1620_Users_Group_Western_Region_Minutes_of_the_Meeting_196312 1620 Users Group Western Region Minutes Of The Meeting 196312
1620_Users_Group_Western_Region_Minutes_of_the_Meeting_196312 1620_Users_Group_Western_Region_Minutes_of_the_Meeting_196312
User Manual: 1620_Users_Group_Western_Region_Minutes_of_the_Meeting_196312
Open the PDF directly: View PDF .
Page Count: 281
Download | |
Open PDF In Browser | View PDF |
1620 USERS GROUP WESTERN REGION MINUTES OF THE MEETING DECEMBER 11-13, 1963 TEMPE, ARIZONA ROBERT R. WHITE WESTERN REGION SECRETARY I nom C . • •n .. ·"····· ... ·"'' ' ' '".,'' ' ' ' '····,'' ' ' '·U'' ' ' ' ' ' .........._ _ _ _ _ _ __ CONTENTS 1. Letter from Robert A. Ebert to 1620 Users Group 2. Roster of Attendees 3. Minutes of the Seventh Meeting of the Western Region 1620 Users Group 4. Agenda and Abstracts 5. The Impact of Automation on the Professional Engineer; Melford E. Monsees 6. Expotential and Sinusoidal Curve Fitting; E. P. Hilar 7. Automated Design Engineering; W. W. Rogers 8. A Payroll and Labor Distribution Program Package; Elias C. Tonias 9. The lISPIREII System; Gary J. Reed 10. A 1620 Program for Minimization of Boolean FUnctions, Expressed as Sums of Minterms; Thomas R. Hoffman 11. Critical Speed, Stress, and Bearing Reaction Calculations for a General Shaft, Using Numerical Integration; Ralph B. Bates 12. Three Dimensional Surface Fit; David G. Kitzinger 13. Maximum Likelihood Resolution of tion; Reimut Wette, D.Sc. 14. Comparison of Two Methods of Finding Significant Contributors in Multiple Regression; M. J. Garber 15. Network Analysis; H. N. TYson, Jr. 16. An Integrated Earth-work System; Cecil L. Ashley ~~o Mixed Normal Distribu- 17. Hydro System Daily Operation Analysis; C. R. Hebble o fI' ,I,l,l"!i.'"&un,!,n '''WLWN IMiL 1620 PRESIDENT SECRETARY-TREASURER us ERS J. L. DAVIDSON .(lJ Long Island Lighting Co. 175 Old Country Road MARLENE T. METZNER Pratt & Whitney Aircraft Co. Fla. Research & Dev. Center West Palm Beach, Florida GROUP~ Hicksville, New York WESTERN REGION PRESIDENT MID-WESTERN REGION PRESIDENT EASTERN REGION PRESIDENT ROBERT EBERT Spectrol Electronics Corp. 1704 S. Del Mar Avenue San Gabriel, California W. A. BURROWS Dravo Corporation Neville Island Pittsburgh 25, Pennsylvania J. R. OLIVER Univ. of Southwestern La. Box 133, USL Station laFayette, Louisiana CANADIAN REGION PRESIDENT EUROPEAN REGION PRESIDENT D. A. JARDINE Dupont of Canada Research Center Kingston, Ontario H. TOMPA European Research Associates 95 Rue Gatti De Gamond Bruxelles 18, Belgium December 26, 1963 To the Members of the 1620 Users Group: It is with sincere regret that I must announce my resignation, effective during this meeting, as President of the Western Region. The executive council of the group, acting in accord with the by-laws, has appointed Paul Bickford to serve the remainder of the current termo Paul's. appointment was made upon my recommendation, and I feel that his experience with the group will make him eminently qualified to continue its growth and activity. In turning the administration of the Western Region over to Paul, I can only say that I shall sincerely miss working with all of you, I express my thanks and appreciation to those of you, too numerous to mention individually, who have worked to make the meetings which I have conducted successful, and I extend my best wishes to the officers and members of the 1620 Users Group for the continued growth and success which you have enjoyed over the last frour years o Sincerely, ~ /? f'" ,.,"'.--, Cr;~~?1~ i ~" 'Rob'ert A,. Ebert o 3 -'" ... . 1620 USERS GROUP MEETING TEMPE, ARIZONA DECEMBER 11-12,13. 1963 ROSTER OF ATTENDEES USERS GROUP NO. 1118 1118 INSTALLATION REPRESENTATIVE NANCY PAQUIN U.S. PUBLIC HEALTH SERVICE ROCKVILLE, MARYLAND G. ROBERT ORNDORFF U.S. PUBLIC HEALTH SERVICE ROCKVILLE, MARYLAND 1216 CARLIS TAYLOR A F R R I BETHESDA, MARYLAND 1302 T. R. HOFFMAN UNION COLLEGE SCHENECTADY. N. V. 1334 DR. REIMUT WETTE M.,O. ANDERSON HOSPITAL. AND TUMOR INST. HOUSTON, TEXAS 1346 ELIAS C. TONIAS ERDMAN AND ANTHONY - CONS. ENGRS. ROCHESTER, NEW YORK 3082 PAUL BICKFORD OKLAHOMA UNIV. NED. RESEARCH COMP. CTR. OKLAHOMA CITY. OKLA,. 3082 CARA MITCHELL OKLAHOMA UNIV. MEO. RESEARCH CQIIP,. CTR.. OKLAHOMA CITY, OKLA. 3166 OR. HERMAN 8. WEISSMAN UN1VERS I TV OF I L.L lNO I S CHICAGO, ILLINOIS 3175 HELEN LtGON BAYLOR UNIVERSITY WACO. TEXAS 3182 '3240 ROBERT C;. LANGE AUTOMATIC ELECTRIC LABORATORJ'£'S. ' INC. NORTHLAKE, ILLINOIS MELFORD E. MONSEES U.S. ARMY ENGR. DISTRICT KANSAS C,I TY • 3261 M I SSOUR I GREGORY ..J. SHANAHAN CECO STEEL PRODUCTS CORP. CICERO. ILLINOIS . .' USERS GROUP NO. 3273 I NSTAU...AT ION REPRESENTATIVE R08ERT C. BABIONE ACIC ST. 3273 L.OVI,S. MISSOURI CHARL.ES WEISS USAF AERONAUT leAL, CHART AND I NFO CENTER ST. LOUJ S,. M 1 SSOUR 1 5001 R. C. WENRICK ACF INOUSTRI'ES. INC. ALBUQUERQUE. NEW MEX ICO 5001 G. ..J. REED ACF lNOUSTRlES. INC. AL.BUQUERQUE. NEW MEX l'CO R. e. WE'AY'ER BEAR CREEk MINING co. SALT LAKE CITY,UTAH 5014 WALTER DAVIS GENERAL OYNAMI:'CS/ASTRONAUT lC5 SAN D'IEGO. CAL.IFORN IA 5016 RICHARD W. PUGSLEY COMPUTEAMAT INC LOS ANGELES. CALIFORNIA 5019 EDGAR M. BLIZZARD .JET PROPULSJONLAB PASADENA, CALIFORNI,A 5020 MARI:LYN DOIG COLORADO STATE UN IVERS I TV FORT COt.LJ:'N$,. COL:QRADO 5021 ..J. W. LAFON us ARMY EN'GINEER D:ISTRICT AL.BUQUERQUE ALBUQUERQUE,. NEW MEX I CO 5027 MARVIN RUBINSTEIN ELECTRO OPTICAL SYSTEMS. INC. PASADENA, CAL.IFORNIA 5032 BOB 'MAMtING GOOD'Y~AR AEROSPACE LITCHF'-ELO PARK, ARIZONA 5032 N. A. KUFFEL GOODYEAR AEROSPACE LITCHFIELD PARK, ARIZONA 5032 oJ. MOSS GOODYEAR AEROSPACE LITCfPIELD PARK, ARIZONA S- • o USERS GROuP MO. 5032 INSTALLATION REPRESENTATIVE DAVI'D H. o I HERREN GOODYEAR AEROSPACE CO L.ITCtFl£LD PAAc: ARIZ 5041 DAVID KEY MOTOAOLA PHOENIX. ARIZQNA 5045 w. 5053 ROBERT L.. SHUTT SACRAMENTO PEAK WILCOXSON U.S. NAVAL CIVIL ENGR. LAB. PORT HUEPEME. CAL IFORNIA OBSERVATORY SUNSPOT. NEW MEXICO 5053 FRANK BJ,RI) SACRAMENTO PEAK OBSERVATORY SUNSPOT. NEW MEXICO 5057 DAVID G. K1TZINGER SAND'IACORP ALBUQUERQUE, NEW MEXICO 5057 EL.IZABETH L.FROST SAN))A CORP. ALBUQUERQUE. NEW MEXICO 5058 BOB BABCOCK SUNDSTRAN) AVIATION DENVER DENVER. COLORADO 5064 ALLENL. GRAVITT SIGNAL., OIL AND GAS CO. LOS ANGELES. CALIFORNIA 5065 ROBERT W. WILLSON SALT RIVER PROJECT PHOE,NIX. ARIZONA 5065 EAJEST NICHOLS SALT RIVER PROJECT PHCENIX. ARIZONA 5065 MAX A. MAYES SALT RIVER PRGUECT PHOENIX. ARIZONA 5078 o OR. MOAR I S ..J. GARBER UN I V. OF CAL t;PORN tA AT RIVERSIDE RIVERSIDE. CALlFORNIA 5078 THOMAS M. LITTLE UNIVERSITY OF CALIFORNIA AT RivERSIDE RIVERSIDE. CALIFORNIA b USERS INSTALLATION GROUP NO. REPRESENTATIVE 5079 5084 SAMUEL K. PRINGLE MAGNOLJ'A P'IPE LINE CO. DAL.LAS,. TEXAS E. B. LOOP UNION OIL CO .OP" CALIF RODEO. CALIFa.tI,A 508S ROBERT 0. MOFP'l TT U.S. AMY ENGR. DIVISION. N.P. PORTLAND. OREGON 5089 DONALD ..I. MART I N U.S. PUBLIC MEALTH SERVICE LAS VEGAS. NEVADA !5089 DAVft) L. BAER u.s. PUBLIC-MEAL.TH SERVICE LAS VEGAS,. NEVADA 5095 HARRY D,. RENICK WEYERHAEUSER CO TACOMA .WASttINGTON 5096 5096 e,. MOIiiIR I S BUREAU orr REQ.AMATION SACRAMEMTO,. CALIFORNIA BOYD u.s. R., SRUCE 8AOW~I" U.S. BUREAU OF REQ..AMATIQN SA'CRAMENTO. CALlfrOANIA 5099 BEVERLY DOl,S WSMR WHITE SAM>S. NEW NEXt,cO 5'104 5120 ~. A. GUI ... TEXAS COLLEGE OF ARTS +I'NoUSTRI£S I( I NG;SV ILLE • TEXAS ' WILLI.b' L. REUTER ,SO. DAJ«)TA SCHOOL OF MINES AND TECM_ RAPID CITY. SOUTH DAKOTA 5'125 5133 CHIN MO LEE UTAH POWER AND LIGHT CO. SALTLAKE CITY. UTAH -'OSE RAMIREZ MASON AND • .NGER - S fLAS MASON CO.. t HC o AMARILLO. TEXAS 51,39 RICHARD A. HARR I S NORTH TEXAS STATE UNIVERSITY DENTON. TEXAS 7 . _ - ... --_._ .. USERS c INSTALLATION REPRESENTATIVE GROUP NO. 5144 .JAMES ..J. STANLEY U S WEATHER BUREAU RFC SACRAMENTO. CALIFORNIA 5146 a I L.L. WOLLENiAUPT GOLDSTONE TRACkI." STATION BARSTOW. CALIFORNIA JOE SIBL!:Y GOLDSTONE TRACKING STATION BARSTOW. CALlfI'ORNIA 5150 SAM THOMPSON 51'!50 GEORGE A. LARCAOE HALL I BURTON CO DUNCAN. OICLAHOMA 5158 HALL IBURTON co· DUNCAN. OKLAHOMA A.-6AALAN) NECHES BUTANE PAODUCTS PORT NECHES. TEXAS 5165 co'. CHARLES R. HE88l.E CORPS OFENGJNEERS WAU-A WALLA. WASHt·NGTON 5165 cec 1 L. L. ASHLEY U S ARMY CORPS OF ENG.1ME£R WALLA WALLA WASH 5179 5181 L. E. MARYEY fl'OOTHt u... CGLLEGE LOS AI... TOS H'ILLS. CAL·J'FOANIA ROBERT R. _ITE LOS ANGELES DEPT. OF WATER AND POWER LOS ANGELES. CALIFORNIA 5183 JOSEPH P. SNOW UNIVERSITY OF WYOMlHG LARAMIE. WYOMING 5190 WILLIAM G. LANE CHICO STATE COLLEGE CHICO, CALIFORNIA 5195 o 5199 S.V·. BURKS,. JR. PITTSBURGH PLATE GLASS CORPUS CHRISTI. TEXAS CHARLES S. WALKER SCHOOL. OF ENGIN!ERtNG TEMPE. AR J ZONA co. - CHIEM. DIV. ., USI!RS INSTALLATION GROUP NO. REPRESENTATIVE 5210 A....L. NIEHUES PALO ALTO UNI:FI"I!D SCHOOL. DISTRICT PALO ALTO. CALIFORNIA !5210 ARt.. I'HE K. ICAPPHAHN PALO ALTO UNIPIED SCHOOL DISTRICT PALO ALTO. CALIFORNIA 5211 KEe.tETH ICR lEGE CAL.IFORNIA STATE .....VTECHN·IC CCLLEGE POJiIONA t CAL..I;P'ORN IA 5215 AOSE'MARYPETERSEN UCLA - WESTERN DATA PROCESSING CENTER LOS ANGELES. CALlFORNIA 7007 A. C. R. NEWBERRY UHIV£RSITV OF ALBERTA CALGARY ALBERT A. CANADA SETH P. EVANS PHOENI X COLLEGE PHOENIX. ARIZONA WARREN BUXTOH PHOEN I X COL.L.EE PHOENIX,. ARIZ"A .JOHN P. MCCALL! STER U S WEATHER BUREAU FT. WORTH. TEXAS IBM W. H. DUKELOW IBM KANSAS C.ITY. MISSOURI IBM CHARLES E·. BEARY IBM - WESTERN REGION OFFICE L.OS ANGEL.ES. CALliFORNIA IBM ROBERT A. EBERT IBM - WESTERN DATA PROCESSING CENTER LOS ANGELES. CALIFORNIA IBM JAMES E. MCRGAN IBM - WESTERNRSIOH OFFICE LOS ANGELES-. CALIFORNIA IBM BRUCE ~. SOCKS 1.814 CORP. CHICAGO. I t.L INC! S IBM ANGELO ARENA IBM WHITE PLAINS. N.V. o I I, , 'ttWNHt'" IW!' 