00i FIGURE CP6.2 K = 1 is stable;K = 2 is marginally stable; and K = 5 is unstable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 271 Computer Problems CP6.3 The closed-loop transfer function and the roots of the characteristic equation are shown in Figure CP6.3. Transfer function: s+1 ---------------------s^3 + 4 s^2 + 7 s + 11 numg=[1 1]; deng=[1 4 6 10]; sysg = tf(numg,deng); sys = feedback(sysg,[1]) r=pole(sys) r= -2.8946 -0.5527 + 1.8694i -0.5527 - 1.8694i FIGURE CP6.3 Closed-loop transfer function and roots. CP6.4 There are no poles in the right half-plane, but the system is unstable since there are multiple poles on the jω-axis at s = ±j and s = ±j (see Figure CP6.4). Step Response From: U(1) 25 20 -2.0000 0.0000 + 1.0000i 0.0000 - 1.0000i -0.0000 + 1.0000i -0.0000 - 1.0000i 10 5 To: Y(1) ans = 15 Amplitude num=[1]; den=[1 2 2 4 1 2]; sys = tf(num,den); pole(sys) t = 0:0.1:100; step(sys,t) 0 -5 -10 -15 -20 -25 0 10 20 30 40 50 Time (sec.) FIGURE CP6.4 Unstable system step response. 60 70 80 90 100 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 272 CHAPTER 6 CP6.5 The Stability of Linear Feedback Systems The closed-loop system poles for the slow/fast pilots are shown in Figure CP6.5. The maximum allowable time delay is 0.2045seconds. At the maximum allowable time delay, the system has roots on the jω-axis at s = ±2.6j. The slow pilot destabilizes the aircraft. nume=[-10]; dene=[1 10]; syse = tf(nume,dene); numg=[-1 -6]; deng=[1 3 6 0]; sysg = tf(numg,deng); % % Fast pilot % tau=0.1; tau1=2; K=1; tau2=0.5; nump=-K*[tau1*tau tau-2*tau1 -2]; denp=[tau2*tau tau+2*tau2 2]; sysp = tf(nump,denp); sysa = series(sysp,syse); sysb = series(sysa, sysg); sys = feedback(sysb,[1]); fast_pilot=pole(sys) % % Slow pilot % tau=0.6; tau1=2; K=1; tau2=0.5; nump=-K*[tau1*tau tau-2*tau1 -2]; denp=[tau2*tau tau+2*tau2 2]; sysp = tf(nump,denp); sysa = series(sysp,syse); sysb = series(sysa, sysg); sys = feedback(sysb,[1]); slow_pilot = pole(sys) % % Maximum pilot time delay, tau = 0.2045 sec % tau=0.2045; tau1=2; K=1; tau2=0.5; nump=-K*[tau1*tau tau-2*tau1 -2]; denp=[tau2*tau tau+2*tau2 2]; sysp = tf(nump,denp); sysa = series(sysp,syse); sysb = series(sysa, sysg); sys = feedback(sysb,[1]); max_pilot_delay=pole(sys) closed-loop system poles fast_pilot = -19.6267 -10.7712 -3.8885 -0.1697 + 2.7880i -0.1697 - 2.7880i -0.3742 slow_pilot = -9.4526 -4.5228 + 2.2595i -4.5228 - 2.2595i 0.2793 + 2.0314i 0.2793 - 2.0314i -0.3937 max_pilot_delay = -10.0433 + 2.2684i -10.0433 - 2.2684i -4.3153 0.0001 + 2.6040i 0.0001 - 2.6040i -0.3783 FIGURE CP6.5 Closed-loop system poles for an aircraft with a pilot in-the-loop. CP6.6 The closed-loop transfer function is T (s) = s3 + 5s2 1 . + (K − 3)s + K + 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 273 Computer Problems Utilizing the Routh-Hurwitz approach, for stability we determine that K>4. When K = 4, the roots of the characteristic equation are s1 = −5 and s2,3 = ±j . The m-file script which generates a plot of the roots of the characteristic equation as a function of K is shown in Figure CP6.6. K=[0:0.1:5]; n=length(K); for i=1:n numg=[1]; deng=[1 5 K(i)-3 K(i)]; sys_o = tf(numg,deng); sys_cl = feedback(sys_o,[1]); p(:,i)=pole(sys_cl); end plot(real(p),imag(p),'x'), grid text(-0.9,0.95,'K=4 -->'); text(-0.2,1.3,'K=5'); text(0,0.2,'K=0') % From a Routh-Hurwitz analysis we find that % minimum K for stability is K=4 Kmax=4; numg=[1]; deng=[1 5 Kmax-3 Kmax]; sysg = tf(numg,deng); sys_cl = feedback(sysg,[1]); pole(sys_cl) 1.5 K=5 1 K=4 --> 0.5 K=0 0 -0.5 -1 -1.5 -6 -5 -4 -3 -2 -1 0 FIGURE CP6.6 Roots of the characteristic equation as a function of K, where 0 < K < 5. 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 274 CHAPTER 6 CP6.7 The Stability of Linear Feedback Systems The characteristic equation is p(s) = s3 + 10s2 + 14s + 12 . A=[0 1 0;0 0 1;-12 -14 -10]; b=[0;0;12]; c=[1 1 0]; d=[0]; sys = ss(A,b,c,d); % % Part (a) % p=poly(A) % % Part (b) % roots(p) % % Part (c) % step(sys) p= 1.0000 10.0000 14.0000 12.0000 ans = -8.5225 -0.7387 + 0.9286i -0.7387 - 0.9286i Step Response 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 Time (sec) 6 7 8 9 FIGURE CP6.7 Characteristic equation from the state-space representation using the poly function. The roots of the characteristic equation are s1 = −8.5225 and s2,3 = −0.7387 ± 0.9286j . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 275 Computer Problems The system is stable since all roots of the characteristic equation are in the left half-plane. The unit step response and associated m-file script are shown in Figure CP6.7. CP6.8 The characteristic equation is s3 + 10s2 + 10s + 5K1 = 0 . (a) The Routh array is s3 1 10 s2 10 5K1 s1 100−5K1 10 so 5K1 From the Routh-Hurwitz criterion, we obtain the limits 0 < K1 < 20 for stability. (b) The plot of the pole locations is 0 < K1 < 30 is shown in Figure CP6.8. As seen in Figure CP6.8, when K1 > 20, the pole locations move into the right half-plane. Root Locus 4 3 Imaginary Axi s 2 1 k=20 0 ?-1 ?-2 ?-3 ?-4 ?-12 ?-10 ?-8 ?-6 ?-4 ?-2 0 Real Axi s FIGURE CP6.8 Pole locations for 0 < K1 < 30. CP6.9 (a) The characteristic equation is s3 + 2s2 + s + k − 4 = 0 . 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 6 The Stability of Linear Feedback Systems The Routh array is s3 1 1 s2 2 s1 6−k 2 k−4 so k−4 For stability, we obtain 4 < k < 6. (b) The pole locations for 0 < k < 10 are shown in Figure CP6.9. We see that for 0 < k < 4 the system is unstable. Similarly, for 6 < k < 10, the system is unstable. Root Locus 2 k=10 pole locations when k=0 1.5 1 0.5 pole location when k=0 increasing k k=10 0 k=6 inc rea ?-1 ?-1.5 ?-2 ?-3 k=4 gk ?-0.5 sin Imaginary Axi s 276 ?-2 ?-1 Real Axi s FIGURE CP6.9 Pole locations for 0 < k < 10. 0 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 7 The Root Locus Method Exercises (a) For the characteristic equation 1+K s(s + 4) =0, + 2s + 2 s2 the root locus is shown in Figure E7.1. 4 3 2 1 Imag Axis E7.1 0 x o o -1 x -2 -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE E7.1 s(s+4) Root locus for 1 + K s2 +2s+2 = 0. 277 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 278 CHAPTER 7 The Root Locus Method (b) The system characteristic equation can be written as (1 + K)s2 + (2 + 4K)s + 2 = 0 . Solving for s yields −(1 + 2K) s= ± (1 + K) p (2 + 4K)2 − 8(1 + K) . 2(1 + K) When (2 + 4K)2 − 8(1 + K) = 0 , then we have two roots at s1,2 = − (1+2K) 1+K . Solving for K yields K = 0.31. (c) When K = 0.31, the roots are s1,2 = −(1 + 0.62) = −1.24 . (1.31) (d) When K = 0.31, the characterisitc equation is s2 + 2.472s + 1.528 = (s + 1.24)2 = 0 . Thus, ωn = 1.24 and ζ = 1, the system is critically damped. The settling time is Ts ≈ 4 sec. E7.2 (a) The root locus is shown in Figure E7.2. When K = 6.5, the roots of the characteristic equation are s1,2 = −2.65 ± j1.23 and s3,4 = −0.35 ± j0.8 . The real part of the dominant root is 8 times smaller than the other two roots. (b) The dominant roots are (s + 0.35 + j0.8)(s + 0.35 − j0.8) = s2 + 0.7s + 0.7625 . From this we determine that ωn = 0.873 and ζ = 0.7 = 0.40 . 2(0.873) Thus, the settling time is 4 4 = = 11.43 sec . ζωn 0.35 √ 2 The percent overshoot is P.O. = e−πζ/ 1−ζ = 25.4%. Ts = © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 279 Exercises 4 3 * K=6.5 2 * Imag Axis 1 x * 0 x -1 x * x * -2 -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE E7.2 Root locus for 1 + K s(s+2)(s12 +4s+5) = 0. The root locus is shown in Figure E7.3. The roots are s1 = −8.7, s2,3 = −1.3 ± j2.2 when K = 7.35 and ζ = 0.5. 4 zeta=0.5 3 2 o * <----- K=7.35 1 Imag Axis E7.3 0 x * x -1 -2 o * -3 -4 -10 -8 -6 -4 -2 Real Axis FIGURE E7.3 2 Root locus for 1 + K ss2+4s+8 = 0. (s+4) 0 2 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 280 CHAPTER 7 E7.4 The Root Locus Method The root locus is shown in Figure E7.4. 2 1.5 1 x Imag Axis 0.5 0 o -0.5 -1 x -1.5 -2 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 Real Axis FIGURE E7.4 s+1 = 0. Root locus for 1 + K s2 +4s+5 The departure angles and entry points are θd = 225o , −225o and σb = −2.4 . E7.5 (a) The root locus is in Figure E7.5. The breakaway points are σb1 = −13.0 , σb2 = −5.89 . (b) The asymptote centroid is σcent = −18 , and φasym = ±90o . (c) The gains are K1 = 1.57 and K2 = 2.14 at the breakaway points. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 281 Exercises 40 30 20 Imag Axis 10 0 -10 < asymptote -20 -30 -40 -15 -10 -5 0 5 Real Axis FIGURE E7.5 s2 +2s+10 Root locus for 1 + K (s4 +38s3 +515s 2 +2950s+6000) = 0. The system is unstable for K > 75. 10 8 System: sys Gain: 75 Pole: −0.000981 + 8.66i Damping: 0.000113 Overshoot (%): 100 Frequency (rad/sec): 8.66 6 4 Imaginary Axis E7.6 2 0 −2 −4 −6 −8 −10 −10 −8 −6 FIGURE E7.6 15K Root locus for 1 + K s(s2 +15s+75) = 0. −4 Real Axis −2 0 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 282 CHAPTER 7 E7.7 The Root Locus Method The root locus is shown in Figure E7.7. The characteristic equation has 20 15 asymptote −−−> Imaginary Axis 10 5 0 System: sys Gain: 27.3 Pole: −1.44 + 1.11i Damping: 0.792 Overshoot (%): 1.7 Frequency (rad/sec): 1.81 −5 −10 −15 −20 −25 −20 −15 −10 −5 Real Axis 0 5 10 FIGURE E7.7 s+8 = 0. Root locus for 1 + K s(s+4)(s+6)(s+9) 4 poles and 1 zero. The asymptote angles are φ = +60o , −60o , −180o centered at σcent = −3.7. When K = 27.35 then ζ = 0.8 for the complex roots. E7.8 The characteristic equation is 1+K (s + 1) =0, s2 (s + 9) or s3 + 9s2 + Ks + K = 0 . For all the roots to be equal and real, we require (s + r)3 = s3 + 3rs2 + 3r 2 s + r 3 = 0 . Equating terms and solving for K yields K = 27. All three roots are equal at s = −3, when K = 27. The root locus is shown in Figure E7.8. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 283 Exercises 8 6 4 Imag Axis 2 3 roots at s=-3 0 x o x -2 -4 -6 -8 -15 -10 -5 0 5 Real Axis FIGURE E7.8 Root locus for 1 + K s2s+1 (s+9) = 0. E7.9 The characteristic equation is 1+K 1 =0 s(s2 + 2s + 5) or s3 + 2s2 + 5s + K = 0 . (a) The system has three poles at s = 0 and −1 ± j2. The number of asymptotes is np − nz = 3 centered at σcent = −2/3, and the angles are φasymp at ±60o , 180o . (b) The angle of departure, θd , is 90o +θd +116.6o = 180o , so θd = −26.6o . (c) The Routh array is s3 1 5 s2 2 K s1 b so K where b = 5 − K/2. So, when K = 10 the roots lie on the imaginary © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 284 CHAPTER 7 The Root Locus Method axis. The auxilary equation is 2s2 + 10 = 0 √ s1,2 = ±j 5 . which implies (d) The root locus is shown in Figure E7.9. 4 3 2 x Imag Axis 1 asymptote ---> 0 x -1 -2 x -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE E7.9 1 = 0. Root locus for 1 + K s(s2 +2s+5) E7.10 (a) The characteristic equation is 1+ K(s + 2) =0. s(s + 1) Therefore, K=− (s2 + s) , (s + 2) and dK s2 + 4s + 2 =− =0. ds (s + 2)2 Solving s2 +4s+2 = 0 yields s = −0.586 and −3.414. Thus, the system © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 285 Exercises breakaway and entry points are at s = −0.586 and s = −3.414. (b) The desired characteristic polynomial is (s + 2 + aj)(s + 2 − aj) = s2 + 4s + 4 + a2 = 0 , where a is not specified. The actual characteristic polynomial is s2 + (1 + K)s + 2K = 0 . Equating coefficients and solving for K yields√K = 3 and a = Thus, when K = 3, the roots are s1,2 = −2 ± 2j. √ 2. (c) The root locus is shown in Figure E7.10. 2 K=3, s=-2+1.414j 1.5 * 1 Imag Axis 0.5 s=-3.41 0 o x s=-0.58 x -0.5 -1 * -1.5 -2 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 Real Axis FIGURE E7.10 s+2 Root locus for 1 + K s(s+1) = 0. E7.11 The root locus is shown in Figure E7.11 for the characteristic equation 1+ K(s + 2.5) =0. (s2 + 2s + 2)(s2 + 4s + 5) From the root locus we see that we can only achieve ζ = 0.707 when K = 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 286 CHAPTER 7 The Root Locus Method 5 4 3 2 Imag Axis 1 0 x x x x -2 -1 <---- zeta=0.707 & K=0 o -1 -2 -3 -4 -5 -5 -4 -3 0 1 2 3 4 5 Real Axis FIGURE E7.11 Root locus for 1 + E7.12 K(s+2.5) (s2 +2s+2)(s2 +4s+5) = 0. (a) The root locus is shown in Figure E7.12 for the characteristic equation 1+ K(s + 1) =0. s(s2 + 6s + 18) (b) The roots of the characteristic equation are (i) K = 10: s1,2 = −2.8064 ± 4.2368j and s3 = −0.3872 (ii) K = 20: s1,2 = −2.7134 ± 5.2466j and s3 = −0.5732 (c) The step response performance of the system is summarized in Table E7.12. K 10 20 Ts (sec) 9.0 5.5 0 0 4.8 2.6 P.O. Tr (sec) TABLE E7.12 System performance when K = 10 and K = 20. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 287 Exercises Root Locus 15 10 Imaginary Axis 5 0 −5 −10 −15 −3.5 FIGURE E7.12 Root locus for 1 + E7.13 −3 −2.5 K(s+1) s(s2 +6s+18) −2 −1.5 Real Axis −1 −0.5 0 0.5 = 0. (a) The characteristic equation is s(s + 1)(s + 3) + 4s + 4z = 0 . Rewriting with z as the parameter of interest yields 1+z 4 =0. s(s + 1)(s + 3) + 4s The root locus is shown in Figure E7.13a. (b) The root locations for z = 0.6 , 2.0 , and 4.0 are shown in Figure E7.13a. When z = 0.6, we have ζ = 0.76 and ωn = 2.33. Therefore, the predicted step response is P.O. = 2.4% and Ts = 2.3 sec (ζ = 0.6) . When z = 2.0, we have ζ = 0.42 and ωn = 1.79. Therefore, the predicted step response is P.O. = 23% and Ts = 5.3 sec (ζ = 2.0) . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method 4 3 * z=0.6 o z=2.0 + z=4.0 + 2 x o * Imag Axis 1 0 + o x * -1 x * o -2 + -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE E7.13 4 = 0. (a) Root locus for 1 + z s(s+1)(s+3)+4s Finally, when z = 4.0, we have ζ = 0.15 and ωn = 2.19. Therefore, the predicted step response is P.O. = 62% and Ts = 12 sec. (c) The actual step responses are shown in Figure E7.13b. 1.6 1.4 ___ z=0.6 - - - z=2.0 1.2 ..... z=4.0 1 y(t) 288 0.8 0.6 0.4 0.2 0 0 2 4 6 8 time (sec) 10 FIGURE E7.13 CONTINUED: (b) Step Responses for z = 0.6, 2.0, and 4.0. 12 14 16 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 289 Exercises (a) The root locus is shown in Figure E7.14 for the characteristic equation K(s + 10) =0. s(s + 5) 1+ The breakaway point is sb = −2.93; the entry point is se = −17.1. 10 8 6 K=5, s=-5+5j * 4 2 Imag Axis E7.14 0 s=-17.1 o x -10 -5 s=-2.93 x -2 -4 -6 -8 -10 -20 -15 0 5 Real Axis FIGURE E7.14 Root locus for 1 + K(s+10) s(s+5) = 0. √ (b) We desire ζ = 1/ 2 = 0.707. So, the desired characteristic polynomial is 1 s2 + 2 √ ωn s + ωn2 = 0 . 2 Comparing the desired characteristic polynomial to the actual we find the relationships √ 2ωn = 5 + K . ωn2 = 10K and Solving for K and ωn yields K = 5 and ωn = 7.07. The roots are s1,2 = −5 ± j5 when K = 5. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 290 CHAPTER 7 (a) The characteristic equation 1+K (s + 10)(s + 2) =0 s3 has the root locus in Figure E7.15. Root Locus 15 10 Imaginary Axi s E7.15 The Root Locus Method 5 K=1.67 0 ?-5 ?-10 ?-15 ?-25 ?-20 ?-15 ?-10 -5 0 5 Real Axi s FIGURE E7.15 Root locus for 1 + K(s+10)(s+2) s3 = 0. (b) The Routh array is s3 1 12K s2 K 20K s1 b so 20K when b = 12K − 20. For stability, we require all elements in the first column to be positive. Therefore, K > 1.67 . (c) When K > 3/4, we have ess = lim sE(s) = lim s s→0 s→0 1 1 s2 · 2 = lim 3 =0. s→0 s + K(s + 1)(s + 3) 1 + GH(s) s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 291 Exercises The expansion for e−T s is e−T s = 1 − T s + (T s)2 − ... 2! If (T s) << 1, then e−T s ≈ 1 − T s = a + bs , c + ds where a, b, c and d are constants to be determined. Using long division, 40 30 K=21 20 * 10 Imag Axis E7.16 0 x x o -10 -20 -30 -40 -40 -30 -20 -10 0 10 20 30 40 50 60 Real Axis FIGURE E7.16 Root locus for 1 + K(20−s) (s+1)(20+s) = 0. we expand (a + bs)/(c + ds) and match as many coefficients as possible. In this case, we determine that a = c = (2/T ) and also that b = −d = −1. In this case, with T = 0.1, we have e−T s = 20 − s −(s − 20) = . 20 + s (s + 20) So, the characteristic equation is 1+ −K(s − 20) , (s + 1)(s + 20) and the root locus is shown in Figure E7.16. Using a Routh-Hurwitz © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 292 CHAPTER 7 The Root Locus Method analysis with the characteristic polynomial s2 + (21 − K)s + 20 + 20K = 0 , we determine that the system is stable for −1 < K < 21. (a) The root locus is in Figure E7.17a. 2 1.5 1 0.5 Imag Axis E7.17 0 x x -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 Real Axis FIGURE E7.17 (a) Root locus for 1 + K s(s−1) = 0. The root locus is always in the right half-plane; the system is unstable for K > 0. (b) The characteristic equation is 1+ K(s + 2) =0, s(s − 1)(s + 20) and the root locus is shown in Figure E7.17b. The system is stable for K > 22.3 and when K = 22.3, the roots are s1,2 = ±j1.53 and s3 = −19 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 293 Exercises 10 8 6 4 Imag Axis 2 * 0 x o K=22.3 x x -2 -4 -6 -8 -10 -30 -25 -20 -15 -10 -5 0 5 10 Real Axis FIGURE E7.17 CONTINUED: (b) Root locus for 1 + = 0. The root locus is shown in Figure E7.18. 6 4 2 Imag Axis E7.18 K(s+2) s(s+20)(s−1) x + 0 x + K=8.15 x + x + -2 -4 -6 -6 -4 -2 0 2 4 6 Real Axis FIGURE E7.18 Root locus for 1 + K s(s+3)(s2 +2s+2) = 0. When K = 8.15, the roots are s1,2 = ±j1.095 and s3,4 = −2.5 ± j0.74. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 294 CHAPTER 7 E7.19 The Root Locus Method The characteristic equation is 1+ K =0, + 6s + 64) 3)(s2 s(s + and the root locus is shown in Figure E7.19. When K = 1292.5, the roots are s1,2 = ±j4.62 and s3,4 = −4.49 ± j6.36 . 15 10 x + Imag Axis 5 + 0 x K=1292.5 x + -5 + x -10 -15 -15 -10 -5 0 5 10 15 Real Axis FIGURE E7.19 Root locus for 1 + E7.20 K s(s+3)(s2 +6s+64) = 0. The characteristic equation is 1+ K(s + 1) =0, s(s − 2)(s + 6) and the root locus is shown in Figure E7.20. The system is stable for K > 16 . The maximum damping ratio of the stable complex roots is ζ = 0.25 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 295 Exercises 20 15 10 Imaginary Axis 5 ζmax = 0.25 0 0 1 2 Real Axis FIGURE E7.20 Root locus for 1 + = 0. The gain is K = 10.8 when the complex roots have ζ = 0.66. 10 5 K=10.8 Imag Axis E7.21 K(s+1) s(s−2)(s+6) + 0 x + x o + x -5 -10 -10 -5 0 Real Axis FIGURE E7.21 Root locus for 1 + Ks s3 +5s2 +10 = 0. 5 10 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 296 CHAPTER 7 E7.22 The Root Locus Method The root locus is shown in Figure E7.22. The characteristic equation is 1+ K(s2 + 18)(s + 2) =0. (s2 − 2)(s + 12) Root Locus 5 4 3 Imaginary Axis 2 1 0 −1 −2 −3 −4 −5 −14 FIGURE E7.22 Root locus for 1 + E7.23 −12 −10 K(s2 +18)(s+2) (s2 −2)(s+12) −8 −6 Real Axis −4 −2 0 = 0. The characteristic equation is 5s2 + as + 4 = 0 , which can rewritten as 1+ as =0. +4 5s2 The roots locus (with a as the parameter) is shown in Figure E7.23. 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 297 Exercises 1.5 1 x Imag Axis 0.5 0 o -0.5 x -1 -1.5 -1.5 -1 -0.5 0 0.5 1 1.5 Real Axis FIGURE E7.23 Root locus for 1 + E7.24 as 5s2 +4 = 0. The transfer function is G(s) = C(sI − A)−1 B + D = [ 1 0 ] = s2 s −1 4 s+k 1 . + ks + 4 −1 0 1 Therefore, the characteristic equation is s2 + ks + 4 = 0 , or 1+k s2 s =0. +4 The root locus for 0 < k < ∞ is shown in Figure E7.24. The closed-loop system is stable for all 0 < k < ∞. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 298 CHAPTER 7 The Root Locus Method 2.5 2 1.5 Imaginary Axis 1 0.5 0 5 5 5 5 5 5 5 0 0.5 Real Axis FIGURE E7.24 Root locus for 1 + k s2s+4 = 0. The characteristic equation is 1+K 10 =0. s(s + 25) The root locus shown in Figure E7.25 is stable for all 0 < K < ∞. 15 10 5 Imaginary Axis E7.25 0 Real Axis FIGURE E7.25 10 Root locus for 1 + K s(s+25) = 0. 0 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 299 Exercises E7.26 The characteristic polynomial is det or s −1 s+K −3 s+K +2 1+K =0 s+1 =0. s2 + 2s − 3 The root locus shown in Figure E7.26 is stable for all 0 < K < 3. Root Locus 0.8 0.6 Imaginary Axis 0.4 0.2 0 −0.2 −0.4 −0.6 −0.8 −12 −10 −8 −6 −4 Real Axis −2 0 2 FIGURE E7.26 s+1 Root locus for 1 + K s2 +2s−3 = 0. E7.27 The characteristic equation is 1+p s2 s =0. + 4s + 40 The root locus shown in Figure E7.27 is stable for all 0 < p < ∞. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 300 CHAPTER 7 The Root Locus Method 8 6 Imaginary Axis 4 2 0 0 2 Real Axis FIGURE E7.27 s = 0. Root locus for 1 + p s2 +4s+40 The characteristic equation is 1+K s(s2 s−1 =0. + 2s + 2) The system is stable for −1.33 < K < 0. 1.5 1 0.5 Imaginary Axis E7.28 0 #5 #5 ! " Real Axis FIGURE E7.28 Root locus for 1 + K s(s2s−1 = 0. +2s+2) 0 2 4 6 8 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 301 Problems Problems P7.1 Root Locus 30 20 Imaginary Axis 10 0 −10 −20 −30 −50 −40 −30 −20 −10 Real Axis 0 10 20 Root Locus 5 4 3 Imaginary Axis 2 1 0 −1 −2 −3 −4 −5 −7 FIGURE P7.1 (a) Root locus for 1 + −6 −5 K s(s+10)(s+8) −4 −3 −2 Real Axis = 0, and (b) 1 + −1 0 1 K (s2 +2s+2)(s+1) 2 = 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 302 CHAPTER 7 The Root Locus Method Root Locus 40 30 Imaginary Axis 20 10 0 −10 −20 −30 −40 −12 −10 −8 −6 −4 Real Axis −2 0 2 Root Locus 4 3 Imaginary Axis 2 1 0 −1 −2 −3 −4 −4 −3.5 −3 −2.5 FIGURE P7.1 CONTINUED: (c) Root locus for 1 + P7.2 −2 −1.5 Real Axis K(s+5) s(s+2)(s+7) −1 −0.5 = 0, and (d)1 + 0 0.5 K(s2 +4s+8) s2 (s+7) = 0. The root locus is shown in Figure P7.2 for the characteristic equation 1+ 10Kv (s + 10) =0. s(s + 1)(s + 100) The damping ratio is ζ = 0.6 when Kv = 0.8, 135 and 648. The roots of the characteristic equation are: (a) Kv = 0.8 : s1 = −99.9, s2,3 = −0.54 ± j0.71 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 303 Problems (b) Kv = 135 : s1 = −85.9, s2,3 = −7.5 ± j10 (c) Kv = 648 : s1 = −11.7, s2,3 = −44.6 ± j59.5 30 20 Imag Axis 10 0 x o xx -10 -20 -30 -100 -80 -60 -40 -20 Real Axis FIGURE P7.2 Root locus for 1 + P7.3 10Kv (s+10) s(s+1)(s+100) = 0. (a) The breakaway point is s = −0.88 at K = 4.06. (b) The characteristic equation can be written as s(s + 2)(s + 5) + K = 0 . The Routh array is s3 1 10 s2 7 K s1 b 0 so K where b= 70 − K . 7 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 304 CHAPTER 7 The Root Locus Method √ When K = 70, the system has roots on jω-axis at s = ±j 10. (c) When K = 6, the roots are s1,2 = −0.83 ± j0.66, s3 = −5.34. (d) The characteristic equation 1+ K =0 s(s + 2)(s + 5) has the root locus shown in Figure P7.3. 10 8 6 4 Imag Axis 2 0 -2 -4 -6 -8 -10 -10 -8 FIGURE P7.3 Root locus for 1 + P7.4 -6 -4 K s(s+2)(s+5) -2 0 Real Axis 2 4 6 8 10 = 0. The characteristic equation for the large antenna is 1 + G1 G(s) = 1 + 100ka =0, (0.1s + 1)(s2 + 14.4s + 100) or 1+ 1000ka =0. (s + 10)(s2 + 14.4s + 100) The root locus is shown in Figure P7.4. Using Routh’s criteria, we find that the system is stable for −1 < ka < 4.83 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 305 Problems 20 * 15 <-- K=4.827 10 x Imag Axis 5 0 x -5 x -10 -15 * -20 -20 -15 -10 -5 0 5 10 15 20 Real Axis FIGURE P7.4 Root locus for 1 + 1000ka (s2 +14.14s+100)(s+10) = 0. When ka = 4.83, we have s1,2 = ±j15.53. P7.5 (a) The characteristic equation for hands-off control is 1+ 25K2 (s + 0.03)(s + 1) =0. (s + 0.4)(s2 − 0.36s + 0.16)(s + 9) The root locus is shown in Figure P7.5a. The damping ratio is ζ = 0.707 when K2 = 1.6 or K2 = 0.74. (b) The transfer function from Td (s) to Y (s) is Y (s) = G2 (s)Td (s) , 1 + G2 (s)Gf (s) where Gf (s) = K2 (s + 1) . s+9 Using the final value theorem, we determine that yss = lim s s→0 G2 (s) 1 11.7 = 3.8 , = 1 + G2 (s)Gf (s) s 1 + 11.7 K92 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method 5 4 <-- K=1.6 3 2 <-- K=0.74 Imag Axis 1 x 0 x o x o x -1 -2 -3 -4 -5 -10 -8 -6 -4 -2 0 2 4 Real Axis FIGURE P7.5 (a) Root locus for 1 + 25K2 (s+0.03)(s+1) (s+9)(s2 −0.36s+0.16)(s+0.4) = 0. 20 15 10 5 Imag Axis 306 x 0 x o x xo x -5 -10 -15 -20 -20 -15 -10 -5 0 5 10 15 20 Real Axis FIGURE P7.5 CONTINUED: (b) Root locus for 1 + 25K1 (s+0.03)(s+9) (s+0.045)(s2 +12s+1)(s+1.33)(s2 +7.66s+29.78) = 0. where we have selected K2 = 1.6. For K2 = 0.74, we find that yss = 5.96. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 307 Problems (c) The closed-loop characteristic equation with the pilot loop added is 1+ 25K1 (s + 0.03)(s + 9) =0. (s + 0.045)(s + 1.33)(s2 + 7.66s + 29.78)(s2 + 12s + 1) The root locus is shown in Figure P7.5b. (d) Using K1 = 2, we determine that ess = 0.44 . (a) The characteristic equation is 1+ K(s + 0.20)(s2 + 4s + 6.25) =0. (s + 0.9)(s − 0.6)(s − 0.1)(s + 4) The root locus is shown in Figure P7.6. 4 zeta=0.5 zeta*wn=-1/3 3 2 o K=4 --> 1 Imag Axis P7.6 0 x x o x x -1 o -2 -3 -4 -6 -5 -4 -3 -2 -1 0 1 2 Real Axis FIGURE P7.6 Root locus for 1 + K(s+0.2)(s2 +4s+6.25) (s+0.9)(s−0.6)(s−0.1)(s+4) = 0. (b) For Ts < 12 sec, we require ζωn > 1/3. Also, we want ζ > 0.5. So, we seek roots for a stable system with ζωn > 1/3 and ζ > 0.5. This occurs when K > 4. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 308 CHAPTER 7 (a) The characteristic equation for the speed control system is 1+ K =0, (s + 4)2 (s + δ) where K= 0.004 R and δ= 0.75 = 0.0001875 . 4000 The root locus is shown in Figure P7.7. At ζ = 0.6, we have K = 19.1, 6 4 2 <-- K=19.1 Imag Axis P7.7 The Root Locus Method 0 x x -2 -4 -6 -6 -5 -4 -3 -2 -1 0 1 2 3 Real Axis FIGURE P7.7 Root locus for 1 + K (s+4)2 (s+1.875e−04) = 0. therefore R = 0.00021 . When K = 19.1 the roots are s1,2 = −1.1 ± j1.43 and s3 = −5.80 . (b) The steady-state error is lim s∆ω(s) = lim s s→0 s→0 (0.25s + 1)2 ∆L(s) (0.25s + 1)2 (Js + b) + 1/R 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 309 Problems = 1 ∆L ≈ ∆LR , b + 1/R when R < 0.1. (a) The characteristic equation for the speed control system with the hydroturbine is 1+ K(−s + 1) =0, (s + 4)(s + 2)(s + δ) where K= 0.002 R and δ= 0.75 = 0.0001875 . 4000 The root locus is shown in Figure P7.8. At ζ = 0.6, we have K = 2.85, 2 1.5 1 K=2.85 --> 0.5 Imag Axis P7.8 0 x x x o 0 1 -0.5 -1 -1.5 -2 -6 -5 -4 -3 -2 -1 2 Real Axis FIGURE P7.8 Root locus for 1 + K(−s+1) (s+4)(s+2)(s+δ) = 0. therefore R = 0.0007 . When K = 2.85 the roots are −0.45 ± j0.60, and -5.1. 3 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 310 CHAPTER 7 The Root Locus Method (b) The steady-state error is lim s∆ω(s) = lim s s→0 s→0 = (0.25s + 1)(0.5s + 1) ∆L(s) (0.25s + 1)(0.5s + 1)(Js + f ) + (−s + 1)/R 1 ∆L ≈ ∆LR , f + 1/R when R < 0.1. The characteristic equation is 1+K (s + 0.5)(s + 0.1)(s2 + 2s + 289) =0 s(s + 30)2 (s − 0.4)(s + 0.8)(s2 + 1.45s + 361) where K = K1 K2 . The root locus is shown in Figure P7.9. When K = 4000 , the roots are s1,2 = −0.82 ± j19.4 50 40 30 20 10 Imag Axis P7.9 0 -10 -20 -30 -40 -50 -35 FIGURE P7.9 Root locus for 1 + -30 -25 -20 -15 Real Axis -10 K(s+0.5)(s+0.1)(s2 +2s+289) s(s+30)2 (s−0.4)(s+0.8)(s2 +1.45s+361) -5 = 0. 0 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 311 Problems s3 s4 s5 s6 s7 (a) The characteristic equation is 1+ K1 K2 (s + 2)2 =0. (s + 10)(s + 100)(s2 + 1.5s + 6.25) The root locus is shown in Figure P7.10. 10 8 6 4 x 2 Imag Axis P7.10 = −39.8 = −14.9 = −5.0 = −0.38 = −0.14 . 0 x x -2 o x -4 -6 -8 -10 -120 -100 -80 -60 -40 Real Axis FIGURE P7.10 Root locus for 1 + K1 K2 (s+2)2 (s+10)(s+100)(s2 +1.5s+6.25) = 0. (b) The gain K1 K2 = 1620 when ζ = 0.707. Therefore, K2 = 81000 , -20 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 312 CHAPTER 7 The Root Locus Method since K1 = 0.02 at medium weight cruise condition. (c) At lightweight cruise condition K1 = 0.2 . Using K2 = 81000, we find the roots are s1,2 = −54 ± j119 s3,4 = −2 ± j0.6 . The roots s3,4 become negligible and the roots at s1,2 become highly oscillatory. Hence, in this case ζ = 0.41 . (a) The closed-loop characteristic equation is 1+ 20Ka (s2 + s + 0.02) =0, s(s + 1)2 (s2 + 2s + 0.8) where K2 = 10 . Then, the root locus is shown in Figure P7.11a. 3 2 1 Ka=0.035 --> Imag Axis P7.11 0 x xo x ox -1 -2 -3 -3 -2 -1 0 1 2 Real Axis FIGURE P7.11 20s2 +20s+0.4 (a) Root locus for 1 + Ka s(s+1) 2 (s2 +2s+0.8) = 0, where K2 = 10. 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 313 Problems (b) When Ka < 0.035 , all the roots have a damping greater than or equal to 0.60. (c) Select Ka = 0.035 . Then, the characteristic equation with K2 as the parameter is 1 + K2 0.07(s2 + s) =0. s5 + 4s4 + 5.8s3 + 3.6s2 + 0.8s + 0.014 The root locus is shown in Figure P7.11b. 3 Ka=0.035 2 Imag Axis 1 0 x x o x x xo -1 -2 -3 -3 -2 -1 0 1 2 3 Real Axis FIGURE P7.11 0.07s(s+1) CONTINUED: (b) Root locus for 1+K2 s(s+1)2 (s2 +2s+0.8)+0.014 = 0, where Ka = 0.035. P7.12 (a) The closed-loop transfer function is T (s) = 1.8s2 (s Ka Km (s + 25)(s + 15) . + 2) + Ka Km (s + 25)(s + 15) + 1.6Km s(s + 2) So, with E(s) = R(s) − Y (s), we have E(s) = (1 − T (s))R(s) and ess = lim sE(s) = 1 − T (0) = 0 . s→0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method Therefore, when the system is stable, it has zero steady-state error. (b) The characteristic equation is s3 + (3.6 + Ka )s2 + (3.2 + 40Ka )s + 375Ka . The Routh array is s3 1 3.2 + 40Ka s2 3.6 + Ka 375Ka s1 b so 375K Solving for b > 0 leads to 0 < Ka < 0.05 or Ka > 5.64 for stability. (c) The characteristic equation can be written as 1+ Ka (s + 25)(s + 15) =0. s(s + 2)(s + 1.6) The root locus is shown in Figure P7.12. (d) When K > 40 , 40 30 20 10 Imag Axis 314 0 o o xx x -10 -20 -30 -40 -70 -60 -50 -40 -30 -20 Real Axis FIGURE P7.12 (s+25)(s+15) Root locus for 1 + Ka s(s+2)(s+1.6) = 0, where Km = 1.8. -10 0 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 315 Problems the roots are s1 = −123 and s2,3 = −15.6 ± j31.2 . From the step response we find P.O. = 5% Tp = 0.67 sec Ts = 0.25 sec . (a) The characteristic equation is 1+ s(s + 3)(s2 K =0. + 4s + 7.84) The root locus is shown in Figure P7.13. The breakaway point is s = −1.09 at K = 9.72. (b) When K = 13.5, the roots are s1,2 = −0.84 ± j0.84 s3,4 = −2.66 ± j1.55 . 6 4 2 x + Imag Axis P7.13 + 0 x x + + -2 x -4 -6 -6 -4 -2 0 Real Axis FIGURE P7.13 Root locus for 1 + K s(s+3)(s2 +4s+7.84) = 0. 2 4 6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 316 CHAPTER 7 The Root Locus Method (c) The roots s = −0.84 ± j0.84 are dominant roots. (d) For the dominant roots, we determine that ζ = 0.7 and ωn = 1.19. Therefore, the settling time is Ts = sec . The characteristic equation is 1+ K(s + 2.5)(s + 3.2) =0. + 1)(s + 10)(s + 30) s2 (s The root locus is shown in Figure P7.14. When K = 559.3, the roots are s1 = −30.75 s2 = −8.48 s3 = −1.78 s4,5 = ±j3.11 . s3 = −2.21 s4,5 = ±j10.23 . When K = 4321, the roots are s1 = −34.45 s2 = −4.35 The crossover points are s = ±j3.11 and s = ±j10.23 . 25 20 15 10 Imaginary Axis P7.14 4 = 4.8 ζωn 5 0 $& $)' $)& $(' $(& $%& $%' $(& $(' $)& Real Axis FIGURE P7.14 (s+2.5)(s+3.5) Root locus for 1 + K s2 (s+1)(s+10)(s+30) = 0. $)' $& 0 5 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 317 Problems Therefore, the system is stable for 559.3 < K < 4321 . The characteristic equation is 1+ K(s2 + 30s + 625) . s(s + 20)(s2 + 20s + 200)(s2 + 60s + 3400) The root locus is shown in Figure P7.15. When K = 30000, the roots are s1 = −18.5 s2 = −1.69 s3,4 = −9.8±j8.9 s5,6 = −30.1±j49.9. The real root near the origin dominates, and the step response is overdamped. 100 80 60 x 40 20 Imag Axis P7.15 o x 0 x x x -20 o -40 x -60 -80 -100 -100 -80 -60 -40 -20 0 20 40 Real Axis FIGURE P7.15 s2 +30s+625 Root locus for 1 + K s(s+20)(s2 +20s+200)(s 2 +60s+3400) = 0. 60 80 100 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 318 CHAPTER 7 (a) Let τ = 0. Then, first reduce the motor and rolls to an equivalent G(s) as follows: G(s) = 1 0.25 s(s+1) 0.25 + s(s+1) = 0.25 0.25 = . s(s + 1) + 0.25 (s + 0.5)2 The loop transfer function is then L(s) = 2(s + 0.5)Ka (0.25) 0.5Ka = . 2 2 s(s + 1) (s + 0.5) s(s + 1)2 (s + 0.5) The characteristic equation is 1 + Ka 0.5 =0. s(s + 1)2 (s + 0.5) The root locus is shown in Figure P7.16. 2 1.5 1 0.5 Imag Axis P7.16 The Root Locus Method + + 0 x x x + + -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 Real Axis FIGURE P7.16 Root locus for 1 + 0.5Ka s(s+1)2 (s+0.5) = 0. (b) When K = 0.123, the roots of the characteristic equation are s1,2 = −1.1 ± j0.27 s3,4 = −0.15 ± j0.15 . 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 319 Problems The roots at s = −0.15 ± j0.15 have a damping ratio of ζ = 0.707. (c) When τ becomes nonnegligible, the root locus will have an additional pole, and the root locus will change accordingly. The characteristic equation is 2 (M1 s2 + bs + k1 + k12 )(M2 s2 + k12 ) − k12 =0. 2 is negligible If we let M1 = k1 = b = 1, and assume k12 < 1 so that k12 and k1 + k12 ≈ k1 , then the characteristic equation is (s2 + s + 1)(M2 s2 + k12 ) = 0 or 1+ k =0, s2 where k= k12 . M2 The root locus is shown in Figure P7.17. All the roots lie on the jω axis. If we select s k12 = ωo , M2 then we cancel the vibration. 3 2 root locus --> 1 Imag Axis P7.17 0 x -1 -2 -3 -3 -2.5 -2 -1.5 -1 Real Axis FIGURE P7.17 Root locus for 1 + k s2 = 0. -0.5 0 0.5 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 320 CHAPTER 7 The characteristic equation is βs3 + (1 + 2β)s2 + (2 + 4α)s + 4 = 0 . When β = 0 we have 1+ s2 4αs =0. + 2s + 4 The root locus for β = 0 is shown in Figure P7.18. 3 <-- zeta=0.6 beta=0 2 x * 1 Imag Axis P7.18 The Root Locus Method 0 o -1 * x -2 -3 -3 -2 -1 0 1 Real Axis FIGURE P7.18 4s = 0, where β = 0. Root locus for 1 + α s2 +2s+4 For α = 0.3, the poles are s = −1.6 ± j1.2 . Then, we have 1+ β(s + 2)s2 =0. s2 + (2 + 4α)s + 4 When β = 0.121 s1,2 = −1.51 ± j1.51 s3 = −7.24 . 2 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 321 Problems Thus, ζ = 0.707 and ζωn = 1.5 . So, the performance specs are met. Also, Gc (s) = P7.19 0.3s + 1 2.48(s + 3.33) = . 0.121s + 1 (s + 8.26) The characteristic equation is 1+ Ka (s2 + 4s + 100) =0. s(s + 2)(s + 6) The root locus is shown in Figure P7.19. 10 8 6 Imaginary Axis 4 2 0 −2 −4 −6 −8 −10 −10 −8 −6 −4 Real Axis −2 0 FIGURE P7.19 s2 +4s+100 Root locus for 1 + Ka s(s+2)(s+6) = 0. When Ka = 0.094;, the roots are s1,2 = −0.85 ± j0.85 s3 = −6.38 . Thus, the complex roots have a damping ratio of ζ = 0.707. P7.20 The characteristic equation is s3 + (2 + 1 2 2 4 )s + ( + K)s + = 0 , β β β 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 322 CHAPTER 7 The Root Locus Method where K= 4α β α = 0.3 β = 0.121 . The root sensitivity to changes in K is found to be r1 ∼ SK = ∆r1 = 1.186 −149.75o . ∆K/K The root sensitivity to changes in the pole at s = −2 is found to be r1 ∼ ∆r1 S∆ = 1.656 −137o , = ∆/2 (a) Let the pole be (s + 4 + ∆) and neglect ∆2 terms. Then, the characteristic equation is 1+∆ 2s2 + (8 + 2δ)s + 8δ =0 s3 + (8 + δ)s2 + (16 + 8δ)s + 16δ + K where δ = 0.000788 and K = 19.1. 3 2 x 1 Imag Axis P7.21 where the pole is s + 2 + ∆. 0 x o o -1 x -2 -3 -10 -8 -6 -4 -2 0 2 Real Axis FIGURE P7.21 2s2 +(8+2δ)s+8δ Root locus for 1 + ∆ s3 +(8+δ)s2 +(16+8δ)s+16δ+K = 0, (δ = 0.000788 and K = 19.1). The root sensitivity is determined to be r1 ∼ ∆r1 S∆ = 3.3146 −132o . = ∆/4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 323 Problems (b) Let R = Ro + ∆R, where R = 0.00021. Then, r1 ∼ ∆r1 SR = 1.31 6 −107o . = ∆R/R P7.22 The characteristic equation is s3 + 2s2 + s + K , where K = 0.24 for ζ = 0.707. The root sensitivity to changes in the pole at s = −1 is found to be r1 ∼ ∆r1 S∆ = 0.956 −126o , = ∆ where the pole is s + 1 + ∆. P7.23 The characteristic equation is s3 + 5s2 + (6 + K)s + K , where K = 6.3 for ζ = 0.707. The root sensitivity to changes in the pole at s = −2 is found to be r1 ∼ ∆r1 S∆ = 1.256 −169.4o , = ∆/2 where the pole is s + 2 + ∆. The root sensitivity to changes in the zero at s = −1 is found to be r1 ∼ ∆r1 S∆ = 0.556 34.4o , = ∆ where the zero is s + 1 + ∆. P7.24 The root locus for each of the four cases shown is shown in Figure P7.24. The four open-loop transfer functions are (a) KF (s) = s+8 + + 296s3 + 1170s2 + 1575s 1 s2 + 6s + 6.75 KF (s) = 6 (d) KF (s) = 3 5 4 s + 2s + s s + 5s2 + 4s (b) KF (s) = (c) s2 + 7s + 8.25 s3 + 6s2 + 5s s5 30s4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 324 CHAPTER 7 The Root Locus Method (a) 5 (b) 10 0 ox Imag Axis Imag Axis 5 ox x 0 x ox x x x -5 -5 -10 -5 Real Axis -10 -20 0 (c) 2 5 -10 Real Axis 0 (d) 0 x Imag Axis Imag Axis 1 x 0 ox ox x -1 -2 -2 0 -5 -10 2 Real Axis -5 0 Real Axis FIGURE P7.24 Root locus for the four cases. P7.25 The characteristic equation is 1 + KGc (s)G(s) = 0 , therefore, KGc (s)G(s) = −1 . Squaring both sides yields K 2 G2c (s)G2 (s) = 1 and 1 − K 2 G2c (s)G2 (s) = 0 . The root locus with 0 < K 2 < ∞ is shown in Figure P7.25. The value of K 2 for which the locus crosses the imaginary axis is K 2 = 2/3 , p therefore K = 2/3 = 0.8165 corresponds to the jω-axis crossing (at s = 0). You can check that 1 + KGc (s)G(s) = 0 for K = 0.8165 and s = 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 325 Problems 3 2 Imag Axis 1 0 -1 -2 -3 -3 -2 -1 0 Real Axis 1 2 3 FIGURE P7.25 Root locus for the equation 1 − K 2 G2c (s)G2 (s) = 0. P7.26 (a) The characteristic equation is 1+ K(s + 2)2 =0. s(s2 + 1)(s + 8) The root locus is shown in Figure P7.26. (b) Using Routh’s criteria, we determine that K > 14 for stability. (c) From the Routh array, we determine that for K = 14, we have two purely imaginary poles at √ s = ±j 8 . (d) When K > 50, the real part of the complex roots is approximately equal to the real part of the two real roots and therefore the complex roots are not dominant roots. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 326 CHAPTER 7 The Root Locus Method 15 10 Imag Axis 5 0 -5 -10 -15 -15 -10 -5 0 Real Axis 5 10 15 FIGURE P7.26 (s+2)2 Root locus for 1 + K s(s2 +1)(s+8) = 0. P7.27 The characteristic equation is 1+ K(s2 + 0.1) =0. s(s2 + 2) The root locus is shown in Figure P7.27a. The locus enters the axis at s = −1.26 and leaves the axis at s = −0.36 . Define p(s) = K = −(s3 + 2s) . s2 + 0.1 Then, a plot of p(s) vs s is shown in Figure P7.27b, where it can be seen that p(s) has two inflection points at s = −1.28 and s = −0.36 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 327 Problems Root Locus 2 1.5 Imaginary Axis 1 0.5 0 −0.5 −1 −1.5 −2 −3 −2.5 −2 −1.5 −1 Real Axis −0.5 0 0.5 3.5 3 2.5 p(s) 2 1.5 1 0.5 0 −2 −1.8 −1.6 −1.4 −1.2 −1 s −0.8 −0.6 −0.4 −0.2 0 FIGURE P7.27 s2 +0.1 s3 +2s (a) Root locus for 1 + K s(s 2 +2) = 0. (b) Plot of p(s) = − s2 +0.1 versus s. P7.28 The characteristic equation is 1 + L(s) = 1 + K(s2 + 12s + 20) =0. s3 + 10s2 + 25s The root locus is shown in Figure P7.28. The breakaway point is s = −5.0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 328 CHAPTER 7 The Root Locus Method 6 4 Imag Axis 2 0 -2 -4 -6 -20 -15 -10 -5 0 Real Axis FIGURE P7.28 (s2 +12s+20) Root locus for 1 + K s3 +10s2 +25s = 0. and the entry point is s = −15.6. When K = 2, the roots are s1 = −1.07 s2,3 = −5.46 ± j2.75 . When K = 2, the roots are s1 = −1.07 s2,3 = −4.36 ± j1.68 . The predicted step response when K = 2 is Ts = 9 sec and P O ≈ 0%. P7.29 The characteristic equation is 1+K s2 + 10s + 30 =0. s2 (s + 10) The root locus is shown in Figure P7.29. When ζ = 0.707, the necessary gain is K = 16. The corresponding roots are s1 = −18.87 and s2,3 = −3.56 ± j3.56. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 329 Problems Root Locus 4 System: sys Gain: 16 Pole: −3.56 + 3.57i Damping: 0.707 Overshoot (%): 4.34 Frequency (rad/sec): 5.04 3 Imaginary Axis 2 1 0 −1 −2 −3 −4 −20 −15 −10 −5 0 5 Real Axis FIGURE P7.29 2 = 0. Root locus for 1 + K ss+10s+30 2 (s+10) P7.30 The transfer function is Z(s) = LCRs2 + Ls Rs2 + s = . LCs2 + CRs + 1 s2 + Rs + 1 So, R r1 = − + 2 !1 R2 −1 4 2 . Thus, the nominal r1o = − 21 . Simultaneously, R r2 = − − 2 !1 R2 −1 4 2 . Thus, the nominal r2o = −2. We see that there is a difference by a factor of 4. Also, ri SR ∂r1 = ∂R Ro Ro Ro2 · Ro = − + 2 4 !− 1 Ro2 −1 4 2 = 5 , 6 where Ro = 2.5. And r2 SR ∂r2 = ∂R Ro Ro2 Ro = − − 2 4 Ro !− 1 Ro2 −1 4 2 = −10 . 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 330 CHAPTER 7 The Root Locus Method r2 r1 So, the magnitude of |SR | = 4|SR |. P7.31 The characteristic equation is 1+K s(s + 0.16)(s2 s+4 =0. + 14.6s + 148.999) The root locus is shown in Figure P7.31. When K = 1350, the roots are 20 15 10 (+) K=326 --> x * + <-- K=1350 (*) Imag Axis 5 + * 0 o * xx + -5 + x -10 * -15 -20 -20 -15 -10 -5 0 5 10 Real Axis FIGURE P7.31 s+4 Root locus for 1 + K s(s+0.16)(s2 +14.6s+148.999) = 0. s1,2 = ±j9.6 s3,4 = −7.4 ± j1.9 . When K = 326, the roots are s1,2 = −6.5 ± j8.7 P7.32 s3,4 = −0.9 ± j3.2 . The characteristic equation is 1+ K(s + 1)(s + 5) =0. s(s + 1.5)(s + 2) 15 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 331 Problems 4 3 * 2 * Imag Axis 1 0 o x x o** x -1 -2 * -3 * -4 -10 -8 -6 -4 -2 0 2 4 Real Axis FIGURE P7.32 (s+1)(s+5) Root locus for 1 + K s(s+1.5)(s+2) = 0. K TABLE P7.32 ζ Ts (sec) P.O. (%) 1.57 0.707 0.98 1.4 3.48 0.707 1.1 5.8 2.35 0.69 1.3 4.0 Step Response Results for K = 1.57, K = 3.48, and K = 2.35. (a) The breakaway point is s = −1.73; the entry point is s = −8.62. (b) The damping ratio ζ = 0.707 when K = 1.57 and again when K = 3.46. (c) The minimum damping ratio ζ = 0.69 is achieved when K = 2.35. (d) The results are summarized in Table P7.32.The best choice of gain is K = 1.57. P7.33 (a) The root locus for the V-22 is shown in Figure P7.33a. The system is stable when 0 < K < 0.48 and K > 136.5. (b) The unit step input response (for K = 280) is shown in Figure P7.33b. The step response has a P.O. = 90% and Ts ≈ 50 sec. (c) The plot of y(t) for a unit step disturbance is shown in Figure P7.33b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method 2 1.5 1 Imag Axis 0.5 0 x o o -1 -0.5 xxx -0.5 -1 -1.5 -2 -3 -2.5 -2 -1.5 0 0.5 1 Real Axis (i) Unit step input response 2 y(t) w/o prefilter .... (dotted line) y(t) with prefilter ____ (solid line) y(t) 1.5 1 0.5 0 4 0 x 10 10 20 10 20 -3 30 40 50 Time (sec) (ii) Unit step disturbance response 60 70 80 60 70 80 3 y(t) 332 2 1 0 -1 0 30 40 Time (sec) 50 FIGURE P7.33 s2 +1.5s+0.5 (a) Root locus for 1 + K s(20s+1)(10s+1)(0.5s+1) = 0. (b) (i) Unit step input response with and without prefilter; (ii) Unit step disturbance response. The response to the disturbance is oscillatory, but the maximum value of oscillation is about 0.003; so it is negligible. (d) The effect of adding a prefilter can be seen in Figure P7.33b. With © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 333 Problems the prefilter we find P O = 7% and Ts ≈ 40 sec. P7.34 The characteristic equation is 1+ K(s + 2) =0. (s + 1)(s + 2.5)(s + 4)(s + 10) The root locus is shown in Figure P7.34a. The roots, predicted and actual percent overshoot for K = 400, 500, and 600 are summarized in Table P7.34. The actual unit step input responses are shown in Figure P7.34b. roots ζ predicted P.O. (%) actual P.O. (%) 400 -13.5,-1.00 ± 5.71j,-1.98 0.173 57.6 51.6 500 -14.0,-0.75 ± 6.24j,-1.98 0.120 68.4 61.2 600 -14.4,-0.53 ± 6.71j,-1.98 0.079 77.9 69.6 TABLE P7.34 Summary for K = 400, 500, 600. Root Locus 20 15 10 Imaginary Axis K 5 0 −5 −10 −15 −20 −30 −25 −20 −15 −10 Real Axis −5 FIGURE P7.34 s+2 (a) Root locus for 1 + K (s+1)(s+2.5)(s+4)(s+10) = 0. 0 5 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 334 CHAPTER 7 The Root Locus Method 1.6 K=400 .... (dotted line) 1.4 K=500 −−− (dashed line) K=600 ___ (solid line) 1.2 y(t) 1 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 Time (sec) 12 14 16 18 20 FIGURE P7.34 CONTINUED (b) Unit step input responses for K = 400, 500, 600. (a) The root locus is shown in Figure P7.35 for the characteristic equation 1+ K(s + 1)2 =0. s(s2 + 1) 3 K=4.52 2 * 1 Imag Axis P7.35 x 0 o -1 x x -2 -3 -5 * * -4 -3 -2 Real Axis FIGURE P7.35 (s+1)2 Root locus for 1 + K s(s2 +1) = 0. -1 0 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 335 Problems (b) When K = 4.52, the roots are s1 = −0.58 s2,3 = −1.96 ± j1.96 . The complex roots have ζ = 0.707. (c) The entry point is s = −3.38 when K = 7.41. (d) The predicted P.O. = 4.5% The characteristic equation is 1+ K(s + 1)(s + 2)(s + 3) =0. s3 (s − 1) (a) The root locus is shown in Figure P7.36. 8 6 4 2 Imag Axis P7.36 (ζ = 0.707) and the actual P.O. = 17%. 0 o o o x x -2 -4 -6 -8 -10 -8 -6 -4 -2 Real Axis FIGURE P7.36 (s+1)(s+2)(s+3) Root locus for 1 + K = 0. s3 (s−1) (b) When K = 2.96, the roots are s1,2 = ±j4.08 s3,4 = −0.98 ± j0.33 . 0 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 336 CHAPTER 7 The Root Locus Method (c) When K = 20, the roots are s1 = −1.46 s2 = −1.07 s3,4 = −8.23 ± j2.99 . When K = 100, the roots are s1 s2 s3 s4 = −92.65 = −3.51 = −1.82 = −1.01 . (d) When K = 20, the damping ratio is ζ = 0.94. Therefore, the predicted P.O. = 0.02%. The actual overshoot is P.O. = 23%. P7.37 Since we know that ess = 0 for a step input, we know that a = 0 or b = 0. Select a = 0. Also, ωn = 2π/T = 20 rad/sec. The desired characteristic polynomial is (s + r1 )(s + j20)(s − j20) = s3 + r1 s2 + 400s + 400r1 = 0 . The actual characteristic polynomial is 1+ 2K =0, s(s + b)(s + 40) s3 + (40 + b)s2 + 40bs + 2K = 0 . or Comparing the coefficients in the desired and actual characteristic polynomials, we determine that b = 10, r1 = 50, and K = 10000. P7.38 (a) The characteristic equation is 1+ K(s + 1) =0. s(s − 3) √ The system is stable for K > 3. When K = 3, the roots are s = ±j 3. (b) The root locus is shown in Figure P7.38a. (c) When K = 10 , the roots are s1 = −2 s2 = −5 . Since both roots are real and stable, we expect that there will be zero overshoot. The actual response has a 40% overshoot, as seen in Figure P7.38b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 337 Problems 6 4 Imag Axis 2 0 o x x -2 -4 -6 -6 -4 -2 0 2 4 6 2 2.5 3 Real Axis 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 Time (secs) FIGURE P7.38 s+1 (a) Root locus for 1 + K s(s−3) = 0. (b) Unit step response. P7.39 The loop transfer function is Gc (s)G(s) = 22K . (s + 1)(s2 + 8s + 22) When K = 0.529, the closed-loop poles are s1,2 = −3.34 ± 1.83j and s3 = −2.32 and have the maximum damping ζ = 0.877. The root locus is shown in Figure P7.39a. The step response is shown in Figure P7.39b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method Root Locus 10 8 6 Imaginary Axis 4 2 0 −2 −4 −6 −8 −10 −14 −12 −10 −8 −6 Real Axis −4 −2 0 2 Step Response 0.35 0.3 0.25 Amplitude 338 0.2 0.15 0.1 0.05 0 0 FIGURE P7.39 (a) Root locus for 0.5 1 22K (s+1)(s2 +8s+22) 1.5 2 Time (sec) 2.5 = 0. (b) Unit step response. 