'II I" fl'''l I! IUII'fllli"!, USERS GROUP NO. l·eM INSTALLATION REPRESENT AT I VE GERALD R _HOGSETT IBM - WESTERN REGION OFFICE LOS ANGELES. CALIFORNIA o /0 o til t WESTERN REGION 1620 USERS GROUP MINUTES OF THE SEVENTH MEETING The group assembled on the campus of Arizona State University in Tempe, Arizona, December 11, 12, and 13, 1963. All papers listed In the agenda were presented as scheduled except paper A-l, Generalized SPS Routines for Handling Simple Problems, by Thomas L. Yates, which was not given, paper A-2, Expotential and Sinusoidal Curve Fitting, by E. P. Hilar, which was given at Technical Session "Gil, and paper G-I, Ray Trace Program for a General Lens System, by D. H. 0 'Herren, which was given at Technical Session II A" • Paper E-1, Network Analysis, was presented by Mr. Gerald Hogsett, IBM, Los Angeles. Paper C-2 was read by Bob Ebert. Copies of the abstracts and papers presented are enclosed with the exception of Paper G-l which will be included with minutes of a later meeting. The general meeting was opened on December 11 by President Bob Ebert and after opening remarks was turned over to Jim Morgan of IBM who introduced the other IBM representatives to the group. The sound-off session was moved up to this time to fill in the time for paper A-I. A summary of questions raised and answers to them follows the minutes of the business meeting. Business Meeting December the floor in order referred The business meeting was opened by Bob Ebert on 12 and under' old business, the request was made from that the Roster of Members be arranged alphabetically of the name of the installation. The request was to Angelo Arena of IBM. The first item of new business was the resignation of Bob Ebert as president of the Western Region 1620 Users Group. This was necessary under the by-laws as he is no longer connected with a 1620 installation. A letter by Bob to the Users Group is included with the minutes of the meeting. The executive council, acting under the provisions of the, by~laws, apPointed Paul Bickford, Western Region Secretary, to complete the unexpired term of office of President, and he apPointed Robert R. White to fill the vacant position of Secretary. o With Paul now presiding as President the new business continued with the selection of Denver, Colorado, as the site, II .... '\ 2 and the third week in June, 1964, as the probable date of the next meeting. It was announced that the next fall meeting will be a joint meeting with the Midwestern Region, and will probably be held on the campus of Oklahoma University at Norman, Oklahoma, on November 9 or 22. The program committee will distribute the information regarding dates and agenda as early as possible, and they request abstracts of papers be sent in as soon as possible to aid them. A list of pre-registrants will be available early on the first day of the meeting. Ci Paul then announced results of the Executive Council meeting at Pittsburgh and they are as follows: 1. Jim Davidson is soliciting suggestions for methods for removing programs from the library. A criterion for value of the program and a means for the removal of inadequate programs is needed. 2. The cost of running the users group has risen. Dues paid by each region to the National 1620 Users Group were $.25 per registrant at each regional meeting. This has now been raised to $.50. Most of this increased cost has been in publication of the Newsletter. This must now be distributed to over 1,000 members while the average number of registrants at each meeting has not risen. It was stated that the registration fees for the meetings may have to be increased. This concluded the new business and the busfness meeting was adjourned. After an excellent luncheon, Mr. Melford E. Monsees, ADP Co-ordinator" U.S. Army Engineer District, Kansas City, Missouri, spoke on "The Impact of Automation on Professional Engineering." A copy of the talk is enclosed. Two special interest groups met for discussion of mutual problems. The Civil Engineering group was presided over by Elias C. Tonias of Erdman and Anthony, Consulting Engineers, Rochester, New York, and minutes of the discussion are enclosed. The Power Utilities Engineer group was very informal but of benefit to all participants. Sound-Off Session and Comments It was announced that Version 2 of FORTRAN/FORMAT is available and has these advantages over Version 1 1. Multiple Format Specifications (also in the Pre-Compiler) I~ o I I, [, .,1 . ii""""',',',.\0 c ".'.,"~o.',\!on" ..... j,t~ ••11:.+....... , .... /w'·..JJ,,'W·"'JliJ.l,MlIMu.". ,,··Jill' . '" "1""'9: i 1\· "·· .. ill:: ,.. '4¥i+M::j'··. .·iI+:w,w,WfW·rr···,;!WIfli,·Tiji"n·-·.5!""·'m- 3 2. Source decks are available (on magnetiC tape) 3. There are more subroutine options. Ninety hours or more are required for field installation of 1311 disk drives. Some operating difficulties with the disk were discussed and these pOints were brought out. 1. In order to use a FORTRAN lID or SPS lID object deck the Monitor I must be on a working drive. This is because no loader or subroutines are punched. 2. There is no current provision in Monitor I to load object decks to disk unless they are Monitor I compiled. It is therefore necessary to recompile all programs. There is a great deal of interest in FORTRAN/FORMAT. Requests were made for a disk version and a version to batch compile. It can now be changed to allow free form input similar to the first 1620 FORTRAN compiler. Chuck Berr.y of IBM Wilshire office can supply instructions for this change. Requests were made for the following from IBM: c 1. Relocatable I/O and arithmetic subroutines for FORTRAN so only those used would be called. 2. A reduction in nOise level on the Model 2 1622. 3. A FORTRAN pre-compiler which would give indication of the amount of storage the compiled program takes. 4. A 5. A loader for Systems Output Format decks punched under Monitor I control. 6. FORTRAN processors to take advantage of extended machine capabilities. . 7. A Report Program Generator for the 1620-14431311 combination. 8. COBOL for 1620 9. ALGOL tor 1620 10. Tabulate command in FORTRAN. FORTRAN IV. 13 --~-----------.------ --------- ------~- ",'. 4 Some of these requests were answered in the comments session by Jim Morgan and the other IBM representatives. The questions answered are: 1. Processors specifically designed for larger machine configurations are being investigated. 2. There are no present plans for IBM written COBOL, ALGOL, or Report Program Generators, however, there is an ALGOL processor available from Southern Illinois University. 3. The request for FORTRAN IV was noted and Jim asked for more opinions on the need for 1t. The new hardware announced for the 1620 since the last Western Region meeting is the CalComp Plotter in two models, binary capabilities for the Model 2 - 1620, and index registers for the Model 2 - 1620. It was noted that the FORTRAN II sine subroutine does not handle small angles correctly. IBM programming systems is working on the problem now. Angelo Arena discussed the program library and said that the KWIC index is still in the process of being developed and may be changed even further, so that it will be easier to find programs by number. It will be published every six months with supplements issued every month. It will be 3-hole punched. Angelo requested that when programs are ordered, unless the user 1s sure the program f1ts his need, he order only the documentations. IBM can supply this quickly while program decks and tapes take longer. More descriptive titles for programs would also help this p~oblem. There is now available from IBM a manual which lists the RPQ1s available (A26-5799). Some of these mentioned were: 1. The ab1lity to punch one character on paper tape. 2. A real time clock. The clock IBM is now installing on all equipment is an elapsed time clock only and is not addressable. 3. An future. addressable IR-2. Version 2 of FORTRAN II will be available in the near The 1443 print commands are handled in the same manner as 7000 series machines with Column 1 being used for carriage control. In order to get IBM publications without fail, have the IBM Systems Eng1neer put the installation name on the SRL o 5 list. Mailing is then automatic. Be sure that the installation receives the 1620 Bibliography, A26-5692. The meeting was concluded with a demonstration of a Model 2 - 1620 with a 1311 disk file, on the evening of December 12, at the IBM Phoenix office and tutorial sessions on December 13 for MONITOR 1, FORTRAN II, and SPS 1620/1710. o "'. o ;", . ", .!!C .. rj.t·H··d·-..,.:.····++w.:Jki .. ~·-+I,j·r:jWWi:·'· j,J+jHj...... j _. ·--·I!ii"[·"j·jjlw·· w-··'f'lnTmY"jW,ljrm,. .. o 1620 USERS GROUP WESTERN MEETING PHOENIX, ARIZONA, WINTER 1963 CIVIL ENGINEERING SESSION WEDNESDAY. DECEMBER 11. 1963. 3:30 PM The highlights of this session were discussions on the disc file, the use of Fortran VB. SPS in Civil Engineering programs, the introduction of COGO and the use of the CALCOMP plotter. It was the general consensus of the participants that SPS provided a better and faster object program than Fortran; Fortran however may be considered as the language for installations where speed in programming is preferred over speed in program runs. The use of COGO in Civil Engineering installations should be reserved to engineers who are neophytes in the field of electronic programming and should not by any means replace existing programs or prevent the development of special programs. Also of great interest was the use of the CALCOMP plotter now adopted by IBM. Successful programs have been written for contour plotting either with or without photogrammetric equipment and in plotting cross sections and profiles. It has also successfully been used in structural design in such cases as in the plotting of influence lines. Another field of successful use is that of hydraulic engineering and hydrology. The use of other plotting de.vic:es such as the Digital Scale and the Wilde T7 has tremendously helped in reducing conventional type of work and in tightening the interrelationship of photogrammetry, plotting and computing. ELIAS C. TONIAS o 1& c o • o 1620 USERS GROUP WESTERN REGION AGENDA FOR 1963 WINTER MEETING ARIZONA STATE UNIVERSITY TEMPE, ARIZONA DECEMBER 11, 12, 13, 1963 WEDNESDAY -- DECEMBER II, 1963 8:00 - 10:00 9:00 Welcome and Opening remarks - R. Ebert, Regional President, and IBM Representatives 10:15 Coffee Break and informal Ifget acquainted" session 11:00 Technical Session "Alf A-I Generalized SPS Routines for Handling Simple Problems. Thomas L. Yates, Director, Statistics Computing Lab., Oregon State University, Corvallis, Ore. A-2 12:00 1:30 3:00 o Late Registration Exponential and 3inusoidal Curve Fitting. E. P. Hilar, Goodyear Aerospace, Litchfield Park, Ariz. Break for Lunch Technical Session liB" B-1 Automated Design Engineering. W. W. Rogers, IBM, Los Angeles, Calif. B-2 A Payroll and Labor Distribution Program Package. Elias C. Tonias, Erdman and Anthony, Consulting Engrs., Rochester, N.Y. Richard C. Devereaux, IBM, Rochester, N.Y. B-3 The SPIRE System - Salaried Personnel Information REtrieval. Gary J. Reed, Proj. Engr., ACF Industries, Albuquerque, New Mexico Coffee Break Technical Session "c" 0-1 A 1620 Program for Minimization of Boolean Functions, Expressed as Sums of Minterms. Thomas R. Hoffman, Prof. of Elect. Engr., Union College, Schenectady, N.Y. 17 --- ............. _.. __ ..- _._"----_._-_. __ .__ ....................__......... __. Agenda - Tempe - Page 2 WEDNESDAY -- DECEMBER 11, 1963 (Cont'd.) 5:00 C-2 Critical Speed, Stress, and Bearing Reaction Calculations for a General Shaft, Using Numerical Integration. Ralph B. Bates, Mgr. of Engr. Computing, Industrial Div. of American Standard, Detroit, Mich. C-3 Three Dimensional Surface Fit. David G. Kitzinger, ACF Industries, Albuquerque, New Mexico Adjournment of Day's Sessions. 8:00 p.m. New Users Meeting, t~J:)~ followed by Sound-Off Session at Approximately 8:30. THURSDAY -- DECEMBER 12, 1963 9:00 Technical Session liD" D-l Maximum Likelihood Resolution of Two Mixed Normal Distributions. Reimut Wette, D.Sc., Asst. Biometrician, The University of Texas, M.D. Anderson Hospital and Tumor Inst., Houston, Texas D-2 Comparison of Two Methods of Finding Significant. Contributors in Multiple Regression. M. J. Garber, Director, Computing Ctr., University of California, Riverside, Calif. 10:15 Coffee Break 10:45 Technical Session "E" E-l Network Analysis. H. N. Tyson, Jr., IBM, Los Angeles, Calif. 11:15 BUSINESS MEETING 12:00 LUNCHEON "The Impact of Automation on the Professional Engineer" MelfordE. Monsees, ADP Co-ordinator, U.S. Army, Corps of Engineers, Kansas City, Mo. 1:30 Technical Session "F" F-1 An Integrated Earth Work System. Cecil L. Ashley, ADP Co-ordinator, U.S. Army Engineering District, Walla Walla, Wash. F-2 Hydro-System, A Daily Operation Analysis. Charles R. Hebble, Civil Engr., U.S. Army Corps of Engineers, Walla Walla, Wash. I~ o ii, . Agenda - Tempe - Page 3 ~ THURSDAY -- DECEMBER 12, 1963 (Cont l d.) 3:00 Coffee Break 3:30 Technical Session uG" G-l Ray Trace Program for a General Lens System. D. H. 01Herren, Goodyear Aerospace, Litchfield Park, Arizona 4:00 IBM Reports and Discussion of Sound-Off Session. 5:00 Adjournment of Day's Sessions. 8:00 p.m. FRIDAY o Demonstration of the 1620, Model II, with 1311 Disc Pack at the Phoenix IBM Branch Office. (Transportation will be arranged). DECEMBER 13, 1963 9:00 Workshop Session "A" A-I MONITOR I 1:00 Workshop Session "B" B-1 FORTRAN II B-2 SPS 1620/1710 4:00 Conclusion of WorkshOp Sessions, and Final Adjournment of Meeting. '0 ft.. o I" 1620 USERS GROUP WESTERN REGION ABSTRACTS FOR 1963 WINTER MEETING ARIZONA STATE UNIVERSITY TEMPE, ARIZONA DECEMBER 11, 12, 13, 1963 NO. TITLE, AUTHOR, ABSTRACT A-l GENERALIZED SPS ROUTINES FOR HANDLING SOME SIMPLE PROBLEMS. Thomas L. Yates, Director, Statistics Computing tab., Oregon State University, Corvallis, Ore. Two-way frequency distributions; output editing and formattingi card reproducing. (No further abstract available.) A-2 EXPONENTIAL AND SINUSOIDAL CURVE FITTING. E. P. H1lar, Goodyear Aerospace Corp., Litchfield Park, Arizona. A specified number of exponentials are fitted to given equally spaced data pOints. The frequencies of the exponentials can be a mixture of real frequencies and complex conjugate pairs. The least mean square error criterion is used to find both the frequencies and the coefficients of the exponentials. The program is written in FOR'1'RAN language. The limitations of the program will be discussed. o B-1 AUTOMATED DESIGN ENGINEERING. Rogers, IBM Corp., LOs Angeles, Californ1a A presentation of Automated Design Engineering and how to achieve it. The role of a computer, ability to capture design logic, use of decision tables, comparison to manual design methods, and considerations involved with establishing Automated Design Engineering will be discussed. B-2 A PAYROLL AND LABOR DISTRIBUTION PROGRAM PACKAGE. Elias c. Tonias, Head of Data Proc. Dept., Erdman & Anthony, Consulting Engineers, Rochester, N.Y., and Richard C. Devereaux, IBM Corp., Rochester, N. Y. The objective of this paper is to demonstrate how some free 1620 time may be utilized in a relatively small scientific or engineering installation through the use of a package of commercial programs. The successful operation of such a program package since the first of this year (1963) has helped to justify the installation of a 1620 in another firm. This package, designed for use with the basic 20K 1620 w. w. (Cont Id.) Abstracts - Tempe - Page 2 Computer, with paper tape 1/0 and without any peripheral equipment, produces the payroll reports, and complete labor cost distribution and breakdown reports. Written in FORTRAN for easy maintenance, the ideas from this program package might prove a worthwhile tool in justifying the installation, or in increasing the production ratio of a 1620 in a small scientific or engineering account. With few or even no alterations, parts of the package may be used to handle other various time and cost distributions. B-3 C-I THE· It SPIRE" SYSTEM - SALARIED PERSONNEL INFORMATION RETRIEVAL. Gary J. Reed, Project Engineer, ACF Industries, Albuquerque, New Mexico The Albuquerque Division of/ACF Industries has developed a powerful management tool/for use in personnel administration. This tool is the SPIRE system, utilizing an IBM 1620. The system involves over 30 programs designed to utilize information from a master file of information maintained on magnetic tape. This master file contains over 300,000 different items of information. By judicious choice of information blocking, search time has been kept to a minimum for all applications. The SPIRE system current applications include: inplant recruitment, providing resumes of personnel qualified to-fill vacant positions; preparation of quarterly reports containing information on employees eligible for merit cons.1derations; creation of summary reports on salary increases for any time period; man power inventories; prOjections for salary budgeting; preparation of data for salary surveys; current salary status reports; and many other similar reports. The information contained in the SPIRE system has proved to be complete for all applications thus far. The approach and system layout have provided an economic way of producing information that no hand techniques could supply at any cost. In the applications where the ~esults could be obtained by clerical methods, enormous cost savings have resulted. The SPIRE system has been operational since February 1963. The use of existing programs and the creation of new applications consistently increase reflecting overwhelming management acceptance. A 1620 PROGRAM FOR MINIMIZATION OF BOOLEAN FUNCTIONS, EXPRESSED AS sill4S OF MINTERMS. Thomas R. Hoffman, Prof. of Elect. Engr., Union College, Schenectady, N.Y. A FORTRAN program implements a major portion of the Boolean minimization problem by reducing a sum of minterms to a logically equivalent set of prime.implicants. Minterms are entered as 3-digit octal-coded fixed pOint numbers. With the help of a special table, digit comparisons at the octal level reveal terms combinable according to the Boolean identity XA + X1t = X. Systematic (Cont I d.) ~I o Abstracts - Tempe - Page 3 o determination of all possible combinations of this type, coupled with a bookkeeping system that keeps track of the different subsets produced, leads to the desired result. Prime implicants are printed out in octal code, to complete the processing. The program can handle minterms having as many as nine variables, although the 1620 memory capacity (20K) may be exceeded in problems involving both large numbers of variables and long lists of minterms. C-2 CRITICAL SPEED~ STRESS~ AND BEARING REACTION CALCULATIONS FOR A GENERA! HAFT, U ING NUMERICAL INTEGRATION. Ralph B. Bates, Mgr. of Engr. Comput1ng, Industrial Div. of American Standard, Detroit, Michigan Critical speed, stresses, and bearing reactions can be calculated on a digital computer for a general twobearing shaft, any number of cross sections, variable loading, and any length. Besides eliminating the tedious labor of the calculations, the computer provides flexibility. A number of calculations may be rapidly made to optimize design, or to check out application variations on a standard design. The method 1s illustrated 1n detail by an example calculation. Briefly it consists of dividing the shaft into increments of length, determining the load and shaft moments of inertia in each imrement and the computer calculates critical speed, stresses, and bearing reactions. C-3 D-I M-15l, THREE-DIMENSIONAL SURFACE FIT. David G. Kitzinger, ACF Industries, Albuquerque, New Mexico This code uses multiple interpolation techniques in combination with extenSive transformation of variables to effect accurate fits of most smooth functions of three variables. With limited experience in selecting the form of curve fits in two dimensions, the programmer can fit complex three-dimensional surfaces. Output is designed to facilitate successive better approximations to the function in terms of additional transformation of variables. Second and third order fitting of functions is aided b~ a statistical error analysis. Two magnetiC tapes and a 60K memory are reqUired, although modifications can be easily made for smaller machines. MAXIMUM LIKELIHOOD RESOLUTION OF TWO MIXED NORMAL DIS1'RlliUTIoNS • o Reimut Wette, D.Sc., Asst. BiometriCian, Univ. of Texas, M.D. Anderson Hospital and Tumor Institute, Houston, Texas Iterative estimation of the five parameters from a sample taken from two normal distributions by the method of maximum liklihood seems preferable over the method of moments, because generalization to more than two parent distributions appears easier. Two approaches to estimate the maximum likelihood information matrix were abandoned in favor of the third, which cut down on computing time (Cont 1 d.) 014 Abstracts - Tempe - Page 4 (1620 FORTRAN II) for the complete estimation procedure considerably. Computing time is, aside from variations due to behavior of the procedure depending on goodness of initial estimates and structure of the sample, directly proportional to the size of the sample. Therefore, a grouping-of-data program was developed specifically for this problem,/and large samples, preceding the estimation procedure and reducing computing time to reasonable limits. D-2 COMPARISON OF TWO METHODS OF FINDING SIGNIFICANT CON- TRIBUTORS IN MULTIPLE REGRESSION. M. J. Garber, Director, Computing Ctr., University of California, Riverside, Calif. (No abstract available.) E-1 NETWORK ANALYSIS. H. N. TYson, Jr., IBM Corp., Los Angeles, Calif. Discussion will center about the general technique of network analysis, and the capabilities of a system of the 1620 programs utilizing this technique for the analysis of electronics circuits. The programs operate under MONITOR I on a 40K 1620. The capabilities allow for AC, DC, and transient analysis. F-l AN INTEGRATED EARTH-WORK SYSTEM. Cecil t. Ashley, ADp Co-ordinator, U.S. Army Engineering District, Walla Walla, Washington (No Abstract Available.) F-2 HYDRO SYSTEM DAILY OPERATION ANALYSIS C. E. Hildebrand, L. A. DUristan, c. R. Hebble, R. D. Moffitt, U.S. Army Engr. District, Walla Walla, Washington The program is a mathematical model of a system of hydro-electric projects. It is a general program applicable to any river system and scheme of development. It is capable of accurately simulating the hour-by-hour operation of a hydro system for as long a real time period as desired. It will determine the effects produced by eXisting hydro stations in regards to reservoir levels, river stages, and alternative distributions of system load among a group of hydraulically and electrically integrated hydro stations. The program makes it possible to determine the operating characteristics of planned future projects in regards to backwater encroachment on upstream reservoirs, pondage reqUirements, the effect of peaking discharges on downstream river stages and reserVOirs, and effects of added power installations. Two versions of the program exist: One for an IBM 1620 with 40K memory, and one for an IBM 1620 System with 60K in the IBM 1620 and 4K in the IBM 1401. o _j_gn o --'n--"-"- -y"_n_-TET e iMififiF" --_ .. --.U"ilritflrr:f2H.:iiriiti6fbHb.iitrttrlH t"""li6- Ittd' - Irm t"-,) * \ ""Y-" e n!lf "f!,! - "'""il "'["Z'I" WtWH.e\i"jjI"aHt.w, !'