3 3.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 339 Advanced Problems Advanced Problems The characteristic equation is 1+K s+6 =0. s(s + 4)(s2 + 4s + 8) The root locus is shown in Figure AP7.1. The gain at maximum ζ is 10 5 Imag Axis AP7.1 x + 0 o x+ + x + x -5 -10 -10 -5 0 5 10 Real Axis FIGURE AP7.1 s(s+4) Root locus for 1 + K s2 +2s+2 = 0. K = 3.7 . The roots at K = 3.7 are s1 = −3.6424 s2,3 = −1.3395 ± +1.3553j s4 = −1.6786 . Using Figure 5.13 in Dorf & Bishop, the predicted percent overshoot and settling time are P.O. = 5% and Ts = 3 sec , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 340 CHAPTER 7 The Root Locus Method since ζ = 0.7 and a 6 = = 4.5 . ωn ζ 1.9(0.7) The actual percent overshoot and settling time are P.O. = 1% and Ts = 2.8 sec. The characteristic equation is 1+K (s + 1)(s + 4) =0. s(s − 1)(s + 5)(s + 10) The root locus is shown in Figure AP7.2a. The selected gain is K = 43.7. 15 Imaginary Axis 10 5 0 −5 −10 −15 −12 −10 −8 −6 −4 Real Axis −2 0 2 1.5 System: syscl Peak amplitude: 1.48 Overshoot (%): 48.3 At time (sec): 0.857 System: syscl Settling Time (sec): 2.31 1 Amplitude AP7.2 0.5 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 FIGURE AP7.2 (s+1)(s+4) (a) Root locus for 1 + K s(s−1)(s+5)(s+10) = 0; (b) Step response for K = 43.7. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 341 Advanced Problems The actual percent overshoot (see Figure AP7.2b) is P.O. = 48.3%. AP7.3 The characteristic equation (with p as the parameter) is 1+p s3 s(s + 1) =0. + s2 + 10 The root locus is shown in Figure AP7.3. 5 4 3 2 x Imag Axis 1 + 0 x o o + -1 x -2 -3 -4 -5 -5 -4 -3 -2 -1 0 1 2 3 4 5 Real Axis FIGURE AP7.3 s(s+1) Root locus for 1 + p s3 +s2 +10 = 0. When p = 21 the dominant roots have a damping ratio of ζ = 0.707. AP7.4 The characteristic equation (with α as the parameter) is 1+α s(s + 1) =0. + s2 + 1 s3 The root locus is shown in Figure AP7.4a. The steady-state error is 1 =1−α . s→0 1 + G(s) ess = lim sE(s) = lim s→0 To meet the steady-state error specification, we require 0.9 < α < 1.1 . The step responses for α = 0.9, 1 and 1.1 are shown in Figure AP7.4b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 342 CHAPTER 7 The Root Locus Method 3 2 Imag Axis 1 x 0 x o o x -1 -2 -3 -3 -2 -1 0 1 2 3 Real Axis alpha=0.9 (solid); alpha=1.0 (dashed); alpha=1.1 (dotted) 1.8 1.6 1.4 Amplitude 1.2 1 0.8 0.6 0.4 0.2 0 0 5 10 15 20 25 30 35 40 45 50 Time (sec) FIGURE AP7.4 s(s+1) (a) Root locus for 1 + p s3 +s2 +10 = 0. (b) Step responses for α = 0.9, 1 and 1.1. AP7.5 The root locus is shown in Figure AP7.5. When K = 20.45, ζ = 0.707. The r1 ∼ root sensitivity is SK = ∆r1 /(∆K/20.45) = 3.156 87.76o . When K = 88, the complex roots lie on the jω-axis—a 330% increase in the gain. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 343 Advanced Problems 5 4 3 2 Imag Axis 1 + 0 +x x x + -1 -2 -3 -4 -5 -15 -10 -5 0 5 Real Axis FIGURE AP7.5 Root locus for 1 + K s3 +10s21+7s−18 = 0. A gain of K = 13 provides an acceptable response of Ts < 1 and P.O. < 7.5%. The root locus is shown in Figure AP7.6. Root Locus 2.5 2 1.5 1 Imaginary Axis AP7.6 0.5 0 −0.5 −1 −1.5 −2 −2.5 −3 −2.5 −2 FIGURE AP7.6 s2 +3s+6 Root locus for 1 + K s3 +2s 2 +3s+1 = 0. −1.5 −1 Real Axis −0.5 0 0.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 344 CHAPTER 7 AP7.7 The Root Locus Method The root locus for the positive feedback system is shown in Figure AP7.7. 15 10 Imag Axis 5 0 x x -5 -10 -15 -15 -10 -5 0 5 10 15 Real Axis FIGURE AP7.7 −1 Root locus for 1 + K s2 +12s+32 = 0. The root locus is shown in Figure AP7.8a. When k = 0.448, all the roots 30 20 10 Imag Axis AP7.8 x 0 x o x -10 -20 -30 -30 -20 -10 0 Real Axis FIGURE AP7.8 (a) Root locus for 1 + k s3 +19s120s 2 +34s+120 = 0. 10 20 30 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 345 Advanced Problems of the characteristic equation are real—the step response is shown in Figure AP7.8b. 1 0.9 0.8 0.7 Amplitude 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1 1.5 2 2.5 3 Time (secs) FIGURE AP7.8 CONTINUED (b) Step response with k = 0.448. The root locus for each controller is shown in Figure AP7.9. AP7.9 (a) (b) 5 Imaginary Axis Imaginary Axis 5 0 −5 −15 −10 −5 Real Axis 0 0 −5 −15 5 −10 (c) 0 5 0 5 (d) 15 5 10 Imaginary Axis Imaginary Axis −5 Real Axis 5 0 −5 0 −10 −15 −15 −10 −5 Real Axis FIGURE AP7.9 Root locus for the various controllers. 0 5 −5 −15 −10 −5 Real Axis © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 346 CHAPTER 7 AP7.10 The Root Locus Method The characteristic equation (with K as the parameter) is 1+K s2 + 7s + 20 =0. s(s2 + 7s + 10) The root locus is shown in Figure AP7.10. The steady-state value of the 10 8 6 4 Imag Axis 2 0 -2 -4 -6 -8 -10 -10 -8 -6 -4 -2 0 Real Axis 2 4 6 8 10 FIGURE AP7.10 s2 +7s+20 Root locus for 1 + K s(s 2 +7s+10) = 0. step response for any K is 0.5. With K = 15 the closed-loop transfer function is T (s) = 10s + 150 . s3 + 22s2 + 115s + 300 The step response has the following characteristics: P.O. = 4.8% AP7.11 and Ts = 2 seconds . The root locus is shown in Figure AP7.11a. A suitable gain is K = 500. The step response is shown in Figure AP7.11b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 347 Advanced Problems Root Locus 80 60 Imaginary Axis 40 20 0 −20 −40 −60 −80 −100 −80 −60 −40 −20 Real Axis 0 20 40 FIGURE AP7.11 (s+2)2 (a) Root locus for 1 + K s(s+10)(s+20)(s2 +3s+3.5) = 0. Step Response 1.4 System: sys_cl Peak amplitude: 1.09 Overshoot (%): 9.01 At time (sec): 0.945 1.2 Amplitude 1 System: sys_cl Settling Time (sec): 2.39 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 Time (sec) 2.5 3 3.5 4 FIGURE AP7.11 CONTINUED: (b) Step response with K = 500. AP7.12 The root locus is shown in Figure AP7.12a. The PI controller can be written as Gc (s) = Kp s + KI s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method 8 6 4 Imag Axis 2 0 -2 -4 -6 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 Real Axis Step Response From: U(1) 1.4 1.2 1 0.8 To: Y(1) Amplitude 348 0.6 0.4 0.2 0 0 5 10 15 Time (sec.) FIGURE AP7.12 (s+0.2) (a) Root locus for 1 + Kp s(s2 +7s+10) = 0. (b) Step response with Kp = 5.54. and setting KI = 0.2Kp , the characteristic equation can be written as 1 + Kp (s + 0.2) =0 + 7s + 10) s(s2 A suitable gain is Kp = 5.55. The step response is shown in Figure AP7.12b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 349 Advanced Problems AP7.13 The characteristic equation is 1 + K1 K2 1 = 0. (s + 5)(s − 1) The root locus is shown in Figure AP7.12a. The fastest expected settling Root Locus 4 3 Imaginary Axis 2 1 0 −1 −2 −3 −4 −6 −5 −4 −3 −2 Real Axis −1 0 1 2 FIGURE AP7.13 1 Root locus for 1 + K1 K2 (s+5)(s−1) = 0. time is Ts = 4/ωn ζ = 2 seconds since maximum |ωn ζ| = 2. AP7.14 The root locus of the uncompensated transfer function is shown in Figure AP7.14a. It can be seen that the system is unstable for Ku = 131.25 with a period of Tu = 0.72, as illustrated in FigureAP7.14b. Using the Ziegler-Nichols design formulas yields KP = 0.6Ku = 78.75, KI = 1.2Ku /Tu = 218.75, and KD = 0.6Ku Tu = 7.0875 where © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method Root Locus 30 System: sysg Gain: 131 Pole: 0.0153 + 8.66i Damping: −0.00176 Overshoot (%): 101 Frequency (rad/sec): 8.66 Imaginary Axis 20 10 0 −10 −20 −30 −40 −30 −20 −10 Real Axis 0 10 20 8 10 12 FIGURE AP7.14 10 = 0. (a) Root locus for 1 + Ku s(s+10)(s+7.5) Step Response 2 1.8 1.6 1.4 Amplitude 350 1.2 1 0.8 0.6 0.4 0.2 0 0 2 4 6 Time (sec) FIGURE AP7.14 CONTIUED: (b) Step response at the ultimate gain Ku = 131. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 351 Advanced Problems Step Response 1.6 System: sys_cl Peak amplitude: 1.6 Overshoot (%): 59.5 At time (sec): 0.445 1.4 Amplitude 1.2 1 System: sys_cl Settling Time (sec): 2.1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 Time (sec) 2.5 3 3.5 FIGURE AP7.14 CONTINUED: (c) Step response with the Ziegler-Nichols tuned PID controller. Step Response −3 14 x 10 12 10 Amplitude 8 6 4 2 0 −2 −4 0 0.5 1 1.5 2 Time (sec) 2.5 3 3.5 FIGURE AP7.14 CONTINUED: (d) Disturbance response with the Ziegler-Nichols tuned PID controller. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 352 CHAPTER 7 The Root Locus Method Design Problems CDP7.1 The closed-loop transfer function from the input to the output is 26.035Ka θ(s) = 2 , R(s) s + (33.1415 + 26.035Ka K1 )s + 26.035Ka where we consider for the first time the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). The characteristic equation is 1 + K1 26.035Ka s =0. s2 + 33.1415s + 26.035Ka The root locus is shown below. In accordance with the discussion in Chap30 20 Imag Axis 10 0 -10 -20 -30 -30 -20 -10 0 Real Axis 10 20 30 ter 5, we continue to use Ka = 22. This allows us to meet the overshoot specification (P.O. < 5%) without the tachometer feedback and to provides good steady-state tracking errors to a step input. To meet the design specifications of both P.O. and Ts we want the closed-loop poles to the left of −ζω = −4/0.3 = −13.33 and ζ > 0.69. A reasonable selection is K1 = 0.012. This places the closed-loop poles at s = −20 ± j13. DP7.1 (a) The characteristic equation is 1+ (s2 18K(s + 0.015)(s + 0.45) =0. + 1.2s + 12)(s2 + 0.01s + 0.0025) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 353 Design Problems Since we want a negative feedback system, we have Gc (s) = −K. When ωn > 2 and ζ = 0.15, the gain K = 0.12. The root locus is shown in Figure DP7.1a. 6 4 x Imag Axis 2 o oxx -0.5 0 0 -2 x -4 -6 -4 -3.5 -3 -2.5 -2 -1.5 -1 0.5 1 Real Axis FIGURE DP7.1 18(s+0.015)(s+0.45) (a) Root locus for 1 + K (s2 +1.2s+12)(s2 +0.01s+0.0025) = 0. (b) The unit step response is shown in Figure DP7.1b. The percent overshoot is P.O. = 100% . (c) The characteristic equation with the anticipatory controller is 1+ 18K(s + 2)(s + 0.015)(s + 0.45) =0. (s2 + 1.2s + 12)(s2 + 0.01s + 0.002s) The root locus is shown in Figure DP7.1c. If we select K = 9.2/18 , then the complex roots have a damping ζ = 0.90. The roots are at s1 = −0.253 s2 = −0.019 s3,4 = −5.07 ± j2.50 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method 0.7 0.6 Amplitude 0.5 0.4 0.3 0.2 0.1 0 0 20 40 60 80 100 120 140 160 180 200 Time (secs) FIGURE DP7.1 CONTINUED: (b) Unit step response for gain controller. 6 4 x 2 Imag Axis 354 0 xx o o o -2 x -4 -6 -6 -4 -2 0 2 4 6 Real Axis FIGURE DP7.1 18(s+2)(s+0.015)(s+0.45) CONTINUED: (c) Root locus for 1 + K (s2 +1.2s+12)(s2 +0.01s+0.0025) = 0. (d) The unit step response for the system with the anticipatory controller is shown in Figure DP7.1d. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 355 Design Problems 1 0.9 0.8 Amplitude 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 20 40 60 80 100 120 140 160 180 200 Time (secs) FIGURE DP7.1 CONTINUED: (d) Unit step response for anticipatory controller. DP7.2 The characteristic equation is 1+ 10K(s + 1) =0. s(s2 + 4.5s + 9) (a) The root locus is shown in Figure DP7.2a. When K = 0.435, we have ζ = 0.6 and the roots are s1 = −0.368 s2,3 = −2.1 ± j2.75 . (b) The response to a step input is shown in Figure DP7.2b. The performance results are P.O. = 0% Tss = 10 sec ess = 0 . (c) We have ζ = 0.41 when K = 1.51. The step response is shown in Figure DP7.2b. The performance results to the step input are P.O. = 0% Ts = 4 sec ess = 0 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 356 CHAPTER 7 The Root Locus Method 5 4 3 2 x Imag Axis 1 0 o x -1 0 -1 -2 x -3 -4 -5 -5 -4 -3 -2 1 2 3 4 5 Real Axis FIGURE DP7.2 10(s+1) (a) Root locus for 1 + K s(s2 +4.5s+9) = 0. 1 0.9 K=0.435 ____ (solid line) 0.8 K=1.510 ---- (dashed line) Amplitude 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 2 4 6 8 10 12 Time (sec) FIGURE DP7.2 CONTINUED: (b) Unit step responses for K = 0.425, 1.51. DP7.3 The characteristic equation is 1+ K(s2 + 6.5s + 12) =0. s(s + 1)(s + 2) 14 16 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 357 Design Problems (a) The root locus is shown in Figure DP7.3. 6 4 2 Imag Axis o 0 x x x -2 -1 0 o -2 -4 -6 -6 -5 -4 -3 1 Real Axis FIGURE DP7.3 s2 +6.5s+12 = 0. Root locus for 1 + K s(s+1)(s+2) When K = 41, the roots are s1 = −37.12 and s2,3 = −3.44 ± j1.19 . (b) The percent overshoot is P.O. ≈ 1% when ζ = 0.82 at K = 0.062. (c) Select K > 300. DP7.4 The characteristic equation is 1+K 10(0.01s + 1) =0. s(s2 + 10s + 10K1 ) If we choose K1 = 2.5, then the root locus will start at s = 0, −5 and -5. This is shown in Figure DP7.4. The root locus then has a nice shape so that we can select K to place the complex poles where desired and the one real root will be farther in the left half-plane; thus the notion of dominant poles will be valid. So, if we desire a P.O. < 5%, we want ζ > 0.69. This occurs when K ≈ 3. Thus, our design is K1 = 2.5 and K = 3 . The unit step response is shown in Figure DP7.4. The settling time is less than 3.5 sec and the P O < 4%. The response to a unit step disturbance is also shown in Figure DP7.4. The steady-state error magnitude to the disturbance is 0.33. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 358 CHAPTER 7 The Root Locus Method 4 3 2 K=3 --> Imag Axis 1 0 -1 -2 -3 -4 -20 -15 -10 -5 Real Axis 0 5 10 1.4 Input step response Disturbance step response 1.2 1 y(t) 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 5 FIGURE DP7.4 10(0.01s+1) (a) Root locus for 1 + K s(s2 +10s+25) = 0. (b) System response to step input and disturbance. DP7.5 The characteristic equation is 1+K s+1 =0. s(s − 0.1)(s2 + 10s + 41) The root locus is shown in Figure DP7.5a. The system is stable for 5 < K < 300. The step response with K = © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 359 Design Problems Root Locus 10 8 System: sysgc Gain: 90.5 Pole: −1.42 + 2.24i Damping: 0.536 Overshoot (%): 13.6 Frequency (rad/sec): 2.66 6 Imaginary Axis 4 2 0 −2 −4 −6 −8 −10 −12 −10 −8 −6 −4 Real Axis −2 0 2 Step Response 1.6 System: sys_cl Peak amplitude: 1.57 Overshoot (%): 57 At time (sec): 1.24 1.4 Amplitude 1.2 System: sys_cl Time (sec): 3.39 Amplitude: 0.98 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 FIGURE DP7.5 s+1 (a) Root locus for 1 + K s(s−0.1)(s 2 +10s+41) = 0. (b) Step response with K = 875. 90.5 is shown in Figure DP7.5b. We choose K = 90.5 to minimize the settling time. The damping of the dominant poles is ζ = 0.54, so that the estimated percent overshoot is P.O. = 13%. The actual percent overshoot and settling time are P.O. = 57% and Ts = 3.4 seconds. The match between the actual and predicted percent overshoot can be improved by selecting a much higher gain K, but then the step response becomes overy oscillatory and the settling time increases too much for a typical high-performance aircraft. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 360 CHAPTER 7 The characteristic equation is 1+K s+2 =0. s(s + 10)(s − 1) The maximum damping is ζ = 0.46 at K = 55. The root locus is shown in Figure DP7.6a; the step response is shown in Figure DP7.6b. The percent overshoot and settling time are P.O. = 61.3% and Ts = 2 seconds. 20 15 10 Imag Axis 5 + 0 x + o x x + -5 -10 -15 -20 -20 -15 -10 -5 0 5 10 15 20 Real Axis 1.8 1.6 1.4 1.2 Amplitude DP7.6 The Root Locus Method 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 Time (secs) FIGURE DP7.6 s+2 (a) Root locus for 1 + K s(s+10)(s−1) = 0. (b) Step response with K = 55. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 361 Design Problems DP7.7 The loop transfer function is Gc (s)G(s) = KP s + KI . s(s + 1)(0.5s + 1) One possible set of PI controller gains are KP = 0.82 and KI = 0.9. The step response is shown in Figure DP7.7. Step Response 1.4 System: syscl Peak amplitude: 1.05 Overshoot (%): 4.59 At time (sec): 3.57 1.2 Amplitude 1 System: syscl Settling Time (sec): 4.94 0.8 0.6 0.4 0.2 0 0 1 2 3 4 Time (sec) 5 6 7 8 FIGURE DP7.7 Step response for with PI controller Gc (s) = (0.82s + 0.9)/s. DP7.8 The closed-loop transfer function is T (s) = Vo (s) G(s) = . V (s) 1 + KG(s) The dc gain is T (0) = G(0) 1 ≈ . 1 + KG(0) K The root locus is shown in Figure DP7.8. The maximum value of K for stability is K = 0.062 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 362 CHAPTER 7 The Root Locus Method x10 7 2 1.5 1 + Imag Axis 0.5 0+ x x -0.5 -1 + -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 Real Axis 2 x10 7 FIGURE DP7.8 3.142K1 ×1017 Root locus for 1 + K (s+3142)(s+10 7 )2 = 0. Therefore, the minimum dc gain is about 1/0.062=16. Selecting K = 0.05 and R1 = 10 K yields R2 = 19R1 = 190 K . DP7.9 The closed-loop transfer function (with Gp (s) = 1 and K = 1) is T (s) = 2s3 + 6s2 + 14s + 10 . s4 + 6s3 + 13s2 + 26s + 6 So, if we select Gp (s) = 1/T (0) = 0.6, the step response (with K = 1) will have a zero steady-state tracking error. The root locus is shown in Figure DP7.9a. The step responses for K = 1, 1.5 and 2.85 are shown in Figure DP7.9b. For K = 1, we have P.O. = 0%, Tr = 7.8 and Ts = 13.9; for K = 1.5, we have P.O. = 0%, Tr = 5.4 and Ts = 9.6; and for K = 2.85, we have P.O. = 5.2%, Tr = 0.5 and Ts = 7.3. The best gain selection is K = 2.85. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 363 Design Problems 8 6 4 Imag Axis 2 x 0 x o -2 x x -4 -6 -8 -8 -6 -4 -2 0 2 4 6 8 Real Axis K=1 (solid); K=1.5 (dashed); K=2.85 (dotted) 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 20 Time (sec) FIGURE DP7.9 6(s+1) (a) Root locus for 1 + K s(s+4)(s2 +2s+5) = 0. (b) Step responses with K = 1, 1.5, 2.85. DP7.10 A suitable selection of the various parameters is ζ = 0.5 and q = 3/5 . With q = 3/5, the open-loop zeros are real and equal. Then, it follows that λ= 2q =3. 1−q © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 364 CHAPTER 7 The Root Locus Method The root locus is shown in Figure DP7.10. A reasonable choice of gain is K = 30 . The resulting step response is extremely fast with no overshoot. The closed-loop transfer function is approximately given by T (s) ≈ 1923 . s + 1923 6 4 x Imag Axis 2 0 o x -2 x -4 -6 -6 -4 -2 0 2 4 6 Real Axis FIGURE DP7.10 4s2 +4s+1 Root locus for 1 + K 0.0625s 3 +0.25s2 +s = 0. DP7.11 The characteristic equation (with K as the parameter) is 1+K 10(s2 + 10) =0. s3 + 20s The root locus is shown in Figure DP7.11a. To maximize the closed-loop system damping we choose K = 0.513. The step response is shown in Figure DP7.11b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 365 Design Problems 5 4 3 2 Imag Axis 1 0 -1 -2 -3 -4 -5 -2 -1.5 -1 -0.5 Real Axis 0 0.5 1 Step Response From: U(1) 1.4 1.2 0.8 To: Y(1) Amplitude 1 0.6 0.4 0.2 0 0 1 2 3 4 5 Time (sec.) FIGURE DP7.11 10(s2 +10) (a) Root locus for 1 + K s3 +20s = 0. (b) Step response with K = 0.513. DP7.12 The characteristic equation is 1+K s + 1.5 =0. (s + 1)(s + 2)(s + 4)(s + 10) The root locus is shown in Figure DP7.12a. 6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method 10 8 6 4 Imag Axis 2 0 x x x o x -2 -4 -6 -8 -10 -15 -10 -5 0 5 Real Axis K=100 (solid); K=300 (dashed); K=600 (dotted) 1.6 1.4 1.2 1 Amplitude 366 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 6 7 8 9 10 Time (sec) FIGURE DP7.12 s+1.5 (a) Root locus for 1 + K (s+1)(s+2)(s+4)(s+10) = 0. (b) Step response with K = 100, 300, 600. The closed-loop system roots are: K = 100 : s1 = −11.38 K = 300 : s1 = −12.94 K = 600 : s1 = −14.44 s2,3 = −2.09 ± 3.10j s2,3 = −1.29 ± 5.10j s2,3 = −0.53 ± 6.72j The step responses are shown in Figure DP7.12b. s4 = −1.45 s4 = −1.48 s4 = −1.49 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 367 Design Problems DP7.13 The closed-loop transfer function is T (s) = s3 Ka . + Ka K2 s + Ka s2 + A suitable choice of gains is Ka = 0.52 and K2 = 3 . The step response is shown in Figure DP7.13. 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 20 Time (secs) FIGURE DP7.13 Step response with Ka = 0.52 and K2 = 3. DP7.14 The characteristic equation is s2 + 10KD s + 10(KP + 1) = 0 . In the Evans form we have 1 + KD 10(s + τ ) =0. s2 + 10 The root locus is shown in Figure DP7.14 for τ = 6. As τ → 0,√the dominant closed-loop pole approaches s = 0 as KD√→ ∞. As τ → 10, the dominant closed-loop pole approaches s = − 10 as KD → ∞. A viable controller is KP = 72 and KD = 12 when τ = 6. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method Root Locus 8 6 4 Imaginary Axis 368 2 0 −2 −4 −6 −8 −25 FIGURE DP7.14 Root locus when τ = 6. −20 −15 −10 Real Axis −5 0 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 369 Computer Problems Computer Problems The root locus for parts (a)-(d) are shown in Figures CP7.1a - CP7.1d. num=[30]; den=[1 14 43 30]; rlocus(sys) 30 Imaginary Axis 20 10 0 −10 −20 −30 −40 −30 −20 −10 Real Axis 0 10 20 num=[1 20]; den=[1 4 20]; rlocus(sys) 20 15 10 Imaginary Axis CP7.1 5 0 −5 −10 −15 −20 −70 −60 −50 −40 −30 Real Axis −20 −10 0 10 FIGURE CP7.1 s+20 (a) Root locus for 1 + k s3 +14s230 = 0. (b) Root locus for 1 + k s2 +4s+20 = 0. +43s+30 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method num=[1 1 2]; den=[1 6 10 0]; rlocus(sys) 1.5 Imaginary Axis 1 0.5 0 −0.5 −1 −1.5 −6 −5 −4 −3 −2 Real Axis −1 0 1 num=[1 4 6 10 6 4]; den=[1 4 4 1 1 10 1]; rlocus(sys) 1.5 1 Imaginary Axis 370 0.5 0 −0.5 −1 −1.5 −7 −6 −5 −4 −3 Real Axis −2 −1 0 1 FIGURE CP7.1 2 +s+2 CONTINUED: (c) Root locus for 1 + k s(ss2 +6s+10) = 0. (d) Root locus for 1 + 5 4 3 2 +4s +6s +10s +6s+4 k s6s+4s 5 +4s4 +s3 +s2 +10s+1 = 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 371 Computer Problems CP7.2 The maximum value of the gain for stability is k = 0.791. The m-file script and root locus is shown in Figure CP7.2. Select a point in the graphics window num=[1 -2 2]; den=[1 3 2 0]; sys = tf(num,den); selected_point = rlocus(sys) rlocfind(sys) -0.0025 + 0.6550i ans = 0.8008 1 0.8 0.6 0.4 Imag Axis 0.2 0 −0.2 −0.4 −0.6 −0.8 −1 −3 −2.5 −2 −1.5 −1 −0.5 Real Axis 0 0.5 1 1.5 2 FIGURE CP7.2 Using the rlocfind function. The value of k = 0.8008 selected by the rlocfind function is not exact since you cannot select the jω-axis crossing precisely. The actual value is determined using Routh-Hurwitz analysis. CP7.3 The partial fraction expansion of Y (s) is Y (s) = s(s2 s+6 0.1667 1.6667 1.5 = − + . + 5s + 4) s+4 s+1 s The m-file script and output is shown in Figure CP7.3. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 372 CHAPTER 7 The Root Locus Method r= 0.1667 -1.6667 1.5000 p= num=[1 6]; den=[1 5 4 0]; [r,p,k]=residue(num,den) -4 -1 0 k= [] FIGURE CP7.3 Using the residue function. The characteristic equation is 1+p s−1 =0. s2 + 5s + 10 The root locus is shown in Figure CP7.4. The closed-loop system is stable for 0 < p < 10 . n*+,-. /.01 23n,-. 4 .501 6789*:;:<:= 2.5 2 1.5 1 Imaginary Axis CP7.4 0.5 0 −0.5 −1 −1.5 −2 −2.5 −8 −7 −6 FIGURE CP7.4 s−1 Root locus for 1 + p s2 +5s+10 = 0. −5 −4 −3 −2 Real Axis −1 0 1 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 373 Computer Problems CP7.5 The characteristic equation is 1+k s+1 =0. s2 The root locus is shown in Figure CP7.5. For k = 2 we obtain s1,2 = −1 ± j, that is, we have ζ = 0.707. num=[1 1]; den=[1 0 0]; sys = tf(num,den); hold off, clf rlocus(sys); hold on plot([0 -2],[0 2*tan(acos(0.707))],'--') plot([0 -2],[0 -2*tan(acos(0.707))],'--') plot([-1 -1],[1 -1],'*') 2.5 2 1.5 1 Imag Axis 0.5 0 -0.5 -1 -1.5 -2 -2.5 -3 -2.5 -2 -1.5 -1 Real Axis -0.5 0 0.5 1 FIGURE CP7.5 Root locus for 1 + k s+1 = 0. s2 CP7.6 We choose a controller with two real poles and two real zeros selected to meet the steady-state specification. The characteristic equation is 1+K (s + 5.5)(s + 0.01) 10 =0. (s + 6.5)(s + 0.0001) s3 + 15s2 + 50s The m-file and root locus is shown in Figure CP7.4a. From the root locus we can select the value of the gain K that results in an estimated P.O. ≤ 5% and a ζωn ≤ −2 to meet the settling time specification. We © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 374 CHAPTER 7 The Root Locus Method select K = 8.58. The step response is shown in Figure CP7.6b showing the percent overshoot and settling time specifications are satisfied. The velocity constant is kv = 145.2 which implies a steady-state error to a ramp input of ess = 1/kv = 0.0069. 25 20 15 System: untitled1 Gain: 8.58 Pole: Damping: 0.807 Overshoot (%): 1.37 Frequency (rad/sec): 2.58 10 Imaginary Axis ng=10; dg=conv([1 10 0],[1 5]); s ysg=tf(ng,dg); nh=conv([1 0.01],[1 5.5]); dh=conv([1 6.5],[1 0.0001]); sysh=tf(nh,dh); figure(1) rlocus(sysg*sysh) K=8.58; sysh=tf(K*nh,dh); sys=series(sysg,sysh);syscl=feedback(sys,1) figure(2) subplot(121) step(syscl); Kv=10*8.58*0.01*5.5/10/6.5/0.0001/5 systd=feedback(sysg,sysh); subplot(122) step(systd) 5 0 −5 −10 −15 −20 −25 −40 −35 −30 −25 Step Response −15 −10 Real Axis −5 0 5 10 Disturbance Response 1.4 0.14 System: syscl Peak 1.2amplitude: 1.02 Overshoot (%): 1.77 At time (sec): 2.14 0.12 1 0.1 System: syscl Settling Time (sec): 1.51 0.8 y(t)/Q Amplitude −20 0.08 0.6 0.06 0.4 0.04 0.2 0.02 0 0 1 2 3 Time (sec) 4 5 0 0 200 400 Time (sec) 600 FIGURE CP7.6 (a) Root locus. (b) Step response and disturbance response. CP7.7 The m-file script to generate the root locus for each controller in parts (a)-(c) is shown in Figure CP7.7. The performance region is indicated on each root locus in Figures CP7.7b - CP7.7d. For part (a), the controller gain is found to be Gc (s) = 11.3920. The integral controller in part (b) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 375 Computer Problems numg=[1]; deng=[1 5 6]; sysg = tf(numg,deng); t=[0:0.1:15]; % % Part (a) % ÈSelect a point in the graphics window sys1 = sysg; rlocus(sys1), grid selected_point = hold on plot([-0.4 -0.4],[-6 6],'--',... -2.5030 + 3.3380i [0 -6*tan(36.2*pi/180)],[0 6],'--',... [0 -6*tan(36.2*pi/180)],[0 -6],'--') ans = hold off [kp,poles] = rlocfind(sys1) 11.3920 % % Part (b) % numc=[1]; denc=[1 0]; sysc = tf(numc,denc); sys2 = series(sysc,sysg); Select a point in the graphics window figure rlocus(sys2), grid selected_point = hold on plot([-0.4 -0.4],[-6 6],'--',... -0.6690 + 0.8210i [0 -6*tan(36.2*pi/180)],[0 6],'--',... [0 -6*tan(36.2*pi/180)],[0 -6],'--') ans = hold off [ki,poles] = rlocfind(sys2) 4.0930 % % Part (c) % Plot performance region boundaries on graph. figure numc=[1 1]; denc=[1 0]; sysc = tf(numc,denc); Select a point in the graphics window sys3 = series(sysc,sysg); rlocus(sys3), grid selected_point = hold on plot([-0.4 -0.4],[-6 6],'--',... -2.0695+ 2.7387i [0 -6*tan(36.2*pi/180)],[0 6],'--',... [0 -6*tan(36.2*pi/180)],[0 -6],'--') ans = hold off [kpi,poles] = rlocfind(sys3) 9.2516 % % Part (d) % figure sys1_o = kp*sys1; sys1_cl = feedback(sys1_o,[1]); sys2_o = ki*sys2; sys2_cl = feedback(sys2_o,[1]); sys3_o = kpi*sys3; sys3_cl = feedback(sys3_o,[1]); [y1,t]=step(sys1_cl,t); [y2,t]=step(sys2_cl,t); [y3,t]=step(sys3_cl,t); plot(t,y1,t,y2,'--',t,y3,':'),grid xlabel('time [sec]'),ylabel('y(t)') title('Gc(s): proportional (solid), integral (dashed) & PI (dotted)') FIGURE CP7.7 (a) Script to generate the root locus for each controller. is determined to be Gc (s) = 4.093 . s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method 6 4 Imag Axis 2 0 -2 -4 -6 -4 -3 -2 -1 Real Axis 0 1 2 FIGURE CP7.7 CONTINUED: (b) Root locus for proportional controller with selected K = 11.3920. The proportional integral (PI) controller in part (c) is 6 4 2 Imag Axis 376 0 -2 -4 -6 -4 -3 -2 -1 Real Axis 0 1 2 FIGURE CP7.7 CONTINUED: (c) Root locus for integral controller with selected K = 4.0930. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 377 Computer Problems Gc (s) = 9.2516(s + 1) . s The proportional controller is stable for all K > 0 but has a significant 6 4 Imag Axis 2 0 -2 -4 -6 -4 -3 -2 -1 Real Axis 0 1 2 FIGURE CP7.7 CONTINUED: (d) Root locus for PI controller with selected K = 9.2516. steady-state error. The integral controller has no steady-state error, but is stable only for K < 30. The PI controller has zero steady-state error and is stable for all K > 0. Additionally, the PI controller has a fast transient response. The step responses for each controller is shown in Figure CP7.7e. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 378 CHAPTER 7 The Root Locus Method Gc(s): proportional (solid), integral (dashed) & PI (dotted) 1.4 1.2 1 y(t) 0.8 0.6 0.4 0.2 0 0 5 10 15 time [sec] FIGURE CP7.7 CONTINUED: (e) Step responses for each controller. CP7.8 The loop transfer function can be written as Gc (s)G(s) = K1 + K2 s s+5 = K̄2 2 2 Js s where K̄2 = K2 /J . The parameter of interest for the root locus is K̄2 . The root locus is shown in Figure CP7.8. The selected value of K̄2 = 7.1075 . Therefore, K2 = 7.1075 J and K1 = 35.5375 . J © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 379 Computer Problems num=[1 5]; den=[1 0 0]; sys=tf(num,den); rlocus(sys); rlocfind(sys) 10 8 6 + 4 Imag Axis 2 0 o x -2 -4 + -6 -8 -10 -10 -8 -6 -4 -2 0 2 4 6 8 10 Real Axis FIGURE CP7.8 Root locus to determine K̄2 . The value of K that results in a damping ratio of ζ = 0.707 is K = 5.2. The root locus is shown in Figure CP7.9. Root Locus 5 4 3 2 Imaginary Axis CP7.9 s = -0.68 + 0.68j 1 0 −1 s = -6.63 s = -0.68 - 0.68j −2 −3 −4 −5 −10 −5 0 Real Axis FIGURE CP7.9 Root locus for 1 + K s3 +8s21+10s+1 = 0. 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 380 CHAPTER 7 (a) The characteristic equation is s3 + (2 + k)s2 + 5s + 1 = 0 . (b) The Routh array is s3 1 5 s2 2+k 1 s1 5k+9 2+k so 1 Root Locus 2 1.5 1 Imaginary Axi s CP7.10 The Root Locus Method 0.5 0 ?-0.5 ?-1 ?-1.5 ?-2 ?-2.5 ?-2 ?-1.5 ?-1 ?-0.5 Real Axi s FIGURE CP7.10 2 Root locus for 1 + k s3 +2ss2 +5s+1 = 0. For stability we require 2 + k > 0 or k > −2 and 5k + 9 > 0 or k > −9/5 . Therefore, the stability region is defined by k > −1.8 . 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 381 Computer Problems (c) Rearranging the characteristic equation yields 1+k s2 s3 + 2s2 + 5s + 1 = 0 . The root locus is shown in Figure CP7.10. We see that the system is stable for all k > 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 8 Frequency Response Methods Exercises E8.1 Given the loop transfer function L(s) = 4 , (s + 2)2 we determine that |L(jω)| = 4 4 + ω2 and φ(ω) = −2 tan−1 ω/2 . The frequency response is shown in Figure E8.1. Bode Diagram Magnitude (dB) 0 −20 −40 −60 Phase (deg) −80 0 −45 −90 −135 −180 −2 10 FIGURE E8.1 Frequency response for L(s) = 382 −1 10 0 10 Frequency (rad/sec) 4 . (s+2)2 1 10 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 383 Exercises The magnitude and phase angle for ω = 0, 0.5, 1, 2, 4, ∞ are summarized in Table E8.1. ω 0 0.5 1 2 4 ∞ |L(jω)| 1 0.94 0.80 0.50 0.20 0 φ (deg) 0 -28.07 -53.13 -90 –126.87 -180 TABLE E8.1 E8.2 Magnitude and phase for L(s) = 4 . (s+2)2 The transfer function is G(s) = 5000 . (s + 70)(s + 500) The frequency response plot is shown in Figure E8.2. The phase angle is computed from φ = − tan−1 ω ω − tan−1 . 70 500 The phase angles for ω = 10, 100 and 700 are summarized in Table E8.2. ω TABLE E8.2 10 200 700 |G(jω)| -16.99 -27.17 -41.66 φ (deg) -9.28 -92.51 -138.75 Magnitude and phase for G(s) = 5000 . (s+70)(s+500) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 384 CHAPTER 8 Frequency Response Methods Bode Diagram Magnitude (dB) 0 −20 −40 −60 −80 Phase (deg) −100 0 −45 −90 −135 −180 0 10 1 10 FIGURE E8.2 Frequency response for G(s) = E8.3 2 3 10 Frequency (rad/sec) 4 10 10 5000 (s+70)(s+500) . The loop transfer function is L(s) = 300(s + 100) . s(s + 10)(s + 40) The phase angle is computed via φ(ω) = −90o − tan−1 ω ω ω − tan−1 + tan−1 . 10 40 100 At ω = 28.3, we determine that φ = −90o − 70.5o − 35.3o + 15.8o = 180o . Computing the magnitude yields 1 |L(jω)| = ω 2 2 300(100)(1 + ( 100 ) ) 1 1 ω 2 2 ω 2 2 ω10(1 + ( 10 ) ) 40(1 + ( 40 ) ) when ω = 28.3. We can also rewrite L(s) as L(s) = s 75( 100 + 1) . s s s( 10 + 1)( 40 + 1) = 0.75 , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 385 Exercises Then, the magnitude in dB is ω 2 ω ) ) − 10 log 10 (1 + ( )2 ) 100 10 ω − 10 log 10 (1 + ( )2 ) − 20 log10 ω = −2.5 dB , 40 20 log 10 |L| = 20 log 10 (75) + 10 log 10 (1 + ( at ω = 28.3. E8.4 The transfer function is G(s) = Ks . (s + a)(s + 10)2 Note that φ = 0o at ω = 3, and that φ = +90o − tan−1 ω ω − 2 tan−1 . a 10 Substituting ω = 3 and solving for a yields a=2. Similarly, from the magnitude relationship we determine that K = 400 . E8.5 The lower portion for ω < 2 is 20 log K = 0 dB , ω at ω = 8. Therefore, 20 log K = 0 dB 8 which occurs when K=8. We have a zero at ω = 2 and another zero at ω = 4. The zero at ω = 4 yields a = 0.25 . We also have a pole at ω = 8, and a second pole at ω = 24. The pole at ω = 24 yields b = 1/24 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 386 CHAPTER 8 Frequency Response Methods Therefore, G(s) = E8.6 8(1 + s/2)(1 + s/4) . s(1 + s/8)(1 + s/24)(1 + s/36) The loop transfer function is L(s) = 10 . s(s/5 + 1)(s/100 + 1) The Bode diagram is shown in Figure E8.6. When 20 log 10 |L(jω)| = 0 dB, we have ω = 9.4 rad/sec . Bode Diagram Magnitude (dB) 50 0 −50 −100 Phase (deg) −150 −90 −135 −180 −225 −270 −1 10 0 10 FIGURE E8.6 Bode Diagram for L(s) = E8.7 1 2 10 10 Frequency (rad/sec) 3 10 4 10 10 . s(s/5+1)(s/100+1) The transfer function is T (s) = 4 . (s2 + s + 1)(s2 + 0.4s + 4) (a) The frequency response magnitude is shown in Figure E8.7. The frequency response has two resonant peaks at ωr1 = 0.8 rad/sec and ωr2 = 1.9 rad/sec . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 387 Exercises 10 Gain dB 5 0 -5 -10 10-1 100 Frequency (rad/sec) 101 Amplitude 1.5 1 0.5 0 0 2 4 FIGURE E8.7 (a) Bode Diagram for T (s) = 6 8 10 12 Time (secs) 4 (s2 +s+1)(s2 +0.4s+4) . 14 16 18 20 (b) Unit step response. (b) The percent overshoot is P.O. = 35% , and the settling time is Ts ≈ 16 sec . (c) The step response is shown in Figure E8.7. E8.8 (a) The break frequencies are ω1 = 1 rad/sec, ω2 = 5 rad/sec, and ω3 = 20 rad/sec . (b) The slope of the asymptotic plot at low frequencies is 0 dB/decade. And at high frequencies the slope of the asymptotic plot is -20 dB/decade. (c) The Bode plot is shown in Figure E8.8. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 388 CHAPTER 8 Frequency Response Methods Bode Diagram 20 Magnitude (dB) 10 0 −10 −20 −30 180 Phase (deg) 135 90 45 0 −45 −90 −2 10 −1 10 FIGURE E8.8 Bode Diagram for Gc (s)G(s) = 1 2 10 3 10 100(s−1) . s2 +25s+100 The Bode diagram for G(s) is shown in Figure E8.9. 40 Gain dB 20 0 -20 -40 10-1 100 101 102 103 102 103 Frequency (rad/sec) 50 Phase deg E8.9 0 10 10 Frequency (rad/sec) 0 -50 10-1 100 101 Frequency (rad/sec]) FIGURE E8.9 Bode Diagram for G(s) = (s/5+1)(s/20+1) . (s+1)(s/80+1) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 389 Exercises E8.10 The frequency response has two peaks; the first peak at f ≈ 1.8 and the second peak at f ≈ 3.1. One possible G(jω) is 1 G(jω) = (jωτ + 1) 1 + 2ζ1 ωn1 jω + jω ωn1 2 1+ 2ζ2 ωn2 jω + jω ωn2 2 , where τ= 1 , 2π(0.2) ωn1 = 2π(1.8 × 103 ) ζ1 = 0.15; ζ2 = 0.15; ωn2 = 2π(3.1 × 103 ) . The damping ratios are estimated using Figure 8.10 in Dorf & Bishop. E8.11 The Bode plot is shown in Figure E8.11. The frequency when 20 log10 |GC G(ω)| = 0 is ω = 9.9 rad/sec. Bode Diagram 20 Magnitude (dB) 0 −20 −40 −60 −80 −100 −120 0 Phase (deg) −45 −90 −135 −180 −225 −270 −1 10 0 10 FIGURE E8.11 Bode Diagram for Gc (s)G(s) = E8.12 1 10 Frequency (rad/sec) 2 10 1000 . (s2 +10s+100)(s+2) (a) The transfer function is G(s) = C(sI − A)−1 B + D = (b) The Bode plot is shown in Figure E8.12. −5(s − 1) . s2 + 3s + 2 3 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 390 CHAPTER 8 Frequency Response Methods Bode Diagram Phase (deg) Magnitude (dB) 10 0 -10 -20 270 180 90 -2 10 10 FIGURE E8.12 Bode Diagram for G(s) = 0 10 Frequency (rad/sec) 10 1 10 2 −5(s−‘1) . s2 +3s+2 The closed-loop transfer function is T (s) = s3 + 100 . + 20s + 110 11s2 The Bode plot of T (s) is shown in Figure E8.13, where ωB = 4.9 rad/sec. Bode Diagram Magnitude (dB) 50 Phase (deg) E8.13 -1 0 -3 dB -50 -100 -45 -90 -135 -180 -225 -270 -1 10 FIGURE E8.13 Bode Diagram for T (s) = ωb=4.9 10 0 1 10 Frequency (rad/sec) 100 s3 +11s2 +20s+110 . 10 2 10 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 391 Exercises E8.14 The loop transfer function is L(s) = 20 . (s2 + 1.4s + 1)(s + 10) The Bode plot of L(s) is shown in Figure E8.14. The frequency when 20 log 10 |L(ω)| = 0 is ω = 1.32 rad/sec. Bode Diagram Magnitude (dB) 50 0 −50 −100 −150 0 Phase (deg) −45 −90 −135 −180 −225 −270 −2 10 FIGURE E8.14 Bode Diagram for L(s) = E8.15 −1 0 10 1 10 10 Frequency (rad/sec) 2 10 3 10 20 . (s2 +1.4s+1)(s+10) The closed-loop transfer function is T (s) = 3s + 5 . s2 + s + K + 6 The bandwidth as a function of K is shown in Figure E8.15. The bandwidth as a function of K is: (a) K = 1 and ωb = 7.0 rad/sec. (b) K = 2 and ωb = 7.9 rad/sec. (c) K = 10 and ωb = 14.7 rad/sec. The bandwidth increases as K increases. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 8 Frequency Response Methods 24 22 20 18 ωb (rad/s) 392 16 14 12 10 8 6 0 FIGURE E8.15 Bandwith of T (s) = 2 4 3s+5 s2 +s+K+6 . 6 8 10 K 12 14 16 18 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 393 Problems Problems (a) The transfer function is 1 , (1 + 0.25s)(1 + 3s) Gc (s)G(s) = and 1 . (1 − 0.75ω 2 ) + j3.25ω Gc (jω)G(jω) = The polar plot is shown in Figure P8.1a. A summary of the magnitude and phase angles for ω = 0, 0.5, 1, 2, 5 and ∞ can be found in Table P8.1a. Nyquist Diagram 0.8 0.6 0.4 Imaginary Axis P8.1 0.2 0 −0.2 −0.4 −0.6 −0.8 −1 −0.8 −0.6 FIGURE P8.1 (a) Polar plot for Gc (s)G(s) = −0.4 −0.2 0 Real Axis 0.2 0.4 0.6 0.8 1 1 . (1+0.25s)(1+3s) ω 0 0.5 1 2 5 ∞ |Gc (jω)G(jω)| (dB) 1.00 0.55 0.31 0.15 0.04 0 φ (deg) 0 -63.4 -85.6 -107.1 -137.51 -180 TABLE P8.1 (a) Magnitudes and phase angles for Gc (s)G(s) = 1 . (1+0.25s)(1+3s) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 394 CHAPTER 8 Frequency Response Methods (b) The transfer function is Gc (s)G(s) = 5(s2 + 1.4s + 1) (s − 1)2 and 5 (1 − ω 2 ) + 1.4jω Gc (jω)G(jω) = . (1 − ω 2 ) − 2jω The polar plot is shown in Figure P8.1b. A summary of the magnitude and phase angles for ω = 0, 0.25, 0.5, 1, 2, 8, 16 and ∞ can be found in Table P8.1b. Nyquist Diagram 5 4 3 Imaginary Axis 2 1 0 −1 −2 −3 −4 −5 −4 −3 −2 FIGURE P8.1 CONTINUED: (b) Polar plot for Gc (s)G(s) = −1 0 1 Real Axis 2 3 4 5 6 5(s2 +1.4s+1) . (s−1)2 ω 0 0.25 0.5 1 2 8 16 ∞ |Gc (jω)G(jω)| (dB) 5.00 4.71 4.10 3.50 4.10 4.92 4.98 5.00 φ (deg) 0 48.5 96.1 180.0 -96.2 -24.3 -12.2 0 TABLE P8.1 CONTINUED: (b) Magnitudes and phase angles for Gc (s)G(s) = 5(s2 +1.4s+1) . (s−1)2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 395 Problems (c) The transfer function is Gc (s)G(s) = (s2 (s − 8 . + 6s + 8) The polar plot is shown in Figure P8.1c. A summary of the magnitude and phase angles for ω = 0, 1, 2, 3, 4, 5, 6, ∞ can be found in Table P8.1c. Nyquist Diagram 0.8 0.6 Imaginary Axis 0.4 0.2 0 −0.2 −0.4 −0.6 −0.8 −1 −0.8 −0.6 FIGURE P8.1 CONTINUED: (c) Polar plot for Gc (s)G(s) = −0.4 −0.2 Real Axis 0 0.2 0.4 s−8 . s2 +6s+8 ω 0 1 2 3 4 5 6 ∞ |Gc (jω)G(jω)| (dB) 1.00 0.87 0.65 0.47 0.35 0.27 0.22 0.00 φ (deg) 180.0 132.3 94.4 66.3 45.0 28.5 15.3 -90.0 TABLE P8.1 CONTINUED: (c) Magnitudes and phase angles for Gc (s)G(s) = s−8 . s2 +6s+8 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 396 CHAPTER 8 Frequency Response Methods (d) The transfer function is Gc (s)G(s) = 20(s + 8) . s(s + 2)(s + 4) The polar plot is shown in Figure P8.1d. A summary of the magnitude and phase angles for ω = 1, 0.1, 0.8, 1.6, 3.2, 12.8, ∞ can be found in Table P8.1d. Nyquist Diagram 20 15 Imaginary Axis 10 5 0 −5 −10 −15 −20 −20 −15 −10 −5 0 5 Real Axis FIGURE P8.1 CONTINUED: (d) Polar plot for Gc (s)G(s) = 20(s+8) . s(s+2)(s+4) ω 0 0.1 0.8 1.6 3.2 12.8 ∞ |Gc (jω)G(jω)| (dB) ∞ 199.70 22.87 9.24 2.79 0.14 0.00 φ (deg) 0 -93.6 -117.4 -139.1 -164.8 174.3 180.0 TABLE P8.1 CONTINUED: (d) Magnitudes and phase angles for Gc (s)G(s) = 20(s+8) . s(s+2)(s+4) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 397 Problems P8.2 (a) The Bode plot is shown in Figure P8.2a. A summary of the magnitude and phase angles for ω = 0.25, 0.5, 1, 2, 4, 8, 16 can be found in Table P8.2a. Bode Diagram Magnitude (dB) 0 −20 −40 −60 Phase (deg) −80 0 −45 −90 −135 −180 −2 10 FIGURE P8.2 (a) Bode plot for Gc (s)G(s) = −1 10 0 1 10 Frequency (rad/sec) 2 10 10 1 . (1+0.25s)(1+3s) ω 0.25 0.5 1.0 2.0 4.0 8.0 16.0 |Gc (jω)G(jω)| (dB) -1.95 -5.19 -10.26 -16.65 -24.62 -34.60 -45.93 φ (deg) -40.5 -63.4 -85.6 -107.1 -130.2 -151.0 -164.8 TABLE P8.2 (a) Magnitudes and phase angles for Gc (s)G(s) = 1 . (1+0.25s)(1+3s) (b) The transfer function is Gc (s)G(s) = 5(s2 + 1.4s + 1) (s − 1)2 The Bode plot is shown in Figure P8.2b. A summary of the magnitude © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 398 CHAPTER 8 Frequency Response Methods and phase angles for ω = 0.25, 0.5, 1, 2, 4, 8, 16 can be found in Table P8.2b. Bode Diagram Magnitude (dB) 14 13 12 11 Phase (deg) 10 0 −90 −180 −270 −360 −1 10 0 1 10 Frequency (rad/sec) FIGURE P8.2 CONTINUED: (b) Bode plot for Gc (s)G(s) = 10 5(s2 +1.4s+1) . (s−1)2 ω 0.25 0.5 1.0 2.0 4.0 8.0 16.0 |Gc (jω)G(jω)| (dB) 13.46 12.26 10.88 12.26 13.46 13.84 13.95 φ (deg) 48.5 96.2 180.0 -96.2 -48.5 -24.3 -12.2 TABLE P8.2 CONTINUED: (b) Magnitudes and phase angles for Gc (s)G(s) = 5(s2 +1.4s+1) . (s−1)2 (c) The transfer function is Gc (s)G(s) = (s2 (s − 8) . + 6s + 8) The Bode plot is shown in Figure P8.2c. A summary of the magnitude and phase angles for ω = 0.6, 1, 2, 3, 4, 5, 6, ∞ can be found in Table P8.2c. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 399 Problems Bode Diagram 0 Magnitude (dB) −10 −20 −30 −40 −50 −60 180 Phase (deg) 135 90 45 0 −45 −90 −2 10 −1 0 10 1 2 10 10 Frequency (rad/sec) FIGURE P8.2 CONTINUED: (c) Bode plot for Gc (s)G(s) = 3 10 10 s−8 s2 +6s+8 . ω 0.6 1 2 3 4 5 6 ∞ |Gc (jω)G(jω)| (dB) -0.45 -1.17 -3.72 -6.49 -9.03 -11.26 -13.18 -120.00 φ (deg) 150.5 132.3 94.4 66.3 45.0 28.5 15.3 -90.0 TABLE P8.2 CONTINUED: (c) Magnitudes and phase angles for Gc (s)G(s) = s−8 s2 +6s+8 . (d) A summary of the magnitude and phase angles for ω = 0.2, 0.8, 3.2, 6.4, 12.8, 25.6, 51.2 can be found in Table P8.2d. The Bode plot is shown in Figure P8.2d. ω 0.2 0.8 3.2 6.4 12.8 25.6 51.2 |Gc (jω)G(jω)| (dB) 39.95 27.19 8.90 -3.98 -17.35 -30.0355 -42.28 φ (deg) -97.1 -117.4 -164.8 178.0 174.2 176.0 177.8 TABLE P8.2 CONTINUED: (d) Magnitudes and phase angles for Gc (s)G(s) = 20(s+8) . s(s+2)(s+4) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 400 CHAPTER 8 Frequency Response Methods Bode Diagram 60 Magnitude (dB) 40 20 0 −20 −40 Phase (deg) −60 −90 −135 −180 −225 −1 10 0 1 10 2 10 10 Frequency (rad/sec) FIGURE P8.2 CONTINUED: (d) Bode plot for Gc (s)G(s) = P8.3 20(s+8) . s(s+2)(s+4) (a) The bridged-T network we found has zeros at s = ±jωn and poles at s=− q ωn ± ωn 1/Q2 − 1 . Q The frequency response is shown in Figure P8.3 for Q = 10. (b) For the twin-T network, we evaluate the magnitude at ω = 1.1ωn or 10% from the center frequency (see Example 8.4 in Dorf & Bishop). This yields |G| ≈ 2.1 × 0.1 3.9 × 1.1 = 0.05 . Similarly, for the bridged-T network |G| = 2.1 × 0.1 2.1 × 0.14 = 0.707 . The bridged-T network possesses a narrower band than the twin-T network. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 401 Problems 0 Gain dB -10 -20 -30 -40 10-1 100 101 w/wn Phase deg 100 50 0 -50 -100 10-1 100 101 w/wn FIGURE P8.3 Bode plot for G(s) = 2 s2 +ωn 2 , s2 +(2ωn /Q)s+ωn where ζ = 1/Q = 0.1. The transfer function is P8.4 1 s 30000(2s + 1) = . s(s + 10)(s + 20)(s2 + 15s + 150) G(s) = Gc G1 H(s) A summary of the magnitude and phase angles can be found in Table P8.4. The Bode plot is shown in Figure P8.4. ω 1 3 5 8 10 15 24 |G(jω| dB 6.95 5.78 5.08 3.38 1.59 -5.01 -17.56 −40.89o −52.39o −77.28o −118.41o −145.99o −203.52o −258.57o φ(deg) TABLE P8.4 Magnitudes and phase angles for GH(s) = 30000(2s+1) . s(s+10)(s+20)(s2 +15s+150) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 402 CHAPTER 8 Frequency Response Methods Bode Diagram Magnitude (dB) 50 0 -50 -100 Phase (deg) -150 0 -90 -180 -270 -360 -2 10 FIGURE P8.4 Bode plot for GH(s) = -1 0 1 10 10 Frequency (rad/sec) 10 2 10 30000(2s+1) . s(s+10)(s+20)(s2 +15s+150) The Bode plot is shown in Figure P8.5. Bode Diagram 50 Magnitude (dB) 0 −50 −100 −150 −200 −250 0 Phase (deg) P8.5 10 −90 −180 −270 −360 −2 10 FIGURE P8.5 Bode plot for G(s) = −1 10 0 10 1 2 10 10 Frequency (rad/sec) 10 . (s/4+1)(s+1)(s/20+1)(s/80+1) 3 10 4 10 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 403 Problems (a) The transfer function is GH(s) = 3.98(1 + s/1) . s(1 + s/10)2 We have a zero at ω = 1 and two poles at ω = 10.0. The low frequency approximation is K/s and at ω = 1 we have K 20 log ω = 12dB . Therefore, K = 3.98 at ω = 1 (an approximation). The phase plot is shown in Figure P8.6a. (a) -40 -60 Phase deg -80 -100 -120 -140 -160 -180 -2 10 10 -1 10 0 10 1 10 2 (b) 100 50 Phase deg P8.6 0 -50 -100 -1 10 FIGURE P8.6 Phase plots for (a) G(s) = 10 0 3.98(s/1+1) . s(s/10+1)2 1 10 Frequency (rad/sec) (b) G(s) = 10 2 10 3 s . (s/10+1)(s/50+1) (b) The transfer function is GH(s) = s . (1 + s/10)(1 + s/50) The poles are located by noting that the slope is ±20 dB/dec. The © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 404 CHAPTER 8 Frequency Response Methods low frequency approximation is Ks, so 20 log Kω = 0dB . At ω = 1 we determine that K=1. The phase plot is shown in Figure P8.6b. The loop transfer function is L(s) = Kv . s(s/π + 1)2 (a) Set Kv = 2π. The Bode plot is shown in Figure P8.7a. Gain dB 40 20 0 -20 10-1 100 Frequency (rad/sec) 101 -80 -100 Phase deg P8.7 -120 -140 -160 -180 10-1 100 101 Frequency (rad/sec) FIGURE P8.7 (a) Bode plot for L(s) = Kv , s(s/π+1) where Kv = 2π. (b) The logarithmic magnitude versus the phase angle is shown in Figure P8.7b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 405 Problems 40 30 Gain dB 20 10 0 -10 -20 -170 -160 -150 -140 -130 -120 -110 -100 -90 Phase deg FIGURE P8.7 CONTINUED: (b) Log-magnitude-phase curve for L(jω). P8.8 The transfer function is T (s) = s2 K . + 10s + K (a) When P.O. = 15%, we determine that ζ = 0.517 by solving √ 2 15 = 100e−πζ/ 1−ζ . So, 2ζωn = 10 implies that ωn = 9.67, hence K = ωn2 = 93.53. Also, q Mpω = (2ζ 1 − ζ 2 )−1 = 1.13 . (b) For second-order systems we have q ωr = ωn 1 − 2ζ 2 = 6.59 when ζ = 0.517 and ωn = 9.67. (c) We estimate ωB to be ωB ≈ (−1.19ζ + 1.85)ωn = 11.94 rad/s . P8.9 The log-magnitude phase curves are shown in Figure P8.9. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 406 CHAPTER 8 Frequency Response Methods (a) (b) 0 40 -5 30 -10 20 -20 10 Gain dB Gain dB -15 -25 0 -30 -10 -35 -20 -40 -45 -200 -150 -100 -50 -30 -180 0 -160 Phase deg -120 -100 Phase deg FIGURE P8.9 Log-magnitude-phase curve for (a) Gc (s)G(s) = 1+0.5s . s2 P8.10 -140 1 (1+0.5s)(1+2s) and (b) Gc (s)G(s) = The governing equations of motion are F (s) = Kf If (s) and If (s) = Vf (s) . Rf + L f s Without loss of generality we can let Kf = 1.0. Also, we have F (s) = (M s2 + bs + K)Y (s) . Therefore, the transfer function is GH(s) = KKf 50K = . 2 (Rf + Lf s)(M s + bs + K) (s + 0.5)(s2 + 2s + 4) This is a type 0 system, therefore Kp = 25K. (a) If we allow a 1% error , we have ess = |R|/(1 + Kp ) = 0.01|R|. Thus Kp = 25K = 99. Select K=4. (b) The Bode plot is shown in Figure P8.10a. (c) The log-magnitude phase curve is shown in Figure P8.10b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 407 Problems Gain dB 40 20 0 -20 10-2 10-1 100 101 100 101 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-2 10-1 Frequency (rad/sec) FIGURE P8.10 (a) Bode plot for GH(s) = 200 . (s2 +2s+4)(s+0.5) 40 30 Gain dB 20 10 0 -10 -20 -300 -250 -200 -150 -100 -50 0 Phase deg FIGURE P8.10 CONTINUED: (b) Log-magnitude-phase curve for GH(s) = 200 . (s2 +2s+4)(s+0.5) (d) The closed-loop transfer function Bode plot is shown in Figure P8.10c. We determine from the plot that Mpω = 1.6, ωr = 4.4 and ωB = 6.8. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 408 CHAPTER 8 Frequency Response Methods 5 Gain dB 0 -5 -10 -15 10-1 100 101 Frequency (rad/sec) Phase deg 100 0 -100 -200 10-1 100 101 Frequency (rad/sec) FIGURE P8.10 CONTINUED: (c) Bode plot for closed-loop T (s) = Y (s)/R(s). The Bode plot is shown in Figure P8.11. Gain dB 200 100 0 -100 10-4 10-3 10-2 10-1 100 101 100 101 Frequency (rad/sec) 100 Phase deg P8.11 0 -100 -200 10-4 10-3 10-2 10-1 Frequency (rad/sec) FIGURE P8.11 Bode plot for G(s) = 0.164(s+0.2)(−s+0.32) . s2 (s+0.25)(s−0.009) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 409 Problems P8.12 The three transfer functions are G1 (s) = 10 G2 (s) = 1 s(s/0.6 + 1) G3 (s) = 3s . (a) When G3 (s) is out of the loop, the characteristic equation is 10 =0 s(s/0.6 + 1) √ or s2 + 0.6s + 6 = 0. Thus, ζ = 0.6/(2 6) = 0.12. (b) With G3 (s), the characteristic equation is 1 + G1 G2 (s) = 1 + 1 + G1 G2 (s) + G2 G3 (s) = 1 + 1.85 6 + =0, s(s + 0.6) s(s + 0.6) or s2 + 2.4s + 6 = 0 . √ Thus, ζ = 2.4/(2 6) = 0.49. P8.13 By inspection of the frequency response, we determine L(s) = Gc (s)G(s)H(s) = K . s(s/100 + 1)(s/1000 + 1)2 For small ω, we have 20 log K/ω = 40 dB at ω = 10. So, K = 1000. P8.14 The data we have are R1 = R2 = 1000Ω, c1 = 10−7 farad and c2 = 10−6 farad. The governing equations are 1 V2 (s) C1 s = , V1 (s) R1 + C11 s and Vo (s) KR2 = . V2 (s) R2 + C12 s So Vo (s) KR2 C2 s 109 s = = . V1 (s) (R1 C1 s + 1)(R2 C2 s + 1) (s + 107 )(s + 1000) (a) The Bode plot is shown in Figure P8.14. (b) The mid-band gain is = 40 dB. (c) The -3 dB points are (rad/sec): ωlow ≈ 7 and ωhigh ≈ 1.5 × 109 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 410 CHAPTER 8 Frequency Response Methods Gain dB 40 20 0 -20 100 101 102 103 104 105 106 107 108 109 1010 107 108 109 1010 Frequency (rad/sec) Phase deg 100 0 -100 -200 100 101 102 103 104 105 106 Frequency (rad/sec) FIGURE P8.14 Bode plot for G(s) = P8.15 109 s . (s+107 )(s+103 ) The data are plotted in Figure P8.15, denoted by an asterisk (*). 50 * * 0 * * * * * * * * * -50 -100 10-1 100 101 102 -50 -100 * * * * * * * * -150 * -200 -250 -300 10-1 FIGURE P8.15 Bode plot for G(s) = * 100 809.7 ; s(s2 +6.35s+161.3) 101 * 102 tabular data is indicated by an asterick (*). The low frequency slope is -20 dB/dec and the initial low frequency φ is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 411 Problems −90o , so we have an integrator of the form K/s. The initial phase is −90o and the final phase −270o , so we have a minimum phase G(s). Now, |G| is 0.97 at ω = 8 and ω = 10 indicating two complex poles. We postulate a transfer function of the form G(s) = K s s2 2 ωn + 2ζs ωn . +1 The phase angle φ = −180o at ω = ωn . Then, from Figure 8.10 in Dorf & Bishop, we determine that ωn = 12.7. At ω = 8, ωωn = 0.63 and φ, due to the complex poles is −30o (subtract −90o due to the integrator). Again, from Figure 8.10 in Dorf & Bishop, we estimate ζ = 0.25. To determine K, note that when ωωn ≤ 0.1, the effect of the complex poles on magnitude is negligible, so at ω = 1 we have K |G| = 5.02 ∼ . = j1 So K = 5.02. Therefore, G(s) = s s2 161.3 + 0.5s 12.7 = +1 809.7 . s(s2 + 6.35s + 161.3) (a) The unit step input response is shown in Figure P8.16. The step Step Response 1.4 1.2 1 Amplitude P8.16 5.02 0.8 0.6 0.4 0.2 0 0 0.1 0.2 0.3 FIGURE P8.16 Unit step input response for T (s) = 0.4 0.5 0.6 Time (sec) 60.2 s2 +12.1s+60.2 . 0.7 0.8 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 412 CHAPTER 8 Frequency Response Methods response is given by y(t) = 1 − e−6.05t (cos 4.85t + 1.25 sin 4.85t) . (b) The system bandwidth is ωB = 4.95 rad/sec. P8.17 The transfer function is Gc (s)G(s) = 4(0.5s + 1) . s(2s + 1)(s2 /64 + s/20 + 1) (a) The Bode plot is shown in Figure P8.17. Gain dB 50 0 -50 -100 10-1 100 101 102 101 102 Frequency (rad/sec) -50 Phase deg -100 -150 -200 -250 -300 10-1 100 Frequency (rad/sec) FIGURE P8.17 Bode plot for Gs (s)G(s) = 4(0.5s+1) . s(2s+1)(s2 /64+s/20+1) (b) When the magnitude is 0 dB, we have ω1 = 1.6 rad/sec and when φ = −180o , we have ω2 = 7.7 rad/sec . P8.18 The transfer function is Gc (s)G(s) = 12(s + 0.5) 0.2(2s + 1) = . (s + 3)(s + 10) (s/3 + 1)(s/10 + 1) The Bode plot is shown in Figure P8.18. Near 0 dB, the frequency is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 413 Problems ω = 5.4 rad/sec. 0 Gain dB -5 -10 -15 -20 10-1 100 101 102 101 102 Frequency (rad/sec) 50 Phase deg 0 -50 -100 -150 -200 10-1 100 Frequency (rad/sec) FIGURE P8.18 Bode plot for Gc (s)G(s) = P8.19 12(s+0.5) s2 +13s+30 . Examining the frequency response, we postulate a second-order transfer function θ(s) ωn2 = 2 . I(s) s + 2ζωn s + ωn2 From the data we see that φ = −90o at ω = 2. Using Figure 8.10 in Dorf & Bishop, we determine that ωn = ω = 2. We also estimate ζ = 0.4 from Figure 8.10. Thus, θ(s) 4 = 2 . I(s) s + 1.6s + 4 P8.20 The transfer function is Gc (s)G(s) = 823(s + 9.8) . + 22s + 471 s2 The Bode plot is shown in Figure P8.20. The maximum value of 20 log10 |Gc (jω)G(jω)| = 32.3 dB occurs at ω = 20.6 rad/sec and the corresponding phase is φ = −19.6o . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 414 CHAPTER 8 Frequency Response Methods Bode Diagram Magnitude (dB) 35 30 25 20 Phase (deg) 15 45 0 −45 −90 −1 10 0 1 10 10 2 10 Frequency (rad/sec) FIGURE P8.20 Bode plot for Gc (s)G(s) = The Bode plot is shown in Figure P8.21. The gain is 24 dB when φ = −180o 40 Gain dB 20 0 -20 -40 10-1 100 101 102 101 102 Frequency (rad/sec) 0 Phase deg P8.21 832(s+9.8) s2 +22s+471 . -100 -200 -300 10-1 100 Frequency (rad/sec) FIGURE P8.21 Bode plot for Gc (s)G(s) = −200s2 . s3 +14s2 +44s+40 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 415 Problems P8.22 The transfer function is G(s) = P8.23 10000(s + 1)(s + 80) . s(s + 300)(s + 9000) The transfer function is G(s) = 100(s + 20)(s + 8000) . (s + 1)(s + 80)(s + 500) The system is type 0 and the steady-state error to a unit step input is ess = 1 = 0.0025 1 + Kp since Kp = lim G(s) = 400 . s→0 P8.24 (a) From the Bode plot we see that 20 log10 Mpω = 12 or Mpω = 3.981. For a second-order system we know that Mpω = (2ζ q 1 − ζ 2 )−1 . Solving for ζ (with Mpω = 3.981) yields ζ = 0.12. Also, from the Bode plot, ωr = 0.9rad/sec . So, ωr ωn = p = 0.91 . 1 − 2ζ 2 Therefore, the second-order approximate transfer function is T (s) = ωn2 0.83 = 2 . s2 + 2ζωn s + ωn2 s + 0.22s + 0.83 (b) The predicted overshoot and settling time are P.O. = 68% and Ts = 37 sec. P8.25 The transfer function is G(s) = 100(s + 10) . s2 (s + 100) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 416 CHAPTER 8 P8.26 Frequency Response Methods The transfer function is T (s) = Vo (s) 1 + R1 /R2 = . V (s) 1 + RCs Substituting R = 10kΩ, C = 1µF , R1 = 9kΩ, and R2 = 1kΩ yields T (s) = 10 . 1 + 0.01s The frequency response is shown in Figure P8.26. Bode Diagrams 20 15 Phase (deg); Magnitude (dB) 10 5 0 0 -20 -40 -60 -80 -100 1 10 10 2 10 Frequency (rad/sec) FIGURE P8.26 Bode plot for T (s) = P8.27 1+R1 /R2 1+RCs The frequency response is shown in Figure P8.27. TABLE P8.27 K 0.75 1 10 |L(jω)|jω=0 , dB 3.52 12.04 26.02 ωb , rad/s 8.3 14.0 33.4 ωc , rad/s 3.5 8.7 22.9 System performance as K varies. 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 417 Problems Bode Diagram 10 Magnitude (dB) 0 K increases −10 −20 −30 −40 K decreases −50 −60 0 Phase (deg) Phase plot remains unchanged as K varies −45 −90 −135 −1 10 FIGURE P8.27 Bode plot for K = 1 10 0 1 10 Frequency (rad/sec) 10 2 10 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 418 CHAPTER 8 Frequency Response Methods Advanced Problems AP8.1 The spring-mass-damper system is described by mẍ + bẋ + kx = p . Taking the Laplace transform (with zero initial conditions) yields X(s) 1 = . 2 P (s) ms + bs + k From Figure AP8.1(b) in Dorf & Bishop, we determine that 20 log 1 X(j0) = 20 log = −26dB . P (j0) k Solving for k yields k = 19.96 N/m . Also, ωn2 = k/m implies m = k/ωn2 , where ωn = corner frequency = 3.2 rad/sec. So, m = 1.949 kg . Comparing Figure AP8.1(b) in Dorf & Bishop to the known standard Bode plot of a second-order system, we estimate ζ ≈ 0.32. Therefore, b = 2mζωn = 2(1.949)(0.32)(3.2) = 3.992 N − s/m . AP8.2 The closed-loop transfer function is T (s) = Y (s) Kb = . R(s) s + 1 + 0.2Kb WIth K = 5, we have T (s) = 5b . s+1+b The sensitivity is SbT = ∂T b s+1 = . ∂b T s+1+b With the nominal value of b = 4, we have SbT = s+1 . s+5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 419 Advanced Problems The sensitivity plot is shown in Figure AP8.2. 0 -2 20*log(mag) (dB) -4 -6 -8 -10 -12 -14 10-1 100 101 102 Frequency (rad/sec) FIGURE AP8.2 Bode plot for SbT (s) = AP8.3 s+1 s+5 . The equation of motion is mẍ + bẋ + Kx = bṙ + Kr . Taking Laplace transforms yields X(s) bs + K = . 2 R(s) ms + bs + K Then, given the various system parameters m = 1 kg, b = 4 Ns/m, K = 18 N/m, we obtain the transfer function: X(s) 4s + 18 = 2 . R(s) s + 4s + 18 √ p Also, ωn = corner frequency = K/m = 18 = 4.243 rad/s and ζ = damping ratio = b/m 4 = = 0.471 . 2ωn 2(4.243) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 420 CHAPTER 8 Frequency Response Methods The Bode plot is shown in Figure AP8.3. 10 Gain dB 0 -10 -20 -30 10-1 100 101 102 101 102 Frequency (rad/sec) Phase deg 0 -50 -100 -150 -200 10-1 100 Frequency (rad/sec) FIGURE AP8.3 Bode plot for G(s) = The Bode plot is shown in Figure AP8.4. Bode Diagram Magnitude (dB) −20 −40 −60 −80 −100 −120 0 −45 Phase (deg) AP8.4 4s+18 . s2 +4s+18 −90 −135 −180 −225 −270 −1 10 0 10 Frequency (rad/sec) FIGURE AP8.4 Bode plot for L(s) = 1 . (0.4s+1)(s2 +3.9s+15) 1 10 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 421 Advanced Problems The closed-loop transfer function with unity feedback is given by T (s) = 10(s + 1) Gc (s)G(s) = 2 . 1 + Gc (s)G(s) s + 9s + 10 (a) Solving for Gc (s)G(s) yields Gc (s)G(s) = 10(s + 1) . s(s − 1) (b) A summary of the plot data (see Figure AP8.5) is presented in Table AP8.5. (c) The open-loop system is unstable; the closed-loop system is stable. 40 30 20 20 log|GcG(j ω)|, dB AP8.5 10 0 −10 −20 −30 −40 100 120 140 160 180 200 Phase, degrees 220 240 260 280 FIGURE AP8.5 Log-magnitude-phase curve for Gc G(jω). ω 1 10 50 110 500 20 log |Gc G| 40 4.85 -13.33 -20.61 -33.94 phase (deg) 101.42 250.17 267.53 268.93 269.77 TABLE AP8.5 Summary of magnitude and phase for ω = 1, 10, 50, 110, 500. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 422 CHAPTER 8 The transfer function is given by T (s) = 1/m . + (b/m)s + (k/m) s2 Selecting k = 1 and b = 2 results in the Bode plot magnitude always √ less than 0 dB. Choosing b = 2/2 leads to a peak response with a sinusoidal input at ω = 0.66 rad/s. Figure AP8.6a shows the Bode plot and Figure AP8.6b shows the response to a sinusiodal input with frequency ω = 1 rad/s is less than 1 in the steady-state, as desired. Bode Diagram 10 System: sys Peak gain (dB): 6.3 At frequency (rad/sec): 0.661 0 Magnitude (dB) −10 −20 −30 −40 −50 −2 10 −1 0 10 1 10 10 Frequency (rad/sec) Impulse Response 1 0.5 Amplitude AP8.6 Frequency Response Methods 0 −0.5 −1 −1.5 0 100 200 300 400 Time (sec) 500 600 700 800 FIGURE AP8.6 (a) Bode plot for b/m = 1 and k/m = 1. (b) Response to a sinusiodal input. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 423 Advanced Problems The transfer function is G(s) = Vo (s) 1 + R2 C 2 s = . Vi (s) 1 + R1 C 1 s Substituting C1 = 0.1 µF ,C2 = 1 mF , R1 = 10 kΩ, and R2 = 10 Ω yields G(s) = 0.01s + 1 . 0.001s + 1 The frequency response is shown in Figure AP8.7. Bode Diagram Magnitude (dB) 20 15 10 5 0 60 Phase (deg) AP8.7 30 0 0 10 FIGURE AP8.7 Bode plot for G(s) = 1 10 0.01s+1 0.001s+1 2 3 10 10 Frequency (rad/sec) 4 10 5 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 424 CHAPTER 8 Frequency Response Methods Design Problems CDP8.1 With the PI controller in the loop, the closed-loop transfer function from the input to the output is 26.035K(s + 2) θ(s) = 2 , R(s) s + (33.1415 + 26.035K)s + 52.07K where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). The Bode plot is shown below for K = 40. From the step response we determine that P.O. = 0 and Ts = 0.19. With K = 40, the closed-loop poles are both real roots with values of s1 = −1072.6 and s2 = −1.9. 60 Gain dB 40 20 0 -1 10 10 0 10 1 10 2 Frequency (rad/sec) Phase deg 0 -30 -60 -90 10 -1 10 0 10 1 10 2 Frequency (rad/sec) DP8.1 The loop transfer function is L(s) = Gc (s)G(s) = K(s + 2) . s2 (s + 12) (a,b) Let K = 1. The Bode plot of the loop transfer function and the closed-loop transfer functions are shown in Figure DP8.1a and Figure DP8.1b, respectively. (c) Let K = 50. The Bode plot of the loop transfer function and the closed-loop transfer functions are shown in Figure DP8.1c and Figure DP8.1d, respectively. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 425 Design Problems 50 Gain dB 0 -50 -100 -1 10 10 0 10 1 10 2 Frequency (rad/sec) Phase deg -120 -140 -160 -180 -1 10 10 0 10 1 10 2 Frequency (rad/sec) FIGURE DP8.1 (a) Bode plot for the loop transfer function Gc (s)G(s) = (s+2) . s2 (s+12) 50 Gain dB 0 -50 -100 -2 10 10 -1 0 10 Frequency (rad/sec) 10 1 10 2 Phase deg 0 -90 -180 10 -2 10 -1 0 10 Frequency (rad/sec) FIGURE DP8.1 CONTINUED: (b) Bode plot for the closed-loop T (s) = 10 1 (s+2) . s3 +12s2 +s+2 10 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 8 Frequency Response Methods 100 Gain dB 50 0 -50 -1 10 10 0 10 1 10 2 Frequency (rad/sec) Phase deg -120 -140 -160 -180 -1 10 10 0 10 1 10 2 Frequency (rad/sec) FIGURE DP8.1 CONTINUED: (c) Bode plot for the loop transfer function Gc (s)G(s) = 50(s+2) . s2 (s+12) 20 Gain dB 0 -20 -40 -60 -1 10 10 0 10 1 10 2 Frequency (rad/sec) 0 Phase deg 426 -90 -180 10 -1 10 0 10 1 10 Frequency (rad/sec) FIGURE DP8.1 CONTINUED: (d) Bode plot for the closed-loop T (s) = 50(s+2) . s3 +12s2 +50s+100 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 427 Design Problems (d) The peak value of Mp ≤ 2 occurs for 14 ≤ K ≤ 350. The maximum bandwidth is achieved for the largest gain K. Thus, we select K = 350 and the corresponding bandwidth is ωB = 29 rad/sec. (e) The system is type 2—the steady-state error is zero for a ramp input. The open-loop transfer function is 20(s + 1) . s(s + 4)(s2 + 2s + 8) Gc (s)G(s) = (a) The phase angle is φ = −180o when ω = 3.54 rad/sec. The magnitude is 0 dB when ω = 0.87 rad/sec. (b) The closed-loop transfer function is T (s) = s4 + 6s3 20(s + 1) . + 16s2 + 52s + 20 The closed-loop Bode plot is shown in Figure DP8.2. Bode Diagram Gm = 6.71 dB (at 3.54 rad/sec) , Pm = 105 deg (at 0.869 rad/sec) 20 Magnitude (dB) 0 −20 −40 −60 −80 −100 −45 −90 Phase (deg) DP8.2 −135 −180 −225 −270 −1 10 0 1 10 10 2 10 Frequency (rad/sec) FIGURE DP8.2 Bode plot for closed-loop T (s) = 20(s+1) . s4 +6s3 +16s2 +52s+20 (c) When K = 22, we have Mpω = 4.84dB , ωr = 3.11 , and ωB = 3.78 rad/sec . When K = 25, we have Mpω = 7.18 dB , ωr = 3.18 rad/sec , and ωB = 3.94 rad/sec . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 428 CHAPTER 8 Frequency Response Methods (d) Select K = 22. DP8.3 The closed-loop transfer function is T (s) = s3 + 7s2 K(s + 5) . + 12s + 10 + 5K When K = 4.2, we have 10 log 10 Mpω = 3 dB. The system bandwidth is ωb = 3.7178 rad/sec. The steady-state tracking error to a unit step input is ess = lim sE(s) = lim 1 − T (s) . s→0 s→0 So, ess = 1 − 5K = 0.322 , 10 + 5K when K = 4.2. Since the system is unstable when K > 14.8, the steadystate error does not exist after K = 14.8. The Bode plot is shown in Figure DP8.3. 20 Gain dB 0 -20 -40 -60 -80 10-1 100 101 102 101 102 Frequency (rad/sec) Phase deg 0 -50 -100 -150 -200 10-1 100 Frequency (rad/sec) FIGURE DP8.3 Bode plot for closed-loop T (s) = DP8.4 K(s+5) , s3 +7s2 +12s+10+5K where K = 4.2. We have a second-order loop transfer function Gc (s)G(s) = K . (0.3s + 1)(0.6s + 1) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 429 Design Problems With Mpω = 1.5, we determine that q Mpω = (2ζ 1 − ζ 2 )−1 or ζ = 0.3568 . Now the characteristic equation is s2 + 5s + 5.56(1 + K) = 0 . So, solving 2ζωn = 5 yields ωn = 7. Therefore, K = 0.18ωn2 − 1 = 7.82 . The closed-loop transfer function is T (s) = K 5.56(K + 1) . 2 K + 1 s + 5s + 5.56(K + 1) So, the overall gain of the standard second-order system will be attenuated by the factor K/(K + 1). To compensate, we amplify the gain by a small factor. Thus we choose K = 10. The bandwidth is ωb = 11.25 rad/sec and the peak magnitude is Mpω =1.5. DP8.5 From the Bode plot of G(s) we find that there exists two pnoles, at approximately ω = 1 rad/sec and ω = 10 rad/sec. Then, by examining the Bode plot we estimate G(s) = 10 . (s + 1)(s + 10) We use a scale factor of 10 because at low frequency the Bode plot has magnitude 0 dB (or a DC gain of 1). With G(s) as above, we can utilize the controller Gc (s) = 500 s + 20 yielding a crossover ωc = 12.9 rad/sec and a magnitude of at least 25 dB for ω < 0.1 rad/sec. Figure DP8.5 shows the compensator Bode plot of Gc (s)G(s). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 430 CHAPTER 8 Frequency Response Methods Bode Diagram Phase (deg) Magnitude (dB) 50 -50 -100 -150 0 -45 -90 -135 -180 -225 10 -2 10 -1 FIGURE DP8.5 Bode Diagram for G(s)Gc (s) = DP8.6 ωc=12.9 25 dB 0 0 1 10 10 Frequency (rad/sec) 10 2 10 3 5000 . (s+1)(s+10)(s+20) Let K = −1 to meet the steady-state tracking error requirement and p = 2ζ, where ζ = 0.69 to obtain a 5% overshoot. The system is given by ẋ = Ax + Bu where A= 0 1 −1 −1.38 , B= The characteristic polynomial is −1 0 , and C= 0 1 . s2 + 1.38s + 1 = 0 . The associated damping ratio is ζ = 0.69 and the natural frequency is ωn = 1 rad/s. Using the approximation ωb = (−1.19ζ + 1.85)ωn we obtain ωb ≈ 1.028 rad/s. The Bode plot is shown in Figure DP8.6. The bandwidth is ωb = 1.023 rad/s. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 431 Design Problems Bode Diagram 20 Magnitude (dB) 0 −20 −40 −60 −80 0 Phase (deg) −45 −90 −135 −180 −2 10 −1 0 10 10 Frequency (rad/sec) 1 10 2 10 FIGURE DP8.6 Bode diagram for K = −1 and p = 1.38. DP8.7 A viable controler is Gc (s) = KP + 3.33 KI + KD s = 5.5 + + 3.5s. s s The loop transfer function is Gc (s)G(s) = 10.5s2 + 16.5s + 10 s2 (s2 + 4s + 5) and computing Ka yields Ka = lim s2 Gc (s)G(s) = s→0 10 = 2, 5 as desired. The phase margin is P.O. = 44.35◦ and the bandwidth is ωb = 4.5 rad/sec. The step response is shown in Figure DP8.7. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 8 Frequency Response Methods Step Response 1.4 System: sys_cl Peak amplitude: 1.32 Overshoot (%): 32.1 At time (sec): 1.11 1.2 System: sys_cl Settling Time (sec): 3.93 1 Amplitude 432 0.8 0.6 0.4 0.2 0 0 1 2 3 4 Time (sec) FIGURE DP8.7 Step response for KP = 5.5, KI = 3.33, and KD = 3.5. 5 6 7 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 433 Computer Problems Computer Problems CP8.1 The m-file script and Bode plot are shown in Figure CP8.1. The script automatically computes Mpω and ωr . num=[25]; den=[1 1 25]; sys = tf(num,den); w=logspace(0,1,400); [mag,phase]=bode(sys,w); [y,l]=max(mag); mp=20*log10(y), wr=w(l) bode(sys,w); mp = 14.0228 wr = 4.9458 Bode Diagrams From: U(1) 15 10 5 Phase (deg); Magnitude (dB) 0 -5 - 10 0 To: Y(1) - 50 - 100 - 150 - 200 0 10 10 1 Frequency (rad/sec) FIGURE CP8.1 Generating a Bode plot with the bode function. CP8.2 The m-file script to generate the Bode plots is shown in Figure CP8.2a. The Bode plots are presented in Figures CP8.2b-CP8.2e. The transfer functions are (a) : G(s) = 1000 ; (s + 10)(s + 100) (b) : G(s) = s + 100 ; (s + 2)(s + 25) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 8 Frequency Response Methods (c) : G(s) = s2 100 ; + 2s + 50 (d) : G(s) = s−6 . (s + 3)(s2 + 12s + 50) % Part (a) num=[1000]; den=conv([1 10],[1 100]); sys1=tf(num,den); sys = tf(num,den); figure(1), bode(sys1), grid % Part (b) num=[1 100]; den=conv([1 2],[1 25]); sys2=tf(num,den); sys = tf(num,den); figure(2), bode(sys2), grid % Part (c) num=[100]; den=[1 2 50]; sys3=tf(num,den); sys = tf(num,den); figure(3), bode(sys3), grid % Part (d) num=[1 -6]; den=conv([1 3],[1 12 50]); sys4=tf(num,den); sys = tf(sys); figure(4), bode(sys4), grid FIGURE CP8.2 (a) Script to generate the four Bode plots. Bode Diagram Magnitude (dB) 0 −20 −40 −60 −80 −100 0 Phase (deg) 434 −45 −90 −135 −180 −1 10 0 10 1 2 10 10 Frequency (rad/sec) FIGURE CP8.2 CONTINUED: (b) Bode plot for G(s) = 1000 . (s+10)(s+100) 3 10 4 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 435 Computer Problems Bode Diagram Magnitude (dB) 20 0 −20 −40 −60 Phase (deg) −80 0 −45 −90 −135 −1 10 0 10 1 2 10 10 Frequency (rad/sec) FIGURE CP8.2 CONTINUED: (c) Bode plot for G(s) = 3 4 10 10 s+100 . (s+2)(s+25) Bode Diagram 20 Magnitude (dB) 10 0 −10 −20 −30 Phase (deg) −40 0 −45 −90 −135 −180 0 10 1 10 Frequency (rad/sec) FIGURE CP8.2 CONTINUED: (d) Bode plot for G(s) = 100 s2 +2s+50 . 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 436 CHAPTER 8 Frequency Response Methods Bode Diagram −20 Magnitude (dB) −30 −40 −50 −60 −70 Phase (deg) −80 180 90 0 −90 −180 −1 10 0 1 10 2 10 10 Frequency (rad/sec) FIGURE CP8.2 CONTINUED: (e) Bode plot for G(s) = The Bode plots are shown in Figure CP8.3(a-d) with the transfer functions listed in the caption. The crossover frequency for (a) is 17 rad/sec. Bode Diagram 20 Magnitude (dB) 0 −20 −40 −60 −80 −100 0 Phase (deg) CP8.3 s−6 . (s+3)(s2 +12s+50) −45 −90 −135 −180 −1 10 FIGURE CP8.3 (a) Bode plot for G(s) = 0 10 1 2 10 10 Frequency (rad/sec) 2000 . (s+10)(s+100) 3 10 4 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 437 Computer Problems The crossover frequency for (b) is 0.99 rad/sec. Bode Diagram 20 Magnitude (dB) 0 −20 −40 −60 −80 0 Phase (deg) −45 −90 −135 −180 −225 −270 −2 10 −1 10 0 10 Frequency (rad/sec) FIGURE CP8.3 CONTINUED: (b) Bode plot for G(s) = 1 2 10 10 100 . (s+1)(s2 +10s+2) The crossover frequency for (c) is 70.7 rad/sec. Bode Diagram 40 Magnitude (dB) 30 20 10 0 −10 −20 Phase (deg) −30 0 −45 −90 −135 −2 10 −1 10 0 1 10 10 Frequency (rad/sec) FIGURE CP8.3 CONTINUED: (c) Bode plot for G(s) = 50(s+100) . (s+1)(s+50) The crossover frequency for (d) is 3.1 rad/sec. 2 10 3 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 438 CHAPTER 8 Frequency Response Methods Bode Diagram 20 Magnitude (dB) 10 0 −10 −20 −30 Phase (deg) −40 0 −45 −90 −1 10 0 1 10 FIGURE CP8.3 CONTINUED: (d) Bode plot for G(s) = 3 10 4 10 100(s2 +14s+50) . (s+1)(s+2)(s+500) The m-file script and Bode plot are shown in Figure CP8.4a and b. The bandwidth is ωb = 10 rad/sec. Bandwidth=10.0394 rad/sec 10 Magnitude (dB) 0 −10 −20 −30 −40 −50 0 Phase (deg) CP8.4 2 10 10 Frequency (rad/sec) −45 −90 −135 −180 −1 10 0 10 Frequency (rad/sec) FIGURE CP8.4 (a) Bode plot for T (s) = 54 s2 +6s+54 . 1 10 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 439 Computer Problems numg=[54]; deng=[1 6 0]; sys_o = tf(numg,deng); sys_cl = feedback(sys_o,[1]) wb=bandwidth(sys_cl) bode(sys_cl), grid titlename=strcat('Bandwidth= ', num2str(wb), ' rad/sec') title(titlename) FIGURE CP8.4 CONTINUED: (b) M-file script to obtain the closed-loop Bode plot. The Bode plot of the closed-loop system is shown in Figure CP8.5. The closed-loop transfer function is T (s) = s2 100 . + 6s + 100 (a) From the Bode plot we determine that Mpω ≈ 5 dB and ωr ≈ 9 rad/sec . Bode Diagrams From: U(1) 20 0 - 20 - 40 Phase (deg); Magnitude (dB) - 60 - 80 0 - 50 To: Y(1) CP8.5 - 100 - 150 - 200 -1 10 10 0 10 1 Frequency (rad/sec) FIGURE CP8.5 Closed-loop system Bode plot. 10 2 10 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 440 CHAPTER 8 Frequency Response Methods (b) From Equations (8.36) and (8.37) in Dorf & Bishop, we find that ζ ≈ 0.28 and ωr /ωn ≈ 0.92 which implies that ωn = ωr /0.92 = 9.8 rad/sec . (c) From T (s) we find that ωn = 10 rad/sec and ζ = 0.3 . The actual values and the estimated values compare very well. The open-loop and closed-loop Bode plots are shown in Figure CP8.6a and b. The open-loop and closed-loop transfers functions are Gc (s)G(s) = 25 s3 + 3s2 + 27s + 25 and T (s) = Gc (s)G(s) 25 = 3 . 1 + Gc (s)G(s) s + 3s2 + 27s + 50 Loop transfer function; bode(syso) Magnitude (dB) 0 −20 −40 −60 −80 −100 0 −45 Phase (deg) CP8.6 −90 −135 −180 −225 −270 −2 10 −1 10 0 10 Frequency (rad/sec) FIGURE CP8.6 (a) Open-loop system Bode plot for Gc (s)G(s) = 1 10 25 . s3 +3s2 +27s+25 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 441 Computer Problems Closed−loop system; bode(syscl) Magnitude (dB) 0 −20 −40 −60 −80 −100 0 Phase (deg) −45 −90 −135 −180 −225 −270 −1 10 0 10 1 2 10 10 Frequency (rad/sec) FIGURE CP8.6 CONTINUED: (b) Closed-loop system Bode plot T (s) = CP8.7 25 . s3 +3s2 +27s+50 The m-file script and plot of ωb versus p are shown in Figure CP8.7a and b. p=[0:0.001:1]; w=logspace(-1,1,1000); n=length(p); for i=1:n num=[1]; den=[1 2*p(i) 0]; sys = tf(num,den); sys_cl = feedback(sys,[1]); [mag,phase,w]=bode(sys_cl,w); a=find(mag<0.707); wb(i)=w(a(1)); end plot(p,wb) xlabel('p'), ylabel('Bandwidth (rad/sec)') FIGURE CP8.7 (a) M-file script to generate plot of ωb versus p. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 442 CHAPTER 8 Frequency Response Methods 1.6 1.5 1.4 Bandwidth (rad/sec) 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0 0.1 0.2 0.3 0.4 0.5 p 0.6 0.7 0.8 0.9 1 FIGURE CP8.7 CONTINUED: (b) Plot of ωb versus p. CP8.8 The transfer function from Td (s) to θ(s) is θ(s)/Td (s) = s3 + 10s2 −0.01(s + 10) . + (0.01K − 10.791)s − 107.91 + 0.05K Using the final value theorem and Td (s) = 1/s, we determine that lim sθ(s) = s→0 −0.1 . −107.91 + 0.05K The design specifications require that |ess | < 0.1o . So, solving for K yields K > 3300 . We can select K = 3300 as the initial value of K for the design. The m-file script is shown in Figure CP8.8a. For the design shown, the final selection for the gain is K = 6000. The disturbance response is shown in Figure CP8.8b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 443 Computer Problems Mb=100; Ms=10; L=1; g=9.81; a=5; b=10; % K=6000; % Final design value of K % numg=[-1/Mb/L]; deng=[1 0 -(Mb+Ms)*g/Mb/L]; sysg = tf(numg,deng); numc=-K*[1 a]; denc=[1 b]; sysc = tf(numc,denc); % % Part (a) % sys = feedback(sysg,sysc); w=logspace(0,1,400); bode(sys,w) [mag,phase]=bode(sys,w); [M,l]=max(mag); MpDb=20*log10(M)-20*log10(mag(1)) % Mpw in decibels wr=w(l) % Mpw and peak frequency % % Part (b) % % From Eqs. (8.35) and (8.37) Mpw=10^(MpDb/20);zeta=sqrt((1-sqrt(1-(1/Mpw^2)))/2); wn=wr/sqrt(1-2*zeta^2); ts=4/zeta/wn po=100*exp(-zeta*pi/sqrt(1-zeta^2)) % % Part (c) % t=[0:0.1:10]; [y,x]=step(sys,t); plot(t,y*180/pi) xlabel('time [sec]') ylabel('theta [deg]') grid MpDb = 4.0003 wr = 4.7226 meets specs ts = 2.23 po = 32.75 0 -0.005 -0.01 theta [deg] -0.015 -0.02 -0.025 -0.03 -0.035 -0.04 0 1 2 3 4 5 6 7 time [sec] FIGURE CP8.8 (a) Design script. (b) Disturbance response - meets all specs! 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 444 CHAPTER 8 A viable filter is G(s) = 0.7 (s + 1000)(s + 1) . (s + 100)(s + 10) The Bode plot is shown in Figure CP8.9 Bode Diagram Magnitude (dB) 20 15 10 5 0 -5 90 Phase (deg) CP8.9 Frequency Response Methods 45 0 -45 -90 -2 10 10 0 10 2 Frequency (rad/sec) FIGURE CP8.9 (s+1000)(s+1) Bode plot for G(s) = 0.7 (s+100)(s+10) . 10 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 9 Stability in the Frequency Domain Exercises The Bode plot for the transfer function Gc (s)G(s) is shown in Figure E9.1, where Gc (s)G(s) = 2(1 + s/10) . s(1 + 5s)(1 + s/9 + s2 /81) The gain and phase margins are P.M. = 17.5o . G.M. = 26.2 dB and Bode Diagram Gm = 26.2 dB (at 2.99 rad/sec) , Pm = 17.5 deg (at 0.618 rad/sec) Magnitude (dB) 50 0 −50 −100 −150 −90 −135 Phase (deg) E9.1 −180 −225 −270 −315 −2 10 FIGURE E9.1 Bode Diagram for Gc (s)G(s) = −1 10 0 10 Frequency (rad/sec) 1 10 2 10 2(1+s/10) . s(1+5s)(1+s/9+s2 /81) 445 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 446 CHAPTER 9 E9.2 Stability in the Frequency Domain The loop transfer function is Gc (s)G(s) = 10.5(1 + s/5) . s(1 + s/2)(1 + s/10) The Bode plot is shown in Figure E9.2. The phase margin is P.M. = 40.4o at ωc = 4.96 rad/sec. Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 40.4 deg (at 4.96 rad/sec) Magnitude (dB) 50 0 −50 Phase (deg) −100 −90 −135 −180 −1 10 0 10 FIGURE E9.2 Bode Diagram for Gc (s)G(s) = 1 10 Frequency (rad/sec) 2 10 3 10 10.5(1+s/5) . s(1+s/2)(1+s/10) E9.3 The phase margin P.M. ≈ 75o at 200 kHz. We estimate the −180o phase angle at 2 MHz, so the gain margin is G.M. ≈ 25 dB. E9.4 The loop transfer function is Gc (s)G(s) = 100 . s(s + 10) The Nichols diagram is shown in Figure E9.4. When the gain is raised by 4.6 dB, Mpω = 3 and the resonant frequency is ωR = 11 rad/sec. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 447 Exercises 40 0 0.25 30 0.5 1 20 3 6 10 Gain dB -1 -3 -6 0 -10 K=171 ------ -12 ------ K=100 -20 -20 -30 -40 -350 -300 -250 -200 -150 -100 -40 0 -50 Phase (deg) FIGURE E9.4 Nichols Diagram for Gc (s)G(s) = K) , s(s+10) where K = 100 and K = 171. E9.5 (a) The G.M. ≈ 5 dB and the P.M. ≈ 10o . (b) Lower the gain by 10 dB to obtain P.M. ≈ 60o . E9.6 The Bode plot of the closed-loop transfer function is shown in Figure E9.6. The value of Mpω = 3 dB. The phase margin is P.M. = 40o when K = 50. 5 0 -5 -10 Gain dB -15 -20 -25 -30 -35 -40 -45 10-1 100 101 Frequency (rad/sec) FIGURE E9.6 Closed-loop Bode Diagram for T (s) = 50(s+100) s3 +50s2 +450s+5000 . 102 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 448 CHAPTER 9 E9.7 Stability in the Frequency Domain The Nyquist plot is shown in Figure E9.7 for K = 5; the plot is a circle with diameter= K/5. For K > 5, we have P = 1 and N = −1 (ccw as Nyquist Diagram 0.5 0.4 0.3 Imaginary Axis 0.2 0.1 0 −0.1 −0.2 −0.3 −0.4 −0.5 −1 −0.8 −0.6 FIGURE E9.7 Nyquist Diagram for Gc (s)G(s) = −0.4 −0.2 Real Axis K s−5 , 0 0.2 0.4 where K = 5. shown). So Z = N + P = −1 + 1 = 0 and the system is stable for K > 5. (a) When K = 4, the G.M. = 3.5 dB. This is illustrated in Figure E9.8. Bode Diagram Gm = 3.52 dB (at 1.41 rad/sec) , Pm = 11.4 deg (at 1.14 rad/sec) Magnitude (dB) 50 0 −50 −100 −150 −90 −135 Phase (deg) E9.8 −180 −225 −270 −1 10 0 1 10 10 Frequency (rad/sec) FIGURE E9.8 Bode Diagram for Gc (s)G(s) = K , s(s+1)(s+2) where K = 4. 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 449 Exercises (b) The new gain should be K = 1 for a gain margin G.M. = 16 dB. E9.9 For K = 5, the phase margin P.M. = 5o as shown in Figure E9.9. Bode Diagram Gm = 1.58 dB (at 1.41 rad/sec) , Pm = 5.02 deg (at 1.29 rad/sec) 100 Magnitude (dB) 50 0 -50 -100 -150 -90 Phase (deg) -135 -180 -225 -270 -2 10 10 FIGURE E9.9 Bode Diagram for Gc (s)G(s) = 0 10 Frequency (rad/sec) K , s(s+1)(s+2) 10 1 10 2 where K = 5. The Bode plot is shown in Figure E9.10a. The closed-loop frequency 100 Gain dB 50 0 GM=12.35 dB -50 -100 10-2 10-1 100 Frequency (rad/sec) 101 102 0 Phase deg E9.10 -1 -100 PM=23.14 deg -200 -300 10-2 10-1 FIGURE E9.10 (a) Bode Diagram for Gc (s)G(s) = 100 Frequency (rad/sec) 101 326s+1304 s4 +14.76s3 +151.3s2 +23.84s . 102 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 450 CHAPTER 9 Stability in the Frequency Domain 10 0 -10 Gain dB -20 -30 -40 -50 -60 -70 10-1 100 101 102 Frequency (rad/sec) FIGURE E9.10 CONTINUED: (b) Closed-loop frequency response: ωB = 6 rad/sec. response is shown in Figure E9.10b. The bandwidth is ωB = 6 rad/sec. The Bode plot is shown in Figure E9.11. The system is stable. Bode Diagram Gm = 3.91 dB (at 3.74 rad/sec) , Pm = 14.4 deg (at 2.76 rad/sec) Magnitude (dB) 100 50 0 −50 −100 −90 −135 Phase (deg) E9.11 −180 −225 −270 −2 10 −1 10 FIGURE E9.11 Bode Diagram for Gc (s)G(s) = 0 10 Frequency (rad/sec) 10(1+0.4s) . s(1+2s)(1+0.24s+0.04s2 ) 1 10 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 451 Exercises E9.12 We select the gain K = 10 to meet the 10% steady-state tracking error specification for a ramp input. The Bode plot and Nichols chart are shown in Figures E9.12a and E9.12b, respectively. 50 Gain dB 0 GM=14.82 dB -50 -100 -150 10-1 100 101 Frequency (rad/sec) 102 103 102 103 Phase deg 0 -100 -200 PM=31.79 deg -300 10-1 100 101 Frequency (rad/sec) 40 0 0.25 30 0.5 1 20 10 Gain dB -1 3 6 8 -3 -6 0 -12 -10 -20 -20 -30 -40 -350 -300 -250 -200 -150 -100 -50 -40 0 Phase (deg) FIGURE E9.12 (a) Bode Diagram for Gc (s)G(s) = 10 . s(0.02s+1)(0.2s+1) E9.13 10 . s(0.02s+1)(0.2s+1) (b) Nichols chart for Gc (s)G(s) = (a) The Nichols diagram is shown in Figure E9.13a and Mpω = 7.97 dB. (b) The closed-loop Bode plot is shown in Figure E9.13b. The bandwidth ωB = 18.65 rad/sec and the resonant frequency is ωr = 11.69 rad/sec. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 452 CHAPTER 9 Stability in the Frequency Domain 40 0 0.25 30 0.5 1 20 Gain dB 10 0 -1 3 6 -3 8 -6 -12 -10 -20 -20 -30 -40 -350 -300 -250 -200 -150 -100 -50 -40 0 Phase (deg) 10 Gain dB 0 -10 -20 -30 -40 10-1 100 101 102 101 102 Frequency (rad/sec) Phase deg 0 -50 -100 -150 -200 10-1 100 Frequency (rad/sec) FIGURE E9.13 (a) Nichols Diagram for Gc (s)G(s) = 150 . s2 +5s+150 150 . s(s+5) (b) Closed-loop Bode Diagram for T (s) = (c) From Mpω = 8 dB, we estimate ζ = 0.2, so the expected P.O. = 52%. E9.14 (a) The peak resonance Mpω = 6 dB. (b) The resonant frequency is ωr = ω2 = 3 rad/sec. (c) The bandwidth is ωB = ω4 = 10 rad/sec. (d) The phase margin is P.M. = 30o . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 453 Exercises E9.15 The loop transfer function is Gc (s)G(s) = 100 , s(s + 20) and the closed-loop transfer function is T (s) = 100 . s2 + 20s + 100 The magnitude plot for the closed-loop system is shown in Figure E9.15. With bandwidth defined as frequency at which the magnitude is reduced Bode Diagram 0 −1 Magnitude (dB) −2 −3 −4 −5 −6 −7 −1 10 0 10 Frequency (rad/sec) FIGURE E9.15 Magnitude plot for the closed-loop T (s) = 1 10 100 . s2 +20s+100 -3 dB from the dc value, we determine that ωB = 6.4 rad/sec. E9.16 The transfer function of the approximation is G(jω) = 1 − jω/10 , 1 + jω/10 and the magnitude is |G(jω)| = 1 − jω/10 =1, 1 + jω/10 which is equivalent to the actual time delay magnitude. The phase ap- © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 454 CHAPTER 9 Stability in the Frequency Domain proximation is φ = − tan−1 ω/10 + tan−1 (−ω/10) = −2 tan−1 ω/10 and the actual phase is φ = −0.2ω . The phase plots are shown in Figure E9.16. The approximation is accurate for ω < 3 rad/sec. Actual _______ & Approximation −−−−−−− 0 −20 Phase deg −40 −60 −80 −100 −120 −2 10 −1 0 10 1 10 10 Frequency (rad/sec) FIGURE E9.16 Phase plots for time delay actual vs approximation. E9.17 (a,b) The phase angle for P.M. = 30 is φ = −90o + tan−1 ω 2ω − tan−1 = −150o . 2 15 − ω 2 Solving for ω yields ω = 4.7. Then, at ω = 4.7, we have K = 10.82 when 1 |Gc G(jω)| = K(ω 2 + 4) 2 1 ω((2ω 2 )2 + (15 − ω 2 )2 ) 2 The Bode plot is shown in Figure E9.17. =1. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 455 Exercises Bode Diagrams Gm=3.5545 dB (at 4.3301 rad/sec), Pm=40 deg. (at 3.5147 rad/sec) 50 Phase (deg); Magnitude (dB) 0 - 50 - 100 - 50 - 100 - 150 - 200 - 250 -1 10 10 0 10 1 10 2 Frequency (rad/sec) FIGURE E9.17 Bode Diagram for Gc (s)G(s) = K(s+2) , s3 +2s2 +15s where K = 10.82. (c) The steady-state error for a ramp is ess = A A = 10K = 0.60A , Kv 15 where R(s) = A/s2 . E9.18 (a) The gain crossover is at ωc = 486 Hz, and the phase margin P.O. = 36.2o . So, ζ ≈ 0.36. Then, the expected percent overshoot to a step input is √ 2 P.O. = 100e−ζπ/ 1−ζ = 30% , where ζ = 0.36 . (b) The estimated bandwidth is ωB ≈ 2π(600). (c) Approximate ωn ≈ ωr = 2π(480) . Then, Ts = 4 4 = ≈ 4 ms . ζωn (0.36)2π(480) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 456 CHAPTER 9 E9.19 Stability in the Frequency Domain The Bode plot is shown in Figure E9.19 for K = 16.75. The phase and gain margins are P M = 50.0o and GM = 2.72 dB. Gm=2.7233 dB (at 20.618 ad/sec), r Pm=50 deg . (at 13.434ad/sec) r 10 0 - 10 Phase (deg); Mag nitude (dB) - 20 - 30 - 40 0 - 100 - 200 - 300 - 400 - 500 0 10 1 2 10 3 10 10 Frequency (rad/sec) FIGURE E9.19 −0.1s Bode Diagram for Gc (s)G(s) = K es+10 , where K = 16.75. The system response for both drivers is shown in Figure E9.20. T=1 sec (solid line) & T=1.5 sec (dashed line) 1 0 -1 Automobile velocity change E9.20 -2 -3 -4 -5 -6 -7 -8 0 1 2 3 4 5 6 7 Time (sec) FIGURE E9.20 Change in automobile velocity due to braking for two drivers. 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 457 Exercises E9.21 The Bode plot is shown in Figure E9.21. 50 Gain dB 0 -50 GM=12.04 dB -100 -150 10-1 100 101 Frequency (rad/sec) 102 103 102 103 Phase deg 0 -100 -200 PM=16.85 deg -300 10-1 100 101 Frequency (rad/sec) FIGURE E9.21 Bode Diagram for Gc (s)G(s) = 1300 . s(s+2)(s+50) E9.22 When K = 10, the P.M. = 36.9o ; the system is stable. Decreasing the gain to K = 4 results in a P.M. = 60o . E9.23 The Nichols chart is shown in Figure E9.23. 40 0 0.25 30 0.5 1 20 10 Gain dB -1 3 6 8 -3 -6 0 -12 -10 -20 -20 -30 -40 -350 -300 -250 -200 -150 Phase (deg) FIGURE E9.23 Nichols chart for Gc (s)G(s) = 438 . s(s+2)(s+50) -100 -50 -40 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 458 CHAPTER 9 Stability in the Frequency Domain The actual values are Mpω = 1.6598 (4.4 dB) ωr = 2.4228 rad/sec ωB = 4.5834 rad/sec . E9.24 Using the Nyquist criterion, we have P = 1 and N = 0 which implies Z = N +P = 1 . Therefore, the system has one root in the right half-plane. E9.25 The Bode plot is shown in Figure E9.25. PM=27.73 deg at wc=8.29 rad/sec Gain dB 50 0 -50 -100 10-1 100 101 102 101 102 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-1 100 Frequency (rad/sec) FIGURE E9.25 Bode plot for Gc (s)G(s) = E9.26 11.7 . s(0.05s+1)(0.1s+1) The Nichols chart for Gc (s)G(s) = 11.7 s(0.05s + 1)(0.1s + 1) is shown in Figure E9.26, where we find that Mpω = 6.76 dB ωr = 8.96 rad/sec ωB = 13.73 rad/sec . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 459 Exercises 40 0 0.25 30 0.5 1 20 3 6 8 10 Gain dB -1 -3 -6 0 -12 -10 -20 -20 -30 -40 -350 -300 -250 -200 -150 -100 -50 -40 0 Phase (deg) FIGURE E9.26 Nichols chart for Gc (s)G(s) = The Bode plot for G(s) with K = 122.62 is shown in Figure E9.27. K=122.63 Gm=10.938 dB (at 6 rad/sec), Pm=40 deg. (at 2.7978 rad/sec) 50 0 Phase (deg); Magnitude (dB) E9.27 11.7 . s(0.05s+1)(0.1s+1) - 50 - 100 - 50 - 100 - 150 - 200 - 250 - 300 -1 10 10 0 10 Frequency (rad/sec) FIGURE E9.27 Bode plot for Gc (s)G(s) = K , s(s+6)2 with K = 122.62. 1 10 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 460 CHAPTER 9 Stability in the Frequency Domain The phase margin is P.M. = 40.0o and the gain margin is G.M. = 10.94 dB . E9.28 The phase margin is P.M. = 28o . The estimated damping is ζ= P.M. = 0.28 . 100 The estimated percent overshoot is √ 2 P.O. = 100e−πζ/ 1−ζ = 40% . The actual overshoot is P.O. = 44.43%. E9.29 The F (s)-plane contour is shown in Figure E9.29, where F (s) = 1 + Gc (s)G(s) = s+3 . s+2 F(s)-plane 0.6 * 0.4 0.2 * Im * 0 * * * -0.2 * -0.4 * -0.6 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 Re FIGURE E9.29 F (s)-plane contour, where F (s) = 1 + Gc (s)G(s) = E9.30 The Bode plot is shown in Figure E9.30. s+3 s+2 . 1.8 1.9 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 461 Exercises Bode Diagram Magnitude (dB) 50 0 -50 Phase (deg) -100 0 -45 -90 -135 -180 10 -2 10 0 10 2 10 4 Frequency (rad/sec) FIGURE E9.30 Bode plot for G(s) = C [sI − A]−1 B + D = The Bode plot is shown in Figure E9.31. The phase margin is P.M. = 50.6 deg. Bode Diagram Gm = Inf , Pm = 50.6 deg (at 0.341 rad/sec) Magnitude (dB) 80 60 40 20 0 -20 -40 -90 Phase (deg) E9.31 1000 . s2 +100s+10 -120 -150 -3 10 10 -2 10 -1 Frequency (rad/sec) FIGURE E9.31 Bode plot for L(s) = G(s)H(s) = 2s+1 10s2 +s . 10 0 10 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 462 CHAPTER 9 E9.32 Stability in the Frequency Domain The Bode plot is shown in Figure E9.32. The phase margin is P.M. = 29◦ . Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 29 deg (at 3.1 rad/sec) Magnitude (dB) 20 0 −20 −40 −60 Phase (deg) −80 0 −45 −90 −135 −180 −1 10 0 1 10 2 10 10 Frequency (rad/sec) FIGURE E9.32 Bode plot for G(s) = C [sI − A]−1 B + D = The Bode plot is shown in Figure E9.33. The phase margin is P.M. = 17.7◦ and the gain margin is G.M. = 5.45 dB. Bode Diagram Gm = 5.45 dB (at 5.68 rad/sec) , Pm = 17.7 deg (at 4.24 rad/sec) Magnitude (dB) 50 0 −50 −100 −150 0 −45 Phase (deg) E9.33 6.4 s2 +s+4 . −90 −135 −180 −225 −270 −1 10 FIGURE E9.33 Bode plot for L(s) = 0 10 1 10 Frequency (rad/sec) 200 . (s2 +2.83s+4)(s+10) 2 10 3 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 463 Problems Problems P9.1 (a) The loop transfer function is Gc (s)G(s) = 1 . (1 + 0.5s)(1 + 2s) P = 0, N = 0; therefore Z = N +P = 0. The system is stable. (Note: See P8.1 for the polar plots.) (b) The loop transfer function is 1 + 0.5s . s2 P = 0, N = 0, therefore Z = N + P = 0. The system is stable. (c) The loop transfer function is s2 s+4 . + 5s + 25 P = 0, N = 0, Z = N + P = 0. Therefore, the system is stable. (d) The loop transfer function is 30(s + 8) . s(s + 2)(s + 4) P = 0, N = 2 therefore Z = P + N = 2. Therefore, the system has two roots in the right half-plane, and is unstable. P9.2 (a) The loop transfer function is Gc (s)G(s) = s(s2 K , + s + 6) and Gc (jω)G(jω) = K K[−ω 2 − jω(6 − ω 2 )] == . jω(−ω 2 + jω + 6) [(6 − ω 2 )2 ω 2 + ω 4 ] To determine the real axis crossing, we let Im{Gc (jω)G(jω)} = 0 = −Kω(6 − ω 2 ) or ω= √ 6. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 464 CHAPTER 9 Stability in the Frequency Domain Then, Re{Gc (jω)G(jω)}ω=√ 6 = −Kω 2 ω4 √ ω= 6 = −K . 6 So, −K/6 > −1 for stability. Thus K < 6 for a stable system. (b) The loop transfer function is Gc (s)G(s) = K(s + 1) . s2 (s + 6) The polar plot never encircles the -1 point, so the system is stable for all gains K (See Figure 10 in Table 9.6 in Dorf & Bishop). P9.3 (a,b) The suitable contours are shown in Figure P9.3. jw jw Gs Gs q =cos z q r r approaches infinity r s -s 1 r approaches infinity s (b) (a) FIGURE P9.3 Suitable contours Γs for (a) and (b). (c) Rewrite the characteristic equation as 1+ 96 =0. s(s2 + 11s + 56) In this case, −σ1 = −1. Therefore, we have one pole inside the contour at s = 0, so P = 1. The polar plot yields N = −1, so Z = N + P = 0. Therefore, all three roots have real parts less than -1. In fact, the roots are s1 = −3, and s2,3 = −4 ± j4. P9.4 (a) P = 0, N = 2, therefore Z = 2. The system has two roots in the right hand s-plane. (b) In this case, N = +1 − 1 = 0, so Z = 0. Therefore the system is stable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 465 Problems P9.5 (a) The loop transfer function is L(s) = Gc (s)G(s)H(s) = K . (s + 1)(3s + 1)(0.4s + 1) The steady-state error is ess = |R| . 