-.WHI'. Abstracts - Tempe - Page 5 G-l RAY TRACE PROGRAM FOR A GENERAL LENS SYSTEM. D. H. QIHerren, Goodyear Aerospace Corp., Litchfield Park, Arizona This program, written in FORTRAN with Format, and requiring 40K core storage, is designed to trace rays through complex lens systems, outputting intersection points at selected surfaces in the system. Cylindrical, conic, and toric surfaces can be handled and can be arbitrarily oriented with respect to the optical axis. Lens surfaces can be virtually any contour describably by second degree (or less) equations or at least divisable into sections which can be so described. A sample ray tracing problem is given to illustrate the use of the program. The compiler workshOp sessions to be held on Friday will assume that those attending will have a working knowledge of the externals of the programming systems (i.e. how to write programs in FORTRAN or SPS, etc.). The p'urpose of these sessions will be to present some of the "Hows' and "Whys" of the internal workings of the compilers. The session on MONITOR I will include more information on using the monitor, as well as delving into some of the internals. o t- 't'f'ti"¢,#' .. .. o I '1"1I1"!!lltt111l!4\J\*r \. 'w",,' ~ "If" !l~'.'!! 'i"l" ""t I,' h' ''''1'' , • o THE IHPACT OF AUTOM..I-\TION ON PROFESSIONAL ENGINEERING A PAPER PRESENTED AT ARIZONA STATE UNIVERSITY TEMPE, ARIZONA 12 DECEMBER 1963 BY MELFORD E. MONSEES LWP COORDINATOR U. S. L\RMY ENGINEER DISTRICT lZANSAS CITY, MISSOURI o • • I~. V o , '4* L ;"",,1'1 ' 'l/4 'eCL 'W;;' AbW '''.W1t&l! '''lL''g '1Ibt'!\I"\ il"\'/il"IL:"'I'M1i,!!,),Lr,',\\"LL\'/""111'1""/ ,L,L""!"')'\L,'l''''I'", SOCIAL t\ND ECONOHIC ASPECTS OF AUTOH.A.TION There are many differing views regarding the impact that the electronic computer 'will have on the development of our social and economic lives, but most of us can agree that the potentialities of automatic data processing (ADP) are limited only by the boundaries of our individual imaginations. The outlook, therefore, is open to wide-ranging speculations. The truly great impact of digital computation will be a dramatic speedup in the rate of technological progress -- and concurrently in the evolution of our social and economic lives. Electronic data processing is one of the most pO"t-lerful catalysts of technological development yet discovered. This is so because ADP has the ability of extending the capabilities of man's intellect. The human mind is the most powerf~l; most versatile, most useful natural gift bestowed upon man by his Creator. Any instrument that can substantially increase its capabilities is certain to have a profound effect on our future development. Han succeeded in building our technology to an extraordinarily sophisticated level during \.j'orld ~var II without an aid to his bility even remotely approximating the power of ADP. ~intellectual capa- Now, with the aid of computers that increase the productivity of his intellect in many areas by factors running into the tens of thousands, we are certain to see significant adva.nces in the tempo of technological, economic and social progress. o 2 - - - - - - - - .__....... • Another broad area of impact will be in fields of. business, industry and communications. We are all familiar with the progress ADP has already made in automating the business office. And computers are now being used increasingly in manufacturing--not only in process control applications, but also as an aid to the efficient management of the overall manufacturing operation. We are entering a battle for our very survival in the market places of the world. Much emphasis is being placed on economic development. Technological progress, rising productivity and ascending standard of living are the true sources of economic strength. They are vital to national survival in today's competitive world. As we move forward, we will encounter the problems always inherent with social and economic change. These problems should in ·no way warrant artificial restrictions on technological development for this is vital to the success of any business in a free economy. It may well turn out that the efforts of the new technology will be far more lastingly felt in its impact on many of the traditional principles and practices of management. Many traditional personnel practices are obviously going to be automated or abolished and various leader groups will change in power and prestige. While there are no precise means as yet of measuring the speed of technological change, it is reasonable to assume that by the mid-1960's, as those born during the Second World War establish families and the 3 o J7 • o challenge of foreign competition becomes more intense, the present rate of change will increase. At the present time, productivity increases in the nonfarm sector amount to roughly 2.5 percent a year. Even a moderate speedup in this rate would mean that by 1990, a relatively short span of less than 30 years, industry could double its production with the same labor as it employs today.l THE ENGINEERING EMPLOYMENT SITUATION2 Engineering, the second largest profeSSional occupation, is exceeded in size only by teaching; for men, it is the largest profession. The approximately 875,000 engineers in the United States in mid-1960 have made major contributions to the design, construction, and efficient utilization of the machines, equipment, roads, and buildings used by the .Nation's 180 million people. Engineers provide technical, and fre- quently, managerial leadership in industry and Government. They develop new products and processes, design many types of machines and structures, and contribute in countless other ways to the technological progress of the country and to the national defense. The outlook is for continued rapid expansion of the engineering profession. Engineering has been one of the fastest growing professional lJoseph A. Raffaele, "Automation and the Coming Difussion of Power in Industry," Personnel, (May-June 1962), 30. 2U. S. Bureau of Labor Statistics, Employment Outlook for Engineers, (Washington: U. S. Government Printing Office, 1961), pp. 101, 103-104. 0·' r, occupations in the United States in the past 50 years, and there is every indication that the demand for engineers will continue to grow. c As in recent years, there will probably be a particular need for engineers with advanced degrees to teach and to do research. Some of the major factors expected to raise the demand for engineering personnel are: Continued high levels of Government spending for defense, accentuated by the increasingly large amount of engineering time necessary for the development of modern weapons; growth of population and expansion of industry; increasing complexity of industrial technology, as such the trend toward automation of industrial manufacturing processes; and further growth in expenditures for research and development. In particular, the large sums spent for research and development in recent years by both industry and Government -- total research and development expenditures in the United States amounted to more than $13 hi1lion in 1960-61 -- have broadened existing areas of employment for engineers and opened up new ones, such as those concerned with computers, missiles, and nuclear energy. As scientific frontiers are extended, more areas of work for engineers will be provided. 5 o rT'i! "!"'"'H"!\'R!' '9"',," o \'Ilf¥ THE EMERGENCE OF AUTOMATION IN ENGINEERING That the systems concept is having a reaction in the engineering field is generally recognized. The question is how much. The change in a lot of features surrounding the engineering profession is beginning to assume large proportions and that rate of change is increasing rather than leveling off. For instance, we are told that the amount of engineering information and scientific information which is directly relevant to engineering problems has doubled within the last fifteen years. Even now the volume of knowledge related to engineering is so great that no one man can possibly know it all, even though he does nothing but study from boyhood to senility. For us engineers that means two things. We must specialize more than was formerly necessary or desirable, and second, we must be diligent students throughout our active lives. If we fail to study and keep abreast of the developments which directly apply to our chos~n fields of engineering, we shall quickly become back numbers and soon thereafter become useless to an advancing civilization. As a civil engineer, I am more familiar with the developments in that field than in some of the others. I recall the observations I have made and the discussions I have had with various men over the past 25 years, relative to the extreme reluctance of engineers to adapt labor-saving devices in their ~ work. Our fellow engineers were accomplishing much in the industrial world in devising and perfecting labor-saving equipment, 6 o 30 but we civil engineers were extremely slow in demanding any sort of o machinery or equipment that would make our work easiero For example, up to a few years ago, we were using the same methods of surveying that had been devised in ancient times. The transit was a little better than the old surveyor's compass, but the long and tedious process of making ground surveys was basically unchanged for 3,000 years. Now electronic measuring devices for survey parties are becoming standard equipment. The electronic computer is an item which cuts across all the fields of engineering arid many of the areas of science as well. They have been in use now for only about 15 years, but in that time they have increased the computational ability of mankind one million timeso The industry has grown in 15 years from zero to the production of a billion dollars worth of equipment during the year 1960. Since that date there has been an upsurge in the number of manufacturers of electronic computers. Competi- tion has become keener than it was a few years ago and we can confidently look forward to a greater variety of computers, composed of more dependable pieces of equipment, at a cheaper price than they command today.3 The electronic computer systems have eliminated much of the routine drudgery that has long been the bane of an engineering office. Also, they have eliminated the need for those men who were fitted for nothing more than routine work. There will be less and less need for the man who can only run a calculator or a slide rule, and who has been noted in an office 3Murray A. Wilson, "Change or Progress," American Engineer, (June, 1962), 29. o 7 31 • o primarily because he could remember all the formulae that applied to his line of work, able to recall and use them accurately as the occasion demanded. This work, the machine can do better and much more rapidly. The other feature is that the computer can solve the basic, theoretical equation and eliminate the necessity for the approximations and shortcuts that have been in common use in so many phases of design, simply because the basic formulae were so long and complicated that the process of solving them longhand took so long that no one could afford to use them. On the other hand, the new facilities will put a premium on imagination and ingenuity. These have always been desirable qualities in an engineer, but in the past a great many engineers have been kept gainfully employed on jobs that require little of either. This situation is changing and the prizes in the future are going to those men who are well endowed with these two important attributes. EFFECT OF AUTOMATION ON ENGINEERING EDUCATION Actual and tangible changes in the practice of the art and science of engineering are being reflected in our schools and are under constant discussion in publications, so it seems probable that we have created a new concept of the profession of engineering, or if you please, a new image of the engineer. The curricula in our colleges are in a state of fluidity with considerable differences in their means of meeting the challenge of the changing conditions.' On one thing all seem to agree-- that is, that the engineer of tomorrow and the day after, will need to be o 8 • more thoroughly grounded in the basic physical sciences than was thought necessary for his predecessor. Others have felt that an equally important need of the coming engineers is a knowledge of the past such as is gained through a study of the "humanities. 1f In an attempt to make room for these additional courses in a curriculum already overcrowded, something had to give, so the subprofessional subjects such as shop practice, material testing, laboratory work of various kinds, surveying and similar courses are being eliminated. The logical justification of this elimination is that these functions are actually to be performed by subprofessional men anyWay, or as we propose to call them, the engineering technicians. Modern problems of engineering are no respecters of traditional boundaries between the specialities. by schools of engineering. Accordingly, changes are being made It is understood that Dr. Keith Glennan has made broad moves which will go far to establish interdisciplinary approaches in engineering education at Case Institute of Technology. At the under- graduate level he has consolidated the departments of chemical, civil, electrical and mechanical engineering into a single administrative unit the Engineering Division. The engineering faculty are re-grouping in a natural way according to their common professional interests such as systems, design, energy conversion, materials, information processing and other emerging fields. Degrees in electrical, mechanical, civil and chemical engineering will continue to be granted, but a new degree at Case -- probably named Bachelor of Science in Engineering -- will be offered. It will give the student the opportunity to plan his elective program -- with faculty o 9 33 o advice -- to suit his career interests. Such a program can be designed to lead more effectively into advanced work. There is little doubt that the four-year graduate will continue to play an important role in industry, but there is an ever increasing need for the engineer with the depth of knowledge and experience produced by work at the advanced graduate level. Also, Dean B. R. Teare, Jr. of the College of Engineering and Science, Carnegie Institute of Technology, said recently in a letter to me -"The electronic computer has certainly made an impact on the individual courses in our engineering programs but it has not yet been the reason for extensive curriculum changes. Some engineering departments, pressed with the lack of time in a four-year program, have had to decide between continuing courses in engineering graphics and courses in computer 10gic.,,4 The electronic computer is also having its impact on the curriculum at M.I.T. In a letter I received last January from Dean Gordon Brown's office some of the directions in which the computer was leading were outlined. 5 For example the inclusion in the curriculum at M.I.T. of the following subjects: Digital Computer Programming Systems Mathematical Methods in Civil Engineering Digital System Application. 4B• Ro Teare, Jr., Carnegie Institute of Technology, in letter to author dated January 3, 1963. 5Gordon Baty, Adm. Asst. to Dean of school of Engineering, M.I.T., in letter to author dated January 18, 1963. o 10 Other related subjects have been included in the curriculum, but it pointed out that the automatic computations is not presented as an end in itself, but as another tool of analysis in the engineer's kit. The subjects merely focuses upon the techniques available to the engineer for exploiting the power of electronic computations • . Dean Brown's office also advised me that the impact of the computer upon research activities has been enormous. A copy of the semiannual Report, available from the MoI.T. Computation Center, can give you some idea of its magnitude. Yet, however important these computer-related activities have become to the School, there is a danger involved in attributing any of them simply to the availability of computer technology. For this is only one of the influences which have converged to create many of our most exciting research projects and subject offerings. Others include new methods in statistics and operation analysis, systems analysis and synthesis, theory of learning, and the information technologies. TANGIBLE BENEFITS FROM AN ELECTRONIC COMPUTER SYSTEM First hand knowledge of the impact of automation on professional eng ineering has been obtained as a member of the staff of the U. S. Army Engineer District, Kansas City, Missouri. This Corps of Engineers office has civil works engineering and construction in St~tes of Kansas, Missouri, Nebraska, Iowa, and Colorado, and military engineering and construction in the States of Kansas and Missouri. The total work of this district averages about 70 million dollars per year and includes the design and construction 11 O. , 'i, • o of large multipurpose dams, levees, £loodwa11s, channel i~provement works, pumping plants, and necessary utilities, highways, railroads, bridges, etc., required to be relocated in connection with flood control projects. Also, included in the assigned work of the district is the construction of military facilities and structures for the Army and Air Force. Computer facilities for engineering applications have been available since January 1958. The initial computer was the Burroughs E-l02, which, although of limited capability, was used until replaced in December 1960 with an IBM 1620 paper-tape system. Utilization of the paper-tape system for engineering applications increased rapidly and in January 1962, the system was augmented to provide for high speed punched card input-output. During this period, the utilization of electronic computer systems has provided tangible benefits through reduced cost of construction, savings in engineering and clerical manpower, and by providing a superior endproduct or flood control structure. To date the major effort in implementing ADPS procedures has been directed toward high-benefit engineering applica..; tions. Only preliminary phases of planned implementation of ADPS in areas of personnel administration, property accounting, real estate activities and fiscal responsibilities has been possible. As of December 1963, over 75 computer programs were being used in the fields of mechanical engineering, structural engineering, hydraulic engineering, hydrology, reservoir regu1a. tion and earthwork and soil mechanics. The currently installed system in the Kansas City District Office has provided the following: 0·'·" ,J 12 36 f~ \,--",) a. Improvement of the engineering and design product. b. More timely and accurate data which is fully responsive to the needs of management and engineering, including data previously not economically obtainable. c. Savings in costs by maintaining continued evaluation and balance of equipment and personnel. d. Savings in costs and engineering manpower by application of ADP principles. e. Simplified and reduced manual data handling and eliminated dupli- cation of files, reports, and entrance of source data. Further benefits are being obtained through realization of the following objectives: a. Expansion of hydropower and pumped storage-power stu~y programs to be used in connection with reservoir regulation, hydrology and hydraulic engineering programs used for design, construction, and operation of nrultipurpose projects. b. Expansion and refinement of structural design analysis programs used for preliminary and final design of various civil engineering projects. c. Expansion and refinement of earthwork and soils mechanics programs used for quantity computations and stability analysis of large and small embankments. d. Implementation and expansion of Critical Path Scheduling techniques to be used for coordination of construction activities as well as coordination and scheduling of engineering and design programs. 