1+K We require ess = 0.1|R|, so K > 9. (b) Use K = 9. The Nyquist plot is shown in Figure P9.5. We determine that P = 0 and N = 0. Therefore, Z = 0 and the system is stable. 8 6 4 Imag Axis 2 0 -2 -4 -6 -8 -2 0 2 4 6 8 10 Real Axis FIGURE P9.5 Nyquist Diagram for L(s) = Gc (s)G(s)H(s) = 9 . (s+1)(3s+1)(0.4s+1) (c) The phase and gain margins are P.M. = 18o and G.M. = 5 dB. P9.6 The rotational velocity transfer function is ω(s) = G(s) = R(s) 1+ K s 3.7(2π) s 68(2π)+1 . At low frequency, we have the magnitude near 35 dB, so 20 log K = 35 dB © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 466 CHAPTER 9 Stability in the Frequency Domain and K = 56. Since the frequency response plot is for rotational velocity ω(s), and we are interested in position control, we add an integrator. The characteristic equation is 1 56(23)(427) 1 + G(s) = 1 + =0. s s(s + 23)(s + 427) The roots are s1 = −430 and s2,3 = −10 ± j35 . Thus, ωn = 36 and ζ = 0.28. The time constant of the closed-loop system is τ= The loop transfer function is L(s) = Gc (s)G(s)H(s) = 10K1 s(s + 7) . (s + 3)(s2 + 0.36) (a) The Bode plot is shown in Figure P9.7 for K1 = 2. Gain dB 100 50 0 -50 10-1 100 101 102 101 102 Frequency (rad/sec) 100 50 Phase deg P9.7 1 = 99.6 msec . ζωn 0 -50 -100 -150 10-1 100 Frequency (rad/sec) FIGURE P9.7 Bode Diagram for Gc (s)G(s)H(s) = 10K1 s(s+7) , (s+3)(s2 +0.36) where K1 = 2. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 467 Problems (b) The phase margin P.M. = 80o and the gain margin G.M. = ∞, since φ never crosses = −180o . (c) The transfer function from Td (s) to θ(s) is θ(s) = G(s) Td (s) . 1 + Gc (s)G(s)H(s) Then, for a step disturbance θ(∞) = lims→0 sθ(s) = G(0) = 10/0.36 = 27.8, since H(0) = 0. (d) The system is so highly damped, there is very little resonant peak. (e) The estimated ζ = P.M./100 = 0.80. The actual ζ = 0.97. (a) The loop transfer function is Gc (s)G(s)H(s) = s2 ω12 + (0.02s + 1) 2ζ1 s ω1 s2 ω22 +1 + 2ζ2 s ω2 , +1 where ω1 = 20π = 62.8, ω2 = 14π = 43.9, ζ1 = 0.05 and ζ2 = 0.05. The Bode plot is shown in Figure P9.8a. The phase margin is P.M. = −9o . Therefore, the system is unstable. Gain dB 20 0 -20 -40 100 101 102 103 102 103 Frequency (rad/sec) 0 Phase deg P9.8 -50 -100 -150 -200 100 101 Frequency (rad/sec) FIGURE P9.8 (a) Bode Diagram for Gc (s)G(s)H(s) = and ω2 = 14π. s2 /ω12 +(0.1/ω1 )s+1 , (0.02s+1)(s2 /ω22 +(0.1/ω2 )s+1) where ω1 = 20π (b) In this case ζ2 = 0.25, with all other parameters the same as before. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 468 CHAPTER 9 Stability in the Frequency Domain 10 Gain dB 0 -10 -20 -30 -40 100 101 102 103 102 103 Frequency (rad/sec) Phase deg 0 -50 -100 -150 -200 100 101 Frequency (rad/sec) FIGURE P9.8 CONTINUED: (b) Bode Diagram for Gc (s)G(s)H(s) = where ω1 = 20π and ω2 = 14π. s2 /ω12 +(0.1/ω1 )s+1 , (0.02s+1)(s2 /ω22 +(0.5/ω2 )s+1) The Bode plot is shown in Figure P9.8b. The phase margin is P.M. = 86o . Therefore, the system is now stable. P9.9 (a) The Bode plot is shown in Figure P9.9a The phase margin is P.M. = 83o and the gain margin is G.M. = ∞. (b) With the compensator, the loop transfer function is Gc (s)G(s)H(s) = K1 0.30(s + 0.05)(s2 + 1600)(s + 0.5) , s(s2 + 0.05s + 16)(s + 70) where K2 /K1 = 0.5 . Let K1 = 1. The Bode plot is shown in Figure P9.9b. The phase margin is P.M. = 80o and the gain margin is G.M. = ∞, essentially the same as in (a). But the system in (b) is a type one, so that ess = 0 to a step input or disturbance. We cannot achieve a G.M. = 10 dB by increasing or decreasing K1 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 469 Problems 40 Gain dB 20 0 -20 -40 -60 10-3 10-2 10-1 100 101 102 103 101 102 103 102 103 102 103 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-3 10-2 10-1 100 Frequency (rad/sec) FIGURE P9.9 (a) Bode Diagram for Gc (s)G(s)H(s) = 0.3(s+0.05)(s2 +1600) . (s+70)(s2 +0.05s+16) 40 Gain dB 20 0 -20 -40 -60 10-3 10-2 10-1 10-2 10-1 100 101 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-3 100 101 Frequency (rad/sec) FIGURE P9.9 CONTINUED: (b) Bode Diagram for Gs (s)G(s)H(s) = K1 = 1. 0.15K1 (s+0.05)(s2 +1600)(s+0.5) , (s+70)(s2 +0.05s+16) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 470 CHAPTER 9 The equations of motion are F (s) = 3I(s) and I(s) = Eo (s) Eo (s) = . R + Ls 0.1 + 0.2s So, F (s) = 30 Eo (s) . (2s + 1) The actuator without the spring (see Table 2.7, Number 9 in Dorf & Bishop) is modeled via X(s) 1 Ka = = . 2 Y (s) M s + Bs τa s 2 + s With the spring, we have Ka X(s) = 2 Y (s) τa s + s + Ks or GA (s) = 0.4s2 1 . + s + 1.5 Then, the loop transfer function is L(s) = 30K1 . (2s + 1)(0.4s2 + s + 1.5) (a) The Bode plot for K1 = 0.2 in Fig. P9.10 shows the P.M. = 30o . 20 Gain dB 0 -20 -40 -60 10-2 10-1 100 101 100 101 Frequency (rad/sec) 0 Phase deg P9.10 Stability in the Frequency Domain -100 -200 -300 10-2 10-1 Frequency (rad/sec) FIGURE P9.10 Bode Diagram for L(s) = 30K1 , (2s+1)(0.4s2 +s+1.5) where K1 = 0.2. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 471 Problems (b) For K1 = 0.2, we determine that Mpω = 7.8 dB, ωr = 1.9 rad/sec, and ωB = 2.8 rad/sec. (c) The estimated percent overshoot is P.O. = 51% and the estimated settling time is Ts = 10 sec. This is based on ζ = 0.21 and ωn ≈ ωr = 1.9 rad/sec. The loop transfer function is Gc (s)G(s) = 5(K1 s + K2 )e−1.5s . s(5s + 1) (a) Let K1 = K2 = 1. Then Gc (s)G(s) = 5(s + 1) −1.5s e . s(5s + 1) The Bode plot is shown in Figure P9.11a. The phase margin is P.M. = −48o . The system is unstable. (b) Let K1 = 0.1 and K2 = 0.04. Then, the loop transfer function is Gc (s)G(s) = 5(0.1s + 0.04)e−1.5s . s(5s + 1) The Bode plot shown in Figure P9.11b shows P.M. = 45o . Thus, the system is stable. 60 Gain dB 40 20 0 -20 10-2 10-1 100 101 100 101 Frequency (rad/sec) 0 -200 Phase deg P9.11 -400 -600 -800 -1000 10-2 10-1 Frequency (rad/sec) FIGURE P9.11 (a) Bode Diagram for Gc (s)G(s) = 5(s+1)e−sT s(5s+1) , where T = 1.5. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 472 CHAPTER 9 Stability in the Frequency Domain 40 Gain dB 20 0 -20 -40 10-2 10-1 100 101 100 101 Frequency (rad/sec) 0 Phase deg -200 -400 -600 -800 -1000 10-2 10-1 Frequency (rad/sec) FIGURE P9.11 CONTINUED: (b) Bode Diagram for Gc (s)G(s) = 5(0.1s+0.04)e−sT s(5s+1) , where T = 1.5. (c) When K2 = 0.1394, the phase margin is P.M. = 0o and G.M. = 0 dB. So, for stability we require K2 ≤ 0.1394 when K1 = 0. (a) The Bode plot is shown in Figure P9.12. Bode Diagram Gm = 12 dB (at 3.46 rad/sec) , Pm = 67.6 deg (at 1.53 rad/sec) 20 Magnitude (dB) 0 −20 −40 −60 −80 −100 0 −45 Phase (deg) P9.12 −90 −135 −180 −225 −270 −1 10 0 10 Frequency (rad/sec) FIGURE P9.12 Bode Diagram for Gc (s)G(s) = 2 . (0.5s+1)3 1 10 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 473 Problems The loop transfer function (without the time delay) is Gc (s)G(s) = 2 . (0.5s + 1)3 The phase margin is P.M. = 67.6o . (b) With the delay, the loop transfer function is Gc (s)G(s)H(s) = 2e−0.5s . (0.5s + 1)3 The phase margin is now P.M. = 23.7o . So the 0.5 sec time delay has reduced the phase margin by 43.9◦ . The loop transfer function is Gc (s)G(s) = Ka (Ks + 1) −1.2s e . s (a) Let Ka = K = 1. Without the time delay, the system has infinite phase and gain margin. However, with the time delay, the system has a negative gain margin, hence it is unstable. (b) A plot of phase margin versus Ka is shown in Figure P9.13. 100 80 60 Phase margin deg P9.13 40 20 0 -20 -40 -60 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Ka FIGURE P9.13 Phase margin as a function of Ka for Gc (s)G(s) = Ka (s+1)e−1.2s . s 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 474 CHAPTER 9 Stability in the Frequency Domain Let K = 1, and find Ka for a stable system. Then, Gc (s)G(s) = Ka (s + 1)e−1.2s . s If Ka = 0.8, then the phase margin is P.M. = 50o . The loop transfer function is Gc (s)G(s) = Ke−0.2s . s(0.1s + 1) (a) The Nichols diagram is shown in Figure P9.14 for K = 2.5. 40 0 0.25 30 0.5 1 20 -1 2 3 6 10 Gain dB P9.14 -3 -6 0 -12 -10 -20 -20 -30 -40 -350 -300 -250 -200 -150 -100 -50 -40 0 Phase (deg) FIGURE P9.14 Nichols diagram for Gc (s)G(s) = Ke−0.2s , s(0.1s+1) for K = 2.5. It can be seen that Mpω = 2.0 dB . The phase and gain margins are P.M. = 48.5o and G.M. = 7.77 dB. (b) We determine that ζ = 0.43 (based on Mpω = 2 dB) and ζ = 0.48 (based on the phase margin P.M. = 48.5o ). (c) The bandwidth is ωB = 5.4 rad/sec . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 475 Problems (a) The ship transfer function is G(s) = −0.164(s + 0.2)(s − 0.32) . s2 (s + 0.25)(s − 0.009) The closed-loop system is unstable; the roots are s1 = −0.5467 s2,3 = 0.2503 ± 0.1893j s4 = −0.1949 Therefore the ship will not track the straight track. (b) The system cannot be stabilized by lowering the gain; this is verified in the root locus in Figure P9.15, where it is seen that the locus has a branch in the right half-plane for all K > 0. (c) Yes, the system can be stabilized. (d) When the switch is closed, we have a derivative feedback, which adds 90o phase lead. This is not enough to stabilize the system. Additional lead networks are necessary. 0.6 0.4 0.2 Imag Axis P9.15 0 x o xx -0.2 0 o -0.2 -0.4 -0.6 -0.6 -0.4 0.2 Real Axis FIGURE P9.15 −0.164(s+0.2)(s−0.32) Root locus for 1 + GH(s) = 1 + K s2 (s+0.25)(s−0.009) = 0. 0.4 0.6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 476 CHAPTER 9 P9.16 Stability in the Frequency Domain The loop transfer function is Gc (s)G(s) = K . (s/10 + 1)(s2 + s + 2) When K = 3.2, the phase margin is P.M. ≈ 30o . The Bode plot is shown in Figure P9.16. Gm=10.88 dB, (w= 3.464) Pm=29.91 deg. (w=2.083) 50 Gain dB 0 -50 -100 -1 10 10 0 10 1 10 2 Frequency (rad/sec) 0 Phase deg -90 -180 -270 -360 -1 10 10 0 10 1 10 2 Frequency (rad/sec) FIGURE P9.16 Bode plot for Gc (s)G(s) = P9.17 K , (s/10+1)(s2 +s+2) where K = 3.2. (a) We require ess ≤ 0.05A, and we have ess = A < 0.05A 1 + Kp or Kp > 19. But 20K1 s→0 (0.5s + 1) Kp = lim G1 (s)G2 (s)G3 (s)G4 (s) = lim s→0 So, Kp = 0.2K1 > 19, or K1 > 95. 0.1 1 + 4s 2 = 0.2K1 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 477 Problems (b) Given 1 s+1 G1 (s) = K1 (1 + ) = K1 s s , we require 1.05 < MPt < 1.30, or 0.70 > ζ > 0.36, or 70o > P.M. > 36o . Then, G1 (s)G2 (s)G3 (s)G4 (s) = 0.2K1 (s + 1) . s(0.5s + 1)(4s + 1)2 When K1 = 0.8, the P.M. = 40o . The Bode plot is shown in Figure P9.17a. Gain dB 50 0 -50 -100 10-2 10-1 100 101 100 101 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-2 10-1 Frequency (rad/sec) FIGURE P9.17 (a) Bode plot for G1 (s)G2 (s)G3 (s)G4 (s) = P.M. = 40o . 0.2K1 (s+1) , s(0.5s+1)(4s+1)2 where K1 = 0.8 and (c) For part (a), we had G1 (s)G2 (s)G3 (s)G4 (s) = 2.375 . (s + 2)(s + 0.25)2 The characteristic equation is s3 + 2.5s2 + 1.06s + 2.50 = (s + 2.48)(s2 + 0.02s + 1.013) . The dominant complex roots are lightly damped since ζ = 0.01 and ζωn = 0.01. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 9 Stability in the Frequency Domain Thus, Ts = 4 = 400 sec . ζωn For part (b), we had G1 (s)G2 (s)G3 (s)G4 (s) = (0.2)(0.8)(s + 1) . s(0.5s + 1)(4s + 1)2 The characteristic equation is 8s4 + 20s3 + 8.5s2 + 1.16s + 0.16 = 0 . The roots are s1 = −2, s2 = −0.4 and s3,4 = −0.05 ± j0.15. Thus ζ = 0.16 and ζωn = 0.05. So, Ts = 4 4 = = 75 sec . ζωn 0.05 (d) Let U (s) be a unit step disturbance and R(s) = 0. Then Y (s) G3 (s)G4 (s) = = U (s) 1 + G1 (s)G2 (s)G3 (s)G4 (s) 1+ 2 0.1 1+4s 20K1 (s+1) s(0.5s+1)(4s+1)2 The disturbance response is shown in Figure P9.17b. x10 -3 6 5 4 3 Amplitude 478 2 1 0 -1 -2 0 10 20 30 40 50 60 70 Time (secs) FIGURE P9.17 CONTINUED: (b) System response to a unit disturbance U (s). 80 90 100 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 479 Problems P9.18 The transfer function is Gc (s)G(s)H(s) = 5.3(s2 + 0.8s + 0.32)e−T s . s3 The Bode plot is shown in Figure P9.18. 80 Gain dB 60 40 20 0 -20 10-1 100 101 Frequency (rad/sec) Phase deg 300 T=0 solid ___ & T=0.1 dashed ---- & T=0.2674 dotted .... 250 200 150 100 10-1 100 101 Frequency (rad/sec) FIGURE P9.18 K(s2 +0.8s+0.32)e−sT , where T = 0 (solid line), Bode diagram for Gc (s)G(s)H(s) = s3 T = 0.1 (dashed line), and T = 0.2674 (dotted line). The following results are verified in the figure. (a) The phase margin is P.M. = 81o at ω = 5.3 when T = 0. (b) For T = 0.1, the added phase is φ = −T ω (in radians). The phase margin is P.M. = 51o at ω = 5.3 when T = 0.1. (c) The system is borderline stable when T = 0.2674 sec. The phase margin is P.M. = 0o at ω = 5.3. P9.19 The transfer function is Gc (s)G(s) = 0.5 . s(1 + 2s)(4 + s) (a) The Nichols diagram is shown in Figure P9.19. The gain margin is G.M. = 31.4 dB. (b) The phase margin is P.M. = 75o and Mpω = 0 dB. The bandwidth is 0.17 rad/sec. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 480 CHAPTER 9 Stability in the Frequency Domain Phase margin = 180-105=75o Nichols Chart 40 0.25 dB 0.5 dB 1 dB 3 dB 6 dB 20 0 0 dB sys System: Gain (dB): Phase (deg): Frequency (rad/sec): 0.122 AM IW APR IW ARN IW ARN AVN Open AMN AMN IW ATN ATN IW APNN APNN IW APRN APRN IW APVN APVN IW APMN ALMN ALPQ ARS0 ARRQ APTN APLQ O>?@ABCC> DEFG? HI?JK FIGURE P9.19 Nichols diagram for Gc (s)G(s) = AUN Gain margin = 31.4 dB APMN IW 0 AVQ 0.5 . s(2s+1)(s+4) (a) Let K = 100. The Bode plot is shown in Figure P9.20a. The loop transfer function is Gc (s)G(s) = K(s2 + 1.5s + 0.5) . s(20s + 1)(10s + 1)(0.5s + 1) Gain dB 100 50 0 -50 10-3 10-2 10-1 Frequency (rad/sec) 100 101 100 101 0 Phase deg P9.20 AVN IW System: sys Gain (dB): 4 Phase (deg): Frequency (rad/sec): 1.44 -100 -200 -300 10-3 10-2 10-1 Frequency (rad/sec) FIGURE P9.20 (a) Bode diagram for Gc (s)G(s) = K(s2 +1.5s+0.5) , s(20s+1)(10s+1)(0.5s+1) where K = 100. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 481 Problems (b) The phase margin is P.M. = −3.5o and the gain margin is G.M. = 2.7 dB. (c) You must decrease K below 100 to achieve a P.M. = 40o . For K = 0.1, the phase margin P.M. = 37.9o . (d) The step response is shown in Figure P9.20b for K = 0.1. 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 50 100 150 200 250 300 350 400 Time (secs) FIGURE P9.20 CONTINUED: (b) Unit step response K = 0.1. P9.21 The loop transfer function is Gc (s)G(s) = K . s(s + 1)(s + 4) (a) The Bode plot is shown in Figure P9.21 for K = 4. (b) The gain margin is G.M. = 14 dB . (c) When K = 5, the gain margin is G.M. = 12 dB . (d) We require Kv > 3, but Kv = K4 . So, we need K > 12. This gain can be utilized since K < 20 is required for stability. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 482 CHAPTER 9 Stability in the Frequency Domain 50 Gain dB 0 -50 -100 -150 10-1 100 101 102 101 102 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-1 100 Frequency (rad/sec) FIGURE P9.21 Bode diagram for Gc (s)G(s) = P9.22 K , s(s+1)(s+4) where K = 4. (a) The resonant frequency ωr = 5.2 rad/sec is point 6 on the Nichol’s chart. (b) The bandwidth is between points 8 and 9. We estimate the bandwidth to be ωB = 7.5 rad/sec. (c) The phase margin P.M. = 30o . (d) The gain margin G.M. = 8 dB. (e) Since we have P.M. = 30o , then we estimate ζ = 0.3. We can also approximate ωn ≈ ωr = 5.2 .ap9.1 Thus, Ts = P9.23 4 4 = = 2.5sec . ζωn 1.56 The phase margin is P.M. = 60 deg when K = 266. The gain margin is G.M. = 17.2 dB . The Bode plot is shown in Figure P9.23. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 483 Problems Bode Diagram Gm = 17.2 dB (at 9.8 rad/sec) , Pm = 60 deg (at 2.58 rad/sec) Magnitude (dB) 50 0 −50 −100 Phase (deg) −150 −90 −135 −180 −225 −270 −1 10 0 10 FIGURE P9.23 Bode diagram for Gc (s)G(s) = P9.24 1 2 10 Frequency (rad/sec) K , s(s+8)(s+12) 10 3 10 where K = 266. When K = 14.1, then P.M. = 45 deg, G.M. = ∞ dB and ωB = 29.3 rad/sec. Gm=356.59 dB (at 0 rad/sec), Pm=60 deg. (at 17.321 rad/sec) 100 Phase (deg); Magnitude (dB) 50 0 - 50 - 80 - 100 - 120 - 140 - 160 - 180 -1 10 10 0 10 1 10 2 Frequency (rad/sec) FIGURE P9.24 Bode diagram for G(s) = P9.25 K(s+20) , s2 where K = 14.1. The phase margin is P.M. = 60 deg when K = 2.61 and T = 0.2 second. The Bode plot is shown in Figure P9.25. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 484 CHAPTER 9 Stability in the Frequency Domain K=2.61; PM=60.09 at wc=2.61 rad/sec 40 Gain dB 20 0 -20 -40 -1 10 10 0 10 1 10 2 Frequency (rad/sec) Phase deg 0 -500 -1000 -1500 -1 10 10 0 10 1 10 2 Frequency (rad/sec) FIGURE P9.25 Bode diagram for Gc (s)G(s) = P9.26 Ke−0.2s , s where K = 2.61. The loop transfer function is Gc (s)G(s) = K . s(0.25s + 1)(0.1s + 1) The Bode plot is shown in Figure P9.26a for K = 10. The Nichols chart is shown in Figure P9.26b. The phase and gain margins are P.M. = 9o and G.M. = 3 dB . The system bandwidth is ωB = 8 rad/sec. From the P.M. = 9o , we estimate ζ = 0.09. Therefore, the predicted overshoot is √ −πζ/ P.O. = 100e 1−ζ 2 = 75% , where ζ = 0.09 . The resonant peak occurs at ωr = 5.5 rad/sec. If we estimate ωn ≈ ωr = 5.5 rad/sec, then the settling time is Ts = 4 = 8 sec . ζωn © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 485 Problems Gain dB 50 0 -50 -100 10-1 100 101 102 101 102 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-1 100 Frequency (rad/sec) 40 0 0.25 30 0.5 1 20 10 Gain dB -1 2 3 -3 16 0 -6 -12 -10 -20 -20 -30 -40 -350 -300 -250 -200 -150 -100 -50 -40 0 Phase (deg) FIGURE P9.26 (a) Bode diagram for Gc (s)G(s) = for Gc (s)G(s) = P9.27 K , s(0.25s+1)(0.1s+1) K , s(0.25s+1)(0.1s+1) where K = 10. (b) Nichols chart where K = 10. The loop transfer function is L(s) = Gc (s)G(s)H(s) = 4K . (s2 + 2s + 4)(s + 1) The plot of the phase margin versus the gain K is shown in Figure P9.27. As the gain increases towards Kmax = 3.5, the phase margin decreases © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 486 CHAPTER 9 Stability in the Frequency Domain towards zero. 180 160 140 Phase margin (deg) 120 100 80 60 40 20 0 1 1.5 2 2.5 3 3.5 K FIGURE P9.27 Phase margin versus the gain K. P9.28 The loop transfer function is Gc (s)G(s) = KP . s(s + 1) When KP = 1.414, we have P.M. ≈ 45◦ . Using the approximation that ζ ≈ P.M./100 we estimate that ζ = 0.45. Then using the design formula √ 2 P.O. = 100e−πζ/ 1−ζ = 20.5% . The actual overshoot is 23.4%. The step input response is shown in Figure P9.28. The actual damping ratio is ζ = 0.42. This shows that the approximation ζ ≈ P.M./100 is quite applicable and useful in predicting the percent overshoot. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 487 Problems Step Response 1.4 1.2 System: syscl Peak amplitude: 1.23 Overshoot (%): 23.3 At time (sec): 2.97 Amplitude 1 0.8 0.6 0.4 0.2 0 0 5 10 Time (sec) FIGURE P9.28 Step response showing a 23.3% overshoot. 15 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 488 CHAPTER 9 Stability in the Frequency Domain Advanced Problems The loop transfer function is L(s) = Gc (s)G(s)H(s) = 236607.5(s + 10)(s + 5) . s(s + 2)(s2 + 100s + ωn2 )(s + 1) (a) The Bode plot for ωn2 = 15267 is shown in Figure AP9.1a. 150 Gain dB 100 50 0 -50 -100 10-3 10-2 10-1 10-2 10-1 100 101 Frequency (rad/sec) 102 103 102 103 0 Phase deg AP9.1 -100 -200 -300 10-3 100 101 Frequency (rad/sec) FIGURE AP9.1 (a) Bode Diagram for L(s) = 236607.5(s+10)(s+5) 2 )(s+1) , s(s+2)(s2 +100s+ωn 2 where ωn = 15267. The phase and gain margins are P.M. = 48.6o and G.M. = 15.5 dB . (b) The Bode plot for ωn2 = 9500 is shown in Figure AP9.1b. The gain and phase margins are P.M. = 48.5o and G.M. = 10.9 dB . Reducing the natural frequency by 38% has the effect of reducing the gain margin by 30%. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 489 Advanced Problems 150 Gain dB 100 50 0 -50 -100 10-3 10-2 10-1 100 101 102 103 101 102 103 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-3 10-2 10-1 100 Frequency (rad/sec) FIGURE AP9.1 CONTINUED: (b) Bode Diagram for L(s) = 2 where ωn = 9500. (a) The Bode plot with T = 0.05 sec is shown in Figure AP9.2a. The phase margin is P.M. = 47.7o and the gain margin is G.M. = 11.2 dB. 40 Gain dB 20 0 -20 -40 100 101 102 Frequency (rad/s) -100 Phase deg AP9.2 236607.5(s+10)(s+5) 2 )(s+1) , s(s+2)(s2 +100s+ωn -200 -300 -400 100 101 Frequency (rad/s) FIGURE AP9.2 (s+5) (a) Bode Diagram for Gc (s)G(s)H(s) = 8 s(s+2) e−sT , where T = 0.05s. 102 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 490 CHAPTER 9 Stability in the Frequency Domain (b) The Bode plot with T = 0.1 sec is shown in Figure AP9.2b. The 40 Gain dB 20 0 -20 -40 100 101 Frequency (rad/s) 102 Phase deg 0 -200 -400 -600 -800 100 101 102 Frequency (rad/s) FIGURE AP9.2 (s+5) CONTINUED: (b) Bode Diagram for Gc (s)G(s)H(s) = 8 s(s+2) e−sT , where T = 0.1s. phase margin is P.M. = 22.1o and the gain margin is G.M. = 4.18 dB. A 100% increase in time delay T leads to a 50% decrease in phase and gain margins. (c) The damping ratio ζ ≈ P.M./100 and √ 2 P.O. ≈ 100e−πζ/ 1−ζ . So, for T = 0.05 sec, ζ ≈ 0.47 and P.O. ≈ 18.7%. Also, for T = 0.1 sec, ζ ≈ 0.22 and P.O. ≈ 49.2%. AP9.3 The loop transfer function is L(s) = Gc (s)G(s)H(s) = 66K(1 + 0.1s) . (1 + 0.01s)(1 + 0.01s)(1 + 1.5s)(1 + 0.2s) (a) When K = 1, the gain and phase margins are G.M. = 18.4 dB and P.M. = 55o . (b) When K = 1.5, the gain and phase margins are G.M. = 14.9 dB and P.M. = 47.8o . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 491 Advanced Problems (c,d) The bandwidth and settling time with K = 1 are ωB = 233.6 rad/sec and Ts = 0.4 second. When K = 1.5, we determine that ωB = 294.20 rad/sec and Ts = 0.33 second. AP9.4 The loop transfer function is L(s) = Gc (s)G(s) = K(s + 40) . s(s + 15)(s + 10) The gain K = 28.8 satisfies the specifications. The actual gain and phase margins are G.M. = 18.8 dB and P.M. = 45o . The system bandwidth is ωB = 10.3 rad/sec. The step response is shown in Figure AP9.4. Step Response System: sys_cl Peak amplitude: 1.23 1.4 Overshoot (%): 23.4 At time (sec): 0.476 1.2 Amplitude 1 System: sys_cl Settling Time (sec): 1.1 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 1.2 Time (sec) 1.4 1.6 1.8 2 FIGURE AP9.4 Closed-loop system step response. AP9.5 The loop transfer function is L(s) = Gc (s)G(s) = K s + 0.4 . s4 + 9s3 + 18s2 The Bode plot for K = 1 is shown in Figure AP9.5. From the phase response, we determine that the maximum P.M. ≈ 41o . From the magnitude response (for K = 1), we find that the gain needs to be raised to K = 14 to achieve maximum phase margin at ω = 0.826 rad/sec. The © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 492 CHAPTER 9 Stability in the Frequency Domain gain and phase margin with K = 14 are G.M. = 19.3 dB and P.M. = 40.9o . Also, the overshoot is P.O. = 38.3%. Bode Diagram Gm = 42.3 dB (at 3.79 rad/sec) , Pm = 16.7 deg (at 0.154 rad/sec) Magnitude (dB) 50 0 −50 −100 Phase (deg) −150 −135 System: sys Frequency (rad/sec): 0.865 Phase (deg): −139 −180 −225 −270 −2 10 −1 10 0 10 Frequency (rad/sec) 1 10 2 10 FIGURE AP9.5 s+0.4 Bode plot for L(s) = K s4 +9s 3 +18s2 with K = 1. AP9.6 With D > 2m, the gain can be increased up to K = 100, while still retaining stability. AP9.7 The loop transfer function is L(s) = Gc (s)G(s) = K(s + 4) . s2 We select √ K=2 2 for P.M. = 45o . The system bandwidth is ωB = 5.88 rad/sec . The disturbance response is shown in Figure AP9.7. The maximum output due to a disturbance is y(t) = 0.11. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 493 Advanced Problems 0.12 0.1 Amplitude 0.08 0.06 0.04 0.02 0 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (secs) FIGURE AP9.7 Closed-loop system disturbance response. A reasonable choice for the gain is K = 2680. The phase margin is P.M. = 42.8◦ and the percent overshoot is P.O. = 18.9%. The Nichols chart is shown in Figure AP9.8. Nichols Chart 60 40 0.25 dB 0.5 dB 1 dB 3 dB 6 dB 20 Open−Loop Gain (dB) AP9.8 0 dB −1 dB −20 −3 dB −6 dB −12 dB −20 dB −40 −40 dB −60 −60 dB −80 −80 dB −100 −100 dB 0 −120 −360 FIGURE AP9.8 Nichols chart. −315 −270 −225 −180 −135 Open−Loop Phase (deg) −90 −120 dB −45 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 494 CHAPTER 9 AP9.9 Stability in the Frequency Domain The loop transfer function is L(s) = Gc (s)G(s) = Kp (s + 0.2) 2 s (s2 + 7s + 10) . At the maximum phase margin, Kp = 4.9 for P.M. = 48.6o . The Bode diagram is shown in Figure AP9.9. Bode Diagrams Gm=21.788 dB (at 2.9326 rad/sec), Pm=48.457 deg. (at 0.50782 rad/sec) 100 50 0 Phase (deg); Magnitude (dB) -50 -100 -150 -100 -150 -200 -250 -300 -3 10 10 -2 10 -1 10 0 10 1 10 2 Frequency (rad/sec) FIGURE AP9.9 Phase and gain margin. AP9.10 The closed-loop transfer function is T (s) = s2 K . + 3s + 1 We require K = 1 a zero steady-state tracking error to a unit step. The step response is shown in Figure AP9.10. Computing T (jω) = 0.707 it follows that (jω)2 1 = 0.707 + 3jω + 1 or ω 4 + 7ω 2 − 1 = 0 . Solving for ω yields ω = 0.37 rad/s. This is the bandwidth of the system. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 495 Advanced Problems Step Response 1 0.9 0.8 Amplitude 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 2 4 6 8 Time (sec) 10 12 14 16 FIGURE AP9.10 Unit step response. The phase margin versus time delay is shown in Figure AP9.11a. 80 Time Delay=1 70 PM=58.5285 Time Delay=3.0455 60 Phase Margin (deg) AP9.11 PM=0.001 50 40 30 20 10 0 −10 0.5 1 FIGURE AP9.11 Phase margin versus time delay. 1.5 2 Time Delay (s) 2.5 3 3.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 9 Stability in the Frequency Domain The maximum time delay is T = 3.04 s for stability. The step response is shown in Figure AP9.11b. The percent overshoot is P.O. = 7.6%. 1.4 1.2 1 Amplitude 496 0.8 0.6 0.4 0.2 0 0 FIGURE AP9.11 Unit step response. 1 2 3 4 5 6 Time (sec) 7 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 497 Design Problems Design Problems CDP9.1 The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: 26.035 θ(s) = , Va (s) s(s + 33.142) where we neglect the motor inductance Lm and where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). The closedloop system characteristic equation is 1+ 26.035Ka =0. s(s + 33.142) The phase margin is P.M. = 70.4◦ when Ka = 16. The step response with K = 16 is shown below. 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 DP9.1 0 0.05 0.1 0.15 Time (secs) 0.2 0.25 0.3 (a) The gain and phase margins are G.M. = 7 dB and P.M. = 60o . (b) The resonant peak and frequency are Mpω = 2 dB and ωr = 5 rad/sec. (c) We have ωB = 20 rad/sec. From Mpω = 2 dB we estimate ζ = 0.45 (Figure 8.11 in Dorf & Bishop). Also, ωr /ωn = 0.8, so ωn = 6.25. Thus, Ts = 1.4. (d) We need P.O. = 30o or ζ = 0.3 or P.M. ≈ 30o . So, we need to raise the gain by 10 dB or K = 3.2. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 498 CHAPTER 9 The loop transfer function is L(s) = Gc (s)G(s) = K(s + 0.5) . + 7.5s + 9) s2 (s2 When K = 6.25, we have the maximum phase margin. The phase margin maximum is P.M. = 23o . The plot of P.M. versus K is shown in Figure DP9.2a. 24 22 20 18 Phase Margin deg DP9.2 Stability in the Frequency Domain 16 14 12 10 8 6 4 0 1 2 3 4 5 6 7 8 9 K FIGURE DP9.2 (a) Phase margin versus K for L(s) = K(s+0.5) . s2 (s2 +7.5s+9) The predicted damping is ζ = 0.23. It then follows that the predicted percent overshoot is √ 2 P.O. = 100e−πζ/ 1−ζ = 48% . The actual overshoot is 65%. The step input response is shown in Figure DP9.2b. The resonant peak occurs at ωr = 0.75 rad/sec. Approximating ωn ≈ ωr = 0.75 rad/sec, we can estimate the settling time as Ts = 4 = 23 sec . ζωn The actual settling time is 20 sec. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 499 Design Problems 1.8 1.6 1.4 Amplitude 1.2 1 0.8 0.6 0.4 0.2 0 0 5 10 15 20 25 30 Time (secs) FIGURE DP9.2 CONTINUE: (b) Closed-loop unit step response. We want to select the gain K as large as possible to reduce the steady-state error, but we want a minimum phase margin of P.M. = 45o to achieve good dynamic response. A suitable gain is K = 4.2, see Figure DP9.3. K=4.2; PM=45.34 at wc=0.102 rad/sec 20 Gain dB 0 -20 -40 -2 10 10 -1 10 0 10 1 Frequency (rad/sec) 0 -100 Phase deg DP9.3 -200 -300 10 -2 10 -1 10 Frequency (rad/sec) FIGURE DP9.3 Bode plot for G(s) = Ke−10s 40s+1 . 0 10 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 500 CHAPTER 9 DP9.4 Stability in the Frequency Domain We are given the loop transfer function L(s) = Gc (s)G(s) = K s(s + 1)(s + 4) which can be written as Gc (s)G(s) = Kv . s(s + 1)(0.25s + 1) The performance results are summarized in Table DP9.4. Kv TABLE DP9.4 G.M. P.M. ωB P.O. Ts (dB) (deg) (rad/sec) (%) (sec) 0.40 21.9 64.2 0.62 4.4 9.8 0.75 16.5 49.0 1.09 19.0 10.1 Summary for Kv = 0.40 and Kv = 0.75. When Kv = 0.40, we have ess 1 = = 2.5 , A 0.40 or 2 1/2 times the magnitude of the ramp. This system would be acceptable for step inputs, but unacceptable for ramp inputs. DP9.5 (a) With a time delay of T = 0.8 second, we determine that the proportional controller Gc (s) = K = 7 provides a suitable response with P.O. = 8.3 % ess = 12.5 % Ts = 4.38 sec . (b) A suitable proportional, integral controller is Gc (s) = K1 + K2 /s = 6 + 0.6/s . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 501 Design Problems The response to a unit step is P.O. = 5.14 % ess = 0 % Ts = 6.37 sec . The Nichols chart is shown in Figure DP9.5. 40 0 0.25 30 0.5 1 20 3 6 8 10 Gain dB -1 -3 -6 0 -12 -10 -20 -20 -30 -40 -350 -300 -250 -200 -150 -100 -50 -40 0 Phase (deg) FIGURE DP9.5 Nichols chart for Gc (s)G(s) = DP9.6 (K1 s+K2 )e−0.8s , s(10s+1) where K1 = 6 and K2 = 0.6. With K = 170, at the two extreme values of b, we have b = 80 b = 300 P.M. = 91.62o P.M. = 75.23o G.M. = 13.66 dB G.M. = 25.67 dB . Since reducing the value of K only increases the P.M. and G.M., a value of K = 170 is suitable to meet P.M. = 40o and G.M. = 8 dB for the range of b. DP9.7 A suitable gain is K = 0.22 . This results in P.M. = 60.17o and G.M. = 13.39 dB. The step reponse is shown in Figure DP9.7. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 502 CHAPTER 9 Stability in the Frequency Domain 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 20 Time (secs) FIGURE DP9.7 Lunar vehicle step response. A gain of K = 315000 will satisfy the P.O. specification, while giving the fastest response. The step response is shown in Figure DP9.8. 1.2 1 0.8 Amplitude DP9.8 0.6 0.4 0.2 0 0 0.1 0.2 0.3 0.4 0.5 Time (secs) FIGURE DP9.8 Steel rolling mill step response. 0.6 0.7 0.8 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 503 Design Problems The closed-loop transfer function is Ts (2) = Gc (s)G2 (s) G1 (s) To (s) + T2d (s) . 1 + Gc (s)G2 (s) 1 + Gc (s)G2 (s) where 1 (10s + 1)(50s + 1) G1 (s) = and G2 (s) = 0.01 . (10s + 1)(50s + 1) The steady-state error (with Gc (s) = 500) to a unit step 2A (and after the system has settled out subsequent to a step of magnitude A) is ess = 2(0.167) = 0.33 . The step response is shown in Figure DP9.9. Gc=500 (solid); Gc=1/s (dashed); Gc=600+6/s (dotted) 2.5 2 1.5 T2/A DP9.9 1 0.5 0 0 200 400 600 800 1000 Time (sec) FIGURE DP9.9 Two tank temperature control step response. A suitable integral controller is Gc (s) = 1 . s 1200 1400 1600 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 504 CHAPTER 9 Stability in the Frequency Domain In this case, the steady-state tracking error is zero , since the system is a type 1. The system response is shown in Figure DP9.9. With the integral controller, the settling time is about Ts = 438 seconds and the P.O. = 7%. A suitable PI controller is Gc (s) = 600 + 6 . s With the PI controller, the settling time is about Ts = 150 seconds and the P.O. = 10%. DP9.10 The system is given by ẋ = Ax + Br y = Cx where A= 0 1 2 − K1 3 − K2 , B= 0 1 The associated transfer function is T (s) = s2 , and C= 1 0 . 1 . + (K2 − 3)s + K1 − 2 The characteristic polynomial is s2 + (K2 − 3)s + K1 − 2 = 0 . If we select K1 = 3, then we have a zero-steady error to a unit step response R(s) = 1/s, since s2 + (K2 − 3)s =0. s→0 s2 + (K2 − 3)s + K1 − 2 lim s [1 − T (s)] R(s) = lim s→0 Let K= 3 4.3 . The step response is shown in Figure DP9.10a. The bandwidth is ωb = 1.08 rad/s, as seen in Figure DP9.10b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 505 Design Problems Step Response 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 Time (sec) 6 7 8 9 Bode Diagram 5 0 System: sys Frequency (rad/sec): 1.08 Magnitude (dB): −3 −5 Magnitude (dB) −10 −15 −20 −25 −30 −35 −40 −1 10 0 10 Frequency (rad/sec) 1 10 FIGURE DP9.10 Step response with K = [3 4.3] and closed-loop Bode plot. DP9.11 A time domain step response specification P.O. > 10% requires the dominant poles to have a damping ration of ζ = 0.6. This time domain specification can be transformed to a frequency response specification using the approximation P.M. ≈ 100ζ = 60◦ . To keep the problem tractable, we consider the controller with the form Gc (s) = KP + KI 1 = KP + , s s where we let KI = 1. The plot of the P.M. as a function of KP is shown © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 9 Stability in the Frequency Domain in Figure DP9.11a. If we select KP = 0.07 we expect a phase margin of approximately 60◦ , hence a percent overshoot P.O. ≤ 10%. The step response is shown in Figure DP9.11b. The actual phase margin is P.M. = 60.2◦ , the percent overshoot is P.O. = 5.9% and the settling time is Ts = 3.4 sec. 85 Phase Margin (deg) 80 75 70 65 60 55 0 0.05 0.1 0.15 0.2 0.25 KP 0.3 0.35 0.4 0.45 0.5 1.4 1.2 1 Amplitude 506 0.8 0.6 0.4 0.2 0 0 1 2 3 Time (sec) 4 5 6 FIGURE DP9.11 (a) Phase margin versus controller gain KP and KI = 1. (b) Step response with KP = 0.07 and KI = 1. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 507 Computer Problems Computer Problems The m-file script to generate the Bode plot (from which the gain and phase margin can be determined) is shown in Figure CP9.1. The transfer function is G(s) = s2 141 . + 2s + 12 The gain margin is G.M. = ∞ and the phase margin is P.M. = 10o . num=141; den=[1 2 12]; sys = tf(num,den); margin(sys); Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 10 deg (at 12.3 rad/sec) Magnitude (dB) 40 20 0 −20 −40 0 Phase (deg) CP9.1 −45 −90 −135 −180 −1 10 0 10 Frequency (rad/sec) FIGURE CP9.1 Gain and phase margin with the margin function. 1 10 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 508 CHAPTER 9 The Nyquist plots are shown in Figures CP9.2a-c. num=[5]; den=[1 5]; sys=tf(num,den); nyquist(sys) 0.5 0.4 0.3 Imaginary Axis 0.2 0.1 0 −0.1 −0.2 −0.3 −0.4 −0.5 −1 −0.8 −0.6 FIGURE CP9.2 (a) Nyquist plot for G(s) = −0.4 −0.2 0 Real Axis 0.2 0.4 0.6 0.8 1 5 s+5 . num=[50]; den=[1 10 25]; sys=tf(num,den); nyquist(sys) 1.5 1 Imaginary Axis CP9.2 Stability in the Frequency Domain 0.5 0 −0.5 −1 −1.5 −1 −0.5 0 0.5 Real Axis FIGURE CP9.2 CONTINUED: (b) Nyquist plot for G(s) = 50 s2 +10s+25 . 1 1.5 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 509 Computer Problems num=[15]; den=[1 3 3 1]; sys=tf(num,den); nyquist(sys) 15 10 Imaginary Axis 5 0 −5 −10 −15 −5 0 5 Real Axis FIGURE CP9.2 CONTINUED: (c) Nyquist plot for G(s) = CP9.3 10 15 15 s3 +3s2 +3s+1 . The m-file script to generate the Nichols chart for part (a) is shown in Figure CP9.3a. The Nichols charts for (b) and (c) are similiarly generated; all plots are in Figure CP9.3a-c. Nichols Chart 40 0 dB 30 0.25 dB num = [1]; den = [1 0.2]; sys = tf(num,den); nichols(sys) ngrid Open−Loop Gain (dB) 0.5 dB 20 1 dB −1 dB 3 dB 10 −3 dB 6 dB 0 −6 dB −10 −12 dB −20 dB −20 −360 FIGURE CP9.3 (a) M-file script and Nichols chart for G(s) = −315 1 s+0.1 . −270 −225 −180 −135 Open−Loop Phase (deg) −90 −45 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 9 Stability in the Frequency Domain The gain and phase margin for each transfer function are as follows: (a) G.M. = ∞ and P.M. = 102o (b) G.M. = ∞ and P.M. = ∞ (c) G.M. = 20 dB and P.M. = ∞ Nichols Chart 40 0 dB 30 0.25 dB 0.5 dB Open−Loop Gain (dB) 20 1 dB −1 dB 3 dB 6 dB 10 −3 dB 0 −6 dB −10 −12 dB −20 −20 dB −30 −40 dB −40 −50 −60 −360 −315 −270 −225 −180 −135 Open−Loop Phase (deg) FIGURE CP9.3 CONTINUED: (b) Nichols chart for G(s) = −90 −45 −60 dB 0 1 s2 +2s+1 . Nichols Chart 40 0 dB 0.25 dB 0.5 dB 1 dB 20 −1 dB 3 dB 6 dB −3 dB −6 dB 0 Open−Loop Gain (dB) 510 −12 dB −20 −20 dB −40 −40 dB −60 −60 dB −80 −80 dB −100 −360 −315 −270 −225 −180 −135 Open−Loop Phase (deg) FIGURE CP9.3 CONTINUED: (c) Nichols chart for G(s) = −90 24 s3 +9s2 +26s+24 . −45 −100 dB 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 511 Computer Problems CP9.4 To obtain a phase margin P.M. = 40◦ we select K = 15 when T = 0.2 second. The variation in the phase margin for 0 ≤ T ≤ 0.3 is shown in Figure CP9.4. T=[0:0.01:0.3]; K=15; num=K;den=[1 12]; sys = tf(num,den); % w=logspace(-2,1,400); for i=1:length(T) [mag,phase,w]=bode(sys); ph(1:length(phase))=phase(1,1,:); ph=ph'; ph2=ph-w*T(i)*180/pi; [Gm,Pm,Wcg,Wcp]=margin(mag,ph2,w); clear ph ph2 PMo(i)=Pm; end plot(T,PMo), grid xlabel('Time delay (sec)') ylabel('Phase margin (deg)') K=15 160 140 Phase margin (deg) 120 100 80 60 40 20 0 −20 0 0.05 0.1 0.15 0.2 Time delay (sec) 0.25 0.3 FIGURE CP9.4 Variation in the phase margin for 0 ≤ T ≤ 0.3 with K = 15. CP9.5 The loop transfer function is L(s) = Gc (s)G(s) = K(s + 50) . s(s + 20)(s + 10) 0.35 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 512 CHAPTER 9 Stability in the Frequency Domain The plot of system bandwidth versus the gain K is shown in Figure CP9.7. K=[0.1:1:50]; w=logspace(-2,3,2000); den=[1 30 200 0]; for i=1:length(K) num=K(i)*[1 50]; sys = tf(num,den); sys_cl = feedback(sys,[1]); [mag,phase,w]=bode(sys_cl,w); L=find(mag<0.707); wb(i)=w(L(1)); end plot(K,wb), grid xlabel('Gain K') ylabel('Bandwidth (rad/sec)') 15 Bandwidth (rad/sec) 10 5 0 0 5 10 15 20 25 Gain K 30 35 40 45 50 FIGURE CP9.5 Variation in the system bandwidth for 0 ≤ K ≤ 50. CP9.6 The m-file script and Bode plot are shown in Figure CP9.6. The gain and phase margin and ωc are determined to be G.M. = 2.23, P.M. = 26o and ωc = 12.6 rad/sec. So, the maximum value of bo is found to be bomax = 2.13bo = 1.11 . In this problem, there is also a minimum value of bo . Using the Routh- © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 513 Computer Problems gm = 2.2238 numg = -0.5*[1 0 -2500]; deng = [1 47 850 -3000]; sysg = tf(numg,deng); numc = 10*[1 3]; denc = [1 0]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); bode(sys_o) [mag,phase,w] = bode(sys_o); [gm,pm,wg,wc] = margin(mag,phase,w) pm = 26.3187 wg = 26.1155 wc = 12.6487 Gain dB 50 0 -50 10-1 100 101 102 103 102 103 Frequency (rad/sec) Phase deg 250 200 150 100 50 10-1 100 101 Frequency (rad/sec) FIGURE CP9.6 Using the margin function to compute stability margins. Hurwitz method, we determine that (for stability) the range of bo is 0.14 < bo < 1.11 . CP9.7 The m-file script is shown in Figure CP9.7a. Since we do not have a value for J, we write the loop transfer function as Gc (s)G(s) = K̄1 + K̄2 s s2 where K̄1 = K1 /J and K̄2 = K2 /J. We work with K̄1 and K̄2 , then we can always compute K1 and K2 whenever J is specified. A PD controller which meets the specs is given by Gc (s) = 0.04 + 0.3s . The step response is shown in Figure CP9.7b. The Bode plot is shown in © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 514 CHAPTER 9 Stability in the Frequency Domain % % Part (a) % numc = [0.3 0.04]; denc = [1]; sysc = tf(numc,denc); numg = [1]; deng = [1 0 0]; sysg = tf(numg,deng); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); step(sys_cl), pause % % Part (b) % w = logspace(-1,1,400); [mag,phase] = bode(sys_o,w); [gm,pm,w1,w2] = margin(mag,phase,w); margin(mag,phase,w), pause % % Part (c) % T = [1:0.1:5]; for i = 1:length(T) [numd,dend] = pade(T(i),2); sysd = tf(numd,dend); sys_o1 = series(sysd,sys_o); sys_cl1 = feedback(sys_o1,sysd); p(:,i) = pole(sys_cl1); end plot(real(p),imag(p),'*');grid xlabel('Real Axis'); ylabel('Imag Axis') FIGURE CP9.7 Script to assist in all three parts of the problem. Figure CP9.7c. The phase margin is P.M. = 67.7o at ω = 0.32 rad/sec. The loop transfer function is Gc (s)G(s)H(s) = K̄1 + K̄2 s −2T s e s2 where T is the one-way time delay. If the phase lag introduced by the delay is greater than 67.7o at ω = 0.32 rad/sec, then the system will become unstable. So, since the phase lag due to the time delay T̃ is φ(ω) = ω T̃ we have 67.7o π/180 = 0.32(2T ) where T̃ = 2T . Solving for T yields T = 1.82 seconds. This is the maximum allowable one-way time delay. Executing the third part of the m-file script in Figure CP9.7a generates the plot illustrating the movement of the closed-loop system roots as the time delay is varied. The plot is shown in Figure CP9.7d. Examining the root locations, we find that when T = 1.9, the closed-loop roots © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 515 Computer Problems 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 5 10 15 20 25 30 35 40 Time (secs) FIGURE CP9.7 CONTINUED: (b) Step response without time delays meets specs. Gain dB 20 0 -20 -40 10-1 100 101 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-1 100 101 Frequency (rad/sec) FIGURE CP9.7 CONTINUED: (c) System Bode plot shows P.M. = 67.7o . are s1 = −4.56, s2,3 = −0.94 ± 2.02j, s4 = −0.19, and s5,6 = ±0.32j. Therefore, the system is marginally stable when T = 1.9, and is unstable as the time delay increases. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 516 CHAPTER 9 Stability in the Frequency Domain 4 * * 3 * * * * * 2 Imag Axis 1 0 * * * * * ** ** ** ** *** *** **** *** *** * *********** ** ****** ** ********** * * * * * * * * * * * * * * ********************** -1 -2 -3 * * * * * * * *** *** **** * * ** *** ** ** ** * * -4 -8 -7 -6 -5 -4 -3 -2 -1 0 1 Real Axis FIGURE CP9.7 CONTINUED: (d) Closed-loop root locations as the time delay varies. CP9.8 The Nyquist plot and associated m-file code are shown in Figure CP9.8. Nyquist Diagram 150 a=[0 1;-1 -15]; b=[0;30]; c=[8 0]; d=[0]; sys=ss(a,b,c,d); nyquist(sys) Imaginary Axis 100 50 0 −50 −100 −150 −50 FIGURE CP9.8 Using the Nyquist function to obtain a Nyquist plot. 0 50 100 Real Axis 150 200 250 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 517 Computer Problems CP9.9 The Nichols chart is shown in Figure CP9.9. The phase and gain margins are 37.1 degrees and ∞ dB, respectively. a=[0 1;-1 -10]; b=[0;22]; c=[10 0]; d=[0]; sys=ss(a,b,c,d); nichols(sys) ngrid Nichols Chart 60 Open-Loop Gain (dB) 40 0.25 dB 0.5 dB 1 dB 3 dB 6 dB 20 0 dB ?-1 dB ?-3 dB ?-6 dB 0 ?-12 dB -20 ?-20 dB -40 ?-40 dB -60 -360 ?-60 dB -315 -270 -225 -180 -135 -90 -45 0 Open-Loop Phase (deg) FIGURE CP9.9 The Nichols chart for the system in CP9.8. CP9.10 (a) The Nyquist plot is shown in Figure CP9.10. The phase margin is P.M. = 18o . (b) When the time delay is T = 0.05 seconds, the phase margin is P.M. = 9o . (c) When the time delay is T = 0.1 seconds, the system is marginally stable. So, for T > 0.1 seconds, the system is unstable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 9 Stability in the Frequency Domain Nyquist Diagram 200 150 100 Imaginary Axi s 518 50 -1 point 0 -50 -100 -150 -200 -10 -8 -6 -4 Real Axi s FIGURE CP9.10 Nyquist plot for G(s)H(s) = 10 . s(s+1) -2 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 1 0 The Design of Feedback Control Systems Exercises E10.1 From the design specifications, we determine that our desired ζ = 0.69 and ωn = 5.79. The characteristic equation is 1 + Gc (s)G(s) = 1 + K(s + a) =0, s(s + 2) or s2 + (2 + K)s + Ka = 0 . Our desired characteristic polynomial is s2 + 2ζωn s + ωn2 = s2 + 8s + 33.6 = 0 . Thus, K + 2 = 8, or K=6 and Ka = 33.6, so a = 5.6. The actual percent overshoot and settling time will be different from the predicted values due to the presence of the closed-loop system zero at s = −a. In fact, the actual percent overshoot and settling time are P.O. = 12.6% and Ts = 0.87s, respectively. E10.2 The characteristic equation is 400 1 1 + Gc (s)G(s) = 1 + K1 + s(s + 40) s =1+ 400(K1 s + 1) =0, s2 (s + 40) or 1 + K1 400s =0. s3 + 40s2 + 400 519 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 520 CHAPTER 10 The Design of Feedback Control Systems We desire ζ = 0.45 for an overshoot of 20%. The root locus is shown in Figure E10.2. We select a point slightly inside the performance region (defined by ζ = 0.45 ) to account for the zero. Thus, K1 = 0.5 and the closed-loop poles are s1 = −35 s2,3 = −2.7 ± j2 . The actual P.O. = 20.7% . 50 40 30 Imag Axis 20 10 0 x * * * x o x -10 -20 -30 -40 -50 -50 -40 -30 -20 -10 0 Real Axis FIGURE E10.2 400s Root locus for 1 + K1 s3 +40s 2 +400 = 0. E10.3 The step response is shown in Figure E10.3 for τ = 1 and K = 0.5. It can be seen that the P.O. = 4% , so this is a valid solution. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 521 Exercises 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 -0.2 1 0 2 3 4 5 6 7 8 Time (secs) FIGURE E10.3 Step response for K = 0.5 and τ = 1. The Bode plot is shown in Figure E10.4. The phase and gain margins are marked on the plot, where it can be seen that P.M. = 75.4o and G.M. = 28.6 dB. Bode Diagram Gm = 28.6 dB (at 11.8 rad/sec) , Pm = 75.4 deg (at 0.247 rad/sec) Magnitude (dB) 150 100 50 0 -50 -100 -150 -45 Phase (deg) E10.4 -90 -135 -180 -225 -270 -4 10 10 -2 10 0 Frequency (rad/sec) FIGURE E10.4 Bode plot for Gc (s)G(s) = 100(s+0.15)(s+0.7) . s(s+5)(s+10)(s+0.015)(s+7) 10 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 522 CHAPTER 10 E10.5 The Design of Feedback Control Systems We require that Kv ≥ 2.7, ζ = 0.5 and ωn = 3 for the dominant roots. We want to place a zero to left of the pole at -2, so the complex roots will dominate. Set the zero at s = −2.2. Then for the desired roots find the location of pole p in compensator Gc (s) = K1 (s + 2.2) (s + p) to satisfy 180o phase at the desired roots. This yields p = 16.4. Using root locus methods, we find that KK1 = 165.7, so with K1 = 7.53, we determine that K = 22, and Gc (s) = 7.46(s + 2.2) . (s + 16.4) Then Kv = 2.78 . E10.6 The closed-loop transfer function is T (s) = 326(s + 4) Gc (s)G(s) = 4 . 3 1 + Gc (s)G(s) s + 14.76s + 151.3s2 + 349.8s + 1304 The roots are s1,2 = −0.87 ± j3.2 s3,4 = −6.5 ± j8.7 . Assuming s1,2 dominates, then we expect overshoot P.O. = 43% and Ts = 4.6 sec . The discrepencies with the actual P.O. and Ts are due to the poles s3,4 and the zero at s = −4. E10.7 The loop transfer function is L(s) = Ke−0.6s . s(s + 20) A plot of P.M. as a function of K is shown in Figure E10.7. It can be seen that P.M. = 40o when K = 26.93. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 523 Exercises phase margin versus K (PM=40º, K=26.93) 90 80 Phase Margin deg 70 60 50 40 30 20 10 0 0 5 10 15 K 20 25 30 FIGURE E10.7 Plot of phase margin versus K. E10.8 The open-loop transfer function is G(s) = 2257 806071.4 = , s(0.0028s + 1) s(s + 357.14) and the compensator is Gc (s) = K1 (s + z) , s where z = K2 /K1 . The characteristic equation is s3 + 357.14s2 + K1 s + K2 = 0 . Using Routh-Hurwitz methods, the system is stable for 0 < K2 < 357.14 K1 or K2 /K1 < 357.14. Select the zero z at s = −10, then using root locus methods we determine that K1 = 0.08 and K2 = 0.8. The roots of the characteristic equation are s1 = −10.6 and s2,3 = −175 ± j175 , and ζ = 0.707, as desired. The step response is shown in Figure E10.8. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 524 CHAPTER 10 The Design of Feedback Control Systems 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Time (secs) FIGURE E10.8 Step response with K1 = 0.08 and K2 = 0.8. E10.9 The loop transfer function is L(s) = Gc (s)G(s) = K1 (s + K2 /K1 ) , s(s + 1) and Kv = lim sGc (s)G(s) = K2 . s→0 Select K2 = 5. The characteristic equation is s2 + (K1 + 1) + K2 = 0 , and we want s2 + 2ζωn s + ωn2 = 0 . √ √ Equating coefficients yields ωn = K2 = 5. Also, since we want P.O. = 5%, we require ζ = 0.69. Thus, 2ζωn = K1 + 1 implies K1 = 2.08 . √ The step response with K1 = 2.08 and K2 = 5 yields a P.O. > 5%. This © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 525 Exercises is due to the zero at s = −1.08 . So, we raise the gain K1 = 3 and then the P.O. = 5%. The step response is shown in Figure E10.9. 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 Time (secs) FIGURE E10.9 Step response with K1 = 3 and K2 = 5. E10.10 The loop transfer function is L(s) = Gc (s)G(s) = (KP s + KI ) . s(s + 1)(s + 2) Let KI = 2. Then, the plot of the phase margin as a function of KP is shown in Figure E10.10, where it can be seen that P.M. = 71.6o is the maximum achievable phase margin. When KP = 1.54 and KI = 2 we have P.M. = 60o , as desired, and P.O. = 9% and Tp = 3.4 sec. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 526 CHAPTER 10 The Design of Feedback Control Systems 75 Phase Margin (deg) 70 65 60 55 50 1 2 3 4 5 6 7 8 9 10 KP FIGURE E10.10 Phase margin versus KP with KI = 2. The Nichols diagram and the closed-loop Bode plot are shown in Figures E10.11a and E10.11b, respectively. 40 0 0.25 30 0.5 1 20 -1 2.3 10 Gain dB E10.11 -3 -6 0 -12 -10 -20 -20 -30 -40 -350 -300 -250 -200 -150 -100 Phase (deg) FIGURE E10.11 (a) Nichols diagram for Gc (s)G(s) = 1350(1+0.25s) . s(s+2)(s+30)(1+0.025s) -50 -40 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 527 Exercises Bode Diagram 20 Magnitude (dB) 0 −20 −40 −60 −80 −100 0 Phase (deg) −45 −90 −135 −180 −225 −270 0 10 1 2 10 3 10 10 Frequency (rad/sec) FIGURE E10.11 CONTINUED: (b) Closed-loop Bode plot. E10.12 The loop transfer function is L(s) = Gc (s)G(s) = KK1 s + 1 2 s2 (s + 5) . When KK1 = 5.12, the roots are s1,2 = −0.58 ± j0.58 s3 = −3.84 . The complex poles have ζ = 0.707 and the predicted settling time is Ts = 4/0.58 = 6.89 sec . The actual settling time is Ts = 6.22 s. E10.13 For the cascade compensator, we have T1 (s) = Gc (s)G(s) 8.1(s + 1) = , 1 + Gc (s)G(s) (s + r1 )(s + r̂1 )(s + r2 ) where r1 = −1 + j2 and r2 = −1.67. For the feedback compensator, we © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 528 CHAPTER 10 The Design of Feedback Control Systems have T2 (s) = 8.1(s + 3.6) G(s) = , 1 + Gc (s)G(s) (s + r1 )(s + r̂1 )(s + r2 ) where G(s) = 8.1 s2 and Gc (s) = s+1 . s + 3.6 The response of the two systems differ due to different value of the zero of T1 and T2 , however, both systems have the same characteristic equation. E10.14 The Bode plot (with the lag network) is shown in Figure E10.14; the phase margin is P.M. = 46o . Bode Diagram Gm = 21.9 dB (at 1.84 rad/sec) , Pm = 46.4 deg (at 0.344 rad/sec) 100 Magnitude (dB) 50 0 −50 −100 −150 −90 Phase (deg) −135 −180 −225 −270 −4 10 −3 10 FIGURE E10.14 Bode plot for Gc (s)G(s) = E10.15 −2 10 −1 0 10 Frequency (rad/sec) 5(7.5s+1) s(s+1)(0.25s+1)(110s+1) 10 1 10 2 10 = 0. At the desired crossover frequency ωc = 10 rad/sec, we have 20 log |Gc (j10)G(j10)| = −8.1 dB and 6 Gc (j10)G(j10) = −169o . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 529 Exercises Therefore, the phase margin is P.M. = 11o . So, φ = 30o − 11o = 19o and M = 8.1 dB . Since φ > 0 and M > 0, a lead compensator is required. E10.16 At the desired crossover frequency ωc = 2 rad/sec, we have 20 log |Gc (j2)G(j2)| = 17 dB and 6 Gc (j2)G(j2) = −134o . Therefore, the phase margin is P.M. = 46o . So, φ = 30o − 46o = −16o M = −17 dB . Since φ < 0 and M < 0, a lag compensator is required. E10.17 Using a prefilter Gp (s) = KI KP s + KI the closed-loop transfer function is T (s) = s2 KI . + (KP + 1)s + KI The required coefficients for a deadbeat system are α = 1.82 and Ts = 4.82. Therefore, KI = ωn2 KP = αωn − 1 . Since we desired a settling time less than 2 seconds, we determine that ωn = Ts /2 = 4.82/2 = 2.41 . Then, the gains are KP = 3.39 KI = 5.81 . The step response (with the prefilter) is shown in Figure E10.17. The percent overshoot is P.O. = 0.098% and the settling time is Ts = 1.99 seconds. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 530 CHAPTER 10 The Design of Feedback Control Systems 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Time (secs) FIGURE E10.17 Step response for the deadbeat system. E10.18 Consider the PI controller Gc (s) = Kp + Kp s + KI 30s + 300 KI = = s s s and the prefilter Gp (s) = 10 . Then, the closed-loop system is T (s) = s2 300s + 3000 . + 280s + 3000 The percent overshoot is P.O. = 9.2% and the settling time Ts = 0.16 seconds. The steady-state tracking error to a unit step is zero, as desired. E10.19 Consider the PID controller Gc (s) = 29 s2 + 10s + 100 . s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 531 Exercises The closed-loop transfer function is 29(s2 + 10s + 100) . s3 + 24s2 + 290s + 2900 T (s) = The settling time to a unit step is Ts = 0.94 seconds. E10.20 Consider the PD controller Gc (s) = KD s + Kp = 3s + 1 . The loop transfer function is L(s) = Gc (s)G(s) = 3s + 1 . s(s − 2) The Bode plot is shown in Figure E10.20. The phase margin is P.M. = 40.4◦ . This is a situation where decreasing the gain leads to instability. The Bode plot shows a negative gain margin indicating that the system gain can be decreased up to -3.5 dB before the closed-loop becomes unstable. Bode Diagram Gm = −3.52 dB (at 0.816 rad/sec) , Pm = 40.4 deg (at 2.28 rad/sec) Magnitude (dB) 40 20 0 −20 Phase (deg) −40 −90 −135 −180 −225 −270 −2 10 −1 10 0 10 Frequency (rad/sec) FIGURE E10.20 Bode plot for the loop transfer function L(s) = E10.21 3s+1 . s(s−2) The transfer function from Td (s) to Y (s) is T (s) = s2 1 . + 4.4s + K 1 10 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 10 The Design of Feedback Control Systems The tracking error is E(s) = R(s) − Y (s). When R(s) = 0, then E(s) = −Y (s). The final value of the output to a unit step disturbance is ess = 1/K. If we want the tracking error to be less than 0.1, then we require K > 10. When K = 10, we have the disturbance response shown in Figure E10.21. Step Response 0.12 0.1 0.08 Amplitude 532 0.06 0.04 0.02 0 0 0.5 FIGURE E10.21 Disturbance response for K = 10. 1 1.5 Time (sec) 2 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 533 Problems Problems P10.1 (a) The loop transfer function is L(s) = Gc (s)G(s)H(s) = (1 + ατ s)K1 K2 . α(1 + τ s)(Js2 ) We desire ζ = 0.6, Ts ≤ 2.5 or ζωn ≥ 1.6. The uncompensated closedloop system is T (s) = K , s2 + K where K = K1 K2 /J and K = ωn2 . We can select K = 20, and then ζωn > 1.6. First, plot the Bode diagram for G(s)H(s) = 20 s2 where K1 K2 /αJ = 20. The phase margin of the uncompensated system is 0o . We need to add phase at ωc . After several iterations, we choose to add 40o phase at ωc , so sin 40o = α−1 = 0.64 . α+1 Therefore, α = 4.6. Then, 10 log α = 10 log 4.6 = 6.63dB . We determine the frequency where magnitude is -6.63 dB to be ωm = 6.6 rad/sec. Then, √ p = ωn α = 14.1 and z = p/α = 3.07 . The compensated loop transfer function (see Figure P10.1a) is 20 Gc (s)G(s)H(s) = 2 s s 3.07 s 14.1 +1 . +1 (b) Since we desire ζωn ≥ 1.6, we place the compensator zero at z = 1.6. Then, we place the compensator pole far in the left half-plane; in this case, we selected p = 20. Thus, the compensator is Gc (s) = s + 1.6 . s + 20 The root locus is shown in Figure P10.1b. To satisfy the ζ = 0.6 requirement, we find K = 250, and the compensated loop transfer © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 10 The Design of Feedback Control Systems Gain dB 100 50 0 -50 10-1 100 101 102 101 102 Frequency (rad/sec) Phase deg -140 -150 -160 -170 -180 10-1 100 Frequency (rad/sec) FIGURE P10.1 (a) Compensated Bode plot for Gc (s)G(s)H(s) = 20(s/3.07+1) . s2 (s/14.1+1) function is 20 250(s + 1.6) Gc (s)G(s)H(s) = 2 = 2 s (s + 20) s s 1.6 + 1 s 20 + 1 . 20 15 * 10 5 Imag Axis 534 0 x *o x -5 -10 * -15 -20 -25 -20 -15 -10 -5 0 Real Axis FIGURE P10.1 CONTINUED: (b) Root locus for Gc (s)G(s)H(s) = 1 + K s2s+1.6 . (s+20) 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 535 Problems P10.2 The transfer function of the system is G(s) = s3 1.0e + 14 , + 2000s2 + 1e + 11s where we use the system parameters given in P7.11 with the following modifications: τ1 = τ1 = 0 and K1 = 1. Also we have scaled the transfer function so that the time units are seconds. The parameters in P7.11 are given for time in milliseconds. A suitable compensator is Gc (s) = s + 500 . s+1 The closed-loop system response is shown in Figure P10.2. The percent overshoot is P.O. ≈ 20% and the time to settle is Ts < 0.01 second. 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 0.005 0.01 0.015 0.02 0.025 0.03 Time (secs) 0.035 0.04 0.045 0.05 FIGURE P10.2 Step response. P10.3 The loop transfer function is Gc (s)G(s) = 16(s + 1) K(s + z) . s(s2 + 2s + 16) (s + p) We desire dominant roots with Ts < 5 sec and P.O. < 5%, so use ζ = 0.69 and ζωn = 0.8. One solution is to select z = 1.1 (i.e. to the left of the © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 536 CHAPTER 10 The Design of Feedback Control Systems existing zero at s = −1) and determine the pole p and gain K for dominant roots with ζ = 0.69. After iteration, we can select p = 100, so that the root locus has the form shown in Figure P10.3. Then, we select K = 320, 200 150 100 Imag Axis 50 0 -50 -100 -150 -200 -200 -150 -100 -50 0 Real Axis 50 100 150 200 FIGURE P10.3 16(s+1)(s+1.1) Root locus for 1 + K s(s2 +2s+16)(s+100) = 0. so that ζ = 0.69. The final compensator is Gc (s) = 320(s + 1.1) . s + 100 The design specifications are satisfied with this compensator. P10.4 The uncompensated loop transfer function is G(s) = 1 1 s2 ( 40 s + 1) = 40 . + 40) s2 (s We desire 10% < P.O. < 20%, so 0.58 < ζ < 0.65, and Ts < 2 implies ζωn < 2. We will utilize a PD compensator Ka (s + a). We select a = 2, to obtain the root locus shown in Figure P10.4. Then with Ka = 23.5, we have the desired root location, and Gc (s) = 23.5(s + 2) . The design specifications are satisfied with the PD compensator. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 537 Problems 30 + * 20 Imag Axis 10 0 + *o x x -10 -20 + * -30 -50 -40 -30 -20 -10 0 Real Axis FIGURE P10.4 40(s+2) Root locus for 1 + Ka s2 (s+40) = 0. P10.5 We desire P.O. < 10% and Ts < 1.5 sec. The compensator is a PI-type, given by Gc (s) = K2 + K2 s + K3 K2 (s + a) K3 = = s s s where a = K3 /K2 . So, ess = 0 for a step input and G(s) = 3.75Ka 25Ka = . (s + 0.15)(0.15s + 1) (s + 0.15)(s + 6.67) The loop transfer function is Gc (s)G(s) = 25Ka K2 (s + a) . s(s + 0.15)(s + 6.67) Using root locus methods, we select a = 0.2 (after several iterations) and determine Ka K2 to yield ζ = 0.65. This results in Ka K2 = 1. The root locus is shown in Figure P10.5. The design specifications are met. The actual percent overshoot and settling time are P.O. = 7.4% and Ts = 1.3 s. The controller is Gc (s) = 1 + 0.2 . s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 538 CHAPTER 10 The Design of Feedback Control Systems 20 15 10 + * Imag Axis 5 0 x x+ x *o -5 + * -10 -15 -20 -20 -15 -10 -5 0 5 10 15 20 Real Axis FIGURE P10.5 25(s+0.2) Root locus for 1 + Ka K2 s(s+0.15)(s+6.67) = 0. As in P10.5, using root locus we find that placing z = 15 and p = 30 yields a root locus shape (see Figure P10.6) where the loop transfer function is 60 40 20 Imag Axis P10.6 + * 0 x + * o x x + * -20 -40 -60 -60 -40 -20 0 Real Axis FIGURE P10.6 25(s+15) Root locus for 1 + Ka (s+0.15)(s+6.67)(s+30) = 0. 20 40 60 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 539 Problems Gc (s)G(s) = 25Ka (s + z) . (s + p)(s + 0.15)(s + 6.67) and where z, p and Ka are the parameters to be determined. Properly choosing the parameter values allows us to increase ζωn of the dominant roots (compared to the PI compensator of P10.5). Then, with Ka = 3.7, the dominant roots have ζ = 0.65. The design specifications are met with the compensator. P10.7 The plant transfer function is G(s) = e−50s . (40s + 1)2 The steady-state error is ess = A < 0.1A . 1 + Kp Therefore, Kp > 9. Insert an amplifier with the compensator with a dc gain = 9, as follows Gc (s)G(s) = 9e−50s (s + 2) . (40s + 1)2 (s + p) The system is unstable without compensation, and it is very difficult to compensate such a time delay system with a lead compensator. Consider a lag network Gc (s) = s+z s+p where z > p. Let z = 10p. Then, a plot of the P.M. versus p is shown in Figure P10.8a. Suitable system performance can be obtained with P.M. > 45o , so choose p = 0.0001. The Bode plot of the compensated and uncompensated systems is shown in Figure P10.7c, where we have selected z = 0.001 and p = 0.0001. The compensated system has P.M. = 62o and Ts = 9 minutes . The step response is shown in Figure P10.7b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 540 CHAPTER 10 The Design of Feedback Control Systems (a) Phase Margin (deg) 150 100 50 0 -50 0 0.5 1 1.5 p 2.5 3 x10 -3 (b) 1.5 Amplitude 2 1 0.5 0 0 100 200 300 400 500 600 700 800 Time (secs) FIGURE P10.7 (a) Phase margin versus p. (b) Step response with p = 0.0001 and z = 0.001. 20 Gain dB 10 0 -10 -20 -30 10-4 10-3 10-2 10-1 10-2 10-1 Frequency (rad/sec) 0 Phase deg -100 -200 -300 -400 -500 10-4 10-3 Frequency (rad/sec) FIGURE P10.7 CONTINUED: (c) Bode plot for the compensated system (solid lines) and the uncompensated system (dashed line). P10.8 The transfer function is G(s) = 5000 . s(s + 10)2 To meet the steady-state accuracy, we need Kv > 40. The uncompensated © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 541 Problems Kv = 50, so the steady-state accuracy can be met. (a) Using the Bode method, we need P.M. = 70% (to meet P.O. < 5% specification). Let Gc (s) = bs + 1 . as + 1 The plot of P.M. versus b is shown in Figure P10.8a, where we set a = 50b. Choosing b = 20 should satisfy the P.O. specification. The Bode plot is shown in Figure P10.8c. Thus, (a) Phase Margin (deg) 80 70 60 50 40 30 0 5 10 15 b 25 30 (b) 1.5 Amplitude 20 1 0.5 0 0 5 10 15 20 25 30 35 40 45 50 Time (secs) FIGURE P10.8 (a) Phase margin versus b; (b) Step response for lag compensator designed with Bode where a = 1000 and b = 20. Gc (s)G(s) = 5000(20s + 1) . s(s + 10)2 (1000s + 1) The step response is shown in Figure P10.8b. (b) We require that ζ = 0.7 to meet the P.O. specifications. Let Gc (s) = K(bs + 1) . (as + 1) Using root locus methods, we fix a and b, and then determine K for ζ = 0.7. Let a = 50b and select b = 10 (other values will work). The root locus is shown in Figure P10.8d. We find K = 2.5 when ζ = 0.7. Now, Kv = 125, so the steady-state accuracy requirement is satisfied for the step response as shown in Figure P10.8e. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 10 The Design of Feedback Control Systems 150 Gain dB 100 50 0 -50 -100 10-4 10-3 10-2 10-3 10-2 10-1 100 Frequency (rad/sec) 101 102 101 102 Phase deg 0 -100 -200 -300 10-4 10-1 100 Frequency (rad/sec) FIGURE P10.8 CONTINUED: (c) Bode plot for the compensated system with Gc (s) = 20s+1 1000s+1 . 20 15 10 5 Imag Axis 542 * 0 * x *ox * -5 -10 -15 -20 -20 -15 -10 -5 0 5 Real Axis FIGURE P10.8 5000(10s+1) CONTINUED: (d) Root locus for 1 + K s(s+10)2 (500s+1) . 10 15 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 543 Problems 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 20 Time (secs) FIGURE P10.8 CONTINUED: (e) Step response for lag compensator designed with root locus methods, where K = 2.5. We desire a small response for a disturbance at 6 rad/sec. The Bode plot of Gc (s)G(s) is shown in Figure P10.9a where we consider a compensator Gain dB 0 -50 -100 -150 10-1 100 101 Frequency (rad/sec) 102 103 100 101 Frequency (rad/sec) 102 103 0 Phase deg P10.9 -100 -200 -300 10-1 FIGURE P10.9 (a) Bode plot for the compensated system with Gc (s) = 10(s2 +4s+10) . s2 +36 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 544 CHAPTER 10 The Design of Feedback Control Systems of the form Gc (s) = K(s2 + as + b) . s2 + 36 Notice that the magnitude is large at ω = 6, as desired. We select a = 4, b = 10 and K = 10 . The response to a sinusoidal disturbance at 6 rad/sec is shown in Figure P10.9b. Notice that the effect of the disturbance is virtually eliminated in steady-state. 0.02 0.015 0.01 Amplitude 0.005 0 -0.005 -0.01 -0.015 -0.02 0 10 20 30 40 50 60 70 80 90 100 Time (secs) FIGURE P10.9 CONTINUED: (b) Disturbance response for a sinusoidal disturbance at 6 rad/sec. P10.10 The step response with Gc (s) = 1 is shown in Figure P10.10. A suitable lag compensator is Gc (s) = s + 0.05 . s + 0.005 The step response of the compensated system is also shown in Figure P10.10. The settling time of the compensated system is Ts = 28 seconds . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 545 Problems Compensated system (solid) & Uncompensated system (dashed) 30 25 Amplitude 20 15 Input (dotted line) 10 5 0 0 5 10 15 20 25 30 Time (sec) FIGURE P10.10 Step response of uncompensated and compensated systems. The root locus is shown in Figure P10.11 where a suitable lead-lag com- 300 200 100 Imag Axis P10.11 + 0 x + oo +x + -100 -200 -300 -300 -200 -100 0 Real Axis FIGURE P10.11 160(s+17)(s+10) Root locus for 1 + K s2 (s+170)(s+1) = 0. 100 200 300 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 546 CHAPTER 10 The Design of Feedback Control Systems pensator is Gc (s) = K s + 10 s + 17 . s + 1 s + 170 The selected gain is K = 57, so that the damping of the complex roots is about ζ = 0.7. For this particular design, the closed-loop system zeros will affect the system response and the percent overshoot specification may not be satisfied. Some design iteration may be necessary or aprefilter can be utilized. A suitable prefilter is Gp (s) = 17 . s + 17 The acceleration constant is Ka = 9120. We choose K = 10. This yields a velocity constant Kv = 20K = 200, as desired. A suitable two-stage lead compensaator is Gc (s) = (0.05s + 1)(0.05s + 1) . (0.0008s + 1)(0.0008s + 1) The Bode plot is shown in Figure P10.12. The phase margin is P.M. = 75.06o . Phase margin=75.06 deg 100 Gain dB 50 0 -50 -100 10-1 100 101 102 103 104 103 104 Frequency (rad/sec) 0 Phase deg P10.12 -100 -200 -300 10-1 100 101 102 Frequency (rad/sec) FIGURE P10.12 200(0.05s+1)2 Bode plot for s(0.1s+1)(0.05s+1)(0.0008s+1)2 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 547 Problems P10.13 (a) When Gc (s) = K = 0.288 , the phase margin is P.M. = 49.3o and the bandwidth is ωB = 0.95 rad/sec. (b) A suitable lag compensator is Gc (s) = 25s + 1 . 113.6s + 1 The compensated system phase margin is P.M. = 52.21o and Kv = 2, as desired. P10.14 A suitable lead compensator is Gc (s) = 1.155s + 1 . 0.032s + 1 The compensated system phase margin is P.M. = 50o and Kv = 2, as desired. The settling time is Ts = 3.82 seconds. One possible solution is Gc (s) = K (s + 12)(s + 15) , (s + 120)(s + 150) where K = 900. The disturbance response is shown in Figure P10.15. Step Response 0.1 0.08 Amplitude P10.15 0.06 0.04 0.02 0 0 0.1 0.2 0.3 Time (sec ) FIGURE P10.15 Compensated system disturbance response. 0.4 0.5 0.6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 548 CHAPTER 10 The PI controller is given by K(s + b) , s Gc (s) = where K and b are to be determined. To meet the design specifications, we need ζ = 0.6 and ωn = 6.67 rad/sec . The closed-loop transfer function is T (s) = K(s + b) . s2 + Ks + bK Solving for the gains yields K = 2ζωn = 8 and b = ωn2 /K = 5.55. A suitable prefilter is Gp (s) = 5.55 . s + 5.55 The step response, with and without the prefilter, is shown in Figure P10.16. Without prefilter (solid) & with prefilter (dashed) 1.4 1.2 1 Amplitude P10.16 The Design of Feedback Control Systems 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 Time (sec) FIGURE P10.16 Compensated system response with and without a prefilter. 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 549 Problems P10.17 The plant transfer function is G(s) = K . s(s + 10)(s + 50) We desire ζωn > 10 to meet Ts < 0.4 sec and ζ = 0.65 to meet P.O. < 7.5%. Try a pole at s = −120. The root locus is shown in Figure P10.17. The gain K = 6000 for ζ = 0.65. Thus, Gc (s)G(s) = 6000(s/15 + 1) s(s + 10)(s + 50)(s/120 + 1) and Kv = 6000 = 12 . 500 200 150 100 Imag Axis 50 0 *x x * * ox x * -50 -100 -150 -200 -200 -150 -100 -50 0 50 100 150 200 Real Axis FIGURE P10.17 s/15+1 Root locus for 1 + K s(s+10)(s+50)(s/120+1) . P10.18 (a) The loop transfer function is L(s) = K1 e−2T s 0.25s + 1 where T = 1.28. The phase angle is φ = −2.56ω − tan 0.25ω . So, ω = 1.12 rad/sec when φ = −180o . However, the break frequency © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 10 The Design of Feedback Control Systems is 4 rad/sec. Therefore, you cannot achieve P.M. = 30o and have the system be stable for K1 < 1. The steady-state error is ess = A A = 1 + Kp 1 + K1 since K1 = Kp . (b) Set K1 = 20, then Kp = 20 and this yields a 5% steady-state error. Without compensation, the system is now unstable. Let Gc (s) = s/b + 1 s/a + 1 where b = 5 and a = 0.01. Then, the system is stable with P.M. = 63o . The system response is shown in Figure P10.18. 1.2 1 0.8 Amplitude 550 0.6 0.4 0.2 0 -0.2 0 2 4 6 8 10 Time (secs) FIGURE P10.18 Unit step response with Gc (s) = 20(s/5+1) s/0.01+1 . 12 14 16 18 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 551 Problems P10.19 (a) The open-loop transfer function is G(s) = Ke−sT , (s + 1)(s + 3) where T = 0.5 sec. We desire P.O. < 30%, thus ζ > 0.36. We will design for ζ = 0.4, which implies P.M. = 40o . Then φ = − tan−1 ω − tan−1 ω − 0.5ω(57.3o ) . 3 At ωc = 1.75, the phase margin is P.M. = 40o , and solving |G(jω)| = K [(3 − ω 2 )2 1 + (4ω)2 ] 2 =1 at ω = 1.75 yields K = 7. Then ess = 0.3. (b) We want ess < 0.12, so use ess = 0.10 as the goal. Then Gc (s)G(s) = Ke−0.5s (s + 2) , (s + 1)(s + 3)(s + b) and ess = where Kp = 2K 3b . 1 1 + Kp If b = 0.1 then Kp = 6.7K and ess = 1 . 1 + 6.7K So, we need 6.7K = 9, or K = 1.35. We need a lag compensator (i.e. b < 2) to meet ess < 12% and have stability. P10.20 We desire Kv = 20, P.M. = 45o and ωB > 4 rad/sec. Thus, we set K = 20, and G(s) = s s 2 20 +1 s 6 . +1 Then, the Bode plot yields P.M. = −21o uncompensated at ωc = 5.2 rad/sec. The phase lead compensator must add 66 o plus phase lead to account for the shift of the crossover to a higher frequency with the phase lead compensator. Consider Gc (s) = 1 + ατ s 1 + τs 2 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 552 CHAPTER 10 The Design of Feedback Control Systems One solution is to use α = 10 τ = 1/67 . Then Gc (s) = 100(s + 6.7)2 . (s + 67)2 The compensator has two zeros at ω = 6.7, two poles at ω = 67 yielding P.M. = 47o , ωc = 7.3 and ωB = 12 rad/sec. P10.21 We desire Kv = 20, P.M. = 45o and ωB ≥ 2. The lag compensator is Gc (jω) = 1 + jωτ 1 + jωατ where α > 1. From the Bode plot, φ = −135o at ω ∼ = 1.3. So, at ω = 1.3, we need to lower the magnitude by 22 dB to cause ω = 1.3 to be ωc′ , the new crossover frequency. Thus, solving 22 = 20 log α yields α = 14. We select the zero one decade below ωc′ or Therefore, 1 τ = 0.13. 1 0.13 = = 0.0093 . ατ 14 Then, the lag compensator is given by Gc (s) = s 1 + 0.13 s + 0.13 . = s 1 + 0.0093 14(s + 0.0093) The new crossover is ωc′ = 1.3, and ωB = 2.14 rad/sec. P10.22 We desire P.M. = 45o , Kv = 20 and 2 ≤ ωB ≤ 10. The lead-lag compensator is Gc (s) = s 1 + sb 1 + 10a · . s 1 + 10b 1 + as Since ωB ∼ = 1.5ωc , we design for a new crossover frequency ωc′ so that 1.4 < ωc′ < 7 . Try for ωc′ = 4. The phase φ = −190o at ω = 4, so we need to add phase lead of 55o plus phase to account for lag part of network at ωc′ . Use α = 10 and bracket ω = 4 with the lead network. Put the zero at ω = 0.8 = b © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 553 Problems and the pole at ω = 8. For the lag compensator, put the zero at a lower frequency than ωc′ /10. So try a zero at ω = 0.2 = 10a and a pole at ω = 0.02 = a. Then, the lead-lag compensator is s s 1 + 0.8 1 + 0.2 . Gc (s) = s 1 + 8s 1 + 0.02 The compensated Bode plot yields P.M. = 50o ωc′ = 3.5 rad/sec, The steady-state error is 1 1 = = 0.05 . 1 + Kp 1 + K/25 ess = So, we need K/25 ≥ 19 or K ≥ 475. One possible solution is Gc (s) = 4s + 1 12s + 1 and K = 475 . The compensated Bode plot is shown in Figure P10.23. The phase margin is P.M. = 46o . Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 46 deg (at 11.5 rad/sec) Magnitude (dB) 40 20 0 −20 −40 0 Phase (deg) P10.23 and ωB = 6.2 rad/sec . −45 −90 −135 −180 −3 10 FIGURE P10.23 Bode plot for Gc (s)G(s) = −2 10 −1 0 10 10 Frequency (rad/sec) 475(4s+1) . (s+5)2 (12s+1) 1 10 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 554 CHAPTER 10 P10.24 The Design of Feedback Control Systems The arm-rotating dynamics are represented by G(s) = 80 s s2 4900 + s 70 . +1 We desire Kv = 20, and P.O. < 10%. One possible solution is the lead-lag compensator Gc (s) = (s + 50)(s + 0.48) . 4(s + 400)(s + 0.06) With this compensator, we have P.O. = 9.5% P10.25 and Kv = 20 . Neglect the pole of the airgap feedback loop at s = 200. The characteristic equation is 1 + K̄ (s + 20)(s + c) =0, s3 where K K1 + K2 K2 b c= . K1 + K2 K̄ = Choose c = 10 to attain the root locus structure shown in Figure P10.25. The gain K̄ = 38.87 insures the damping ratio of ζ = 0.5. Then, solving for K1 and b yields K1 = K − K2 38.87 and b= 0.1K . 38.87K2 For given values of K and K2 (unspecified in the problem), we can compute K1 and b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 555 Problems 40 30 + 20 Imag Axis 10 0 o o+ x -10 0 -10 -20 + -30 -40 -40 -30 -20 10 20 30 40 Real Axis FIGURE P10.25 (s+20)(s+10) = 0. Root locus for 1 + K̄ s3 P10.26 The loop transfer function is Gc (s)G(s) = 0.15K(10as + 1) , s(s + 1)(5s + 1)(as + 1) where K and a are to be selected to meet the design specifications. Suitable values are K = 6.25 and a = 0.15 . Then, the phase margin is P.M. = 30.79o and the bandwidth is ωB = 0.746 rad/sec. The lead compensator is Gc (s) = 6.25 P10.27 1.5s + 1 . 0.15s + 1 (a) Let Gc (s) = K = 11. Then the phase margin is P.M. = 50o and the performance summary is shown in Table P10.27. (b) Let Gc (s) = K(s + 12) , (s + 20) where K = 32. Then, the phase margin is P.M. = 50o and the performance summary is given in Table P10.27. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 556 CHAPTER 10 compensator The Design of Feedback Control Systems P.M. P.O. Tp Ts Mpω ωB Gc (s) = K = 11 50o 18% 0.34 sec 0.78 sec 1.5 dB 13.9 rad/sec 32(s+12) s+20 50o 18% 0.20 sec 0.47 sec 1.5 dB 26.3 rad/sec TABLE P10.27 Performance Summary. P10.28 The loop transfer function is Gc (s)G(s) = K(as + 1) , s(s + 10)(s + 14)(10as + 1) where K and a are to be selected to meet the design specifications, and we have set α = 10. The root locus is shown in Figure P10.28a. To satisfy 30 20 10 Imag Axis Gc (s) = * 0 * x x *ox * -10 -20 -30 -30 -20 -10 0 Real Axis FIGURE P10.28 1400(s+1) (a) Root locus for 1 + K s(s+10)(s+14)(10s+1) = 0. 10 20 30 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 557 Problems the steady-state tracking error we must select K > 1400 . Suitable values for the lag compensator are K = 4060 and a = 1 . Then, the percent overshoot is P.O. = 31% and the settling time is Ts = 2.34 sec. The lag compensator is Gc (s) = s+1 . 10s + 1 The step response is shown in Figure P10.28b. 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 Time (secs) FIGURE P10.28 CONTINUED: (b) Step response. P10.29 The plant transfer function is G(s) = 10e−0.05s . s2 (s + 10) Gc (s) = 16(s + 0.7) (s + 9) The lead network 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 558 CHAPTER 10 The Design of Feedback Control Systems provides Mpω = 3.4 dB and ωr = 1.39 rad/sec. The step response is shown in Figure P10.29. The overshoot is P.O. = 37% and Ts = 3.5 sec. 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 6 Time (secs) FIGURE P10.29 Unit step response with Gc (s) = P10.30 16(s+0.7) . s+9 The vehicle is represented by G(s) = K K ≈ . s(0.04s + 1)(0.001s + 1) s(0.04s + 1) For a ramp input, we want ess 1 = 0.01 = . A Kv So, let G(s) = 100 . s(0.04s + 1) The uncompensated P.M. = 28o at ωc = 47 rad/sec. We need to add 17o . Case (1) Phase lead compensation: Gc (s) = 1 + 0.021s . 1 + 0.01s The phase margin is P.M. = 45o . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 559 Problems Case (2) Phase lead compensation: Gc (s) = 1 + 0.04s . 1 + 0.005s The phase margin is P.M. = 65o . For Case 1, we have P.O. = 25% , Ts = 0.13 sec and Tp = 0.05 sec . For Case 2, we have P.O. = 4% , P10.31 Ts = 0.04 sec and Tp = 0.03 sec . As in P10.30, the plant is given by G(s) = 100 . s(0.04s + 1) The uncompensated P.M. = 28o . We need P.M. = 50o . The phase lag compensator Gc (s) = 1 + 0.5s 1 + 2.5s results in P.M. = 50o . The P.O. = 21%, Ts = 0.72 sec and Tp = 0.17 sec. P10.32 (a) To obtain Kv = 100, we have Gc (s)G(s) = 43.33(s + 500) . s(s + 0.0325)(s2 + 2.57s + 6667) With K = 43.33, we have P.M. = 1.2o , Mpω = 26 dB , ωr = 1.8 rad/sec and ωB = 3.7 rad/sec . The Bode plot is shown in Figure P10.32. (b) Let Gc (s) = 0.35s + 1 , 0.001s + 1 and K = 43.33 (as before). Then, P.M. = 36o , Mpω = 5.4 dB , ωr = 1.7 rad/sec and ωB = 3.0 rad/sec . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 560 CHAPTER 10 The Design of Feedback Control Systems 100 Gain dB 50 0 -50 -100 -150 10-2 10-1 100 101 102 103 102 103 Frequency (rad/sec) -100 Phase deg -150 -200 -250 -300 -350 10-2 10-1 100 101 Frequency (rad/sec) FIGURE P10.32 Bode plot with Gc (s) = K = 43.33. The step response is shown in Figure P10.33, where Gc (s) = 10(s + 0.71)(s + 0.02) . (s + 0.0017)(s + 10) 1.2 1 0.8 Amplitude P10.33 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 Time (secs) FIGURE P10.33 Step response with the lead-lag compensator Gc (s) = 10(s+0.71)(s+0.02) . (s+0.0017)(s+10) 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 561 Problems Then, Kv = 80 and P.O. = 17%, Ts = 1.8 sec, and ζ = 0.54. The process model is G(s) = s2 (s 1 , + 10) and we consider the lead compensator Gc (s) = K 1 + sατ , 1 + sτ where α = 100, τ = 0.4 and K = 0.5. Then, P.M. = 46.4o . The step response is shown in Figure P10.34. The system performance is P.O. = 22.7% Ts = 5.2 sec Tp = 1.72 sec . 1.4 1.2 1 Amplitude P10.34 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 Time (secs) 6 7 FIGURE P10.34 40s+1 Step response with the lead compensator Gc (s) = 0.5 0.4s+1 . 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 562 CHAPTER 10 P10.35 The Design of Feedback Control Systems The phase margin is shown in Figure P10.35. As the time delay increases, the phase margin decreases. The system is unstable when T > 2.1843 s. 140 120 100 Phase margin (deg) 80 60 40 Stability boundary 20 0 −20 0 0.5 1 1.5 2 Time delay (s) FIGURE P10.35 Step response with Gc (s)G(s) = 2.5 where 0 ≤ T ≤ 2.5. One possible solution is the integral controller Gc (s) = 2/s. The step response is shown in Figure P10.36. The steady-state tracking error to a 1.6 1.4 1.2 1 Amplitude P10.36 2s+0.54 −T s e , s(s+1.76) T=2.1843 s 0.8 0.6 0.4 0.2 0 -0.2 0 1 2 3 4 5 6 Time (secs) FIGURE P10.36 Step response with the integral controller Gc (s) = 2/s. 7 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 563 Problems step input is zero since the system is type-1. The phase margin is P.M. = 32.8◦ and the bandwidth is ωB = 4.3 rad/s . P10.37 One possible solution is Gc (s) = 1600(s + 1) . 25s + 1 The overshoot to a unit step is P.O. = 4.75% and the steady-state error to a step input is ess = 1%. The system bandwidth is ωB = 9.7 rad/sec. P10.38 The lead compensator is Gc (s) = 2.88(s + 2.04) . s + 5.88 The Bode plot is shown in Figure P10.38. The phase margin is P.M. = 30.4o at ωc = 9.95 rad/sec and the bandwidth is ωB = 17.43 rad/sec. 60 Gain dB 40 20 0 -20 -40 10-1 100 101 102 101 102 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-1 100 Frequency (rad/sec) FIGURE P10.38 Bode plot for Gc (s)G(s) = P10.39 115.29(s+2.04) . s(s+2)(s+5.88) The lag compensator is Gc (s) = 1 + 1.48s . 1 + 11.08s The Bode plot is shown in Figure P10.39. The steady-state error specification is satisfied since Kv = 20. The phase margin is P.M. = 28.85o at © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 564 CHAPTER 10 The Design of Feedback Control Systems Gain dB 100 50 0 -50 10-3 10-2 10-1 100 101 100 101 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-3 10-2 10-1 Frequency (rad/sec) FIGURE P10.39 Bode plot for Gc (s)G(s) = 40(1+1.48s) . s(s+2)(1+11.08s) ωc = 2 rad/sec and the bandwidth is ωB = 3.57 rad/sec. P10.40 The lag compensator is Gc (s) = 2.5(1 + 1.64s) . 1 + 30.5s The Bode plot is shown in Figure P10.40. The steady-state error specification is satisfied since Kv = 50 . The phase margin is P.M. = 28.93o at ωc = 1.98 rad/sec and the bandwidth is ωB = 3.59 rad/sec. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 565 Problems Gain dB 100 50 0 -50 10-3 10-2 10-1 100 101 100 101 Frequency (rad/sec) Phase deg 0 -100 -200 -300 10-3 10-2 10-1 Frequency (rad/sec) FIGURE P10.40 Bode plot for Gc (s)G(s) = P10.41 100(1+1.64s) . s(s+2)(1+30.5s) We use Table 10.2 in Dorf & Bishop to determine the required coefficients α = 1.9 and β = 2.2 . Also, ωn Tr = 4.32 implies ωn = 4.32, since we require Tr = 1 second. The characteristic equation is s3 + 8.21s2 + 41.06s + 80.62 = s3 + (1 + p)s2 + (K + p)s + Kz = 0 . Equating coefficients and solving yields p = 7.21 P10.42 K = 33.85 z = 2.38 . From Example 10.4 in Dorf & Bishop, we have the closed-loop transfer function T (s) = (s2 96.5(s + 4) . + 8s + 80)(s + 4.83) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 566 CHAPTER 10 The Design of Feedback Control Systems A suitable prefilter is Gp (s) = 4 . s+4 The step response (with and without the prefilter) is shown in Figure P10.42. With prefilter (solid) & without prefilter (dashed) 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Time (sec) FIGURE P10.42 Step response with and without the prefilter. P10.43 Let K = 100. The Bode plot is shown in Figure P10.43a and the response to a simusoidal noise input with ω = 100 rad/s is shown in Figure P10.43b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 567 Problems Bode Diagram 60 40 Magnitude (dB) 20 0 −20 System: sysg Frequency (rad/sec): 100 Magnitude (dB): −40.1 −40 −60 −80 −1 10 0 10 1 2 10 Frequency (rad/sec) 3 10 10 0.07 0.06 0.05 Amplitude 0.04 0.03 0.02 0.01 0 −0.01 −0.02 0 1 2 3 4 Time (sec) 5 6 7 8 FIGURE P10.43 (a) Bode magnitude plot. (b) Response to a noise input. P10.44 For 0.129 < K ≤ 69.87, the system is unstable. The percent overshoot is shown in Figure P10.44 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 10 The Design of Feedback Control Systems Percent Overshoot 100 50 0 −50 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 85 90 95 100 K 160 Percent Overshoot 568 140 120 100 65 70 75 80 K FIGURE P10.44 Percent overshoot. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 569 Advanced Problems Advanced Problems AP10.1 (a) With Gc (s) = K , the closed-loop transfer function is T (s) = s3 + K . + 4s + K 5s2 When K = 2.05, the characteristic equation is s3 + 5s2 + 4s + 2.05 = 0 with poles at s = −4.1563 and s = −0.4219 ± j0.5615. Therefore ζ = 0.6, and the predicted overshoot is √ P.O. = 100e−π0.6/ 1−0.62 = 9.5% < 13% . The actual overshoot is P.O. = 9.3% and Ts = 8.7 seconds. (b) When Gc (s) = 82.3(s + 1.114) s + 11.46 the closed-loop transfer function is 82.3(s + 1.114) + + 61.3s2 + 128.14s + 91.6822 82.3(s + 1.114) = . (s + 1.196)(s + 12.26)(s + 1.5 ± j2) T (s) = s4 16.46s3 Therefore ζ = 0.6 and the predicted overshoot is P.O. = 9.5% < 13%. The actual overshoot is P.O. = 12% and Ts = 2.5 seconds. AP10.2 The lag network is given by Gc = K(s + a1 ) . s + a2 The closed-loop transfer function is T (s) = K s4 + (5 + a2 )s3 s + a1 . + (4 + 5a2 )s2 + (4a2 + K)s + Ka1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 570 CHAPTER 10 The Design of Feedback Control Systems Computing the steady-state tracking error yields s4 + (5 + a2 )s3 + (4 + 5a2 )s2 + 4a2 s s→0 s5 + (5 + a2 )s4 + (4 + 5a2 )s3 + (4a2 + K)s2 + Ka1 s 4a2 = < 0.125 . a1 K ess = lim If we select K = 2.05 (as in AP10.1), then a1 > 15.61a2 . So, take a2 = a1 /16. The lag compensator can now be written as Gc (s) = 2.05 s + a1 . s + a1 /16 Select a1 = 0.018. Then, the closed-loop transfer function is T (s) = s4 + 5.0011s3 2.05(s + 0.018) . + 4.0056s2 + 2.0545s + 0.0369 The performance results are P.O. = 13% and Ts = 29.6 seconds for a step input, and ess = 0.12 for a ramp input. AP10.3 The plant transfer function is G(s) = 1 s(s + 1)(s + 4) and the PI controller is given by Gc (s) = Kp s + KI . s The closed-loop transfer function is T (s) = s4 + 5s3 Kp s + KI . + 4s2 + Kp s + KI For a unit ramp, the steady-state tracking error is s4 + 5s3 + 4s2 =0. s→0 s5 + 5s4 + 4s3 + Kp s2 + KI s ess = lim Any KI > 0 and Kp > 0 (such that the system is stable) is suitable and will track a ramp with zero steady-state error. Since we want P.O. < 13%, the damping of the dominant roots should be ζ ≈ 0.6. One suitable © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 571 Advanced Problems solution is to place the zero at s = −0.01 and select the PI controller 2.05(s + 0.01) . s Gc (s) = Therefore, Kp = 2.05 and KI = 0.0205. The closed-loop transfer function is T (s) = s4 + 5s3 2.05(s + 0.01) . + 4s2 + 2.05s + 0.0205 The performance results are P.O. = 11.5% and Ts = 9.8 seconds for a step input, and ess = 0 for a unit ramp. AP10.4 The closed-loop transfer function is T (s) = 10K1 . s2 + 10(1 + K1 K2 )s + 10K1 From the performance specifications, we determine that the natural frequency and damping of the dominant poles should be ωn = 5.79 and ζ = 0.69. So, s2 + 8(1 + K1 K2 )s + 8K1 = s2 + 2ζωn s + ωn2 = s2 + 7.99s + 33.52 . Solving for the gains yields K1 = 4.19 and K2 = 0. The closed-loop transfer function is T (s) = s2 33.52 . + 8s + 33.52 The performance results are P.O. = 5% and Ts = 1 second. AP10.5 (a) From the overshoot specification P.O. = 10%. The plant transfer function is G(s) = 1 . s(s + 1)(s + 10) Let Gp = 1. A suitable compensator is Gc = K s + 0.5 . s + 10 Using root locus methods, we determine that K = 45 yields P.O. ≈ 10%. The closed-loop poles are s1,2 = −2.5 ± j5.1, s3 = −15.48, and s4 = −0.45. (b) The closed-loop transfer function is T (s) = s4 + 21s3 450(s + 0.5) . + 120s2 + 550s + 225 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 572 CHAPTER 10 The Design of Feedback Control Systems The step response is shown in Figure AP10.5. The overshoot and settling time are P.O. = 9.5% and Ts = 3.4 seconds. (c) A suitable prefilter is Gp (s) = 0.5 . s + 0.5 The closed-loop transfer function is T (s) = s4 + 21s3 225 . + 120s2 + 550s + 225 The step response is shown in Figure AP10.5. The overshoot and settling time are P.O. = 0% and Ts = 6.85 seconds. 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 1 2 3 4 Time (sec) 5 6 7 8 FIGURE AP10.5 Step response with prefilter (dashed line) and without prefilter (solid line). AP10.6 From Example 10.12 in Dorf & Bishop, we have the relationship ωn Ts = 4.04 . Thereore, minimizing Ts implies maximizing ωn . Using Table 10.2 in Dorf & Bishop, we equate the desired and actual characteristic polynomials q(s) = s3 + 1.9ωn s2 + 2.2ωn2 s + ωn3 = s3 + (1 + p)s2 + (K + p)s + Kz . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 573 Advanced Problems Comparing coefficients yields (1 + p) = 1.9ωn , K + p = 2.2 1+p 1.9 2 Kz = ωn3 . , So, from the first relationship we see that maximizing ωn implies maximizing p. Solving for p while maintaining K < 52 K= 2.2 2 (p + 2p + 1) − p < 52 3.61 we determine that −9.3643 < p < 9.005 . The largest p = 9. Therefore, K = 51.94 and z = 2.81. The step response is shown in Figure AP10.6. The settling time is Ts = 0.77 second. 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Time (secs) FIGURE AP10.6 Step response with minimum settling time. AP10.7 Let Gp = 1. The closed-loop transfer function is T (s) = K(s + 3) . s4 + 38s3 + 296s2 + (K + 448)s + 3K When K = 311, the characteristic equation s4 + 38s3 + 296s2 + 759s + 933 = 0 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 574 CHAPTER 10 The Design of Feedback Control Systems √ has poles at s = −1.619 ± j1.617 (ζ = 1/ 2), s = −6.25, and s = −28.51. (a) When Gp (s) = 1 and K = 311, the closed-loop transfer function is T (s) = 311(s + 3) . s4 + 38s3 + 296s2 + 759s + 933 The step input performance is P.O. = 6.5%, Ts = 2.5 seconds, and Tr = 1.6 seconds. With the prefilter Gp (s) = 3 s+3 and K = 311, the closed-loop transfer function is T (s) = s4 + 38s3 933 . + 296s2 + 759s + 933 In this case, the step response is P.O. = 3.9%, Ts = 2.8 seconds, and Tr = 1.3 seconds. (b) Now, consider the prefilter Gp (s) = 1.8 s + 1.8 and K = 311. The closed-loop transfer function is (s) = s5 + 39.8s4 559.8(s + 3) . + 364.4s3 + 1291.8s2 + 2299.2s + 1679.4 The step input response is P.O. = 0.7%, Ts = 2.14 seconds and Tr = 1.3 seconds. AP10.8 The plant transfer function is G(s) = 250 . s(s + 2)(s + 40)(s + 45) The performance specifications are P.O. < 20%, Tr < 0.5 second, Ts < 1.2 seconds and Kv ≥ 10. A suitable lead compensator is Gc = 1483.7 s + 3.5 . s + 33.75 The closed-loop transfer function is T (s) = 250(1483.7)(s + 35) s(s + 2)(s + 40)(s + 45)(s + 33.75) + 250(1483.7)(s + 3.5) The actual step input performance (see Figure AP10.8) is P.O. = 18%, Ts = 0.88 second, Tr = 0.18 second, and Kv = 10.7. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 575 Advanced Problems 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Time (secs) FIGURE AP10.8 Step response with lead compensator. The frequency response is shown in Figure AP10.9. Bode Diagrams Gm=12.4 dB (Wcg=20.9); Pm=42.0 deg. (Wcp=9.0) 150 100 50 0 Phase (deg); Magnitude (dB) AP10.9 -50 -100 -150 -50 -100 -150 -200 -250 -300 -3 10 10 -2 10 -1 10 0 Frequency (rad/sec) FIGURE AP10.9 Bode plot with Gc (s) = (s+2.5)(s+0.9871) (s+36.54)(s+0.0675) 10 1 10 2 10 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 576 CHAPTER 10 The Design of Feedback Control Systems One lead-lag compensator that satisfies the specifications is Gc (s) = (s + 2.5)(s + 0.9871) . (s + 36.54)(s + 0.0675) The gain and phase margins are Gm = 12.35 dB and P m = 41.8◦ , respectively. The velocity error constant is Kv = 100. Therefore, all specifications are satisfied. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 577 Design Problems Design Problems The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: 26.035 θ(s) = , Va (s) s(s + 33.142) where we neglect the motor inductance Lm and where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). With a PD controller the closed-loop system characteristic equation is s2 + (33.142 + 26.035KD )s + 26.035Kp = 0 . Using Table 10.2 in Dorf and Bishop we determine that for a second-order system with a deadbeat response we have α = 1.82 and ωn Ts = 4.82. Since we desire Ts < 0.25 seconds, we choose ωn = 19.28. Equating the actual characteristic equation with the desired characteristic equation we obtain s2 + ωn αs + ωn2 = s2 + (33.142 + 26.035KD )s + 26.035Kp . Solving for Kp and KD yields the PD controller: Gc (s) = 14.28 + 0.075s . The step response is shown below. The settling time is Ts = 0.24 second. 