13 J7 c '" ·j·.eWH&S5'#RMFfiWibij"j·6"·'·_·-J···_' o ··I--··r ...- [ j. m.li5W··t-·SW 17·'·····"'_· ·--WW·-Tb-r---··-r ... I~· -[I·· I!· wtirf' if ffffdijwW&ii8\i,ifj·jijiH8itftF#ihti#i5·j"¥#r1ftFj·... ··bHW·#*fjdttT···P·-...·SIW#tftiij"j¥€fHfbiWfj"¥##iii·· ·rIll ¥····.±ifril#8fiff"*e·ttH·ritiiRi.i••··fijHiHM;r····.···d±·ijt··j·····WftiF" e. Refinement of personnel administration and reporting procedures. f. Implementation of one additional phase, of the engineering budget management data to eliminate manual posting and to provide more accurate readily accessible data for management and estimating purposes. g. Exploitation of the principle of "management by exception" through the potentialities of data processing equipment and techniques by continued education and training of personnel in the use of machine oriented reporting procedures and elimination of duplication of detail. h. Full utilization of presently installed data processing equipment by insuring that all data processing activities are essentially highbenefit programs. CONCLUDING OBSERVATIONS Last year, 1962, was a year which may very well be recognized as the beginning of the first plateau of maturity for the industries that automatic control has helped to create. Signs of maturity are also evident among the scientists and engineers who created these new industries and who must continue to act as whole partners with management and finance in continuing to create and exploit the new scientific breakthroughs which will firmly establish automatic control in its ultimate pOSition as the greatest servant of mankind. This maturity takes a number of forms: a. The number of engineers and scientists enrolled in post-graduate or extension courses in management is testimony that the importance of o 14 market and management factors in technical decisions is not widely recognized by the technical exper.ts. C) The idea of technical performance, cost, schedule, physical characteristics, an4 reliability has now a factor in technical decisions. beco~ widely accepted as Finally, an increasing number of scientists and engineers has come to realize that their technical brilliance is wasted if their ideas cannot be sold, and that it is a hollow satisfaction to be able to prove that something new and wonderful can be done unless means are found to ensure that it will be done. b. Electronics engineers, dynamicists, and even civil engineers have discovered the benefits of computers; and the electrical, mechanical, and field service engineers are recognizing the importance of automation in the translation of their diagrams and equations into operating realities. c. The overriding importance of reliability in concept, design, manu- facture, operation, and maintenance of automatic control systems has been thoroughly recognized. d. Our new ability to use high-speed, high-capacity digital computers as controllers for automatic control systems, plus the advent of microminaturization, has provided the flexibility for universal application to systems of almost any complexity, involving almost any combination of scientific disciplines. This new flexibility provides the base from which automatic -control can be adapted to applications ranging from space vehicles to automatic factories -- from complex air-traffic-control systems to the most microscopic of biological measurements and processes -- from the unmanned vehicles of oceanological exploitation to the complex man-machine systems of industry and sociology. 15 3~ o I o e. In the areas of technology, today, because of the broadening appli- cation of this industry into all fields, the engineer and the automatic control technologist must be able to understand and communicate with technologists from almost every field of endeavor. This universality of automatic control science application has brought about a lowering of the barriers of disciplinary specialty to permit an intermingling of the most widely diversified technologies. Finally, there is no doubt that the electronic computer has had its effect and will continue to have an impact, not only on professional engineering but on the curriculum and educational program in our schools of engineering throughout the United States. c 16 40 , 4 c o BmLIOGRAPIIY Josepb A. Raffalle. tlAutomation alld the Coming Diffusion of Power in Industry," Personnel, (May - June 1962). Frank W. Reilly. "Policy Dec~isions and EDP Systems in the Federal Government," Public Admini.stration Review, (September 1962). u. S. Bureau of Labor Statistics. Engineers, (Washington: Office, 1961). !n.m.t2L'UlE'~ Outlook for U. S. Government Printing Murray A. Wilson. "Change or Prl.>gress," (June, 1962). ~mer.ican Engineer, B. R. Teare, Jr. Carnegie Institute of Technology, ltr. dated January 3, 1963. PaulO. Roberts. :?Computer MOdels of Future Roads," (The Technology Review, December, 1962). Mart in Greenberger. Management 8.nd the r.ompu ter of the Future, (New York: John Wiley & Sons, Inc., 1962). Glenn Murphy. "Whither Engineering Education," Journal of Engineering Education (November, 1962). R()ger W. Bo1z. UAutomation as a Social Problem, H American Engineer, (March, 1962). o 44 Lfl o o GOOOfiE4R GOODYEAR AEROSPACE COR P () HAT ION ARIIONA OI\ISION LllCHI-'IE.l.D PARK, Expo:m~lTIAL A~JD sr:usotnAl Af-(IIO~~A CrEVE FITTING E. P. Hilar AAP- IRB!.:) o Ncvember 20, 1963 · , o c LIST OF REFERENCES (1) Dr. F. A. Willers, :~actical Analysis, Graphical and ~'!t1m€'r:tcal Methods, Dover Publications, Inc., 19L8 (2) Louis vJeisner, Introduction to the Theory of Equations, The MacMillian Co. (3) 1620 Po~om1al Rootrinder J. W. Wentzien, o ra~ b.Y BarstowB Method, 1620 Library (lt20 - 07.0.oLo) INTRODUCTION This paper presents a program written around the method of fitting exponental curves to data as given by F. A. 1'.'i1lers in his book, ttpractical Analysis .. l • The program fits exponentals with either real or complex conjugate pair fre~lencies. The frequencies and the coefficients of the exponentals are fitted using the leas t mean square error criteria. The rrumber of 'exponentals to be fit.ted is specified as part of the in~t data. The program is written in Fortran language o ---~------------~--- ~~- ~--.~ ~~--~-~~.~~~ o THEORY OF METHOD Given N observations equally spaced in the independent variable :x by an amount h and originating at ~ • 0, it is desired to fit a sum of n exponentals to them. The obBervations will be represented by the dependent variable y. The desired fit will be written n y(x) • Co + L1 where the coefficients, C, are real, (1) 3f!d the frequencies, a, are real or occur in complex conjugate pairs which give rise to terms of the form C en + 1 The derivation of the method of fitting the begin ttl exacUy. ~xponerta1s to the data will assuming that the pain ts actually fit the exponental repreS'entation An expression involvi.ng only the frequencies, a, and the successive diffel'ences of the data points will be derived. The fiction of an exact fit will then be removed by introducing an error term into this relationshipr; The above assumption may be written n y(~) • Co + L (L) 1 where y(m) is the roth data point. Defining the difference between successive data points as D(j) • y(j+l) - y(j) o ( 5) Using Equation (L), Equation (S) may be written as -1- b n D( j ) . Cj ( e U il -1) e c j (m-1) h (6) 1 Defining (7) (3) Equa tion (6) may be written D( j ) . t (U(j)-l) t(j) (9 ) Using Equations (5) throup,h (8) D(m+k) • y(m+k+1) -y(m+k) n ~ f(j) (U(j)-l) U(j)k (10) Now n+l equations of the form (10) will be wTitten out D(m)· L f(j)(U(.j)-l) • • (11) where the sum runs from lto n on j. Under the assumption that t.he curves are an exact fit, all of these equations (11) hold true. Now i f we consider them as n+l equat.ions in the n unknown f( j) then the determinate of the equat;.ons nrust be zero. This restriction on the determinate yields the following equation o -2- WY"·., . b.··· "•.. i "ifF""" . .... F f #It-·" f"i "iF, %&R£&if&:fbtfHf8iHii'*fdfia-Wtiftidw' iP9i&HA-liHHiit8fC"*fWiiWiiWMHi'iiiiT """/T-' rx o _.. . TSWl"""W' ET¥2f""F"""T-- E wPY W··-7 I ¥ - '"""iF n • Tn r···'srzW'fIK"%fZTWmpmarEP?l'Pwrnm ewes n (12) D (m+k) S(n-k) • 0 The functions S( j) are the s:/Il1Tletric functions and are defined in the two ") equivalent fo~~ shown below L • S(o) • 1 n S (1)· -- ~ U( j ) 1 n ~1 (13) • • o or, given an n th order polynomial whose zeros are the U(j)s, then the pc lynomial can be wri t ten as n ~ S(k)Zn-k • 0 (1),) ft. o where 5(0) • 1 In the problem of fitting the exponentals to the data we do not know the value of the symrlletric functions before hand. frequt?ncies could be calculated from them. If we did the required but by removing the fiction of an exact fit throt .gh introrluc.ing an error tenn into Equation (12) and 1 writing it as o n . ~ D(m+k) S(n-k) • f (m) (15) o 47 . -3- C\ the error can be minwzled by the proper choice of the symmetric functions. The means of choosing them will be the least mean square error criteria. (15) and summing over the permissible values of m yields Squaring Equation (~ S~ce it is the symmetric 2 D(m+k) S(n-k») ~lnctions • t(n+l) m f{m) 2 (16) that are being fitted, the partial derivatives with respect to the symmetric funct.ions of the left hand side of Equation (16) are equated to zero yielding n equations of the form n ~ o (L; D(m+{) D(m+k») S(n-k) • 0 (17) where .R.. runs from 1 ton and all s urns on m run from 1 to ~! - ( n+1) unless otherwise noted. The n( j)s are calculated from the data and the n sOO1 taneo11S Equations (17) are used to solve for the bestf1t symmetric functions. Equation (14) is then solved yielding for its roots the U(j)s. For a real root the frequency , i is calculated from in aj • U(j) 1 (18) n For a complex conjugate pair of roots of the form (19) ntib the corresponJi~ng frequencies are calculated from I 1 2 2 a· ~ inCa +b ) ~l ~. (20) ~ ~ 1 (21) • h arctan (b/a) where the corresponding terms in Equation (1) are no~ of the form (2). 0 1, I I -4- -,I o Onoe the frequencies have been found the fitting of ' the coefficients of Equation (1) to the data 1s a straight forward problem using the least mean square error criteria on the following equation: Co + )lJ~ ~ cje(tj(m-l)h -y(m) • €(m) (22) where the coefficients, C, are now to be fitted. The development from this point on is Fortran oriented and will be presented in the discussion of the program itself. o -5- METHOD AND PROCEIlJRE The description of the program will e'mphasize the means used to speed up the calculations of the various portions of the problem. Once these are understood the listing becomes self-explanatory. The first major task is to generate the coefficients of the symmetric functions in Equations (11). To accommodate the DO loops the following definition will be made D~(j) ~ D(m+j-l) • y(m+j+l) - y(m+j) (23 ) In the program the symbol is written DE(j) and the value of m is kept track of by a DO loop. Now defining (2L) The n equatiors (11) may be written 1 ( 25) where j runs from Ito n The desired coefficients are the sums • (26 ) A rapid method of calculating these coefficients will now be described. This method makes use of the following properties which are easily derived from Equations (23) and (2L). ~'o o -6- t?+l j,k . ~ j+l,k+l (28 ) • ~,j (29) E~,k • ~ ~,j (30) The quantities DE, E, and ;; E are stored as they are calculated. The procedure is as follows: m is set to one, and all the required values of the DEs and the Es are calculated and stored. The values of the Es are entered into the coefficient sums of the form (26) as the first term. Then m is set to 2 and the DEs are shifted one s pace to the left using the property shown in Equation (27). The missing term is calculated from Equation (23). The stored Es are shifted up and left one space using the property shown in Equation (28). The missing terms are calculated using Equation (2L). The values of the Es are entered into the coefficient sums as the second term. The program proceeds as for m • 2 until all the terms of the coefficient sums have been entered. A separate but concurrent aspect is that of storage. The properties of Equations (29) and (30) show that only terms on and to the right of the diagonal need be calculated. The remaining space is used for storage. The terms ~ k are stored in E(j,k+l) and the terms ~ ~,k are stored in E(k,j). At the end of the procedure described above Equation (30) is used to complete the arr~ of coefficients. The n simultaneous linear Equations (25) are solved by a Gaussian reduction routine. The solution yields values for the n s.rmmetric functions (13). o The nth order polynomial (le) is now-solved using a stripped down version of the Barstow method3 o The roots of Equation (lL) may be either real or occur in complex conjugate pairs. In the program the roots are stored in the following manner: Two fields R(j) and M(j) are defined where j runs from 1 to n. If the jth root of Equation (lL) is real then it is stored in R(j) and M(j) is set to zero. If the jth and j+l roots are a complex conjugate pair then the real part of the roots is stored in R(j) and M(j) set to one and the imaginary part of the roots is stored in R( j+l) and M( j+l) is 8Pt to two. The field M(j) serves to identifY the contents of the field R(j). SI -7- • t, The prot:ram does not use the variable S(j) directly but st,ore::;: the solutions of the n linear equations in the field ::( j) wl :E're S(j) • x(n+l-j) 0 < j ~. n (31) The entry into tr.Le polynomial solver places the symmetric functions in the field A( j) where A(j) • S(j-l) • x(n+2-j) 1 < j ~ n, A(l) • S(O) • 1 (32) The transfer from x to A is done directly. Once the roots have been found the frequencies may be calculated using the following form of Equations (lS), (20) and (21) aj • ~ .R n ( R( j a 122 2h ,in(R( j) + R( j+l) ) whBn M( j) )) =0 or j • 1 Pj +1 • Ii Jwhen I H(j+l) \ arctan ( . \ R(j) M( j) .. 1 and M( j+1) • 2 J If a real root or the real p~}rt of a complex root is negative then the remainder of the program cmmot give meaningful results and the program halts o The remainder of the program i5:3. straight-forward detE'rmination of the The di~cussion will be concerr1ed with a coefficients of Equat.ion (22). rapid way of ,g€lierating the nt)ecied numbers. n :: J will be considered here. straight-forward. Co + + The generalization of the results is quite Equation (22) is rewritten C e~(m-l)h 1 + e a 2(m-l)h Cos(P3(m-l)h) 2 c c e a 2(m-l)h sin(~J(m-l)ll) 3 The most gerleral form for -y(m) • t(rn) (3L) -8- o o nefinl!i?, 00 loop notation m V(l). 1 V(2) = eal(m-l)h m ~ e a~(m-l)h ) ) V(3) ~ Cos(~3(m-l h VeL) = m v( S) e a 2(m-l)h (35) Sin(~3(m-l)h) =-y(m) x( j+1) • C.:: J wherE-; m is a supers~ript on V. Using Equation?(35) Equation (3L) is "1Titten as m m m m m x( 1) *V(l)+x( 2) ~~V( 2 )+x(3) ~~V( 3 )+x( L) --:v( L)+v( 5) • (36) £ (m) Fittine the Cs or the xs by the least mean squarf) error cr-i teria as in the first part of this re~rt yields n+1 1irlear simultar)eous equations of the form n+l ~ 1 ~ m m V(j)*V(k) *X(k») + ~ m m (37) V(j)*V(n+2)· 0 ( where j runs from 1 to n+ 1 and all the sums on m now run from 1 to N url1ess otherwise noted. Defining m E(j,k) m m • V(j) * V(k) (38) where m is a superscript. o Equations (37) become -9- n+l ? l~ m \~ \ \ m L!!l E E(j,k») *:x(k) + (j,n+2)· (39) 0 where j runs from 1 to n+l. Now the Va may be calculated frclm equations of the form (35) for each m but the amount of time spent doing so is prohibitive. Recursion formulas may be developed for the Vs by noting that Equations (33) lead directly to R(j) • e ajh when M(j) • 0 and R(j) • eajh {When M(j) cos (f3j+lh) • 1 (40) and M( j+l) • 2 R{j+l) • eajh sin (~. lh) J+ Now from Equations (35) and (40) it is obvious that m+l m (41) Vel) • Vel) • 1 m+l m m V(2) • V(2) e~h • V(2)* R(l) (42) I V(2)· (L3) 1 The recursion formulas for V(3) and veL) are obtained wit.h the help of well known trigonometric identities as m+l V(3) • e02 mb (cos(~3(m-l)h) m cos ~3h - sin (~3(m-l)h) sin ~3h) m • V(3)* R(2) - V(L)*R(3) 1 V(3) • 1 (LL) and m m • V( L)-:tR (2) + V(3 ) ~R (3 ) (45) 1 veL) • o 0 The generalized recursion relationships are -10- o m+l Vel) • 1 • V(j)*R(j-l) wl~en M(j-1) • 0 m+l V(j) m m m+l or V(j) m+I V m (46) • V(j)*R(j-1) - V(j+1)*R(j) m m ) when M( .i-I) • (j+l)· V(j+I)*R(.j-l) ... V(j){~(j) ) and M(j) • 0 2 m+l -y(m+l) V(n+2)· and the vaJ.ues for mel are V(l)· 1 V(j) • when M(j-1) 1 • 0, 1 or (41) v( j) =- 0 when M(j-l) • 2 V(n+2) = -y(l) T'nc program calculates the sums The value of m is set to OIle 01, m in Equations (39) somewhat as bofore. and the initi.a1 values of the Vs are calcu- lated from Equations (L7) using M( j) to control the choice of equations. Tho Vs are stored and the Es are calculated 'Using equation (38) and entered dil"e~tly into of l'!l the sum on 'm of equations (39) as th(~ first term. The value is then set to two and Equations (L6) used to calculate the new Va with M(j) as the control. as before. The Es are calculated and entered into their sum The program fJ'oceeds as for m • 2 until the aulTlS are complete. As in the early part of this paper only the terms 01) and to the rirjlt of the ditlf,onal need be calculated fJtep-by-step. The sarna subprogram is used to calculate the xs of ~quations (39) as with o Equation (25). The coefficients of E~ation (1) and form (2) are then calculated using -11- Co • x(l) C j C i • x(j+l) when M(j) • 0 =V when M(j) • 1 x( j )1)2 + x( j+2)2 ¢ j+l • arctan ( x( j+2) ) and M(j+l) - 2 x(j+l) The method of calculating the two variance termfj is discussed in F. A. Willer 's bOOkl. o -12- o AUTOMATED DESIGN ENGINEERING W. W. ROGERS IBM LOS ANGELES, CALIFORNIA o ';iMRit~ ~i·u·wiji·Hi 'j \~it"'%fl o *'¥'5dtw"";;' rB&fill"ifHRfiFi'S6"¥iWw'--' "jW"'" a'SWUTS'T,m E'V""'''W-YW'WP i8 r W·-S····TiF···W'n "WJ-" TriPliWW" Gentlemen: I would like to talk to you today about a new computer application, Automated Design Engineering. A. D. E. is the use of the computer in the design of nonprototype products, the type of engineering commonly referred to as custom engineering, application engineering, or product engineering. Specifically, A. D. E. can benefit your company through significantly reducing your engineering lead time, increasing your engineering productivity, and decreasing your engineering costs. These and the other advantages of the application can, in turn, result in an improved competitive position for your company in your industry. First, I would like to tell you what A. D. E. is. will discuss where A. D. E. applies. on how A. D. E. works. Second, I Third, I will spend a little time Fourth, I would like to cover in more detail the advantages of A. D. E. Computers have been used, in the past, in the design of many products for industry. Computers have been utilized successfully in the design of circuits, missiles, motors, transformers, and telephone equipment, to name just a few. In each case, the use of the computer has resulted in great savings and great increases in efficiency and o productivity for the companies involved. Now a newly develop~d computer application, Automated Design Engineering, vastly increases - - - - - - _ . _ - - - - _ .__._.. --. ----~--.- _..... __.__._.,.,_ ...•.,.,.... .....,....,, .. 2. the power of the computer as an engineering tool. With A. D. E., the ',,; C " -_ computer can now be used to solve more engineering problems on a wider range of products. Conceptually, A. D. E. involves the sto:r:ing of cfesign ~ogic in a computer so that the computer can accept customer orders as input and automatically generate complete designs. The input to the system would be the same type of orders you now receive from your customers. The completed design would contain the same type of information normally given by your engineers to your manufacturing department; such things as product characteristics, part numbers, assembly numbers, bills of material, purchases parts list, drawing numbers and so forth. The principles of A. D. E. can best be described by using an example. An A. D. E. system for a company that manufactures pumps, for instance, would receive customer requirements in the form of orders or requests for bids. In this case, the customer needs 0 a pump to pump 150 Fahrenheit carbonic acid at the rate of 100 galIons per minute with a head of 40 ieet and he wants the pump motor wired for 220 and, 440 volts. These requirements would be entered into the computer and the completed design would be printed out. These requirements would be entered into the computer, would be processed by the computer with the aid of the stored design logic, o .. 