1 0.9 0.8 0.7 0.6 Amplitude CDP10.1 0.5 0.4 0.3 0.2 0.1 0 0 0.05 0.1 0.15 Time (secs) 0.2 0.25 0.3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 578 CHAPTER 10 The plant is given as G(s) = 20 . s (s + 2) One possible lead compensator is Gclead (s) = 50(s + 1) . s + 20 Similarly, a suitable lag compensator is Gclag (s) = s + 0.1 . s + 0.022 The loop transfer function with the lead-lag compensator is Gc (s)G(s) = 1000(s + 1)(s + 0.1) . s (s + 2) (s + 0.022)(s + 20) The step response and ramp response are shown in Figure DP10.1. The velocity constant is Kv = 50, so the steady-state error specification is satisfied. Step response 1.5 1 0.5 0 0 0.1 0.2 0.3 0.4 0.5 0.6 Time (sec) 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 Time (sec) 0.7 0.8 0.9 1 1 Ramp response DP10.1 The Design of Feedback Control Systems 0.8 0.6 0.4 0.2 0 FIGURE DP10.1 Step response and ramp response. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 579 Design Problems (a) When Gc (s) = K, we require K > 20 to meet the steady-state tracking specification of less than 5%. (b) The system is unstable for K > 20. (c) A single stage lead compensator is Gc1 (s) = 1 + 0.49s . 1 + 0.0035s With this compensator, the bandwidth is ωB = 68.9 rad/sec and the phase margin is P.M. = 28.57o . (d) A two stage lead compensator is Gc2 (s) = (1 + 0.0185s)(1 + 0.49s) . (1 + 0.00263s)(1 + 0.0035s) With the two stage compensator, the bandwidth is ωB = 83.6 rad/sec and the phase margin is P.M. = 56.79o . The step response for the two compensators is shown in Figure DP10.2. Single stage (solid) & two stage (dashed) 1.4 1.2 1 Amplitude DP10.2 0.8 0.6 0.4 0.2 0 0 0.1 0.2 0.3 0.4 0.5 0.6 Time (sec) FIGURE DP10.2 Step response for one- and two-stage lead compensators. 0.7 0.8 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 580 CHAPTER 10 The mast flight system is modeled as 6 . s(s + 1.5)(s + 3.9) G(s) = Consider the proportional controller Gc (s) = K = 0.85 . The system step response is shown in Figure DP10.3. The percent overshoot is P.O. = 15.9%, the rise time is Tr = 3.63 seconds, and the phase margin is P.M. = 52o . Step Response 1.4 System: syscl Peak amplitude: 1.16 Overshoot (%): 15.9 At time (sec): 3.63 1.2 1 Amplitude DP10.3 The Design of Feedback Control Systems 0.8 0.6 0.4 0.2 0 0 2 4 FIGURE DP10.3 Step response for the mast flight system. 6 Time (sec) 8 10 12 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 581 Design Problems DP10.4 One possible compensator is Gc (s) = 5682 s + 12.6 . s + 87.3 The step response is shown in Figure DP10.4. The performance results 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Time (sec) FIGURE DP10.4 Step response for the high speed train system. are P.O. = 4.44% DP10.5 Ts = 0.36 sec Kv = 14.1 . The design specifications are Kv > 200; Ts < 12 ms and percent overshoot P.O. < 10%. The step response is shown in Figure DP10.5. A suitable compensator is Gc (s) = K s + 403 , s + 2336 where K = 1.9476e + 13. Then, P.O. = 9.5% Ts = 10 ms Kv = 560 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 582 CHAPTER 10 The Design of Feedback Control Systems 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 20 Time (ms) FIGURE DP10.5 Step response for the tape transport system. A solution to the problem is the PI controller Gc (s) = 4.21s + 1.2 . s The step response is shown in Figure DP10.6. 1.2 1 0.8 Amplitude DP10.6 0.6 0.4 0.2 0 0 1 2 3 Time (sec) FIGURE DP10.6 Step response for the engine control system. 4 5 6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 583 Design Problems The performance results are P.O. = 8.8% and Ts = 2.14 . The system is a type-1, so the steady-state error for a step input is zero, as desired. The jet aircraft roll angle motion is represented by the transfer function G(s) = 10 . (s + 10)(s2 + 2s + 20) A good controls solution is obtained with a PID controller Gc (s) = 10s2 + 20s + 150 . s The system is type-1, so the steady-state tracking error is zero for a step input. The performance results are P.O. = 9.13% and Ts = 1.56 . Step Response 1.4 1.2 1 Amplitude DP10.7 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 Time (sec) FIGURE DP10.7 Step response for the jet aircraft roll control system. 2.5 3 3.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 584 CHAPTER 10 DP10.8 The Design of Feedback Control Systems One good solution is obtained with the following PI controller 27.35(s + 2) . s Gc (s) = The system is type-1, so the steady-state tracking error is zero for a step input. The step response is shown in Figure DP10.8. Step Response From: U(1) 1.4 1.2 0.8 To: Y(1) Amplitude 1 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 1.2 Time (sec.) FIGURE DP10.8 Step response for the windmill radiometer. DP10.9 Consider the PID controller Gc (s) = Kp + KD s + KI 1.554s2 + 1.08s + 1 = s s and the lead-lag controller Gc (s) = K s+a s+b s+c s+d = 6.04 (s + 10)(s + 2) . (s + 1)(s + 5) Both are stabilizing in the presence of a T = 0.1 second time delay. For the PID controller the phase margin is P.M. = 40o . For the lead-lag controller the phase margin is P.M. = 45o . We find (for these particular designs) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 585 Design Problems that the lead-lag controller is more able to remain stable in the process of increasing time delay. For a time-delay of T = 0.2 seconds, the lead-lag compensator has a phase margin of P.M. = 22o , while the PID controller is unstable. DP10.10 One solution is Gc (s) = 50(s + 0.01) . s+2 The Bode magnitude is shown in Figure DP10.10. You want high gain at Bode Diagram 80 60 System: sys Frequency (rad/sec): 0.101 Magnitude (dB): 26.9 40 Magnitude (dB) 20 0 −20 System: sys Frequency (rad/sec): 10 Magnitude (dB): −26.9 −40 −60 −80 −100 −4 10 −3 10 −2 10 −1 10 Frequency (rad/sec) 0 10 1 10 2 10 FIGURE DP10.10 Step response for the windmill radiometer. low frequency to improve disturbance rejection and decrease sensitivity to plant changes and low gain at high frequency to attenuate measurement noise. DP10.11 One solution is the PD controller Gc (s) = 0008(s + 10) . The step response is shown in Figure DP10.11. The closed-loop transfer © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 10 The Design of Feedback Control Systems function is T (s) = s2 4 , + 3.4s + 4 where we use the prefilter Gp (s) = 4 . 0.36s + 3.6 Step Response 1.4 System: sys_cl Peak amplitude: 1.01 Overshoot (%): 0.637 At time (sec): 2.97 1.2 1 Amplitude 586 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 FIGURE DP10.11 Step response for the polymerase chain reaction system. 3.5 4 4.5 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 587 Computer Problems Computer Problems The m-file script and step response is shown in Figure CP10.1. The phase margin and percent overshoot are P.M. = 50o P.O. ≈ 18% , respectively. nnumc=[110]; denc=[1 0]; sysc = tf(numc,denc); numg=[1]; deng=[1 10]; sysg = tf(numg,deng); syss = series(sysc,sysg); [Gm,Pm]=margin(syss); Pm % sys_cl = feedback(syss,1); [y,t]=step(sys_cl); step(sys_cl); grid S=stepinfo(y,t); PO=S.Overshoot Pm = 49.9158 PO = 17.5724 Step Response 1.4 1.2 1 Amplitude CP10.1 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 Time (sec) 0.8 FIGURE CP10.1 Phase margin and step response for the closed-loop system. 1 1.2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 588 CHAPTER 10 CP10.2 The Design of Feedback Control Systems Using a proportional controller the closed-loop characteristic equation is 1+K s2 24.2 . + 8s + 24.2 A simple m-file script which computes the P.M. as a function of the gain K yields the proportional controller gain K = 6. Checking the phase margin of the system reveals that P.M. ≈ 40◦ , as desired. n=24.2; d=[1 8 24.2]; sys = tf(n,d); K=6; margin(K*sys), grid Bode Diagram Gm = Inf dB (at Inf rad/sec) , Pm = 39.9 deg (at 11.6 rad/sec) Magnitude (dB) 20 0 −20 −40 −60 Phase (deg) −80 0 −45 −90 −135 −180 −1 10 0 10 1 10 Frequency (rad/sec) 2 10 3 10 FIGURE CP10.2 Bode plot with a proportional controller K = 6 in the loop. CP10.3 The uncompensated system is type-1. To realize a zero steady-state error to a ramp input we need to increase the system type by one. One controller that does this is the PI controller: Gc (s) = KP s + KD . s The step response is shown in Figure CP10.3 where it can be seen in the tracking error plot that the settling time is Ts < 5 seconds. The actual settling time is Ts = 3.6 seconds . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 589 Computer Problems KP=20; KD=10; nc=[KP KD]; dc=[1 0]; sysc = tf(nc,dc); n=1; d=[1 2 0]; sys = tf(n,d); sys_o = series(sysc,sys); sys_cl = feedback(sys_o,[1]); t=[0:0.001:10]; sys1 = tf([1],[1 0]); sys_cl1 = series(sys_cl,sys1); subplot(121) y=step(sys_cl1,t); plot(t,y,t,t,'--'), grid xlabel(' Time (sec)'), ylabel('Ramp response') e=y-t'; L=find(abs(e)>0.02); Ts=t(L(length(L))) subplot(122) plot(t,e,[0 10],[0.02 0.02],':',[0 10],[-0.02 -0.02],':') xlabel(' Time (sec)'), ylabel(' Track ing error') grid 10 0.1 9 0.05 8 0 6 Tracking error Ramp response 7 5 4 3 -0.05 -0.1 -0.15 2 -0.2 1 0 0 5 Time (sec) 10 FIGURE CP10.3 Ramp response with a PI controller Gc (s) = CP10.4 -0.25 0 20s+10 s 5 Time (sec) 10 in the loop. From the percent overshoot spec we determine that P.O. < 10% implies ζ > 0.6. So, we target a phase margin P.M. = 100ζ = 60o . The m-file script which generates the uncompensated Bode plot is shown in Figure CP10.4a. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 590 CHAPTER 10 The Design of Feedback Control Systems numg = 100*conv([1 1],[1 0.01]); deng = conv([1 10],conv([1 2 2],[1 0.02 0.0101])); sysg = tf(numg,deng) w=logspace(-1,2,200); [mag,phase,w]=bode(sysg,w); [Gm,Pm,Wcg,Wcp]=margin(mag,phase,w); % Phi=60-Pm Pm Phi=(60-Pm)*pi/180; alpha=(1+sin(Phi))/(1-sin(Phi)) M=-10*log10(alpha)*ones(length(w),1); [mag,phase,w]=bode(sysg,w); for i = 1:length(w), magdB(i) = 20*log10(mag(1,1,i)); end semilogx(w,magdB,w,M), grid xlabel('Frequenc y (rad/sec)'), ylabel('mag [dB]') title('Uncompensated Bode Plot') hold on semilogx([.56072 5.6072 56.072 560.72],[20 0 -20 -40],'--') È Phi = 56.2111 Pm = 3.7889 alpha = 10.8408 Uncompensated Bode Plot 60 40 mag [dB] 20 0 -20 -40 -60 -80 10-1 100 101 102 Frequency (rad/sec) FIGURE CP10.4 (a) Uncompensated Bode plot. We assume that K = 1 and raise the gain at a later step to meet settling time requirement. The uncompensated phase margin is P.M. = 3.7o , so that the lead compensator needs to add φ = 56.2o . The script also calculates α = 10.84. Following the design procedure outlined in Dorf & Bishop, we locate the compensator zero at ω = 2 rad/sec (see dashed line in Figure CP10.4a). Then, p = αz implies p = 21.68. After several iter- © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 591 Computer Problems ations, we converge on K = 4 as a “good” value. The lead compensator is Gc (s) = 4 s+2 . s + 22 The step response is shown in Figure CP10.4b. The compensated Bode is shown in Figure CP10.4c. K=4; numg = 100*conv([1 1],[1 0.01]); deng = conv([1 10],conv([1 2 2],[1 0.02 0.0101])); sysg = tf(numg,deng) numc=K*[1 2]; denc=[1 22]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); t=[0:0.01:5]; f=10*pi/180; [y,t,x]=step(f*sys_cl,t); plot(t,y*180/pi), grid xlabel(' Time (sec)') ylabel('Attitude rate (deg/sec)'), pause w=logspace(-1,2,200); [mag,phase,w]=bode(sys_o,w); [Gm,Pm,Wcg,Wcp]=margin(mag,phase,w); bode(sys_o) title(['Gain Margin = ',num2str(Gm),' Phase Margin = ',num2str(Pm)]) 12 Attitude rate (deg/sec) 10 8 6 4 2 0 0 0.5 1 1.5 2 2.5 Time (sec) FIGURE CP10.4 CONTINUED: (b) Step response. 3 3.5 4 4.5 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 592 CHAPTER 10 The Design of Feedback Control Systems Gain Margin = 14.96 Phase Margin = 60.49 Gain dB 50 0 -50 -100 10-3 10-2 10-1 100 Frequency (rad/sec) 101 102 10-2 10-1 101 102 Phase deg 100 0 -100 -200 -300 10-3 100 Frequency (rad/sec) FIGURE CP10.4 CONTINUED: (c) Bode plot with lead compensator. CP10.5 The closed-loop transfer function is θ(s)/θd (s) = s2 K̄1 + K̄2 s + K̄2 s + K̄1 where K̄1 = K1 /J and K̄2 = K2 /J. A percent overshoot P.O. ≤ 20% requires ζ > 0.45. Select as the initial damping ζ = 0.7 (initial selection) . For a second-order system with ζ = 0.7, we find that ω/ωn ≈ 0.9 when |θ(s)/θd (s)| = 0.7. So, we select ωn = ωB /0.9 as a starting choice. Therefore, since ωB = 10, we have ωn = 11 . The m-file script is shown in Figure CP10.5a. After several iterations, we find a set of “good” values for ζ = 0.8 and ωn = 4.5 (final selection) . The step response and closed-loop Bode plot are shown in Figures CP10.5b and CP10.5c. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 593 Computer Problems % Par t (a) wn=4.5; zeta=0.8; K2=2*zeta*wn; K1=wn^2; % Par t (b) num=[K2 K1]; den=[1 0 0]; sys = tf(num,den); sys_cl = feedback(sys,[1]); f=10*pi/180; % set-up for 10 deg step input t=[0:.05:3]; [y,t,x]=step(f*sys_cl,t); plot(t,y*180/pi), xlabel(' time [sec]'), ylabel(' theta [deg]'), grid, pause % Par t (c) w=logspace(-1,2,400); [mag,phase,w]=bode(sys_cl,w); for i = 1:length(w), magdB(i) = 20*log10(mag(1,1,i)); end semilogx(w,magdB,[w(1) w(length(w))],[-3 -3]), grid xlabel('Frequenc y (rad/sec)') ylabel('Gain dB') FIGURE CP10.5 (a) Script to generate the step response and the closed-loop Bode plot. 12 10 theta [deg] 8 6 4 2 0 0 0.5 1 1.5 time [sec] FIGURE CP10.5 CONTINUED: (b) Step response. 2 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 594 CHAPTER 10 The Design of Feedback Control Systems 5 0 Gain dB -5 -10 -15 -20 -25 10-1 100 101 102 Frequency (rad/sec) FIGURE CP10.5 CONTINUED: (c) Closed-loop Bode plot. CP10.6 The settling time and phase margin specifications require that the dominant closed-loop poles have natural frequency and damping of ζ ≥ 0.45 and ωn ≥ 1.78. The uncompensated roots locus is shown in Figure CP10.6a. 10 + K=10 8 6 x 4 Imag Axis numg=[1 10]; deng=[1 2 20]; sysg = tf(numg,deng); axis([-15,1,-10,10]); rlocus(sysg); hold on % zeta=0.45; wn=1.7778; x=[-10:0.1:-zeta*wn]; y=-(sqr t(1-zeta^2)/zeta)*x; xc=[-10:0.1:-zeta*wn]; c=sqr t(wn^2-xc.^2); plot(x,y,':',x,-y,':',xc,c,':',xc,- c,':') rlocfind(sysg), hold off 2 0 o -2 -4 x -6 -8 + -10 -14 -12 -10 -8 Real Axis FIGURE CP10.6 (a) Uncompensated root locus. -6 -4 -2 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 595 Computer Problems From the final value theorem, we determine that lim = sE(s) ≤ 0.1A s→0 implies A = 0.1A . 1 + GGc (s) Therefore, the compensated Kpcomp ≥ 9. With the compensator Gc (s) = K s+z s+p we find that Kpcomp = Kz Kpuncomp . p But Kpuncomp = 0.5 and (from the uncompensated root locus) a gain of K = 10 results in roots of the characteristic equation in the desired region. Solving for z 1 Kpcomp = ≈2. p K Kpuncomp Select z = 0.5 to minimize changing the root locus. Then, p = 0.25, and the compensator is Gc (s) = 10 s + 0.5 . s + 0.25 The compensated root locus is shown in Figure CP10.6b and the step response is shown in Figure CP10.6c. The phase margin of the compensated FIGURE CP10.6 CONTINUED: (b) Compensated root locus. 10 + 8 6 x 4 Imag Axis numg=[1 10]; deng=[1 2 20]; sysg = tf(numg,deng); numc=[1 0.5]; denc=[1 0.25]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); axis([-15,1,-10,10]); rlocus(sys_o); hold on % zeta=0.45; wn=1.7778; x=[-10:0.1:-zeta*wn]; y=-(sqr t(1-zeta^2)/zeta)*x; xc=[-10:0.1:-zeta*wn]; c=sqr t(wn^2-xc.^2); plot(x,y,':',x,-y,':',xc,c,':',xc,- c,':') rlocfind(sys_o) hold off 2 0 o o+ x -2 -4 x -6 -8 + -10 -14 -12 -10 -8 Real Axis -6 -4 -2 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 596 CHAPTER 10 The Design of Feedback Control Systems system is P.M. = 62.3o and the settling time Ts < 5 seconds. numg=[1 10]; deng=[1 2 20]; sysg = tf(numg,deng); numgc=10*[1 0.5]; dengc=[1 0.25]; sysc = tf(numgc,dengc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); t=[0:0.1:5]; step(sys_cl,t) [mag,phase,w]=bode(sys_o); [gm,pm,w1,w2]=margin(mag,phase,w); pm >> pm = 62.3201 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Time (secs) FIGURE CP10.6 CONTINUED: (c) Step response and phase margin verification. CP10.7 Both design specifications can be satisfied with an integral controller Gc (s) = K1 + K2 10 = . s s The simulation results and m-file script are shown in Figures CP10.7a and b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 597 Computer Problems Unit Step Response Phi dot 1.5 1 0.5 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.7 0.8 0.9 1 Time (sec) Unit Ramp Response Tracking error 0 -0.05 -0.1 -0.15 0 0.1 0.2 0.3 0.4 0.5 0.6 Time (sec) FIGURE CP10.7 (a) Simulation results. K1=0; K2=10; numc=[K1 K2]; denc=[1 0]; sysc = tf(numc,denc); numg=[23]; deng=[1 23]; sysg = tf(numg,deng); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); t=[0:0.01:1]; ys=step(sys_cl,t); subplot(211) plot(t,ys), xlabel(' Time (sec)'), ylabel('Phi dot') title('Unit Step Response'), grid u=t; yr=lsim(sys_cl,u,t); subplot(212) plot(t,yr-u','--') xlabel(' Time (sec)'), ylabel(' Track ing error') title('Unit Ramp Response'), grid FIGURE CP10.7 CONTINUED: (b) M-file design script. CP10.8 From Example 10.3, we have that the loop transfer function is Gc (s)G(s) = 8.1(s + z) , s2 (s + 3.6) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 598 CHAPTER 10 The Design of Feedback Control Systems where z = 1. We want to determine a value of z so the the percent overshoot is reduced from 46% to less than 32%. A valid design is Gc (s)G(s) = 8.1(s + 0.45) . s2 (s + 3.6) The m-file script and step response are shown in Figure CP10.8. The percent overshoot is P.O.=27.7 %. Step Response From: U(1) 1.4 1.2 1 0.8 To: Y(1) Amplitude K1 = 8.1; numg = [K1]; deng = [1 0 0]; sysg = tf(numg,deng); numc = [1 0.45]; denc = [1 3.6]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); step(sys_cl) y=step(sys_cl); po=100*(max(y)-1) 0.6 0.4 0.2 0 0 1.6 3.2 4.8 6.4 8 Time (sec.) FIGURE CP10.8 Response of system with new lead compensator design. CP10.9 From AP10.10, we have the transfer function is Vo (s) Vi (s) 1 + R2 C 2 s = . 1 + R1 C 1 s T (s) = Substituting C1 = 0.1 µF ,C2 = 1 mF , R1 = 10 kΩ, and R2 = 10 Ω yields T (s) = 1 + 0.01s . 1 + 0.001s The frequency response is shown in Figure CP10.9. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 599 Computer Problems Bode Diagrams 20 15 c1=0.0000001; c2=0.001; r1=10000; r2=10; n=[c2*r2 1]; d=[c1*r1 1]; sys=tf(n,d) bode(sys) Phase (deg); Magnitude (dB) 10 5 0 60 50 40 30 20 10 0 1 10 10 2 10 3 10 4 Frequency (rad/sec) FIGURE CP10.9 Op-amp circuit frequency response. CP10.10 The plot of K versus phase margin is shown in Figure CP10.10. The value of K that maximizes the phase margin is K = 4.15. 60 FIGURE CP10.10 Plot of K versus phase margin. 55 50 45 40 P.M. K=[0.1:0.01:10]; T=0.2; [np,dp]=pade(T,6); sysp=tf(np,dp); for i=1:length(K) ng=K(i)*[1 0.2]; dg=[1 6 0 0]; sysg=tf(ng,dg); [gm,pm]=margin(sysg*sysp); PM(i)=pm; end plot(K,PM), grid [P,n]=max(PM); K(n) xlabel('K'), ylabel('P.M.') 35 30 25 20 15 0 1 2 3 4 5 K 6 7 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 1 1 The Design of State Variable Feedback Systems Exercises E11.1 The system is given by ẋ = Ax + Bu u = Kx where A= 0 1 −1 0 B= 1 0 0 1 and K= −k 0 0 −2k . Then, with u = Kx, we have ẋ = −k 1 −1 −2k x . The characteristic equation is det[sI − A] = det s+k −1 1 s + 2k = s2 + 3ks + 2k 2 + 1 = s2 + 2ζωn s + ωn2 = 0 . Solving for k where ωn2 = 2k 2 + 1 and ζ = 1 (critical damping) yields k = 2. 600 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 601 Exercises E11.2 Let u = −k1 x1 − k2 x2 + r . Then, ẋ = 0 1 x+ 9 − k1 −k2 0 1 r , and det(sI − A) = s2 + k2 s + k1 − 9 = 0 . We want ζ = 1, so the desired characteristic equation is pd (s) = (s + co )2 , where co is to be determined to meet Ts = 4 and where k2 = 2co and k1 = c2o + 9. Solving for the state response of x1 (t) to a unit step input we find x1 (t) = 1 − e−co t − co te−co t . When t ≥ Ts = 4 sec we want x1 (t) ≥ 0.98. Solving for co at t = Ts yields co = 1.459, E11.3 k1 = 11.13, and k2 = 2.92 . The controllability matrix is Pc = h B AB i = 0 1 1 −3 , and det Pc 6= 0, therefore the system is controllable. The observability matrix is Po = C CA = 0 2 0 −6 , and det Po = 0; therefore the system is unobservable. E11.4 The controllability matrix is Pc = h B AB i = 0 0 2 −4 , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 602 CHAPTER 11 The Design of State Variable Feedback Systems and the det Pc = 0; therefore the system is uncontrollable. The observability matrix is C Po = CA = 1 0 , −10 0 and det Po = 0; therefore the system is also unobservable. E11.5 The controllability matrix is Pc = h B AB i = 1 −2 −2 , 3 and det Pc = −1 6= 0; therefore the system is controllable. The observability matrix is Po = C CA = 1 0 0 1 , and det Po = 1 6= 0; therefore the system is observable. E11.6 The controllability matrix is Pc = h B AB i = 0 1 1 −2 , and det Pc 6= 0; therefore the system is controllable. The observability matrix is Po = C CA = 1 0 0 1 , and det Po 6= 0; therefore the system is observable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 603 Exercises E11.7 The block diagram is shown in Fig. E11.7. 2 U(s) 12 1 s 1 s + - - - + 2 Y(s) 5 3 FIGURE E11.7 The block diagram for E11.7. E11.8 The block diagram is shown in Fig. E11.8. 10 8 U(s) 4 1 s + - - - + - 1 s 1 3 9 FIGURE E11.8 The block diagram for E11.8. 1 s 2 + ++ Y(s) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 604 CHAPTER 11 E11.9 The Design of State Variable Feedback Systems The controllability matrix is Pc = h B AB i = k1 k1 − k2 , k2 −k1 + k2 and det Pc = −k12 + k22 . So, the condition for complete controllability is k12 6= k22 . E11.10 A matrix differential equation representation is ẋ = 0 1 0 0 −10 −6 −3 y = [−3 4 E11.11 0 0 1 x + 0 u 1 2]x + [0]u . The system is given by ẋ = Ax + Bu y = Cx + Du where A= 0 0 h i , B = , C = 1 1 2 0 , and D = [1] . 0 0 1 0 0 −2 0 −7 The controllability matrix is Pc = h 1 2 B AB A B i 0 = 0 0 1 1 −7 , 1 −7 49 and det Pc = −1 6= 0; therefore the system is controllable. The observability matrix is C Po = CA = CA2 1 2 0 1 0 2 , −4 0 −13 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 605 Exercises and det Po = −29 6= 0; therefore the system is observable. The transfer function is G(s) = s2 6 . + 5s + 6 The response of the system to a unit step is y(t) = 1 − 3e−2t + 2e−3t . The step response is shown in Figure E11.12 1 0.9 0.8 0.7 Step Response E11.12 0.6 0.5 0.4 0.3 0.2 0.1 0 0 FIGURE E11.12 Unit step response. 0.5 1 1.5 2 2.5 Time (s) 3 3.5 4 4.5 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 606 CHAPTER 11 The Design of State Variable Feedback Systems Problems P11.1 Consider the system ẋ = x + u u = −kx . So, ẋ = x − kx = (1 − k)x and x(t) = e(1−k)t x(0) . The system is stable if k > 1. Computing the value of J (assuming k > 1) yields J= Z ∞ e2(1−k)t x2 (0)dt = 0 1 . k−1 Thus, J is minimum when k → ∞. This is not physically realizable. Select k = 35. Then, the value of the performance index J is J= 1 . 34 The system is not stable without feedback. P11.2 (a) The performance index is given J= ∞ Z (x2 + λu2 )dt . 0 The system is ẋ = x + u u = −kx . So, J= Z 0 ∞ 2 2 2 (x + λk x )dt = Z ∞ 2 2 2 (1 + λk )x dt = (1 + λk ) 0 Z ∞ x2 dt . 0 Carrying out the integration (assuming k > 1) yields J = (1 + λk 2 ) 1 . k−1 We want to determine k that minimizes J. Taking the partial of J © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 607 Problems with respect to k and setting the result to zero yields ∂J λk 2 − 2λk − 1 = =0, ∂k (k − 1)2 or λk 2 − 2λk − 1 = 0 . Solving for k yields k =1+ r 1 , λ 1+ where we reject the solution k = 1 − q 1 + λ1 , since we require k > 1. (b) For λ = 2, we determine that k = 2.2 and Jmin = 8.9. P11.3 The system is given by ẋ = 1 0 −1 2 x + 1 1 u u = −k(x1 + x2 ) = −k[1 1]x . Then, with feedback applied, the system is ẋ = (1 − k) −k −(1 + k) (2 − k) x . Solving HT P + PH = −I yields 2p11 (1 − k) − 2p12 (k + 1) = −1 p12 (3 − 2k) − p11 k − p22 (k + 1) = 0 −2kp12 + 2p22 (2 − k) = −1 . Solving for p11 , p12 and p22 yields −(2k 2 − 6k + 7) 4(4k 2 − 8k + 3) 2k 2 − 2k − 1 = 4(4k 2 − 8k + 3) p11 = p12 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 608 CHAPTER 11 The Design of State Variable Feedback Systems p22 = −(2k 2 − 6k + 3) . 4(4k 2 − 8k + 3) The performance index is computed to be J = xT (0)Px(0) = p11 + 2p12 + p22 = 1 , 2k − 1 when x(0) = [1 1]T . So as k → ∞, J → 0. The system is unstable without feedback. The performance index is J = xT (0)Px(0) = p11 − 2p12 + p22 . From Example 11.12 in Dorf and Bishop, we determine that J= 2k 2 + 1 . 2k 2 So, when k → ∞, the performance index J → 1. The plot of J versus k is shown in Figure P11.4. 60 50 40 J P11.4 30 20 10 0 0 1 2 3 FIGURE P11.4 The performance index J versus k. 4 5 K 6 7 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 609 Problems P11.5 The system is given by ẋ = 0 1 0 0 u = −kx . x + 0 0 1 1 u The performance index is J= Z ∞ T T (x x + u u)dt = 0 Z ∞ (1 + k 2 )(xT x)dt . 0 First, we solve HT P + PH = −(1 + k 2 )I , yielding, (1 + k 2 ) 2k 3 k + k2 + k + 1 = 2k 2 3 2k + k 2 + 2k + 1 = . 2k p12 = p22 p11 The performance index is then given by J = p11 + 2p12 + p22 = 2k 4 + 4k 3 + 3k 2 + 4k + 1 . 2k 2 Taking the partial of J with respect to k, setting the result to zero and solving for k yields ∂J 2k 4 + 2k 3 − 2k − 1 = =0 ∂k k3 or 2k 4 + 2k 3 − 2k − 1 = 0 . Solving for k yields k = 0.90. The plot of J versus k is shown in Figure P11.5. The value of the performance index is J = 6.95 when k = 0.90. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 610 CHAPTER 11 The Design of State Variable Feedback Systems 8.8 8.6 8.4 8.2 J 8 7.8 7.6 7.4 7.2 7 6.8 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 k FIGURE P11.5 The performance index J versus k. P11.6 (a) For P11.3, we have J= 1 . 2k − 1 So, as k → ∞, then J → 0. But k = ∞ is not a practical solution, so select k = 10. Then, J = 1/19, and ẋ = −9 −10 −11 −8 x = Ax . The closed-loop system roots are determined by solving det[sI − A] = s2 + 17s − 38 = 0 , which yields s = −19 and s2 = 2. The system is unstable. The original system was unstable, and it remains unstable with the feedback. In general, ẋ = (1 − k) −k −(1 + k) (2 − k) x = Ax and det[sI − A] = s2 + s(2k − 3) + (2 − 4k) = 0. A Routh-Hurwitz analysis reveals that the system is unstable for all k. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 611 Problems (b) For P11.4, we have ẋ = 0 1 −k −k x = Ax , and det[sI − A] = s2 + ks + k = 0 . The performance index was found to be J =1+ 4k + 1 . 2k 2 As k → ∞, we have J → 0. But k = ∞ is not a practical choice for k. Select k = 10. Then, det[sI − A] = s2 + 10s + 10 = (s + 1.13)(s + 8.87) . The closed-loop system is stable. (c) In P11.5, we found that k = 0.90 for Jmin . We are given ẋ = 0 1 −k −k x and det[sI − A] = s2 + ks + k = s2 + 0.9s + 0.9 = (s + 0.45 + j0.835)(s + 0.45 − j0.835) . P11.7 The closed-loop system is ẋ = 0 1 −k1 −k2 x = Hx , and det[sI − H] = s2 + k2 s + k1 = s2 + 2ζωn s + ωn2 = 0 . We desire ωn = 2, so set k1 = 4. With xT (0) = [1, 0], we have J = p11 , and solving HT P + PH = −I © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 612 CHAPTER 11 The Design of State Variable Feedback Systems yields 0 −4 1 −k2 p11 p12 p11 p12 −1 0 0 −1 + p12 p22 p12 p22 = and p11 = 0 1 −4 −k2 , 20 k 2 + 20 k2 + = 2 . 8 8k2 8k2 Select k2 = for Jmin , where Jmin = √ 5 2 . √ 20 Then det[sI − H] = s2 + √ 20s + 4 = 0 , and ωn = 2 and ζ = 1.12. The system is overdamped. P11.8 From Example 11.11 in Dorf and Bishop, we have So, P= k22 +2 2k2 1 2 1 2 1 k2 J = xT (0)Px(0) = . k22 + 2 2k2 when xT (0) = [1 0]. Taking the partial of J with respect to k2 and setting the result to zero yields ∂J k2 + 2 =1− 2 2 =0 . ∂k2 2k2 Solving for the optimum value of k2 yields √ k2 = 2 . P11.9 Let x1 = φ and x2 = ω. We have that ω= dφ . dt © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 613 Problems The state equations are x˙1 = x2 x˙2 = Ku . Select a feedback such that u = −x1 − K1 x2 + r when r(t) is the reference input. Then, ẋ = and 0 1 −K −KK1 x+ 0 K r , det[sI − A] = s2 + K1 Ks + K . so that the overshoot is 4%. Since Ts = 1 = ζω4n , we √ require ζωn = 4 or ωn = 4 2. Then, s2 + 8s + 32 = s2 + K1 Ks + K, or 8 K = 32 and K1 = 32 = 41 . We desire ζ = P11.10 √1 , 2 The system with feedback is given by ẋ = Ax = −10 −25 1 0 x , where x1 (0) = 1, and x2 (0) = −1. The characteristic equation is det[sI − A] = det s + 10 25 −1 s = s(s + 10) + 25 = s2 + 10s + 25 = 0 . The roots are s1,2 = −5. The solution is x(t) = φ11 φ12 φ21 φ22 x(0) = φ11 − φ12 φ21 − φ22 since x1 (0) = 1 and x2 (0) = −1. We compute the elements of the state transition matrix as follows: φ22 (t) = (1 + 5t)e−5t and φ21 (t) = te−5t , therefore x2 (t) = −(1 + 4t)e−5t . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 614 CHAPTER 11 The Design of State Variable Feedback Systems Similarly, φ11 (t) = (1 − 5t)e−5t φ12 = −25e−5t . and Therefore, x1 (t) = (1 + 20t)e−5t . P11.11 Let u = −k1 x1 − k2 x2 + αr where r(t) is the command input. A state variable representation of the plant is −5 −2 h 0 1 ẋ = y= 2 0 i x+ x+ h 0 i 0.5 0 u. u The closed-loop transfer function is T (s) = α . s2 + (k1 /2 + 5)s + 4 + k2 To meet the performance specifications we need ωn = 4.8 and ζ = 0.826. Therefore, the desired characteristic polynomial is q(s) = s2 + 2(0.826)4.8s + 23 = s2 + 8s + 23 . Equating coefficients and solving for k1 and k2 yields k2 = 19 and k1 = 6. Select α = 23 to obtain zero steady-state error to a step input. P11.12 A state variable representation of the dc motor is ẋ = −3 −2 −0.75 0 0 1 0 0 0 x + 0 u 0 0 3 0 0 0 0 2 0 0 0 0 1 0 0 0 0 2 0 y = [0 0 0 0 2.75]x . 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 615 Problems The controllability matrix is 1 −3 0 Pc = 0 0 0 3 4.5 −18 3 −9 13.5 18 −18 9 0 6 −18 0 0 6 0 0 0 12 and the det Pc 6= 0, so the system is controllable. The observability matrix is 0 0 Po = 0 0 33 0 0 0 2.75 0 0 5.5 0 0 5.5 0 0 11 0 0 0 0 0 0 0 and the det Po 6= 0, so the system is observable. P11.13 , To meet the Kv = 35 specification, we need K = 2450. A state variable representation is ẋ = 0 1 0 −70 x + 0 2450 u y = [1 0]x . Let u = −k1 x1 − k2 x2 . Then, the closed-loop characteristic equation is q(s) = s2 + (2450k2 + 70)s + 2450k1 = 0 . The desired characteristic polynomial is s2 + 72.73s + 2644.63 = 0 where we select ζ = 0.707 and ωn = 51.42 to meet the performance specifications. Equating coefficients and solving for the gains yields k1 = © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 616 CHAPTER 11 The Design of State Variable Feedback Systems 1.08 and k2 = 0.0011. P11.14 Let u = −k1 x1 − k2 x2 − k3 r where r(t) is the command input. Then, the closed-loop system in state variable form is ẋ = −10 − k1 −k2 1 0 x+ 1 0 r y = [0 1]x . To meet the performance specifications, we want the closed-loop characteristic polynomial to be q(s) = s2 + 8s + 45.96 = 0 where ζ = 0.59 and ωn = 6.78. The actual characteristic polynomial is det(sI − A) = s2 + (10 + k1 )s + k2 = 0 . Equating coefficients and solving for the gains yields k2 = 45.96 and k1 = −2. Select k3 = k2 = 45.96 to obtain a zero steady-state error to a step input. This results in a settling time of Ts = 0.87 s and a percent overshoot of P.O. = 10%. P11.15 The transfer function is G(s) = C(sI − A)−1 B = 1 . s+1 The system is not controllable and not observable. P11.16 Let u = −Kx . Then, Ackermann’s formula is K = [0, 0, ..., 1]P −1 c q(A) where q(s) is the desired characteristic polynomial, which in this case is q(s) = s2 + 2s + 10 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 617 Problems A state-space representation of the limb motion dynamics is −4 ẋ = 0 1 −1 x + 1 u . 0 The controllability matrix is Pc = [B AB] = 1 −4 0 1 and P−1 c = 1 4 0 1 . Also, we have q(A) = A2 + 2A + 10I = 18 0 −3 9 . Using Ackermann’s formula, we have K = [−3 9] . P11.17 The system is either uncontrollable or unobservable if a = 5 or a = 8. Both of these values correspond to system real poles. So, if a takes on either value, a pole-zero cancellation occurs in the transfer function. P11.18 A matrix differential equation representation is ẋ = 0 1 −1 −2 y = [1 x+ 0 1 u 0]x . Let u(t) = −k1 x1 − k2 x2 . Then, the closed-loop characteristic equation is q(s) = s2 + (2 + k2 )s + 1 + k1 = 0 . We desire the characteristic equation √ s2 + 2 2s + 2 = 0 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 618 CHAPTER 11 The Design of State Variable Feedback Systems Equating coefficients and solving for the gains yields k1 = 1 and k2 = √ 2 2 − 2 = 0.828. P11.19 A state space representation is ẋ = 0 1 0 x+ 3 −2 y = [3 1]x . 1 1 r The controllability matrix is 0 Pc = , 1 −2 and det Pc 6= 0, so the system is controllable. The observability matrix is Po = 3 1 3 1 , and the det Po = 0, so the system is not observable. P11.20 The characteristic equation associated with A is s2 (s2 + 0.2s + 0.0015) = 0 . There are two roots at the origin, so the system is unstable. The system can be stabilized with δ = −k1 x1 − k3 x3 = 20x1 − 10x3 . P11.21 (a) Let x1 = i1 , x2 = i2 and u = v. Then, the state equation is ẋ = −(R1 +R3 ) L1 R3 L2 R3 L1 −(R3 +R2 ) L2 x + Also, y = vo , but y = [R3 − R3 ]x . 1 L1 0 u . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 619 Problems (b) The observability matrix is Po = C CA = R3 3 − RL1 R − R32 1 and det Po = 1 L1 + 1 L2 R2 R3 L2 −R3 + R32 1 L1 + 1 L2 R2 R1 − R32 . L2 L1 So, when R2 R1 = , L1 L2 det Po = 0 and the system is not observable. (c) Let a= R1 + R3 , L1 b= R3 + R2 . L2 and Then det[sI − A] = det " (s + a) 3 −R L2 3 −R L1 (s + b) # R32 = (s + a)(s + b) + = (s + r)2 L1 L2 = s2 + (a + b)s + ab + R32 . L1 L2 The system has two equal roots when R32 (a + b) − 4 ab + L1 L2 2 ! or R1 + R3 R3 + R2 + L1 L2 2 −4 (R1 + R3 )(R3 + R2 ) + R32 =0. L1 L2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 620 CHAPTER 11 (a) Without state feedback the state differential equation is given by ẋ = y= h −0.4 −1 1 0 1 0 i x. x+ 1 0 u The step response is shown in Figure P11.22a. (a) Without state feedback 2 x2 1.5 1 0.5 0 0 2 4 6 8 10 12 14 16 18 20 1.4 1.6 1.8 2 Time (sec) (b) With state feedback 1.5 1 x2 P11.22 The Design of State Variable Feedback Systems 0.5 0 0 0.2 0.4 0.6 0.8 1 1.2 Time (sec) FIGURE P11.22 Step response (a) without state feedback, and (b) with state feedback. (b) Consider state feedback u = −K(ax2 + bx1 ) + cr where r is the reference input and K, a, b and c are to be determined. Then, the state differential equation is −0.4 − Kb −1 − Ka h 0 1 ẋ = y= 1 i 0 x, x+ c 0 r and det(sI − A) = s2 + (0.4 + Kb)s + (1 + Ka) = 0. Our specifications 4 are P.O. = 5% and Ts = 1.35 sec. So, ζ = 0.69 and ωn = ζ1.35 = 4.3. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 621 Problems Solving for K, a and b yields Ka = ωn2 − 1 and Kb = 2ζωn − 0.4 . Select K = 1. Then, a = 17.49 and b = 5.53. Select c = 1 + Ka to achieve a zero steady-state tracking error. (c) The step response is shown in Figure P11.22b for the system with state feedback. P11.23 Using the internal model design method for step inputs, we have 0 1 0 0 e ė = 0 0 1 + 0 z ż 0 0 0 1 where we choose w , w = −K1 e − K2 z . To place the poles at s = −10 and s = −2±j we use Ackermann’s formula to compute K1 = 50 K2 = [45 14] . The compensator has the form shown in Figure 11.14 in Dorf and Bishop. P11.24 Using the internal model design method for ramp inputs, we have 0 1 0 0 e ė 0 0 1 0 ë = ė 0 0 0 1 z ż 0 0 0 0 where we choose 0 0 + w 0 1 w = −K1 e − K2 ė − K3 z . To place the poles at s = −20 and s = −2 ± 2j we can use Ackermann’s formula. We also need an additional pole (must be a stable pole); select © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 622 CHAPTER 11 The Design of State Variable Feedback Systems s = −20 as the fourth pole. Then, K1 = 3200 K2 = 1920 K3 = [568 44] . The compensator has the form shown in Figure 11.16 in Dorf and Bishop. P11.25 The observability matrix is Po = C CA = 1 −4 21 −36 , and det Po = 48 6= 0; therefore the system is completely observable. The desired poles of the observer are s1,2 = −1. This implies that the desired characteristic polynomial is pd (s) = s2 + 2s + 1 . The actual characteristic polynomial is det |λI − (A − LC)| = det λ − 1 + L1 −4 − 4L1 5 + L2 λ − 10 − 4L2 = λ2 + (L1 − 4L2 − 11)λ + 10L1 + 8L2 + 30 = 0 . Solving for L1 and L2 yields L= L1 L2 = −0.25 −3.3125 . Checking we find that det(λI − (A − LC)) = s2 + 2s + 1. The response of the estimation error is shown in Figure P11.25, where e(0) = [ 1 1 ]T . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 623 Problems Response to Initial Conditions 2.5 To: Out(1) 2 1.5 1 Amplitude 0.5 0 1 To: Out(2) 0.5 0 -0.5 ?-1 -1.5 0 1 2 3 4 5 6 Time (sec ) FIGURE P11.25 Estimation error response to an initial condition. P11.26 The observability matrix is C 2 Po = CA = 0 CA 2 −4 2 32 20 0 −4 . 14 The det Po = 728 6= 0, hence the system is observable. The gain matrix 0.14 L= −0.93 0.79 results in the observer poles at s1,2 = −1 ± j and s3 = −5, as desired. P11.27 The observability matrix is Po = C CA = 1 0 1 0 . The det Po = 0, hence the system is not completely observable. So, we cannot find an observer gain matrix that places the observer poles at the desired locations. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 624 CHAPTER 11 P11.28 The Design of State Variable Feedback Systems Selecting K = 16 yields a zero steady-state error to a unit step input. The step response is shown in Figure P11.28. Step Response 1 0.9 0.8 Amplitude 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1 1.5 Time (sec) FIGURE P11.28 Estimation error response to an initial condition. P11.29 The system transfer function is Y (s) = 2 U (s) . s+3 The associated state variable model is ẋ = −3x + 2u y=x. 2 2.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 625 Advanced Problems Advanced Problems AP11.1 The closed loop system in state-space form is given by ẋ1 ẋ = 2 ẋ3 0 1 0 0 −1 2 −2KK1 −2KK2 −4 − 2KK3 x1 0 x + 0 u 2 x3 2K x1 h i y= 1 0 0 x2 . The closed-loop transfer function is T (s) = s3 + (2KK3 + 5)s2 x3 4K . + (4KK2 + 2KK3 + 4)s + 4KK1 Setting the steady-state error to zero, we determine that ess = 1 − T (0) = 1 − 1 . K1 Solving for K1 yields K1 = 0.5 . Choosing K2 = 0.5 and K3 = 1.5 results in a percent overshoot of P.O. = 2.82%. AP11.2 A state variable representation is given by ẋ = Ax + Bu where −3 −1 −1 A= 0 0 4 , 0 1 0 3 B= 0 . 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 626 CHAPTER 11 The Design of State Variable Feedback Systems Let u = −Kx . Then, with K= h 4.00 24.33 39.67 i , b1 the closed-loop system poles are s = −4, −5, and −6. AP11.3 Given A= 0 1 −1 −2 , and B = b2 , we compute the determinant of the controllability matrix as det Pc = det[B AB] = − (b1 + b2 ) . The system is controllable if and only if the determinant is non-zero. So, for the system to be controllable, we require that b2 6= −b1 . AP11.4 Consider the state variable feedback law u = −Kx . Using Ackermann’s formula, we determine that K = [−14.2045 − 17.0455 − 94.0045 − 31.0455] results in the closed-loop system characteristic roots at s = −2±j, s = −5 and s = −5. AP11.5 The closed-loop transfer function for the system is T (s) = s3 + (9 + 2K3 )s2 2Kp . + (26 + 2K2 + 10K3 )s + (26 + 6K2 + 12K3 ) Setting the steady-state error for a step input to zero yields ess = 1 − 2Kp =0. 26 + 6K2 + 12K3 Solving for Kp in terms of K2 and K3 yields Kp = 13 + 3K2 + 12K3 . Now, choosing K2 = 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 627 Advanced Problems K3 = 2 results in the closed-loop characteristic roots at s1 = −4 s2 = −4 s3 = −5 . Also, the prefilter gain is Kp = 52 . AP11.6 (a) A state variable representation is given by A= C= h 0 1 −1 −2 1 0 i . , B= 0 1 , Since the determinant of the controllability matrix det[B AB] 6= 0, the system is controllable. (b) The state variable representation is ẋ = Ax + Bu , or ẋ1 ẋ2 = 0 1 −1 −2 x1 x2 + The determinant of the controllability matrix 1 −1 u . det Pc = det[B AB] = 0 . Therefore, the system is uncontrollable. AP11.7 The closed-loop transfer function is T (s) = s3 + (10 + 60K3 )s2 120 . + (16 + 120(K3 + K2 ))s + 120 The state feedback gains K2 = 0.283 and K3 = 0.15 place the poles at the desired locations. The plot of the roll output for a unit step disturbance is shown in Figure AP11.7. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 628 CHAPTER 11 The Design of State Variable Feedback Systems 0.35 0.3 Amplitude 0.25 0.2 0.15 0.1 0.05 0 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (secs) FIGURE AP11.7 Roll angle response to a step disturbance. AP11.8 The state equations are (using the parameters of P3.36 in Dorf and Bishop) 8 1 [80θ − 50h] = −x1 + x2 50 5 θ̇ = ẋ2 = ω = x3 Km Km Kb Km Ka 353 25000 ω̇ = ẋ3 = ia = − ω+ vi = − x3 + vi . J JRa JRa 30 3 ḣ = ẋ1 = In state variable form we have (without feedback) 8 −1 5 ẋ = 0 0 0 1 0 − 353 30 0 x + 0 0 25000 3 vi . (a) In this case we have vi = −kh + ar = −kx1 + ar, where k and a are the parameters to be determined and r is the reference input. With the feedback of h(t) we have ẋ = −1 8 5 0 0 0 1 353 − 25000 3 k 0 − 30 x+ 0 0 a 25000 3 r . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 629 Advanced Problems Since we only have one parameter to adjust, namely k, we will probably not be able to simultaneously meet both design specifications, In fact with k = 0.00056 we obtain the percent overshoot P.O. = 9.89%. The settling time criterion cannot simultaneously be met—the best that can be obtained is Ts ≈ 7.5 seconds. In this case, we choose a = 0.00056 to make the steady-state value of h(t) = 1. (b) In this case we have vi = −k1 h − k2 θ + ar = −k1 x1 − k2 x2 + ar, where k1 , k2 , and a are the parameters to be determined and r is the reference input. Since we have two parameter to adjust, namely k1 and k2 we will probably be able to simultaneously meet both design specifications. In fact with k1 = 0.00056 and k2 = 0.001 we obtain the percent overshoot P.O. = 4.35%. The settling time criterion is easily met— Ts ≈ 5 seconds. In this case, we choose a = 0.0012 to make the steady-state value of h(t) = 1. AP11.9 (a) The state vector differential equation is 0 1 −2 0 ẋ = 0 0 0 0 u , x + 1 0 0 0 1 0 0 1 0 −1 0 1 where x1 = z, x2 = ż, x3 = y and x4 = ẏ. (b) The characteristic equation is s4 + 3s2 + 1 = (s + j0.618)(s − j0.618)(s + j1.618)(s − j1.618) = 0 . So, the system is oscillatory. (c) Let u = −kx4 . Then characteristic equation is s4 + ks3 + 3s2 + 2ks + 1 = 0 which is stable if k > 0. (d) Rewrite the characteristic equation as 1+ ks(s2 + 2) =0. s4 + 3s2 + 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 630 CHAPTER 11 The Design of State Variable Feedback Systems The root locus is shown in Figure AP11.9. A reasonable solution for k is k = 1.35. 3 2 x o 1 Imag Axis x 0 o x -1 o x -2 -3 -3 -2 -1 0 1 2 3 Real Axis FIGURE AP11.9 s(s2 +2) Root locus for 1 + k s4 +3s2 +1 = 0. AP11.10 The state differential equation is ÿ = ky + αu where k and α depend on the system parameters, such as mass and length. The transfer function is y α = 2 u s −k which is unstable at the top of the arc. Since we can only use ẏ for feedback, we have ẏ sα = 2 . u s −k Let Gc (s) = K1 s + K2 . s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 631 Advanced Problems Then GGc (s) = α(K1 s + K2 ) (s2 − k) and the closed-loop characteristic equation is αK1 s + αK2 + s2 − k = 0 or s2 + αK1 s + αK2 − k = 0 . Select αK2 − k > 0 and αK1 > 0 for stability. AP11.11 The state-space representation of the plant is ẋ = Ax + Bu y = Cx where A= 0 1 −2 −3 , B= 0 1 , and C= h 1 0 i . With the intermediate variables defined as z = ẋ and w = u̇ we have where 1 0 0 ė = 0 0 1 ż 0 −2 −3 0 e + 0 w z 1 e=y−r . To meet the design specifications, we require the closed-loop poles to lie to the left of the line in the complex plane defined by s = −0.8. We choose K2 = [10 3] and Gc (s) = 8 . s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 632 CHAPTER 11 The Design of State Variable Feedback Systems This places the closed-loop poles at s = −2, −2 and −2. The closed-loop transfer function with the internal model controller is T (s) = 8 . s3 + 6s2 + 12s + 8 The step response is shown on Figure AP11.11. 1 0.9 0.8 Amplitude 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 2 4 6 8 10 12 14 Time (secs) FIGURE AP11.11 Internal model controller step response. AP11.12 The state-space representation of the plant is ẋ = Ax + Bu y = Cx where A= 0 1 −2 −3 , B= 0 1 , and With the intermediate variables defined as z = ẍ and w = ü C= h 1 0 i . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 633 Advanced Problems we have 0 1 0 0 ė ë = ż where e = y − r. 0 0 e 1 0 + ė 0 0 0 1 z 0 0 −2 −3 0 0 w 0 1 6 5 Amplitude 4 3 2 1 0 0 1 2 3 4 5 6 Time (secs) FIGURE AP11.12 Internal model controller ramp response. To meet the design specifications, we require the closed-loop poles to lie to the left of the line in the complex plane defined by s = −0.67. We choose w = −[K1 K2 e e K3 ] ė = −[16 32 22 5] ė . z Then, Gc (s) = K1 + K2 s 16 + 32s = . 2 s s2 z © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 634 CHAPTER 11 The Design of State Variable Feedback Systems The closed-loop transfer function with the internal model controller is T (s) = s4 + 8s3 32s + 16 . + 24s2 + 32s + 16 This places the closed-loop poles at s = −2, −2, −2 and −2. The ramp response is shown in Figure AP11.12. AP11.13 The controllability matrix is −5 −3 4 −3 22 44 Pc = 1 18 and the observability matrix is Po = . Computing the determinants yields det Pc = −87 6= 0 and det P0 = 242 6= 0 , hence the system is controllable and observable. The controller gain matrix K= h 3.02 6.11 i places the closed-loop poles at the desired locations. Similarly, the observer gain matrix L= 2.38 −1.16 places the observer poles at the desired locations. AP11.14 The controllability matrix is 0 Pc = 0 0 4 4 −12 4 −12 24 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 635 Advanced Problems and the observability matrix is 2 Po = −16 −9 2 29 41 −4 −15 . 120 Computing the determinants yields det Pc = −64 6= 0 and det P0 = 10870 6= 0 , hence the system is controllable and observable. The controller gain matrix K= h −0.5 1.25 0.5 and the observer gain matrix i 57.43 L= −16.11 −104.43 yields the desired closed-loop system poles and observer poles, respectively. AP11.15 The state-variable representation of the system is ẋ = 0 1 −7 −2 x+ 0 1 y = [ 1 4 ]x + [0]u . u The observability matrix is P0 = 1 4 −28 −7 , and det P0 = 105 6= 0, hence the system is observable. The observer gain matrix L= −7.18 6.29 places the observer poles at s1,2 = −10 ± 10, as desired. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 636 CHAPTER 11 The Design of State Variable Feedback Systems Design Problems CDP11.1 A state variable representation is 0 h 1 0 ẋ = y= 1 0 −33.14 i x 0 x+ 0.827 va where x1 = x and x2 = ẋ. Note that we are neglecting the motor inductance and assuming that the position x(t) is the output. Assume that we have available for feedback the angle θ and angle rate θ̇ (see CDP4.1), so that va = − k1 k2 x1 − x2 + au r r where u(t) is the reference input (that is, the desired position x(t)), the gains k1 and k2 and the scaling parameter a are to be determined. Recall that x = rθ = 0.03175θ . With the feedback in the loop we have ẋ = y= h 0 1 −26.03k1 −33.14 − 26.03k2 1 0 i x x+ 0 0.827a Choosing k1 = 50, k2 = 1 and a = 1574.1 results in P.O. = 1.1% and Ts = 0.11 second . The closed-loop poles are s1,2 = −29.59 ± 20.65j. DP11.1 The governing differential equation is ÿ − 2000y = −20i . In state variable form, the system is described by ẋ = 0 1 2000 0 x + 0 −20 i . u © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 637 Design Problems Consider the state feedback i = −k1 x1 − k2 x2 + βr where r(t) is the reference input and k1 , k2 and β are to be determined. Then, the closed-loop system is ẋ = 0 1 2000 − 20k1 −20k2 x+ 0 −20β r . The characteristic equation is s2 + 20k2 s − 2000 + 20k1 = 0 . For stability, let 20k1 − 2000 > 0. Select k1 = 125. Then, ωn = 22.36 rad/sec, and k2 = 2ζωn . 20 Let ζ = 0.59 to meet 10% overshoot specification. Thus, k2 = 2(0.59)(22.36) = 1.32 . 20 The closed-loop transfer function is T (s) = s2 −20β . + 26.4s + 500 s2 500 . + 26.4s + 500 Choose β = −25 so that T (s) = The feedback law is i = 125x1 + 1.32x2 − 25r . DP11.2 The automobile engine control system (see DP10.8 in Dorf and Bishop) is modeled as G(s) = 2e−sT . (0.21s + 1)(4s + 1) In this case, we will assume the delay is negligible. Therefore, T = 0. A © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 11 The Design of State Variable Feedback Systems state variable representation of the system is ẋ = 0 1 −1.19 −5.01 y = [1 0]x . x + 0 1.19 r Let r(t) = −k1 x1 − k2 x2 + k3 u where u(t) is the command input. Using ITAE methods, our desired characteristic polynomial is q(s) = s2 + 1.4ωn s + ωn2 = 0 . Select ωn = 11.315 to obtain a settling time Ts < 0.5 seconds. The characteristic polynomial of the closed-loop system is s2 + (5.01 + 1.19k2 )s + (1.19 + 1.19k1 ) = 0 . Equating coefficients and solving for the gains yields k1 = 106.59 and k2 = 9.235 . Select k3 = 107.59 to yield a zero steady-state error to a step input. 1.2 1 0.8 Amplitude 638 0.6 0.4 0.2 0 0 0.1 0.2 0.3 0.4 Time (secs) FIGURE DP11.2 The step response of the engine control system. 0.5 0.6 0.7 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 639 Design Problems DP11.3 The compensator is x̂˙ = [A − BK − LC] x̂ + Ly + Mr ũ = −Kx̂ where A − BK − LC = N = 363.64 , K= h −28.7 1 −365.19 −20 344.55 15.82 i , M= , and L = 0 200 , 28.7 165.19 . We selected the desired eigenvalues of A − BK at p = −10 ± 10j and the desired eigenvalues of A − LC at q = −20 ± 10j. For initial conditions we let x(0) = [1 1] and x̂(0) = [0 0]. 1.5 Actual x1 x1 1 Estimated x1 0.5 0 0 0.1 0.2 0.3 0.4 0.5 Time (s) 0.6 0.7 0.8 0.9 1 0.4 0.5 Time (s) 0.6 0.7 0.8 0.9 1 6 x2 4 Estimated x2 2 0 −2 Actual x2 0 0.1 0.2 0.3 FIGURE DP11.3 The step response showing the actual and estimated states. DP11.4 The design specifications are (a) Percent overshoot < 20% (b) Ts < 1.5s, and © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 640 CHAPTER 11 The Design of State Variable Feedback Systems (c) steady-state error less than 20% of the input magnitude. The state differential equation is ẋ = Ax + Bu y = Cx where 1 0 0 0 A= 0 −σ1 −α1 = 0 g −α2 −σ2 0 0 B= n = 6.27 g The transfer function is 9.8 9.8 1 −0.415 −1.43 and C = 0 h −0.0111 , −0.0198 1 0 0 i . ns + nσ2 − α1 g θ(s) = 3 δ(s) s + (σ1 + σ2 )s2 + (σ1 σ2 − α1 α2 )s + α1 g 6.27s + 0.0154 = 3 . s + 0.435s2 − 0.0077 + 0.109 Let u = −K1 x1 − K2 x2 − K3 x3 . Then the closed-loop system matrix is 0 A − BK = −nK1 1 −σ1 − nK2 g − gK1 −α2 − gK2 0 −α1 − nK3 , −σ2 − gK3 where K = [K1 K2 K3 ]. From the design specifications, we have the desired roots at s3 +a2 s2 +a1 s+ao = s3 +36s2 +225s+1350 = (s+30)(s+3+j6)(s+3−j6) = 0 . The actual characteristic equation is s3 + (gK3 + K2 n + σ1 + σ2 )s2 + (−α1 α2 − α1 gK2 + K1 n − α2 nK3 + gK3 σ1 + K2 nσ2 + σ1 σ2 )s + α1 g − α1 gK1 + gK3 n + σ2 nK1 = 0 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 641 Design Problems Comparing coefficients yields 0 n n −α1 g + σ2 n g a2 − σ 1 − σ 2 K1 nσ2 − α1 g −α2 n + gσ1 K2 = a1 + α1 α2 − σ1 σ2 0 gn where K3 a0 − α1 g a2 = 36 a1 = 225 a0 = 1350 . The solution for K is K = [53.11 − 28.64 21.96] . DP11.5 The controllability and observability matrices are Pc = P0 = 0.05 −0.04 0.001 −0.001 1 0 −0.8 0.02 Computing the determinants yields det Pc = −1.002e − 05 6= 0 and , respectively. and Po = 0.02 6= 0 , hence the system is controllable and observable. The feedback gain matrix K = [ 3820 −179620 ] yields the desired closed-loop poles. The observer gain matrix L= 120 180000 yields the desired observer poles. The integrated system is shown in Figure DP11.5. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 642 CHAPTER 11 The Design of State Variable Feedback Systems A= -0.8 0.02 -0.02 0 B= 0.05 0.001 System Model u . x=Ax+Bu y C Observer Control Law -K x ^ x . ^ ^ ~ x=Ax+Bu+Ly ~ ^ y=y-Cx + C= 1 0 C K= 3820 -179620 L= 120 180000 FIGURE DP11.5 Integrated controller and observer. DP11.6 (a) The characteristic equation associated with the system matrix is q(s) = s2 + (12 + K2 )s + (36 + K1 ) = 0 , where we have assumed state feedback of the form u = −K1 x1 − K2 x2 . The deadbeat control characteristic equation is s2 + αωn s + ωn2 = 0 , where α = 1.82 and we use ωn = 9.64 to meet the settling time specification. Then, equating coefficients and solving for the gains yields K1 = 56.93 and K2 = 5.54 . (b) Since the closed-loop poles are located at s1,2 = −8.77 ± 4, we can select the observer poles to be about ten times farther in the left-half plane, or s1,2 = −88, −88 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 643 Design Problems Then the observer gains are L= 164 . 5740 (c) The block diagram is shown in Figure DP11.6. 0 1 -36 -12 A= B= 0 1 System Model u y C Observer Control Law -K x . x=Ax+Bu ^ x . ^ ^ ~ x=Ax+Bu+Ly ~ ^ y=y-Cx + C= 1 0 C K= 56.93 5.54 L= 164 5740 FIGURE DP11.6 Block diagram for integrated controller and observer. DP11.7 The compensator is x̂˙ = [A − LC] x̂ + Ly + Bu u = −Kx̂ where −60 A − LC = −1095 1 0 0 1 , −3750 −5 −10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 644 CHAPTER 11 The Design of State Variable Feedback Systems N = 4000 , 60 h i K = 3998 595 30 , and L = 1095 . 3748 We selected the desired eigenvalues of A − BK at p1,2 = −10 ± 10j, p3 = −20 and the desired eigenvalues of A − LC at q1,2 = −20 ± 10j, q3 = −30. For initial conditions we let x(0) = [1 1 1] and x̂(0) = [0 0 0]. The transfer function from r to y is T (s) = 4000s3 + 2.8e05s2 + 6.8e06s + 6e07 . s6 + 110s5 + 5100s4 + 1.29e05s3 + 1.9e06s2 + 1.58e07s + 6e07 The bandwidth is 11.7 rad/s. 1.5 1 0.5 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 20 10 0 −10 200 100 0 −100 −200 FIGURE DP11.7 The step response showing the actual and estimated states. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 645 Computer Problems Computer Problems CP11.1 The controllability and observablity matrices have nonzero determinants, as shown in Figure CP11.1. Therefore, the system is observable and controllable. >> A=[-6 2 0;4 0 7;-10 1 11]; b=[5;0;1]; c=[ 1 2 1]; d=[0]; sys = ss(A,b,c,d); Co=ctrb(sys); dt_Co=det(Co) Ob=obsv(sys); dt_Ob=det(Ob) dt_Co = -84933 dt_Ob = -3.6030e+03 FIGURE CP11.1 Determining controllability and observability. CP11.2 The system is controllable since the determinant of the controllability matrix is nonzero , as shown in Figure CP11.2. a=[0 1;-6 -5]; b=[0;6]; c=[1 0]; d=[0]; sys_ss = ss(a,b,c,d); Pc=ctrb(sys_ss); dt_Pc=det(Pc) Ob=obsv(sys_ss); dt_Ob=det(Ob) sys_tf=tf(sys_ss) dt_Pc = -36 dt_Ob = 1 Transfer function: 6 ------------s^2 + 5 s + 6 FIGURE CP11.2 M-file script to determine controllability and to compute equivalent transfer function model. CP11.3 The gain matrix (computed as shown in Figure CP11.3) is K = a=[0 1;-1 -2]; b=[1;1]; c=[1 -1]; d=[0]; p=[-1;-2]; K=acker(a,b,p) K= 0.5000 0.5000 FIGURE CP11.3 M-file script to place the closed-loop system poles using state feedback. h i 0.5 0.5 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 646 CHAPTER 11 CP11.4 The Design of State Variable Feedback Systems The constant velocity guided missile is not controllable since the controllablity matrix, Co , has a zero determinant, as shown in Figure CP11.4. Using the tf function (see Figure CP11.4), we determine that the transfer function is G(s) = s5 5s . + 0.5s4 + 0.1s3 Cancelling common terms in the transfer function yields G(s) = s4 5 . + 0.5s3 + 0.1s2 Then, using the ss function, we determine a state-space representation of G(s). As shown in Figure CP11.4, the state-space representation is ẋ = Ax + Bu y = Cx A=[0 1 0 0 0;-0.1 -0.5 0 0 0;0.5 0 0 0 0;0 0 10 0 0;0.5 1 0 0 0]; b=[0;1;0;0;0]; c=[0 0 0 1 0]; d=[0]; sys_ss = ss(A,b,c,d); Transfer function: % Part (a) dt_Co = 5s Co=ctrb(sys_ss); dt_Co=det(Co) 0 ----------------------% Part (b) s^5 + 0.5 s^4 + 0.1 s^3 sys_tf = tf(sys_ss) sys_new = minreal(sys_tf ); sys_new_ss=ss(sys_new) a= % Part (c) x1 x2 x3 x4 Co_new=ctrb(sys_new_ss); dt_Co_new=det(Co_new) x1 -0.50000 -0.10000 0 0 % Part (d) x2 1.00000 0 0 0 evalues=eig(sys_new_ss) x3 0 1.00000 0 0 dt_Co_new = x4 0 0 2.00000 0 32 b= u1 evalues = x1 2.00000 0 x2 0 0 x3 0 -0.2500 + 0.1936i x4 0 -0.2500 - 0.1936i c= x1 x2 x3 x4 y1 0 0 0 1.25000 d= u1 y1 0 Continuous-time system. FIGURE CP11.4 Analysis of the constant velocity guided missile state-space model. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 647 Computer Problems where −0.5 −0.1 0 0 1 0 0 0 A= 0 0 1 0 0 0 2 0 2 0 B= 0 and C = h 0 0 0 1.25 i 0 The reduced system is controllable but not stable, since there are two poles at the origin. Systems that are not controllable have too many states. After eliminating unnecessary states, a controllable system of minimal complexity (i.e. states) is obtained. In this case, the number of states is reduced from five to four. CP11.5 The eigenvalues of A are e1 = −2.0727 e2 = −0.2354 e3,4 = 0.2761 ± 0.2593j The system is unstable since there are two eigenvalues in the right halfplane, see Figure CP11.5. The characteristic polynomial is A = [-0.0389 0.0271 0.0188 -0.4555; 0.0482 -1.0100 0.0019 -4.0208; >> 0.1024 0.3681 -0.7070 1.4200; 0 0 1 0]; evalues = b1 = [0.4422;3.5446;-6.0214;0]; 0.2761 + 0.2593i b2 = [0.1291;-7.5922;4.4900;0]; 0.2761 - 0.2593i % Part (a) -0.2354 evalues = eig(A) -2.0727 %part (b) p= p = poly(A) 1.0000 1.7559 -0.6431 0.0618 0.0700 r = roots(p) % Part (c) dt1 = Co1 = ctrb(A,b1); dt1 = det(Co1) -1.8451e+03 Co2 = ctrb(A,b2); dt2 = det(Co2) r= -2.0727 0.2761 + 0.2593i 0.2761 - 0.2593i -0.2354 dt2 = -90.6354 FIGURE CP11.5 Analysis of the VTOL aircraft model. p(s) = s4 + 1.7559s3 − 0.6431s2 + 0.0618s + 0.0700 . The roots of the characteristic equation are the same as the eigenvalues. Also, the system is controllable from either u1 or u2 . If the aircraft should lose the control of the vertical motion through u1 , then the control u2 can . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 648 CHAPTER 11 The Design of State Variable Feedback Systems be used to control both vertical and horizontal motion, and vice versa. CP11.6 The m-file script to analyze the translunar halo orbit problem is shown in Figure CP11.6. The translunar equilibrium point is not a stable point A=[0 0 0 1 0 0;0 0 0 0 1 0; 0 0 0 0 0 1;7.3809 0 0 0 0 -2.1904 0 -2 0 0; 0 0 -3.1904 0 0 0]; c=[0 1 0 0 0 0];d=[0]; b1=[0;0;0;1;0;0]; b2=[0;0;0;0;1;0]; b3=[0;0;0;0;0;1]; sys_ss_1 = ss(A,b1,c,d); sys_ss_2 = ss(A,b2,c,d); dt1 = sys_ss_3 = ss(A,b3,c,d); 0 % Part (a) evalues=eig(A) dt2 = % Part (b) 0 Cb1=ctrb(sys_ss_1); dt1=det(Cb1) dt3 = % Part (c) 0 Cb2=ctrb(sys_ss_2); dt2=det(Cb2) % Part (d) Cb3=ctrb(sys_ss_3); dt3=det(Cb3) % Part (e) sys_tf = tf(sys_ss_2); sys_tf=minreal(sys_tf ) % Part (f ) sys_ss=ss(sys_tf ); Co=ctrb(sys_ss); dt_Co=det(Co) if dt_Co ~= 0 disp('System is completelly Controllable') else disp('System in uncontrollable') end % Part (g) P = [-1+i; -1-i;-10;-10]; [A,B]=ssdata(sys_ss); K = acker(A,B,P) dt_Co = 2 0; evalues = 2.1587 -2.1587 0 + 1.8626i 0 - 1.8626i 0 + 1.7862i 0 - 1.7862i Transfer function: s^2 - 7.381 ---------------------------s^4 - 1.19 s^2 - 16.17 a= x1 x2 x3 x4 x1 x2 x3 x4 0 0.59525 0 2.02089 2.00000 0 0 0 0 2.00000 0 0 0 0 2.00000 0 x1 x2 x3 x4 u1 1.00000 0 0 0 b= c= y1 x1 0 y1 u1 0 x2 0.50000 x3 0 x4 -0.92261 d= 64 System is completelly Controllable FIGURE CP11.6 Analysis of the translunar satellite halo orbit. as evidenced by the eigenvalues of A in the right half-plane; the system is not completely controllable from any ui individually. The transfer function © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 649 Computer Problems from u2 to η is T (s) = s6 s4 − 4.191s2 − 23.55 . + 2s4 − 19.97s2 − 51.58 A careful analysis reveals that T (s) can be reduced by eliminating common factors. The common factors are s2 + 3.1834. The reduced transfer function is T (s) = s2 − 7.3815 . s4 − 1.1837s2 − 16.2030 Using state feedback u2 = −Kx the gain matrix K which places the desired poles (using Ackermann’s formula) is K= CP11.7 h 22 71.56 60 27.02 i . The m-file script to determine the initial state is shown in Figure CP11.7a. Given three data points at t = 0, 2, 4, we construct the three equations A=[0 1 0;0 0 1;-2 -4 -6]; b=[0;0;0]; c=[1 0 0]; d=[0]; sys=ss(A,b,c,d); % % Part (b) v1=c*expm(0*A); v2=c*expm(2*A); v3=c*expm(4*A); V=[v1;v2;v3]; Vi=inv( V ); n=[1;-0.0256;-0.2522]; x0=Vi*n % % Part (c) t=[0:0.1:4]; u=0.0*t; [y,x]=lsim(sys,u,t,x0'); plot(t,y,[0 2 4],[1;-0.0256;-0.2522],'*'), grid xlabel('Time (sec)'), ylabel('y(t)') title('Data points denoted by *') FIGURE CP11.7 (a) Script to determine the initial state from three observations. y(0) = 1 = Ce0A x0 y(2) = −0.0256 = Ce2A x0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 11 The Design of State Variable Feedback Systems y(4) = −0.2522 = Ce4A x0 or, in matrix form Ce0A Ce2A Ce4A 1 x0 = −0.0256 . −0.2522 The problem is solvable if the matrix 0A Ce Ce2A Ce4A is invertible. In this case, the inverse does exist and the solution is x0 = 1 −1 1.9998 . The simulation is shown in Figure CP11.7b. Data points denoted by * 1* 0.8 0.6 0.4 y(t) 650 0.2 0 * -0.2 * -0.4 0 0.5 1 1.5 2 2.5 3 Time (sec) FIGURE CP11.7 CONTINUED: (b) System simulation using computed initial state. 3.5 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 651 Computer Problems CP11.8 Suppose we are given A= 0 1 −1 0 B= 0 1 and the feedback u = −Kx = −[K1 K2 ]x . Solving HT P + PH = −I for P yields K22 + K12 + 3K1 + 2 2(K1 + 1)K2 1 = 2(K1 + 1) K1 + 2 = 2(K1 + 1)K2 p11 = p12 p22 Then, with xo T = [1, 0] we find that J = xo T Pxo = p11 . Computing the partial of J with respect to K2 yields 1 ∂J 1 K1 + 2 = − ∂K2 2 K1 + 1 K22 . Setting ∂J =0 ∂K2 and solving for K2 , we find that K2 = q (K1 + 2)(K1 + 1) . For a given value of K1 , the value of K2 that minimizes J can be computed via the above equation. With K2 given as above, we can compute J to be J= s K1 + 2 . K1 + 1 A plot of J versus K1 (with K2 equal to the minimizing value) is shown in Figure CP11.8. As K1 increases, the performance index J decreases. However, we see that the rate of decrease slows considerably after K1 > 20. Also, K2 increases as K1 increases. We want to keep both gains as © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 652 CHAPTER 11 The Design of State Variable Feedback Systems small as possible, while still having a small J. A reasonable selection is K1 = 20 and K2 = 21.5 . Performance index J versus K1 1.5 1.4 J 1.3 1.2 1.1 1 0 5 10 15 20 25 K1 30 35 40 45 50 30 35 40 45 50 K2 versus K1 60 K2 40 20 0 0 5 10 15 20 25 K1 FIGURE CP11.8 Performance index as a function of K1 and K2 . CP11.9 In this problem, A = −1 and B = 1. Computing Q yields Q = (1 + λ(−k)2 ) = 1 + λk 2 . Define H = A − Bk = −1 − k . Solving H T P + P H = −Q yields p= 1 + λk 2 . 2(k + 1) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 653 Computer Problems 0.5 0.49 0.48 J/x0^2 0.47 0.46 0.45 0.44 0.43 0.42 0.41 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 6 7 8 9 10 k 2.5 2 k min 1.5 1 0.5 0 0 1 2 3 4 5 lambda FIGURE CP11.9 Plot of J/x20 versus k and the minimizing k versus λ. The performance index is J = x20 p which implies J/x20 = 1 + λk 2 . 2(k + 1) The plot of J/x20 versus k is shown in Figure CP11.9. The minimum value is achieved when k = 0.41. To arrive at this result analytically, take the © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 654 CHAPTER 11 The Design of State Variable Feedback Systems partial of J/x20 with respect to k, set the result to zero and solve for k: ∂J/x20 = 0 when k 2 + 2K − 1/λ = 0 . ∂k p Solving for k yields k = −1 ± 1 + 1/λ. So, when λ = 1, k = 0.41. The plot of kmin versus λ is shown in Figure CP11.9. CP11.10 The m-file is shown in Figure CP11.10. A=[0 1;-18.7 -10.4]; B=[10.1; 24.6]; C=[1 0]; D=[0]; % Controller Gains p=[-2;-2 ]; K=acker(A,B,p) >> K= -0.3081 -0.1337 % Observer Gains q=[-20+4*j;-20-4*j]; L = acker(A',C',q); L=L L= 29.6000 89.4600 FIGURE CP11.10 Using the acker function to compute the controller gains and the observer gains. CP11.11 The m-file is shown in Figure CP11.11(a). The compensator can be repA=[0 1 0;0 0 1;-4.3 -1.7 -6.7]; B=[0;0;0.35]; C=[0 1 0]; D=[0]; % Controller Gains p=[-1.4+1.4*j;-1.4-1.4*j;-2]; K=acker(A,B,p) % Observer Gains q=[-18+5*j;-18-5*j;-20]; L = acker(A',C',q); L=L' >> K= 10.1143 22.3429 -5.4286 L= 1.0e+003 * % Simulation of closed-loop system with the observer Ac=[A -B*K;L*C A-B*K-L*C]; Bc=[zeros(6,1)]; Cc=eye(6); Dc=zeros(6,1); sys=ss(Ac,Bc,Cc,Dc); x0=[1;0;0;0.5;0.1;0.1]; t=[0:0.001:3.5]; [y,t]=initial(sys,x0,t); subplot(311) plot(t,y(:,1),t,y(:,4),'--'), grid subplot(312) plot(t,y(:,2),t,y(:,5),'--'), grid subplot(313) plot(t,y(:,3),t,y(:,6),'--'), grid FIGURE CP11.11 (a) M-file. -1.6223 0.0493 0.7370 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 655 Computer Problems resented as x̂˙ = (A − BK − LC)x̂ + Ly and u = −Kx̂ . Since y = Cx, we can write x̂˙ = (A − BK − LC)x̂ + LCx . Similarly, with ẋ = Ax + Bu and u = −Kx̂ we obtain ẋ = Ax − BKx̂ . In matrix form, we have ẋ x̂˙ = −BK A LC A − BK − LC x x̂ , with initial conditions h x(0)T x̂(0)T iT = h 1 0 0 0.5 0.1 0.1 iT . The response of the system is shown in Figure CP11.11(b). 5 Estimated state (dashed line) x1 0 True state 5 0 (solid line) 1 1 2 3 4 0 1 2 3 4 0 1 3 4 x2 0 1 2 x3 0 2 2 Time (sec) FIGURE CP11.11 CONTINUED: (b) Response of system to an initial condition. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 656 CHAPTER 11 CP11.12 The Design of State Variable Feedback Systems The Simulink block diagram is shown in Figure CP11.12. FIGURE CP11.12 Simulink block diagram. CP11.13 The m-file to design the compensator is shown in Figure CP11.13(a). The Simulink simulation is shown in Figure CP11.13(b). The output shown on the x-y graph depicts the state x of the system. The initial conditions selected for the simulation are 1 0 x(0) = 0 0 0.5 0.1 and x̂(0) = . 0.1 0.1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 657 Computer Problems A=[0 1 0 0;0 0 1 0;0 0 0 1;-2 -5 -1 -13]; B=[0;0;0;1]; C=[1 0 0 0]; D=[0]; >> K= % Controller Gains p=[-1.4+1.4*j;-1.4-1.4*j;-2+j;-2-j]; K=acker(A,B,p) 17.6000 24.6800 19.1200 -6.2000 % Observer Gains q=[-18+5*j;-18-5*j;-20;-20]; L = acker(A',C',q); L=L' L= % Simulation of closed-loop system with the observer Ac=[A -B*K;L*C A-B*K-L*C]; Bc=[zeros(8,1)]; Cc=eye(8); Dc=zeros(8,1); sys=ss(Ac,Bc,Cc,Dc); x0=[1;0;0;0;0.5;0.1;0.1;0.1]; t=[0:0.001:10]; [y,t]=initial(sys,x0,t); subplot(311) 100 plot(t,y(:,1),t,y(:,4),'--'), grid subplot(312) 0 plot(t,y(:,2),t,y(:,5),'--'), grid subplot(313) ?100 0 2 plot(t,y(:,3),t,y(:,6),'--'), grid 2 63 1369 10495 1479 4 6 8 10 0 ?2 10 0 2 4 6 8 10 0 2 4 6 8 10 0 ?10 FIGURE CP11.13 (a) M-file to design the compensator, including the observer. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 658 CHAPTER 11 The Design of State Variable Feedback Systems . ^x=[A-BK-LC]x+Ly ^ ^ u=-Kx FIGURE CP11.13 CONTINUED (b) The Simulink simulation. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 1 2 Robust Control Systems Exercises E12.1 The plant transfer function is G(s) = 3 . s+3 Try a PI controller, given by Gc = K1 + K2 . s The ITAE characteristic equation is s2 + 1.4ωn s + ωn2 , where ωn = 30. Then K1 = 13 and K2 = 300 . Without a prefilter, the closed-loop system is Y (s) 39s + 900 = 2 , R(s) s + 42s + 900 and with a prefilter, the closed-loop system is Y (s) 900 = 2 , R(s) s + 42s + 900 where Gp (s) = 23.07 . s + 23.07 The step response, with and without the prefilter, is shown in Figure E12.1. 659 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 660 CHAPTER 12 Robust Control Systems 1.4 Without prefilter With prefilter 1.2 1 y(t) 0.8 0.6 0.4 0.2 0 0 0.05 0.1 0.15 0.2 0.25 Time (sec) 0.3 0.35 0.4 0.45 0.5 FIGURE E12.1 Step response: (a) w/o prefilter (solid line), and (b) w/prefilter (dashed line). The disturbance response is shown in Figure E12.2. 0.05 0.04 0.03 y(t) E12.2 0.02 0.01 0 −0.01 0 0.05 0.1 0.15 0.2 FIGURE E12.2 Disturbance response for system in E12.1. 0.25 Time (sec) 0.3 0.35 0.4 0.45 0.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 661 Exercises E12.3 The closed-loop transfer function is T (s) = s2 25 , + bs + 25 and the sensitivity function is S(s) = s2 + bs , s2 + bs + 25 where b = 8, nominally. The sensitivity of T to changes in b is determined to be SbT = ∂T b −bs = 2 . ∂b T s + bs + 25 The plot of T (s) and S(s) is shown in Figure E12.3, where b = 8. 10 0 Gain dB −10 20log|T| −20 20log|S| −30 −40 −50 −60 −1 10 0 1 10 10 Frequency (rad/sec) FIGURE E12.3 Plot of T (s) and the sensitivity function S(s). E12.4 The plant transfer function is G(s) = 1 , (s + 20)(s + 36) 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 662 CHAPTER 12 Robust Control Systems and the PID controller is given by K3 (s + a)(s + b) . s Gc (s) = Let a=20, b=500, and K3 = 200. Then, the closed-loop system is T (s) = 200s2 + 4000s + 100000 . s3 + 256s2 + 4720s + 100000 The closed-loop poles are s1 =-237.93 and s2,3 = −9.04 ± j18.5 and the zeros are s1,2 = −10±j20. Therefore, there is an approximate cancellation of the complex poles and zeros and the approximate system is T̂ (s) = 238 . s + 238 The actual response and approximation are shown in Figure E12.4. 1.4 1.2 actual 1 approximation y(t) 0.8 0.6 0.4 0.2 0 0 0.05 0.1 0.15 0.2 0.25 0.3 Time (sec) 0.35 0.4 0.45 FIGURE E12.4 Step response for closed-loop actual and approximate transfer functions. E12.5 The loop transfer function is L(s) = Gc (s)G(s) = 10KD (s + KP /KD ) . s(s + 3)(s + 10) 0.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 663 Exercises Select KP /KD = 10. Then L(s) = Gc (s)G(s) = 10KD , s(s + 3) and the closed-loop transfer function is T (s) = 10KD . s2 + 3s + 10KD Let ζ = 0.69, which implies P.O. < 5%. Also, 2ζωn = 3, so ωn = 2.17. Thus, 10KD = ωn2 = 4.72 . Thus, the controller is Gc (s) = 0.47(s + 10). The settling time is Ts = 2.8 s and the percent overshoot is P.O. = 4.6%. As K increases, the percent overshoot increases from 0% to 16% and the settling time generally decreases from 3.8 sec to 2.6 sec. E12.6 The loop transfer function with the PID controller is Gc (s)Gs(s) = KD s2 + KP s + KI 1 . s (s + 5)2 The ITAE step response requires s3 + 1.75ωn s2 + 2.15ωn2 s + ωn3 = s3 + (10 + KD )s2 + (25 + KP )s + KI . For n = 3 we estimate the normalized settling time to be ωn Ts ≈ 8 seconds. Thus, ωn ≈ 6, and KD = 0.5, KP = 52.4, and KI = 216. The step response is shown in Figure E12.6. The transfer function from the disturbance to the output is Y (s) G(s) s = = 3 . 2 Td (s) 1 + Gc (s)G(s) s + 10.5s + 77.4s + 216 The disturbance response is shown in Figure E12.6. The system is effective in reducing the effects of the disturbance, and the maximum output is reduced by 1/100 for a step disturbance. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 664 CHAPTER 12 Robust Control Systems −3 (a) Step response 1.2 10 1 8 0.8 6 x 10 (b) Disturbance response y(t) 12 y(t) 1.4 0.6 4 0.4 2 0.2 0 0 0 0.5 1 TIme (s) 1.5 2 −2 0 0.5 1 TIme (s) 1.5 2 FIGURE E12.6 (a) Step response: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) disturbance response. E12.7 The plant transfer function is G(s) = 1 , (s + 4)2 and the PID controller is Gc (s) = K1 s + K2 + K3 s2 . s Using the ITAE criteria and selecting ωn = 10 yields K3 = 9.5 K2 = 1000 and K1 = 199 . The step response is shown in Figure E12.7. The disturbance response is also shown in Figure E12.7. The maximum y(t) = 0.0041, so the system is effective in rejecting the step disturbance. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 665 Exercises x10 -3 (a) step response 1.4 (b) disturbance 4.5 4 1.2 3.5 1 3 2.5 y(t) y(t) 0.8 0.6 2 1.5 1 0.4 0.5 0.2 0 0 0 0.5 1 1.5 -0.5 2 0 0.5 Time (sec) 1 1.5 2 Time (sec) FIGURE E12.7 (a) Step response: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) disturbance response. The maximum ωn = 60. Then K1 = 3600 and K2 = 80. The maximum control input is max |u(t)| ≈ 80. The plot of the step response and the control input u(t) is shown in Figure E12.8. (a) step response (b) control input u(t) 1.2 90 80 1 70 60 0.8 u(t) 50 y(t) E12.8 0.6 40 30 0.4 20 10 0.2 0 0 0 0.05 0.1 0.15 0.2 -10 0 Time (sec) FIGURE E12.8 Step response w/o prefilter; and (b) control input u(t). 0.05 0.1 Time (sec) 0.15 0.2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 666 CHAPTER 12 E12.9 Robust Control Systems One possible PD controller is Gc (s) = 27.6s + 8.25s . When K=1, the system roots are s1,2 = −3.2 ± j4.3 s3 = −9.5 . The step response is shown in Figure E12.9 for K = 0.5, 1, and 1.5. K=1 (solid); K=0.5 (dashed); and K=1.5 (dotted) 1.4 1.2 1 y(t) 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 1.2 Time (sec) 1.4 1.6 1.8 2 FIGURE E12.9 Step response for K = 0.5, 1, and 1.5. E12.10 One possible PI controller is Gc (s) = 2.2s + 22 . s When K = 1, the system roots are s1,2 = −1.31 ± j1.31, and s3 = −6.37. The step response is shown in Figure E12.10. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 667 Exercises K=1 (solid); K=0.5 (dashed); and K=1.5 (dotted) 1.4 1.2 1 y(t) 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 Time (sec) 3.5 4 4.5 5 FIGURE E12.10 Step response for K = 0.5, 1, and 1.5. The plot is shown in Figure E12.11. 100 90 80 70 60 P.O. (%) E12.11 50 40 30 20 10 0 0 0.5 1 1.5 2 Time (sec) 2.5 3 FIGURE E12.11 Percent overshoot as a function of k in the interval 0.1 ≤ k ≤ 4. 3.5 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 668 CHAPTER 12 The controllability matrix is Pc = and c1 c2 c2 −ac1 − bc2 det Pc = c22 + [bc1 ]c2 + ac21 . For controllability we require det Pc 6= 0, hence c22 + [bc1 ]c2 + ac21 6= 0 implies c2 b q 6= − ± (b/2)2 − a c1 2 where (b/2)2 − a ≥ 0. For real-valued c1 and c2 , if (b/2)2 − a < 0, all real values of c1 and c2 are valid. Valid values of the constants are c1 = 0, c2 = 10, a = 10, and b = 3. The step response is shown in Figure E12.12. Step Response 1.4 1.2 1 Amplitude E12.12 Robust Control Systems 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 Time (sec) 2.5 FIGURE E12.12 Step response with c1 = 0, c2 = 10, a = 10, and b = 3. 3 3.5 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 669 Problems Problems The closed-loop transfer function is T (s) = 4(s + 2) s2 + 4s + 8 and the sensitivity function is S(s) = s2 . s2 + 4s + 8 The plot of 20 log |T | and 20 log |S| is shown in Figure P12.1. The bandwidth is ωB = 6.31 rad/sec . Then T |SK |ωB = 0.98 T ω |SK | B = 0.78 2 T ω |SK | B 4 = 0.30 . 10 0 -10 20log|T| 20log|S| -20 Gain dB P12.1 -30 -40 -50 -60 -1 10 0 10 Frequency (rad/sec) FIGURE P12.1 Plot of T (s) and the sensitivity function S(s). 10 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 670 CHAPTER 12 (a) The loop transfer function is given by Gc (s)G(s) = K . s(0.02s + 1)(0.002s + 1) When K = 100 , the peak magnitude is Mpω = 1.84 . (b) The plot of 20 log |T | and 20 log |S| is shown in Figure P12.2a. 20 20log|S| 0 -20 Gain dB P12.2 Robust Control Systems 20log|T| -40 -60 -80 -100 -120 101 102 103 Frequency (rad/sec) FIGURE P12.2 (a) Plot of T (s) and the sensitivity function S(s). (c) The bandwidth is ωB = 117 rad/sec , and T |SK |ωB = 1.47 T ω |SK | B = 0.39 4 T ω |SK | B = 1.62 . 2 104 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 671 Problems (c) The disturbance response is shown in Figure P12.2b. x10 -7 8 7 6 Amplitude 5 4 3 2 1 0 0 0.05 0.1 0.15 0.2 0.25 0.3 Time (secs) FIGURE P12.2 CONTINUED: (b) Disturbance response for K = 100. P12.3 (a) The loop transfer function is L(s) = Gc (s)G(s) = K(s − 4)(s − 1) . (s + 0.02)(s + 2)2 The characteristic equation is 1 + Gc (s)G(s) = 1 + K (s − 4)(s − 1) =0 (s + 0.02)(s + 2)2 or s3 + (4.02 + K)s2 + (4.08 − 5K)s + 0.08 + 4K = 0 . Using Routh-Hurwitz we find that the system is stable for −4.6987 < K < 0.6947 . (b) The steady-state error is ess = 1 . 1 + 50K Select K = 0.18 to obtain a steady-state error to a unit step of 0.1. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 672 CHAPTER 12 Robust Control Systems (c,d) The plots of y(t) for K = 0.18 K = 0.21 K = 0.15 (nominal) (+15%) (−15%) are shown in Figure P12.3. K=0.18 (solid) & K=0.21 (dashed) & K=0.15 (dotted) 1.2 1 0.8 y(t) 0.6 0.4 0.2 0 −0.2 0 5 10 15 20 25 Time (sec) FIGURE P12.3 Step input response for K = 0.18, K = 0.21 and K = 0.15. P12.4 (a) The plant is given by G= s 1 . +1 s 25 We desire P.O. < 10% and Ts < 100 ms. Using a PD controller Gc (s) = 100 + 2.2s , we determine that P.O. = 7%, Ts < 100 ms and ess = input. The plot of y(t) is shown in Figure P12.4. A 100 for a ramp © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 673 Problems (b) The sensitivity is r |SK | = 27.95 1 when K1 = 1. (c) The plot of y(t) when K1 = 2 (the compensator Gc (s) is unchanged) is shown in Figure P12.4. (d) The disturbance response is shown in Figure P12.4. (b) disturbance 0.012 1 0.01 0.8 0.008 y(t) y(t) (a) step response 1.2 0.6 0.006 0.4 0.004 0.2 0.002 0 0 0.05 0.1 0.15 0 0 0.2 0.05 Time (sec) 0.1 0.15 0.2 Time (sec) FIGURE P12.4 (a) Step response: K1 = 1 (solid line) and K1 = 2 (dashed line); and (b) disturbance response. P12.5 (a) The plant is given by G(s) = 1 s(s + p) where p = 2, nominally. One solution is Gc (s) = 18.7(s + 2.9) . (s + 5.4) Then, T (s) = 18.7(s + 2.9) √ √ . (s + 3.41)(s + 2 + 2 3j)(s + 2 − 2 3j) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 674 CHAPTER 12 Robust Control Systems (b,d) The step responses are shown in Figure P12.5 for p = 2 and p = 1. (c,d) The disturbance responses are shown in Figure P12.5 for p = 2 and p = 1. (a) step response (b) disturbance 1.6 0 1.4 -0.02 1.2 -0.04 1 y(t) y(t) -0.06 0.8 -0.08 0.6 -0.1 0.4 -0.12 0.2 0 0 -0.14 5 0 Time (sec) 5 Time (sec) FIGURE P12.5 (a) Step response: p = 2 (solid line) and p = 1 (dashed line); and (b) disturbance response: p = 2 (solid line) and p = 1 (dashed line). P12.6 (a) The plant is given by G(s) = s(s2 1 , + 4s + 5) and the PID controller is Gc (s) = K(s + z)2 . s When z = 1.25 and K=4, all roots are s = −1 ± j1.22 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 675 Problems Then, the closed-loop transfer function is T (s) = 4(s + 1.25)2 . s4 + 4s3 + 9s2 + 10s + 6.25 (b,c) The step responses with and without a prefilter are shown in Figure P12.6. (d) The disturbance response is shown in Figure P12.6. (a) step response (b) disturbance 1.6 0.02 1.4 0 1.2 -0.02 y(t) y(t) 1 0.8 -0.04 0.6 -0.06 0.4 -0.08 0.2 0 0 5 10 -0.1 0 Time (sec) 5 10 Time (sec) FIGURE P12.6 (a) Step response: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) disturbance response. P12.7 (a) The loop transfer function is Gc (s)G(s) = 10Ka (5s + 500 + 0.0475s2 ) . s3 When Ka = 374.5 , the phase margin is P.M. = 40o . (b) The root locus is shown in Figure P12.7a. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 676 CHAPTER 12 Robust Control Systems 150 * 100 o Imag Axis 50 0 x * -50 o -100 * -150 -150 -100 -50 0 50 100 150 Real Axis FIGURE P12.7 10(0.0475s2 +5s+500) = 0. (a) Root locus for 1 + Ka s3 When Ka = 374.5 , the roots are s1 = −139.8 s2,3 = −19.1 ± j114.2 . (c) The transfer function from Td (s) to Y (s) is Y (s) −s = 3 . 2 Td (s) s + 182s + 19150s + 1915000 The maximum is max |y(t)| = 0.0000389 . (d) The step responses, with and without a prefilter, are shown in Figure P12.7b. P12.8 The polynomial under investigation is s3 + 3s2 + 3s + 4 = 0 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 677 Problems 1.6 1.4 1.2 y(t) 1 0.8 0.6 0.4 0.2 0 0 0.05 0.1 0.15 0.2 0.25 0.3 Time (sec) FIGURE P12.7 CONTINUED: (b) Step response: w/o prefilter (solid line) and w/prefilter (dashed line). From the uncertainty bounds on the coefficients, we define α0 = 4 α1 = 1 α2 = 2 β0 = 5 β1 = 4 β2 = 4 Then, we must examine the four polynomials: s3 + 2s2 + 4s + 5 = 0 s3 + 4s2 + s + 4 = 0 s3 + 4s2 + 4s + 4 = 0 s3 + 2s2 + s + 5 = 0 The fourth polynomial is not stable—therefore, the system is not stable for the uncertain parameters. P12.9 One possible PID controller is Gc (s) = 0.058s2 + 2.17s + 16.95 . s A first-order Pade approximation was used in the design to account for the delay system. The step input response is shown in Figure P12.9. A prefilter should also be used with the PID controller. A suitable prefilter © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 678 CHAPTER 12 Robust Control Systems is Gp (s) = K2 . K3 s2 + K1 s + K2 1.2 1 y(t) 0.8 0.6 0.4 0.2 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Time (sec) FIGURE P12.9 Step response with the PID controller and prefilter. P12.10 The PID controller is given by Gc (s) = KD s2 + KP s + KI . s Using the ITAE method, we desire the characteristic polynomial to be q(s) = s3 + 1.75ωn s2 + 2.15ωn2 s + ωn3 = 0 , where we select ωn = 4 to obtain a peak time of Tp = 1 second. Here we use the approximation for ITAE third-order systems that ωn Tp ≈ 4 from Figure 5.30(c) in Dorf and Bishop. The actual characteristic equation is s3 + 25KD s2 + 25KP s + 25KI = 0 . Equating coefficients and solving for the gains yields KP = 1.376 , KD = 0.28 , and KI = 2.56 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 679 Problems The step response is shown in Figure P12.10, with the prefilter Gp (s) = KD s2 KI . + KP s + KI Step Response 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 Time (sec) 2 2.5 3 FIGURE P12.10 Step response with the PID controller and prefilter. P12.11 We will design for the case where K = 1 and p = 1. The design plant is G(s) = 1 . s(s + 1)(s + 4) The nominal plant is given by G(s) = 2.5 , s(s + 2)(s + 4) and the PID controller is Gc (s) = KD s2 + KP s + KI . s Using the ITAE method, we desire the characteristic polynomial to be q(s) = s4 + 2.1ωn s3 + 3.4ωn2 s2 + 2.7ωn3 s + ωn4 = 0 , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 680 CHAPTER 12 Robust Control Systems where we select ωn = 2.38 to obtain a peak time around Tp = 3 seconds. The actual characteristic equation (with the worst-case plant) is s4 + 5s3 + (4 + KD )s2 + KP s + KI = 0 . Equating coefficients and solving for the gains yields KP = 36.40, KI = 32.08, and KD = 15.26. The step response is shown in Figure P12.11, with the prefilter Gp (s) = KD s2 KI . + KP s + KI 1.4 Worst−case plant Nominal plant 1.2 1 y(t) 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 5 FIGURE P12.11 Step response with the prefilter: nominal plant (dashed line) & worst-case plant (solid line). P12.12 The transfer function is G(s) = C(sI − A)−1 B = h The sensitivity is G SK = 2 0 i s −3 5 s+K ∂G K −Ks . = 2 ∂K G s + Ks + 5 0 1 = s2 −6 . + Ks + 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 681 Advanced Problems Advanced Problems AP12.1 Let Gp (s) = 1. A viable PID controller is Gc (s) = KP + 1000s2 + 3000s + 100 KI + KD s = . s s The loop transfer function is Gc (s)G(s) = 1000s2 + 3000s + 100) . s(50s2 + 1) We can check that Kv = 100, as desired. The step response is shown in Figure AP12.1. Step Response 1.4 System: syscl Peak amplitude: 1.1 Overshoot (%): 9.5 At time (sec): 0.234 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 Time (sec) 0.8 1 1.2 FIGURE AP12.1 Step response with PID controller. AP12.2 For all three controllers, choose K = 1 as the design value. Also, use as the nominal points a = 2 and b = 5 for each design. ITAE methods were employed in all designs, although this did not work well for the PI controller. (a) PI controller: Let Gp (s) = 1 . Not all specifications could be met simultaneously with a PI con- © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 682 CHAPTER 12 Robust Control Systems troller. The best over-all results are achieved when using a = 3 and b = 4.5 as the design values. An acceptable PI controller is Gc (s) = 1.2 + 3.96 . s Controller P.O. Ts Tp |u(t)|max PI 0% 2.29s n.a. 4.43 PD 4.6% 1.72s 1.26s 12.25 PID 1.97% 0.65s 0.47s 37.25 TABLE AP12.2 PI, PD, and PID controller performance summary. The final design is based on root locus methods since the ITAE methods did not produce an effective controller. The closed-loop transfer function is T (s) = 1.2s + 3.96 . s3 + 3s2 + 5.7s + 3.96 (b) PD controller: Let Gp (s) = 12.25 . 7.25 + 2.9s The closed-loop transfer function is T (s) = 7.25 + 2.9s , s2 + 4.9s + 12.25 where the PD controller (based on ITAE methods) is Gc (s) = 7.25 + 2.9s . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 683 Advanced Problems (c) PID controller: Let Gp (s) = 15.5s2 1000 . + 210s + 1000 The closed-loop transfer function is T (s) = 15.5s2 + 210s + 1000 . s3 + 17.5s2 + 215s + 1000 And the PID controller (based on ITAE methods) is Gc (s) = 15.5s2 + 210s + 1000 . s The performance of each controller is summarized in Table AP12.2. AP12.3 (a) The PID controller is Gc (s) = KD s2 + KP KD s + KI KD s . Since we want P.O. < 4% and Ts < 1s, we choose the dominant closed-loop poles to have ωn = 6 and ζ = 0.8. Therefore, we place the zeros at s2 + KP KI s+ = s2 + 10s + 36 . KD KD Solving for the constants yields, KP = 10 , KI KI = 36 . KD Then, using root locus methods, we choose KD = 91 to place the roots near the zeros. The PID controller gains are computed to be KP = 910, KI = 3276 and KD = 91. (b) The loop transfer function is Gc (s)G(s) = KD s2 + KP s + KI . s2 (s2 + 5s + 4) The closed-loop system characteristic equation is s3 + 5s2 + 4s + KD s2 + KP s + KI = 0 . Solving for the PID gains yields KP = 73.4, KI = 216 and KD = 5.5. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 684 CHAPTER 12 Robust Control Systems Therefore, the controller is Gc (s) = 5.5(s2 + 13.35s + 39.3) . s Using the prefilter Gp (s) = s2 39.3 , + 13.35s + 39.3 we obtain the closed-loop transfer function T (s) = s3 + 216 . + 77.4s + 216 10.5s2 The percent overshoot is P.O. ≈ 3.5% and the settling time is Ts ≈ 1.67 sec. The PID controller is Gc (s) = KD s2 + KP KI KD s + KD s . The bounds 1 ≤ a ≤ 2 and 4 ≤ b ≤ 12 imply that 2 ≤ ωn ≤ 3.46 and 0.5 ≤ ζωn ≤ 1. One solution is to place the PID controller zeros at 1. 4 1. 2 1 Amplitude AP12.4 0. 8 0. 6 0. 4 0. 2 0 0 0. 5 1 1. 5 Time (sec) 2 2. 5 3 FIGURE AP12.4 Family of step response with PID controller with nominal case (a, b) = (1.5, 9) denoted by the solid line. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 685 Advanced Problems √ s = −1 ± j 8 (i.e. ζωn = 1 and ωn = 3). So, s2 + KI KP s+ = s2 + 2ζωn s + ωn2 = s2 + 4s + 9 . KD KD The nominal case for design is chosen to be a = 1.5 and b = 9. Using root locus, we select KD = 2.1 to place the closed-loop characteristic roots near the zeros. Then, the PID controller gains are computed to be KP = 8.4, KI = 18.9, and KD = 2.1. The plot of the response to a step input is shown in Figure AP12.4. The off-nominal cases shown in the simulations are (a, b) = (1.2, 4), (1.4, 6), (1.6, 10), and (1.8, 12). AP12.5 To obtain a phase margin of P.M. = 49.77o , select K = 1.5, b = 36 and choose Gp (s) = 1. The PID controller is Gc (s) = 1.5(s2 + 20s + 36) . s When K1 = 0.75, the phase margin is reduced to P.M. = 45.45o ; and when K1 = 1.25, the phase margin is increased to P.M. = 52.75o . AP12.6 With the settling time Ts = 1 and percent overshoot P.O. < 10% specifications, we target for dominant closed-loop poles with ωn = 10. Here we estimate ωn Ts ≈ 10 associated with the ITAE performance. The closedloop transfer function is T (s) = Gp (s) 1.5(KD s2 + KP s + KI ) , (1 + 1.5KD )s2 + 1.5KP s + 1.5KI where we have neglected τ . Using the ITAE method, the desired characteristic polynomial is s2 + √ 2ωn s + ωn2 = s2 + 1.5Kp 1.5KI s+ . 1 + 1.5KD 1 + 1.5KD Let KD = 0.25. Then solving for the remaining PID gains yields KP = 12.96 and KI = 91.67. The pre-filter is Gp (s) = 0.375s2 137.5 . + 19.45s + 137.5 Then the closed-loop transfer function (with τ = 0.001) is T (s) = 0.001s3 137.5 . + 1.375s2 + 19.45s + 137.5 The transfer function from the disturbance to the output is Y (s)/Td (s) = 0.001s3 1.5s . + 1.375s2 + 19.45s + 137.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 686 CHAPTER 12 Robust Control Systems The step input response and disturbance response are shown in Figure AP12.6. (a) (b) 0.05 1.4 1.2 0.04 1 Amplitude Amplitude 0.03 0.8 0.02 0.6 0.01 0.4 0 0.2 0 0 0.2 0.4 0.6 Time (sec) 0.8 1 −0.01 0 0.2 0.4 0.6 Time (sec) 0.8 1 FIGURE AP12.6 (a) Input response; (b) Disturbance response. AP12.7 The PI controller is given by Gc (s) = KP s + KI . s We will also use the prefilter Gp (s) = KI . KP s + KI Using the ITAE method, we determine that √ KP = 2ωn and KI = ωn2 . Let ωn = 2.2. Then KP = 3.11 and KI = 4.8. The step response and control u(t) are shown in Figure AP12.7. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 687 Advanced Problems (a) (b) 1.2 1.2 1 1 0.8 Amplitude Amplitude 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0 5 -0.2 0 5 Time (secs) Time (secs) FIGURE AP12.7 (a) Input response; (b) Control history u(t). AP12.8 (a) A suitable PD controller is given by Gc (s) = 0.6 + 0.4s . The percent overshoot is P.O. = 18.8% and the peak time is Tp = 2.4 sec. (b) A suitable PI controller is given by Gc (s) = 0.15 + 0.01 . s The percent overshoot is P.O. = 23.7% and the peak time is Tp = 7.8 sec. (c) A suitable PID controller is given by Gc (s) = 0.6 + 0.01 + 0.4s . s The percent overshoot is P.O. = 19.9% and the peak time is Tp = 2.5 sec. (d) The PD or PID controllers are the best choices. AP12.9 A robust PID controller designed with ITAE methods will be a suitable controller. From the settling time specification we select ωn = 10, where © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 688 CHAPTER 12 Robust Control Systems we have used ζ = 0.8. The worst case is a = 1 and K=2. The desired closed-loop transfer function is T (s) = ωn3 s3 + 1.75ωn s2 + 2.15ωn2 s + ωn3 and the actual characteristic equation is q(s) = s3 + (2a + KKD )s2 + (a2 + KKP )s + KKI . Equating like terms, we find that KP = 107 KD = 7.75 . We use as the design plant G(s) = s+2 . s(s + 3) 1.4 1.2 1 Amplitude AP12.10 KI = 500 0.8 0.6 0.4 0.2 0 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 Time(sec) FIGURE AP12.10 Family of step responses with the design plant (p, q, r) = (3, 0, 2) denoted by the solid line. Select p1 = 2 and z1 = 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 689 Advanced Problems to cancel a design plant pole and zero. Then, choose p2 = 0 to have zero steady-state error to a unit step. The remaining variables K and z2 are selected based on ITAE methods, where ωn = 100. A suitable compensator is Gc (s) = 141.42(s + 3)(s + 70.71) . s(s + 2) A plot of the step responses for various values of p, q and r is shown in Figure AP12.10. A suitable compensator is Gc (s) = 1000(s + 1.8)(s + 3.5)(s + 5.5) . s(s + 600) 1.4 1.2 1 Amplitude AP12.11 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 6 7 8 9 10 Time (sec) FIGURE AP12.11 Step responses with nominal plant (solid line) and off-nominal plant with all poles reduced by 50% (dashed line). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 690 CHAPTER 12 Robust Control Systems Design Problems The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: θ(s) 26.035 = , Va (s) s(s + 33.142) where we neglect the motor inductance Lm and where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). With a PID controller ,the closed-loop system characteristic equation is s3 + (33.142 + 26.035KD )s2 + 26.035KP s + 26.035KI = 0 . A suitable PID controller is Gc (s) = 50 + s + 0.1 . s This PID controller places the closed-loop system poles to the left of the −ζωn line necessary to meet the settling time requirement. The step response is shown below. The settling time is Ts = 0.12 second. In the steady-state the error due to a step disturbance is zero. 1.2 1 0.8 Amplitude CDP12.1 0.6 0.4 0.2 0 0 0.05 0.1 0.15 0.2 0.25 Time (secs) 0.3 0.35 0.4 0.45 0.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 691 Design Problems The closed-loop transfer function is Y (s) Km Gc (s) . = 2 R(s) s + (2 + Km K1 )s + Gc (s)Km (a) When Gc = K, we have T (s) = s2 15K , + (2 + 15K1 )s + 15K where Km = 15. Using ITAE criteria and ωn = 10, we determine that K1 = 0.81 and K = 6.67. For the disturbance, we have −1 Y (s) = 2 . TL (s) s + 14.14s + 100 The input and disturbance responses are shown in Figure DP12.1, without prefilters. (a) Step response (b) Disturbance response 1.2 0 1 -0.002 0.8 -0.004 y(t) y(t) DP12.1 0.6 -0.006 0.4 -0.008 0.2 -0.01 0 0 0.5 1 Time (sec) -0.012 0 0.5 1 Time (sec) FIGURE DP12.1 (a) Step response: Gc (s) = K (solid line) and Gc (s) = KP + KD s (dashed line); and (b) disturbance response (same for both compensators). (b) When Gc = KP + KD s, we have Y (s) 15(KP + KD s) = 2 . R(s) s + (2 + 15K1 + 15KD )s + 15KP For ωn = 10 and with the ITAE criteria, we determine that (with © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 692 CHAPTER 12 Robust Control Systems KD = 0.1) Y (s) 15(6.67 + 0.1s) = 2 . R(s) s + 14.14s + 100 The nominal plant is given by 1 . s(s + 5) G(s) = The closed-loop transfer function is T (s) = K(KD s2 + KP s + KI ) . s3 + (5 + KKD )s2 + KKP s + KKI Let KP = 450 , KI = 750 , and KD = 150 . A family of responses is shown in Figure DP12.2 a for various values of K. The percent overshoot for 0.1 ≤ K ≤ 2 is shown in Figure DP12.2b. 1.4 1.2 1 Step response DP12.2 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 Time (s) FIGURE DP12.2 (a) Family of step responses for various values of K. 2.5 3 3.5 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 693 Design Problems 9 8 Percent overshoot 7 6 5 4 3 2 1 0 0 0.2 0.4 0.6 0.8 1 K 1.2 1.4 1.6 1.8 FIGURE DP12.2 CONTINUED: (b) Percent overshoot for various values of K. DP12.3 (a) The dexterous hand model is given by G(s) = Km , s(s + 5)(s + 10) where Km = 1, nominally. The PID controller is Gc (s) = KD (s2 + 6s + 18) . s The root locus is shown in Figure DP12.3a. If we select KD = 90 , the roots are s1,2 = −5.47 ± j6.6 s3,4 = −2.03 ± j4.23 . Thus, all roots have ζωn > 4/3 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 12 Robust Control Systems to meet the design specification Ts < 3 sec . (b) The step responses for Km = 1 and Km = 1/2 are shown in Figure DP12.3b. When K = 1/2 , an off-nominal value, the settling time specification is no longer satisfied. 20 15 10 * 5 Imag Axis 694 o 0 x * x x o * -5 * -10 -15 -20 -20 -15 -10 -5 0 Real Axis FIGURE DP12.3 s2 +6s+18 (a) Root locus for 1 + KD s2 (s+5)(s+10) = 0. 5 10 15 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 695 Design Problems 1.6 1.4 1.2 y(t) 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Time (sec) FIGURE DP12.3 CONTINUED: (b) Step response (without prefilters): PID with K3 = 90 and Km = 1 (solid line) and PID with K3 = 90 and Km = 0.5 (dashed line). DP12.4 The nominal plant is G(s) = s(s2 17640 , + 59.4s + 1764) and the PID controller is Gc (s) = KI (τ1 s + 1)(τ2 s + 1) . s (a) Using ITAE methods, we determine that ωn = 28.29, KI = 36.28, τ1 + τ2 = 0.0954 and τ1 τ2 = 0.00149. So, Gc (s) = 36.28(0.00149s2 + 0.0954s + 1) . s (b) The step response for the nominal plant and the PID controller is shown in Figure DP12.4a, with and without a prefilter. (c) The disturbance response is shown in Figure DP12.4b. (d) The off-nominal plant is G(s) = s(s2 16000 . + 40s + 1600) The step response for the off-nominal plant is shown in Figure DP12.4a. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 12 Robust Control Systems (b) off-nominal plant 1.6 1.4 1.4 1.2 1.2 1 1 y(t) y(t) (a) nominal plant 1.6 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0.5 0 0 1 0.5 Time (sec) 1 Time (sec) FIGURE DP12.4 (a) Step response for (i) nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line); and (ii) for off-nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line). disturbance response 0.3 0.25 0.2 0.15 y(t) 696 0.1 0.05 0 -0.05 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Time (sec) FIGURE DP12.4 CONTINUED: (b) Disturbance response for the nominal plant. 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 697 Design Problems DP12.5 One possible solution is Gc (s) = 0.08 (0.01s + 1)(0.99s + 1) . s The phase margin with this controller is P.M. = 45.5o . The step response is shown in Figure DP12.5 for the nominal plant (with and without a prefilter); the step response for the off-nominal plant is also shown in Figure DP12.5. The prefilter is Gp (s) = 13.97s2 1411 . + 1411s + 1411 (a) nominal plant (b) off nominal plant 1. 2 1. 2 1 1 0. 8 0. 8 y(t) 1. 4 y(t) 1. 4 0. 6 0. 6 0. 4 0. 4 0. 2 0. 2 0 0 10 Time (sec) 20 0 0 10 Time (sec) 20 FIGURE DP12.5 (a) Step response for nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) for off-nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line). DP12.6 Using ITAE methods, three controllers are designed for the nominal plant: (i) PID controller: Gc (s) = 0.225s2 + 0.535s + 34.3 s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 698 CHAPTER 12 Robust Control Systems (ii) PI controller: Gc (s) = 0.9s + 22.5 s (iii) PD controller: Gc (s) = 0.9s + 22.5 The step responses for each controller is shown in Figure DP12.6. The responses for the PID and PI controller are the same since the gains were selected to obtain the same ITAE characteristic equation. An appropriate prefilter is used in all cases. (b) off-nominal plant 1.2 1 1 0.8 0.8 y(t) y(t) (a) nominal plant 1.2 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0.5 1 1.5 Time (sec) 2 0 0 0.5 1 1.5 2 Time (sec) FIGURE DP12.6 (a) Step response for nominal plant: PID (solid line); PI (dashed line); and PD (dotted line); (b) for off-nominal plant: PID (solid line); PI (dashed line); and PD (dotted line). DP12.7 The loop transfer function is G(s) = Ka Km K = (0.5s + 1)(τf s + 1)s(s + 1) s(s + 2)(s + 1) since τf is negligible. A suitable PID controller is Gc (s) = 300(s2 + 2.236s + 2.5) KKD (s2 + as + b) = . s s The step response is shown in Figure DP12.7. The percent overshoot is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 699 Design Problems P.O. = 4.6% and the settling time is Ts = 3.74 seconds. 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 Time (secs) FIGURE DP12.7 Step response for the elevator position control. DP12.8 The system transfer function is Y (s) = G(s)Gc (s)Gp (s) R(s) . 1 + G(s)Gc (s) We are given G(s) = e−sT where T = 1 second . Using a second-order Pade approximation yields G(s) ≈ s2 − 6s + 12 . s2 + 6s + 12 Three controllers that meet the specifications are 0.5 (Integral controller) s 0.04s + 0.4 Gc2 (s) = (PI controller) s 0.01s2 + 0.04s + 0.4 Gc3 (s) = (PID controller) . s Gc1 (s) = 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 700 CHAPTER 12 Robust Control Systems In all cases, the steady-state error is zero. Integral PI PID P.O.(%) 4.05 0 0 Ts (sec) 6.03 6.12 6.02 Tp (sec) 4.75 N/A N/A |V (t)|max (volts) 1.04 1 1 The prefilter Gp (s) = 1 is used in all designs. To compute the voltage, the transfer function is V (s) = DP12.9 Gp (s)Gc (s) R(s) . 1 + Gc (s)G(s) The space robot transfer function is G(s) = 1 . s(s + 10) (a) Consider Gc (s) = K. Then T (s) = Gc (s)G(s) K = 2 . 1 + Gc (s)G(s) s + 10s + K We determine that K = 50.73 for ζ = 0.702. Thus, we expect P.O. < 4.5%. So, Gc (s) = 50.73 . (b) Consider the PD controller Gc (s) = KP + KD s . Then T (s) = KP + KD s . s2 + (10 + KD )s + KP Using the ITAE method, we compute KP = 100 and KD = 4 . Thus, Gc (s) = 4s + 100 , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 701 Design Problems and the prefilter is Gp (s) = 100 . 4s + 100 (c) Consider the PI controller Gc (s) = KP + KP s + KI KI = . s s Then, T (s) = s3 KP s + KI . + 10s2 + KP s + KI Using the ITAE method, we have ωn = 5.7 KP = 70.2 and KI = 186.59 . Thus, Gc (s) = 70.2 + 186.59/s , and the prefilter is Gp (s) = 186.59 . 70.2s + 186.59 (d) Consider the PID controller Gc (s) = KD s2 + KP s + KI . s Then, T (s) = KD s2 + KP s + KI . s3 + 10s2 + KD s2 + KP s + KI Using the ITAE method with ωn = 10, we have KD = 7.5 KP = 215 and KI = 1000 . Thus, Gc (s) = 7.5s2 + 215s + 1000 , s and the prefilter is Gp (s) = 7.5s2 1000 . + 215s + 1000 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 702 CHAPTER 12 Robust Control Systems A summary of the performance is given in Table DP12.9. Gc (s) P.O. tp ts yss max|y(t)| K 4.5% 0.62 s 0.84 s 0 0.026 PD 5.2% 0.39 s 0.56s 0 0.010 PI 1.98% 0.81 s 1.32s 0 0.013 PID 1.98% 0.46 s 0.75 s 0 0.004 TABLE DP12.9 A summary of performance to a disturbance input. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 703 Computer Problems Computer Problems The closed-loop transfer function is T (s) = s2 8K , + 2s + 8K T , is and the sensitivity function, SK S(s) = s2 + s . s2 + 2s + 8K The plot of T (s) and S(s) is shown in Figure CP12.1, where K = 10. nt=[80]; dt=[1 2 80]; syst = tf(nt,dt); ns=[1 2 0];ds=[1 2 80]; syss = tf(ns,ds); w=logspace(-1,2,400); [magt,phaset]=bode(syst,w);magtdB(1,:) = 20*log10(magt(1,1,:)); [mags,phases]=bode(syss,w); magsdB(1,:) = 20*log10(mags(1,1,:)); semilogx(w,magtdB,w,magsdB,'--') legend('20log|T|','20log|S|') xlabel('Frequency (rad/sec)') ylabel('Gain dB') grid 20 20log|T| 20log|S| 10 0 −10 Gain dB CP12.1 −20 −30 −40 −50 −60 −1 10 0 1 10 10 Frequency (rad/sec) FIGURE CP12.1 Plot of T (s) and the sensitivity function S(s). 2 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 704 CHAPTER 12 CP12.2 Robust Control Systems A reasonable value of the gain K = 4. The family of step responses is shown in Figure CP12.2. p=[0.5:0.5:20]; K=4; t=[0:0.01:1]; for i=1:length(p) n=[K*p(i)]; d=[1 p(i)]; sys = tf(n,d); sys_cl = feedback(sys,[1]); y=step(sys_cl,t); Y(:,i)=y; [y2,t2]=step(sys_cl); S=stepinfo(y2,t2); Ts(i)=S.SettlingTime; end plot(t,Y) , xlabel('Time (sec)'), ylabel('Step response') 0.9 0.8 0.7 Step response 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.1 0.2 0.3 0.4 0.5 0.6 Time (sec) 0.7 FIGURE CP12.2 Family of step responses for 0.5 < p < 20. CP12.3 The closed-loop characteristic equation is 1 + KD s2 + as + b =0 Js3 where a = KP /KD b = KI /KD . 0.8 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 705 Computer Problems We select a = 1 and b=2 to move the root locus into the left hand-plane (see Figure CP12.3a). Then, we choose KD = 71 from the root locus using the rlocfind function. The closed-loop Bode plot in Figure CP12.3b verifies that the bandwidth ωB < 5 rad/sec. Also, the phase margin is P.M. = 45.7o , which meets the design specification. The plot of phase margin versus J is shown in Figure CP12.3c. We see that as J increases, the phase margin decreases. J=25; a=1; b=2; ng=[1];dg=[J 0 0]; sysg=tf(ng,dg); nc=[1 a b]; dc=[1 0]; sysc=tf(nc,dc); sys=series(sysc,sysg); rlocus(sys) Root Locus 2 1.5 Imaginary Axis 1 0.5 0 −0.5 −1 −1.5 −2 −1.2 −1 −0.8 −0.6 FIGURE CP12.3 2 +s+2 (a) Root locus for 1 + KD s 10s = 0. 3 −0.4 Real Axis −0.2 0 0.2 0.4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 12 Robust Control Systems J=25; a=1; b=2; KD=71; KP=a*KD; KI=b*KD; ng=[1]; dg=[J 0 0]; sysg=tf(ng,dg); nc=[KD KP KI]; dc=[1 0]; sysc = tf(nc,dc); sys=series(sysc,sysg); sys_cl = feedback(sys,[1]); bode(sys_cl); [GM,PM]=margin(sys); PM PM = 45.7093 20 Magnitude (dB) 10 0 −10 −20 −30 XY 10 0 1 10 2 10 10 Frequency (rad/sec) FIGURE CP12.3 CONTINUED: (b) Closed-loop Bode plot with ωB < 5 rad/sec. Ji=[10:1:40]; for i=1:length(Ji) numc=[KD KP KI]; denc=[Ji(i) 0 0 0]; sysc = tf(numc,denc); [gm,pm]=margin(sysc); Pm(i)=pm; end plot(Ji,Pm), grid xlabel('J'), ylabel('Phase Margin (deg)') 90 80 70 60 Phase Margin (deg) 706 50 40 30 20 10 0 −10 0 5 10 15 FIGURE CP12.3 CONTINUED: (c) Phase margin versus J. 20 J 25 30 35 40 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 707 Computer Problems The closed-loop characteristic equation is CP12.4 1+K s2 1 =0 + bs + a where a = 8 and the nominal value of b = 4. The root locus is shown in Figure CP12.4a. clf, hold off a=8; b=4; num=[1]; den=[1 b a]; sys = tf(num,den); rlocus(sys), hold on zeta=0.59; wn=1.35; x=[-10:0.1:-zeta*wn]; y=-(sqrt(1-zeta^2)/zeta)*x; xc=[-10:0.1:-zeta*wn];c=sqrt(wn^2-xc.^2); plot(x,y,':',x,-y,':',xc,c,':',xc,-c,':') rlocfind(sys) ÈSelect a point in the graphics window selected_point = -2.0165 + 2.5426i K ans = 2.4659 4 3 + 2 x Imag Axis 1 0 -1 -2 x + -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE CP12.4 1 (a) Root locus for 1 + K s2 +4s+8 . The performance region is specified by ζ = 0.59 and ωn = 1.35 , which derives from the design specifications Ts < 5 sec and P.O. < 10% . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 708 CHAPTER 12 Robust Control Systems Using an m-file, the value of K = 2.5 is selected with the rlocfind function. The step responses for b = 0, 1, 4 and b = 40 are shown in Figure CP12.4b. When b = 0, the system is marginally stable; b = 1 results in a stable system with unsatisfactory performance. The nominal case b = 4 is stable and all performance specs are satisfied. When b = 40, the system is heavily damped: the percent overshoot specification is satisfied, but the settling time is too long. 0.5 b=0 0.45 0.4 b=1 Amplitude 0.35 0.3 0.25 0.2 0.15 b=4 0.1 0.05 0 0 b=40 1 2 3 4 5 6 7 8 9 10 Time (secs) FIGURE CP12.4 CONTINUED: (b) Step responses for b = 0, 1, 4 and 40. CP12.5 (a) An acceptable lead compensator (designed with root locus methods) is Gc (s) = K s+a s + 0.3 =5 . s+b s+2 The compensated root locus is shown in Figure CP12.5a, where K=5 is selected to place the closed-loop poles in the performance region. (b) The step responses for ζ = 0, 0.005, 0.1 and 1 are shown in Figure CP12.5b. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 709 Computer Problems 4 3 + 2 + x o Imag Axis 1 0 +o x x -1 + -2 o x + -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE CP12.5 (a) Compensated root locus. (c) You would like the actual structural damping to be greater than the design value, if it must be different at all. zeta=0,0.005 (solid); zeta=0.1 (dashed); zeta=1 (dotted) 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 Time (sec) FIGURE CP12.5 CONTINUED: (b) Step responses for ζ = 0, 0.005, 0.1 and 1. 16 18 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 710 CHAPTER 12 CP12.6 Robust Control Systems The m-file script which computes the phase margin as a function of the time delay (using the pade function) is shown in Figure CP12.6. The maximum time delay (for stability) is td = 4.3 seconds. K=5; numg=K*[1]; deng=[1 10 2]; sysg = tf(numg,deng); time delay vector td=[0:0.1:5]; for i=1:length(td) [ndelay,ddelay]=pade(td(i),2); sysd = tf(ndelay,ddelay); sys = series(sysg,sysd); [mag,phase,w]=bode(sys); [gm,pm,w1,w2]=margin(mag,phase,w); pmv(i)=pm; end plot(td,pmv), grid xlabel('time delay [sec]') ylabel('phase margin [deg]') 120 100 phase margin [deg] 80 60 40 20 0 -20 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 time delay [sec] FIGURE CP12.6 Phase margin versus time delay. CP12.7 The m-file script is shown in Figure CP12.7a. The steady-state error (shown in Figure CP12.7b) is zero when a = 0.5 and increases rapidly as a increases past a = 0.5. The maximum initial undershoot is shown in Figure CP12.7c. As a increases, the initial undershoot increases linearly. The gain margin is shown in Figure CP12.7d. It © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 711 Computer Problems can be seen that as a increases, the gain margin decreases very rapidly. a=[0.01:0.01:0.99]; t=[0:0.1:30]; for i=1:length(a) num=a(i)*[1 -0.5]; den=[1 2 1]; sys_o = tf(num,den); [mag,phase,w]=bode(sys_o); [gm,pm,w1,w2]=margin(mag,phase,w); gain margin gmv(i)=gm; sys_cl = feedback(sys_o,[1]); [y,x]=step(-sys_cl,t); negative unit step input yf(i)=1-y(length(t)); steady-state tracking error ym(i)=-min(y)*100; max initial undershoot end figure(1), plot(a,gmv), grid, xlabel('a'), ylabel('gm') figure(2), plot(a,yf ), grid, xlabel('a'), ylabel('steady-state error') figure(3), plot(a,ym), grid, xlabel('a'), ylabel('maximum initial undershoot [%]') FIGURE CP12.7 Script to generate all the plots. 1 0.9 0.8 steady−state error 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.1 0.2 0.3 0.4 0.5 a FIGURE CP12.7 CONTINUED: (b) Steady-state tracking error. 0.6 0.7 0.8 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 12 Robust Control Systems 25 maximum initial undershoot [%] 20 15 10 5 0 0 0.1 0.2 0.3 0.4 0.5 a 0.6 0.7 0.8 0.9 1 0.6 0.7 0.8 0.9 1 FIGURE CP12.7 CONTINUED: (c) Maximum initial undershoot. 250 200 150 gm 712 100 50 0 0 0.1 0.2 FIGURE CP12.7 CONTINUED: (d) Gain margin. 0.3 0.4 0.5 a © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 713 Computer Problems CP12.8 The plant (balloon and canister dynamics plus motor) is given by G(s) = 1 , (s + 2)(s + 4)(s + 10) and the PID controller is Gc (s) = KD (s2 + as + b) . s Let a = 6. Then using the root locus methods, we determine that with KD = 12.5 and b = 10 we have the roots s1 = −8.4 s2 = −4.7 s3,4 = −1.43 ± j1.05 . Thus, ζ = 0.8. The plot of y(t) is shown in Figure CP12.8. The percent overshoot is less that 3%, as desired. 1.4 1.2 With prefilter 1 y(t) 0.8 Without prefilter 0.6 0.4 0.2 0 0 FIGURE CP12.8 Simulation of the GRID device. 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 1 3 Digital Control Systems Exercises E13.1 (a) Elevation contours on a map are discrete signals. (b) Temperature in a room is a continuous signal. (c) A digital clock display is a discrete signal. (d) The score of a basketball game is a discrete signal. (e) The output of a loudspeaker is a continuous signal. E13.2 (a) Using long-division we determine that Y (z) = z −1 + 3z −2 + 7z −3 + 15z −4 + · · · Therefore, with Y (z) = ∞ X y(kT )z −k k=0 we have y(0) = 0 y(T ) = 1 y(2T ) = 3 y(3T ) = 7 y(4T ) = 15 . (b) The exact solution is y(kT ) = ek ln 2 − 1 . E13.3 For the system response y(kT ) = kT where k ≥ 0, we have Y (z) = E13.4 The partial fraction expansion of Y (s) is Y (s) = 714 Tz . (z − 1)2 5 0.25 0.0625 0.3125 = + − . s(s + 2)(s + 10) s s + 10 s+2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 715 Exercises Then, using Table 13.1 in Dorf and Bishop, we determine that z z z + 0.0625 − 0.3125 −10T z −1 z −e z − e−2T z z z = 0.25 + 0.0625 − 0.3125 , z −1 z − 0.135 z − 0.670 Y (z) = 0.25 where T = 0.1. E13.5 The Space Shuttle and robot arm control block diagram is shown in Figure E13.5. The human operator uses information from the computer generated data display and visual sensory data from the TV monitor and by looking out the window. He/she commands the robot arm via a joystick command to the computer. data display measurement digital analog A/D joint angle & rate sensors digital human operator joystick command ref. + Computer digital analog D/A - Robot arm & motors/gears tip position measurement TV monitor & window view FIGURE E13.5 The Space Shuttle/robot arm control block diagram. E13.6 From Section 10.8 in Dorf and Bishop, we find that the design resulted in the compensator Gc (s) = 6.66s + 1 s + 0.15 = 0.1 . 66.6s + 1 s + 0.015 Using the relationships A = e−aT , we compute B = e−bT , and C 1−A a =K , 1−B b A = e−0.15(0.001) = 0.99985 , B = e−0.015(0.001) = 0.999985 , and C = 0.1 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 716 CHAPTER 13 Digital Control Systems Therefore, D(z) = C E13.7 z − 0.99985 z−A = 0.1 . z−B z − 0.999985 Using long-division, we determine that Y (z) = 1 + 3.5z −1 + 5.75z −2 + 6.875z −3 + · · · Therefore, with Y (z) = ∞ X y(kT )z −k k=0 we have y(0) = 1 y(T ) = 3.5 E13.8 y(2T ) = 5.75 y(3T ) = 6.875 . The closed-loop system with T (z) = z2 z + 0.2z − 1.0 is unstable since one of the poles of the transfer function (z = −1.1 and z = 0.90) lies outside the unit circle in the z-plane. E13.9 (a) Using long-division we determine that Y (z) = z −1 + z −2 + z −3 + z −4 + · · · Therefore, with Y (z) = ∞ X y(kT )z −k k=0 we have y(0) = 0 y(T ) = 1 y(2T ) = 1 y(3T ) = 1 y(4T ) = 1 . (b) The exact solution is y(kT ) = 1 − δ(k) where δ(k) = 1 when k = 0 and δ(k) = 0 when k 6= 0. E13.10 We compute T /τ = 1.25. (a) Using Figure 13.19 in Dorf and Bishop, we determine that Kτ = 0.8 which implies K = 100. (b) Using Figure 13.21 in Dorf and Bishop, we determine that ess = 0.75. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 717 Exercises (c) Using Figure 13.20 in Dorf and Bishop, we determine that Kτ = 0.7 which implies K = 88. (a) The transfer function (including the zero-order hold) is Go (s)Gp (s) = 100(1 − e−sT ) . s(s2 + 100) Expanding into partial fractions yields s 1 − 2 G(z) = (1 − z )Z s s + 100 z z(z − cos 10T ) −1 = (1 − z ) − . z − 1 z 2 − 2 cos 10T z + 1 −1 When T = 0.05 we ha,ve G(z) = 0.1224(z + 1) . − 1.7552z + 1 z2 (b) The system is marginally stable since the system poles, z = −0.8776± 0.4794j, are on the unit circle. (c) The impulse response and sinusoidal input response are shown in Figure E13.11. Amplitude 0.5 0 -0.5 2 0 4 6 8 10 12 14 16 No. of Samples 40 Amplitude E13.11 20 0 -20 -40 0 10 20 30 40 50 60 70 No. of Samples FIGURE E13.11 Impulse and sinusoidal (natural frequency) input response. 80 90 100 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 718 CHAPTER 13 E13.12 Digital Control Systems The partial fraction expansion of X(s) is X(s) = s2 s+1 1 2 = − . + 5s + 6 s+3 s+2 Then, with T = 1, we have X(z) = E13.13 z 2z z 2z − = − . −3 −2 z−e z−e z − 0.0498 z − 1353 The root locus is shown in Figure E13.13. For stability: 2.2 < K < 5.8. Root Locus 2 1.5 Imaginary Axis 1 0.5 K=5.8 0 K=2.2 −0.5 −1 −1.5 −2 −2 −1.5 −1 −0.5 0 Real Axis 0.5 1 1.5 2 FIGURE E13.13 Root locus with unit circle (dashed curve). E13.14 Given Gp (s), we determine that (with K = 5) G(z) = 5(1 − e−1 )z . z(z − e−1 ) The closed-loop characteristic equation is z 2 + 1.792z + 0.368 = 0 and the system is unstable, since there is a pole at z = −1.55. The system is stable for 0 < K < 4.32 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 719 Exercises E13.15 The transfer function G(z) is G(z) = z2 The sampling time is T = 1 s. E13.16 0.1289z + 0.02624 . − 0.3862z + 0.006738 The transfer function G(z) is G(z) = 0.2759z + 0.1982 . − 1.368z + 0.3679 z2 The sampling time is T = 0.5 s. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 720 CHAPTER 13 Digital Control Systems Problems P13.1 The plot of the input to the sampler and the output r ∗ (t) is shown in Figure P13.1. 1 0.8 0.6 0.4 r(t), r*(t) 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Time (sec) FIGURE P13.1 Plot of r(t) = sin(ωt) and r ∗ (t). The plot of the input and the output is shown in Figure P13.2. 1 0.8 0.6 0.4 0.2 r(t) P13.2 0 -0.2 -0.4 -0.6 -0.8 -1 0 0.2 0.4 0.6 0.8 1 1.2 Time (sec) FIGURE P13.2 Plot of r(t) = sin(ωt) and output of sample and hold. 1.4 1.6 1.8 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 721 Problems P13.3 The transfer function Y (z)/R∗ (z) = G(z) = z . z − e−T The ramp input is represented by R(z) = Tz . (z − 1)2 The output Y (z) = G(z)R(z) is obtained by long division: h i Y (z) = T z −1 + T (2 + e−T )z −2 − T (1 + 2e−T ) − (2 + e−T )2 z −3 h + T e−T + (1 + 2e−T )(2 + e−T ) − (2 + e−T ) (1 + 2e−T ) − (2 + e−T )2 P13.4 i z −4 + · · · The transfer function Y (s)/R∗ (s) = 1 − e−sT . s(s + 2) The partial fraction expansion (with T = 1) yields G(z) = (1 − z −1 )Z = P13.5 0.5 0.5 − s s+2 0.4323 . z − 0.1353 = (1 − z −1 ) 0.5z 0.5z − z − 1 z − 0.1353 The step input is R(z) = z . z−1 Also, T (z) = G(z) 0.6321 = . 1 + G(z) z + 0.2643 So, Y (z) = T (z)R(z) = 0.6321 z 0.6321z = 2 . z + 0.2643 z − 1 z − 0.7357z − 0.2643 Using long-division we determine that Y (z) = 0.6321z −1 + 0.4650z −2 + 0.5092z −3 + 0.4975z −4 + 0.5006z −5 + · · · © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 722 CHAPTER 13 Digital Control Systems Therefore, with Y (z) = ∞ X y(kT )z −k k=0 we have y(0) = 0, y(T ) = 0.6321, y(2T ) = 0.4650, y(3T ) = 0.5092, y(4T ) = 0.4975, and y(5T ) = 0.5006. P13.6 Using the final value theorem (see Table 13.1 in Dorf and Bishop), we determine that (for a step input) Yss = lim (z − 1)Y (z) = lim (z − 1) z→1 z→1 0.6321 z 0.6321 = = 0.5 . z + 0.2643 z − 1 1.2643 And using the initial value theorem, we compute Yo = lim Y (z) = lim z→∞ z→∞ 0.6321 z =0. z + 0.2643 z − 1 P13.7 Using Figures 13.19 and 13.21 in Dorf and Bishop, we determine that the performance specifications are satisfied when Kτ = 0.5 and Tτ = 2. Computing K and T (with τ = 0.5) yields K = 1 and T = 1. P13.8 We can select K = 1 and r = 0.2. The step responses for the compensated and uncompensated systems are shown in Figure P13.8. 1.2 1 Uncompensated Amplitude 0.8 0.6 Compensated 0.4 0.2 0 0 10 20 30 40 No. of Samples FIGURE P13.8 Plot of compensated and uncompensated systems. 50 60 70 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 723 Problems P13.9 Consider the compensator Gc (s) = K s+a . s+b Then, using Bode methods we can select a = 1, b = 4, and K = 1. The compensated system phase margin is P.M. = 50o and the gain margin is G.M. = 15dB. The crossover frequency is ωc = 2.15 rad/sec. Utilizing the Gc (s)-to-D(z) method and selecting T = 0.01 second, we determine D(z) = C We use the relationships A = e−aT , z − 0.99 z−A = . z−B z − 0.96 B = e−bT , and C 1−A a =K , 1−B b to compute A = e−0.01 = 0.99, B = e−0.04 = 0.96, and C = 1. P13.10 (a) The transfer function G(z)D(z) is G(z)D(z) = K 0.0037z + 0.0026 . − 1.368z + 0.3679 z2 (b) The closed-loop system characteristic equation is 1+K 0.0037z + 0.0026 =0. − 1.368z + 0.3679 z2 (c) Using root locus methods, the maximum value of K is found to be Kmax = 239. (d) Using Figure 13.19 in Dorf and Bishop for T /τ = 1 and a maximum overshoot of 0.3, we find that K = 75. (e) The closed-loop transfer function (with K = 75) is T (z) = 0.2759z + 0.1982 . − 1.092z + 0.5661 z2 The step response is shown in Figure P13.10. (f) The closed-loop poles with K = 119.5 are z = 0.4641 ± 0.6843j. The overshoot is 0.55. (g) The step response is shown in Figure P13.10 (for K = 119.5). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 724 CHAPTER 13 Digital Control Systems K=75 Amplitude 1.5 1 0.5 0 2 0 4 6 8 10 12 14 16 No. of Samples K=119.5 Amplitude 2 1.5 1 0.5 0 0 2 4 6 8 10 12 14 16 18 20 No. of Samples FIGURE P13.10 Step response for K = 75 and K = 119.5. P13.11 (a) Consider the compensator Gc (s) = K s+a . s+b Then, using Bode methods we can select a = 0.7, b = 0.1, and K = 150. The compensated system overshoot and steady-state tracking error (for a ramp input) are P.O. = 30% and ess < 0.01. (b) Utilizing the Gc (s)-to-D(z) method (with T = 0.1 second), we determine D(z) = C We use the relationships A = e−aT , z−A z − 0.9324 = 155.3 . z−B z − 0.99 B = e−bT , to compute A = e−0.007 = 0.9324 , and C 1−A a =K , 1−B b B = e−0.01 = 0.99 , and C = 155.3 . (c) The step response for the continuous system with Gc (s) in part(a) and for the discrete system with D(z) in part (b) is shown in Figure P13.11a. (d) Utilizing the Gc (s)-to-D(z) method (with T = 0.01 second), we de- © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 725 Problems T=0.1 sec 1.8 1.6 1.4 Amplitude 1.2 1 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 20 No. of Samples FIGURE P13.11 (a) Step response for continuous and discrete systems (T=0.1s) in Parts (a) and (b). termine D(z) = C We use the relationships to compute z − 0.993 z−A = 150 . z−B z − 0.999 A = e−aT B = e−bT 1−A a C =K 1−B b A = e−0.07 = 0.993 B = e−0.001 = 0.999 C = 150 . The step response for the continuous system with Gc (s) in and for the discrete system with D(z) in part (d) is shown ure P13.11b. (e) The ramp response for the continuous system with Gc (s) in and for the discrete system with D(z) in part (b) is shown ure P13.11c. part(a) in Figpart(a) in Fig- © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 13 Digital Control Systems T=0.01 sec 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 20 40 60 80 100 120 140 160 180 200 No. of Samples FIGURE P13.11 CONTINUED: (b) Step response for continuous and discrete systems (T=0.01s) in Parts (a) and (d). T=0.1 sec 2 1.8 1.6 1.4 Amplitude 726 1.2 1 0.8 Ramp input (dashed line) 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 20 No. of Samples FIGURE P13.11 CONTINUED: (c) Ramp response for continuous and discrete systems (T=0.1s) in Parts (a) and (b). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 727 Problems P13.12 The root locus is shown in Figure P13.12. For stability: 0 < K < 2. 2 1.5 Unit circle (dashed line) 1 Imaginary Axis 0.5 0 −0.5 −1 −1.5 −2 −2 −1.5 −1 −0.5 0 Real Axis 0.5 1 1.5 2 FIGURE P13.12 z+0.5 = 0. Root locus for 1 + K z(z−1) The root locus is shown in Figure P13.13. When K = 0.027, the characteristic equation has two equal roots: z1,2 = 0.7247 and z3 = 0.2593. 2 Unit circle (dashed line) 1.5 1 0.5 Imag Axis P13.13 0 o o xx x -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 Real Axis FIGURE P13.13 z 2 +1.1206z−0.0364 Root locus for 1 + K z 3 −1.7358z 2 +0.8711z−0.1353 = 0. 1 1.5 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 728 CHAPTER 13 P13.14 Digital Control Systems The root locus is shown in Figure P13.14. When K = 9.5655 × 10−5 , the two real roots break away from the real axis at z = 0.99. For stability: K < 9.7 × 10−5 . 2 Unit circle (dashed line) 1.5 1 Imag Axis 0.5 x 0 o o x x x -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 Real Axis FIGURE P13.14 z 3 +10.3614z 2 +9.758z+0.8353 Root locus for 1 + K z 4 −3.7123z 3 +5.1644z 2 −3.195z+0.7408 = 0. P13.15 Given Gp (s) = 20 s−5 and the sample and hold (T=0.1s) as shown in Figure 13.18 in Dorf and Bishop, we determine that G(z) = 2.595 . z − 1.649 Then, with R(z) = z/(z − 1), we have Y (z) = 2.595z . (z − 1)(z + 0.9462) Therefore, Y (z) = 2.59z −1 + 0.14z −2 + 2.46z −3 + 0.26z −4 + · · ·. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 729 Problems P13.16 Given Gp (s) and the sample and hold (T=1s) as shown in Figure 13.18 in Dorf and Bishop, we determine that 0.22775z + 0.088984 . − 1.0498z + 0.049787 G(z) = z2 Then, with R(z) = z/(z − 1), we have Y (z) = 0.22775z + 0.088984 z . − 0.82203z + 0.13877 z − 1 z2 The plot of y(kT ) is shown in Figure P13.16. 1 0.9 0.8 0.7 y(kT) 0.6 0.5 0.4 0.3 0.2 0.1 0 1 2 3 4 5 6 7 kT FIGURE P13.16 Plot of y(kT ) for a step input. P13.17 The root locus is shown in Figure P13.17 for 1+K 0.39532z + 0.30819 =0. − 1.4724z + 0.47237 z2 The limiting value of the gain for stability is K = 1.71. 8 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 730 CHAPTER 13 Digital Control Systems Root Locus 2 1.5 Imaginary Axis 1 0.5 0 −0.5 −1 Unit circle (dashed line) −1.5 −2 −5 −4 −3 −2 −1 Real Axis 0 1 2 FIGURE P13.17 = 0. Root locus for 1 + K z 20.39532z+0.30819 −1.4724z+0.47237 P13.18 The plot of the step responses for 0 ≤ T ≤ 1.2 is shown in Figure P13.18. The overshoot and settling time summary is given in Table P13.18. T P.O. Ts 0 0.2 0.4 0.6 0.8 1.0 1.2 16.3% 20.6% 25.6% 31.3% 36.9% 40.0% 51.0% 8.1 8.4 8.8 11.4 14.4 16.0 19.2 TABLE P13.18 Performance summary. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 731 Problems 1.6 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 10 20 FIGURE P13.18 Step responses for 0 ≤ T ≤ 1.2. 30 40 50 60 No. of Samples 70 80 90 100 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 732 CHAPTER 13 Digital Control Systems Advanced Problems AP13.1 Given the sample and hold with Gp (s), we determine that G(z) = 10.5K(z − 0.9048) . (z − 1)2 The root locus is shown in Figure AP13.1. For stability: 0 < K < 0.2. 2 Unit circle (dashed line) 1.5 1 Imag Axis 0.5 0 ox -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 Real Axis FIGURE AP13.1 10.5(z−0.9048) = 0 with unit circle (dashed line). Root locus for 1 + K (z−1)2 The root locus is shown in Figure AP13.2a. The loop transfer function is 2 Unit circle (dashed line) 1.5 1 0.5 Imag Axis AP13.2 0 o x x -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 Real Axis FIGURE AP13.2 0.0379z = 0. (a) Root locus for 1 + K (z−1)(z−0.368) 0.5 1 1.5 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 733 Advanced Problems G(z)D(z) = K 0.0379z . (z − 1)(z − 0.368) For stability: Kmax = 72. We select K = 8.2. The step response is shown in Figure AP13.2b. 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 6 7 8 9 10 No. of Samples FIGURE AP13.2 CONTINUED: (b) Step response with K = 8.2. The root locus is shown in Figure AP13.3a. The maximum gain for Root Locus 2 Unit circle (dashed line) 1.5 System: sysz Gain: 5.99 Pole: 0.736 + 0.257i Damping: 0.596 Overshoot (%): 9.74 Frequency (rad/sec): 8.36 1 Imaginary Axis AP13.3 0.5 0 −0.5 −1 −1.5 −2 −3 −2.5 −2 −1.5 −1 −0.5 Real Axis FIGURE AP13.3 (a) Root locus for 1 + K z0.07441z+0.06095 2 −1.474z+0.6098 = 0. 0 0.5 1 1.5 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 734 CHAPTER 13 Digital Control Systems stability is Kmax = 44.3. We select K = 6. The step response is shown in Figure AP13.3b. Step Response 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 Time (sec) 0.8 1 1.2 FIGURE AP13.3 CONTINUED: (b) Step response with K = 6. AP13.4 The loop transfer function is G(z) = 10(1 − e−T ) , z − e−T and the closed-loop transfer function is T (z) = 10(1 − e−T ) . z − (11e−T − 10) For stability, we require |11e−T − 10| < 1 . Solving for T yields 0 < T < 0.2 . Selecting T = 0.1s provides a stable system with rapid response; the settling time is Ts = 0.2s. The step response is shown in Figure AP13.4. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 735 Advanced Problems 1 0.9 0.8 Amplitude 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1 1.5 2 2.5 3 3.5 4 No. of Samples FIGURE AP13.4 Step response with T = 0.1s. The maximum gain for stability is Kmax = 63.15. Root Locus 2 Unit circle (dashed line) 1.5 1 Imaginary Axis AP13.5 0.5 System: sysz Gain: 63.2 Pole: 0.725 − 0.686i Damping: 0.00308 Overshoot (%): 99 Frequency (rad/sec): 7.58 0 −0.5 −1 −1.5 −2 −3 −2.5 −2 −1.5 −1 −0.5 Real Axis FIGURE AP13.5 Root locus for 1 + K 0.004535z+0.004104 z 2 −1.741z+0.7408 = 0. 0 0.5 1 1.5 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 736 CHAPTER 13 Digital Control Systems Design Problems The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: 26.035 , s(s + 33.142) Gp (s) = where we neglect the motor inductance Lm and where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). Letting G(z) = Z G≀ (∫ )G√ (∫ ) we obtain G(z) = 1.2875e − 05(z + 0.989) . (z − 1)(z − 0.9674) A suitable controller is D(z) = 20(z − 0.5) . z + 0.25 The step response is shown below. The settling time is under 250 samples. With each sample being 1 ms this means that Ts < 250 ms, as desired. Also, the percent overshoot is P.O. < 5%. 1.2 1 0.8 Amplitude CDP13.1 0.6 0.4 0.2 0 0 50 100 150 No. of Samples 200 250 300 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 737 Design Problems (a) Given the sample and hold with Gp (s), we determine that KG(z) = K 0.1228 . z − 0.8465 The root locus is shown in Figure DP13.1a. For stablity: 0 ≤ K < 15. Unit circle (dashed line) 1.5 1 0.5 Imag Axis DP13.1 0 x -0.5 -1 -1.5 -1.5 -1 -0.5 0 0.5 1 1.5 Real Axis FIGURE DP13.1 0.1228 (a) Root locus for 1 + K z−0.8465 = 0 with unit circle (dashed line). (b) A suitable compensator is Gc (s) = 15(s + 0.5) . s+5 Utilizing the Gc (s)-to-D(z) method (with T = 0.5 second), we determine D(z) = C z−A z − 0.7788 = 6.22 . z−B z − 0.0821 We use the relationships A = e−aT , B = e−bT , and C 1−A a =K , 1−B b to compute A = e−0.5(0.5) = 0.7788 , B = e−0.5(5) = 0.0821 , and C = 6.22 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 738 CHAPTER 13 Digital Control Systems (c) The step response is shown in Figure DP13.1b. 0.8 0.7 0.6 Amplitude 0.5 0.4 0.3 0.2 0.1 0 0 2 4 6 8 10 12 14 16 18 20 No. of Samples FIGURE DP13.1 CONTINUED: (b) Closed-loop system step response. DP13.2 With the sample and hold (T=10ms), we have G(z) = 0.00044579z + 0.00044453 . z 2 − 1.9136z + 0.99154 A suitable compensator is D(z) = K z − 0.75 , z + 0.5 √ where K is determined so that ζ of the system is 1/ 2. The root locus is shown in Figure DP13.2. We choose K = 1400. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 739 Design Problems Root Locus 1.5 Curve of constant zeta=0.707 (dashed line) 1 Imaginary Axis 0.5 0 −0.5 −1 −1.5 −5 −4 −3 −2 −1 Real Axis 0 1 2 FIGURE DP13.2 0.00044579z+0.00044453 Root locus for 1 + K z−0.75 = 0. z+0.5 z 2 −1.9136z+0.99154 The root locus is shown in Figure DP13.3a. 2 Curve of constant zeta=0.707 (dashed line) 1.5 1 0.5 Imag Axis DP13.3 0 o x x 0.5 1 -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 1.5 2 Real Axis FIGURE DP13.3 z+1 (a) Root locus for 1 + K (z−1)(z−0.5) = 0. The gain for ζ = 0.707 is K = 0.0627. The step response is shown in Figure DP13.3b. The settling time is Ts = 14T = 1.4s and P.O. = 5%. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 740 CHAPTER 13 Digital Control Systems 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 No. of Samples FIGURE DP13.3 CONTINUED: (b) Step response with K = 0.0627. With the sample and hold (T=1s), we have G(z) = 0.484(z + 0.9672) . (z − 1)(z − 0.9048) 2 Curve of constant zeta=0.5 (dashed line) 1.5 1 0.5 Imag Axis DP13.4 0 o x ox x -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 Real Axis FIGURE DP13.4 0.484(z+0.9672) (a) Root locus for 1 + K z−0.88 z+0.5 (z−1)(z−0.9048) = 0. 1 1.5 2 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 741 Design Problems 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 No. of Samples FIGURE DP13.4 CONTINUED: (b) Step response for K = 12.5. A suitable compensator is D(z) = K z − 0.88 , z + 0.5 where K is determined so that ζ of the system is 0.5. The root locus is shown in Figure DP13.4a. We choose K = 12.5. The step response is shown in Figure DP13.4b. Also, Kv = 1, so the steady-state error specification is satisfied. DP13.5 Select T = 1 second. With the sample and hold, we have G(z) = 0.2838z + 0.1485 . − 1.135z + 0.1353 z2 The root locus is shown in Figure DP13.5. To meet the percent overshoot specification, we choose K so that ζ of the system is 0.7. This results in K = 1. The step response has an overshoot of P.O. = 4.6%. Also, from Figure 13.21 in Dorf and Bishop, we determine that the steady-state error to a ramp input is ess = 2 (since T /τ = 2, and Kτ = 0.3). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 742 CHAPTER 13 Digital Control Systems 2 Curve of constant zeta=0.7 (dashed line) 1.5 1 Imag Axis 0.5 0 o x x -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 Real Axis FIGURE DP13.5 Root locus for 1 + K z 20.2838z+0.1485 −1.135z+0.1353 = 0. DP13.6 With the sample and hold at T = 1 , we have G(z) = z2 Consider the digital controller 0.298z + 0.296 . − 1.98z + 0.9802 Dz) = K z − 0.9 . z + 0.6 The root locus is shown in Figure DP13.6. To meet the percent overshoot specification, we choose K so that ζ of the system is greater than 0.52. We select K = 2. The step response has an overshoot of P.O. = 11.9% and the settling time is Ts = 17.8s. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 743 Design Problems Root Locus 1 0.8 0.6 Imaginary Axis 0.4 0.2 0 −0.2 −0.4 −0.6 −0.8 −1 −5 −4 −3 −2 −1 Real Axis 0 1 2 FIGURE DP13.6 0.298z+0.296 Root locus for 1 + K z−0.9 z+0.6 z 2 −1.98z+0.9802 = 0. Step Response 1.4 System: syscl Peak amplitude: 1.12 Overshoot (%): 11.9 At time (sec): 2 1.2 System: syscl Settling Time (sec): 17.8 Amplitude 1 0.8 0.6 0.4 0.2 0 0 5 10 15 20 Time (sec) FIGURE DP13.6 CONTINUED: (b) Step response for K = 2. 25 30 35 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 744 CHAPTER 13 Digital Control Systems Computer Problems CP13.1 The m-file script and unit step response are shown in Figure CP13.1. num=[0.2145 0.1609]; den=[1 -0.75 0.125]; sysd = tf(num,den,1); step(sysd,0:1:50) 1.2 1 Amplitude 0.8 0.6 0.4 0.2 0 0 5 10 15 20 25 30 35 40 45 50 No. of Samples FIGURE CP13.1 Step response. CP13.2 The m-file script utilizing the c2d function is shown in Figure CP13.2. % Part (a) num = [1]; den = [1 0]; T = 1; sys = tf(num,den); sys_d = c2d(sys,T,'zoh') % % Part (b) num = [1 0]; den = [1 0 2]; T = 1; sys = tf(num,den); sys_d=c2d(sys,T,'zoh') FIGURE CP13.2 Script utilizing the c2d function for (a) and (b). Transfer function: 1 ----z-1 Transfer function: 0.6985 z - 0.6985 -----------------z^2 - 0.3119 z + 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 745 Computer Problems % Part (c) num = [1 4]; den = [1 3]; T = 1; sys = tf(num,den); sys_d = c2d(sys,T,'zoh') % % Part (d) num = [1]; den = [1 8 0]; T = 1; sys = tf(num,den); sys_d = c2d(sys,T,'zoh') Transfer function: z + 0.267 ----------z - 0.04979 Transfer function: 0.1094 z + 0.01558 ------------------z^2 - z + 0.0003355 FIGURE CP13.2 CONTINUED: Script utilizing the c2d function for (c) and (d). The continuous system transfer function (with T = 0.1 sec) is T (s) = s2 13.37s + 563.1 . + 6.931s + 567.2 The step response using the dstep function is shown in Figure CP13.3a. The contrinuous system step response is shown in Figure CP13.3b. 1.8 1.6 1.4 1.2 Amplitude CP13.3 1 0.8 0.6 0.4 0.2 0 0 2 4 6 8 No. of Samples FIGURE CP13.3 (a) Unit step response using the dstep function. 10 12 14 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 746 CHAPTER 13 Digital Control Systems 1.8 * 1.6 1.4 * 1.2 * 1 * * * * * * * * * 0.8 * * 0.6 0.4 0.2 0* 0 0.2 0.4 0.6 0.8 1 1.2 1.4 FIGURE CP13.3 CONTINUED: (b) Continuous system step response (* denote sampled-data step response). The root locus in shown in Figure CP13.4. For stability: 0 < K < 2.45. Root Locus 2 1.5 1 Imaginary Axis CP13.4 0.5 0 −0.5 −1 −1.5 −2 −2 −1.5 −1 FIGURE CP13.4 z Root locus for 1 + K z 2 −z+0.45 = 0. −0.5 0 Real Axis 0.5 1 1.5 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 747 Computer Problems CP13.5 The root locus in shown in Figure CP13.5. For stability: 0 < K < ∞. 1 0.8 0.6 Imag Axis 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 Real Axis 0.4 0.6 0.8 1 FIGURE CP13.5 (z−0.2)(z+1) Root locus for 1 + K (z−0.08)(z−1) = 0 The root locus is shown in Figure CP13.6. Root Locus 1.5 1 Imaginary Axis CP13.6 0.5 0 0.5 1 1.5 1.5 1 0.5 0 Real Axis FIGURE CP13.6 Root locus. 0.5 1 1.5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 748 CHAPTER 13 Digital Control Systems We determine the range of K for stability is 0.4 < K < 1.06. % Part (a) num=[1 4 4.25 ]; den=[1 -0.1 -1.5]; sys = tf(num,den); rlocus(sys), hold on xc=[-1:0.1:1];c=sqrt(1-xc.^2); plot(xc,c,':',xc,-c,':') hold off % % Part (b) rlocfind(sys) rlocfind(sys) ÈSelect a point in the graphics window selected_point = -0.8278 + 0.5202i ans = 0.7444 Kmax Select a point in the graphics window selected_point = -0.9745 - 0.0072i ans = 0.3481 Kmin FIGURE CP13.6 CONTINUED: Using the rlocus and rlocfind functions. Using root locus methods, we determine that an acceptable compensator is Gc (s) = 11.7 s+6 . s + 20 With a zero-order hold and T = 0.02 sec, we find that 1.2 1 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 0.8 Amplitude CP13.7 * * * 0.6 * * 0.4 * * 0.2 * * 0* * 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Time (sec) FIGURE CP13.7 System step response (* denotes sampled-data response). 0.8 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 749 Computer Problems D(z) = 11.7z − 10.54 . z − 0.6703 The closed-loop step response is shown in Figure CP13.7.

Source Exif Data:

File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.6 Linearized : Yes Page Mode : UseOutlines XMP Toolkit : 3.1-701 Create Date : 2010:08:05 08:48:05-05:00 Creator Tool : dvips(k) 5.98 Copyright 2009 Radical Eye Software Modify Date : 2010:10:25 08:50:47+02:00 Metadata Date : 2010:10:25 08:50:47+02:00 Producer : GPL Ghostscript 8.70 Format : application/pdf Title : Instructor's Solutions Manual for for Modern Control Systems, 12th Ed Creator : Richard C. Dorf, Robert H. Bishop Subject : 9780136024989 Document ID : uuid:107e56d4-5005-604d-a266-e210e0c618a9 Instance ID : uuid:3e3e4aa5-ad61-4d55-bcd8-16fef4e09972 Has XFA : No Page Count : 754 Page Layout : SinglePage Author : Richard C. Dorf Keywords : 9780136024989EXIF Metadata provided by EXIF.tools