3. o lO and the completed design would be produced. The completed design information can be in many different forms ... such as part numbers, assembly numbers, drawing numbers, manufacturing instructions, etc. In this example the completed design information consists of the pump frame number, FQ 6; the model number CXR; the suction and discharge pipe diameter, 3 and 4 inches, respectively; the impeller diameter, 8 3/4 inches; the motor speed, 1750 RPM; etc. , etc. In short, the completed design would consist of the information needed by manufacturing to build the product. Now that we have discussed what A. D. E. is, where does it apply? Automated Design Engineering applies primarily to "cus- tom" or "product" design which can be defined as "non-prototype custom design variations of a standard product line to meet your customers' requirements on a continuing basis." Even though some of the tools, techniques, and methods available through A. D. E. are applicable to prototype design, the major impact of A.D. E. will be in the custom or product design area. There are many companies in industry today which produce custom designed variations of a standard product line in response to customer orders or requests for bids. "Custom" or "product" de- sign is common in such products as pumps, motors, generators, 4. switch gear, transformers, electrical measuring equipment, indus- o trial furnaces, switchboards, engineering and research instruments, heat exchangers, steam turbines, conveyors, and air compressers, to name just a few. Let us look at the problem of the "custom" or ' 'product" engineer. In industry today, customer orders for products will come to the design engineer. His problem is three-fold: First, he must determine what to build to satisfy customer requirements. Second, he must translate these customer requirements into workable parts and assemblies and their drawings. Third, he must prepare the complete paperwork for manufacturing. It has been determined through actual experience in industry that these problems can be solved on a computer with great savings of time and money and great increases of efficiency and productivity. Now to help you visualize what a system would look like in your company, let us look at a typical Automated Design Engineering System in operation. A customer order coming in would go tb the Engineering Department where two functions would be performed. First, the order would be edited to insure the completeness and validity of the customer order. ing would be performed. Second, any necessary pre-engineer- Once the order editing and pre-engineering l/ o • 4. C") . .1.1 has been completed, the customer order would go to the key punching room where the customer requirements would be transcribed onto cards and fed into the computer. The computer is programmed with design logic and has available to it tables, standards, and reference information. The· output of the system would be completed design which would then be reviewed by the engineering department and passed on to manufacturing. Now that we have covered what A. D. E. is, and where it applies, and what an operating system would look like, how does an Automated Design Engineering System actually work? The key to Automated Design Engineering is the ability to capture the design logic by which customer requirements are translated into product spe cific ations. We have developed new tools, new techniques and new methodologies which will assist you in developing an A. D. E. system for your company. One of the most important tools, and the heart of this new system, is Decision Tables, a technique for capturing the design logic that comprises custom or product design. to capture design logic is the key to A. D. E. The ability Before we discuss Decision Tables, let us clarify what we mean by design logic. 0 ·:'" " • 5. o What is design logic ... ? Design logic is the complex decision-making process through which the engineer proceeds in designing a product to meet a particular customer's requirements. The engineer reads the customer order and then, through the medium of design logic, proceeds to design and select the various parts and assemblies required to satisfy the customer's requirements. The result is the completed design of the product. Let us look at a sample of design logic and how this design logic can be captured using decision tables. example the armature for a voltmeter. We will use as our If we were to ask an engineer working for a company that manufactures voltmeters how he designs the armature for a voltmeter, he might say: "Well: if thre(-customer requires DC service for a speed application and asks for a single phase instrument with villivolt rating, then I know I have to use a moving coil. ~ two windings are needed, which is the case, then the part number will be -12526A. A26A will be used in this case. frequently Drawing number If the rating value specified is be- tween 76 and 200 millivolts and we need a moving coil, the main winding will use Aluminum wire 16 mils in diameter. • The number of turns and the number of layers will come from these two formulas which relate to each other. Based on the case we're using, we'll c !\. . ·,.jij~"5'iil.ff!!1t . "' 1"1' d'IHH t"ji "i'E'iulitl!!w''IWlftl*" "tl 6. o have 13 turns on the main winding and one layer. Again, I have to look up the drawing number ... it's 012526-1A. Now the coil will also need a damper winding which for a 76 to 200 millivolt rating and a scale size specified as four inches, will take 8 mil Copper wire and half the number of turns on the main winding which is also shown on drawing number 012526-1A." This complex decision-making pro- cess can be captured by use of a new tool called decision tables. A decision table consists of four quadrants. In the top two quadrants are the customer requirement names and values. In the lower two quadrants are product specifications, names and values. This decision table is a summary of part of the data given to us by the voltmeter design engineer. In the upper left hand quadrant is a list of customer requirement names ... service, application, rating units and number of phases. Values which the customer might specify for these requirements are given in the upper right hand quadrant ... DC, speed, millivolt, 1, etc. In the lower right hand quadrant are values for these product specifications ..... moving coil, inductive, 1 plus the number of phases, etc. The decision table is read as follows: If the service is DC, and if the application is temperature, then the type of armature needed is amoving coil, and tables number 2 is the table to which we should go next. o b4 Rule 4 would be read as follows. 7. If the service is AC, and if the rating units are milliamps, and if the number of phases is 1, then the type of armature required is inductive, and the number of windings is calculated as being ONE PLUS the number of phases, and the next table is number 2. which can exist is called a rule. sets of conditions or six rules. Each different set of conditions In this case there are six feasible Once the decision table is established, programmed and entered into the computer, the computer will automatically select which rule applies to each individual customer order. It takes hundreds of decision tables to capture and store the design logic for an entire product or product line. When all of the required decision talbes have been programmed and placed in the computer memory, customer orders can be entered into the computer design in a logical stepwise manner. Now that we have covered what A. D. E. is, where it applies, and have gone into a little detail on how it works, let us turn our attention to what an A. D. E. system can do for you, or the advantages of an A. D. E. system. We have already indicated that you can reduce your engineering costs and at the same time, obtain increased productivity and efficiency from your engineering force. Overall lead times can, in addition, be significantly reduced through a reduction of design engineering time, materials procurement time and manufacturing lead time. The design time, for example, can be reduced from weeks or even months to a matter of minutes. This reduction in design time o· • 8. will allow you to give faster response to customer orders and requests o for bids. Faster response to bids, on the one hand, can have a significant effect on your profits if there is a high correlation, as exists in many industries, between speed of your response and acceptance of your bids. Faster response to orders, on the other hand, can result in faster deliveries and improved customer relations. And furthermore, A.utomated Design Engineering can provide a significant increase in your business activity by allowing you to respond to more bids if you now find yourself unable to resp::> nd to all of the requests for bids that are made to your company - - due to the lack of available time in your engineering department. In addition, faster response to bids can have a significant effect on your profits if there is a high correlation, as exists in many industries, between speed of your response and acceptance of your bids. Also, bid costs can be reduced because A. D. E. can reduce the out-of -pocket cost incurred when bids are designed but not won, since the engineering cost per bid can be greatly reduced. Material savings are another potential advantage of Automated Design Engineering. By calculating exactly how much material is required for each job, A. D. E. can reduce overdesign and waste. Another way in which material savings can be realized is through reducing the number of parts in inventory with the same specifications but different part numbers. 0, I As you develop your A. D. E. system, these duplications will become readily apparent and can be eliminated. The consistent I'", use of the best design practices eliminates the proliferation of methods 9. and materials and insures a design commensurate with requirements. o All too often today, the pressures of competition and of constantly shrinking delivery times tend to result in over-design or in picking a design which is perhaps more expensive than the specifications actually call for and the result is loss of profit. Bid and orde'r cost- ing can be incorporated into your Automated Design Engineering system to allow you to not only design but also to price both labor and material for your design. Through the automatic generation of engineering paperwork much of the clerical burden which wastes so much of the time and talent of engineers today can be alleviated. A,utomated Design Engineering also provides an improved error-checking facility. are many. The sources of errors in engineering today These errors can creep in through the customer order, through the sales engineers, through engineering errors or material list preparation. Errors can result in the wrong material, too much material, or too little material being available at assembly points. They result in wrong design and in over-design. With an Automated Design Engineering System these functions are performed automatically. Error checkipg capabilities are built into the computer to catch and eliminate such miscalculations, which are the real problem in industry today. Greater management control is possible because i 67 1,,1 I ~p I' I, o 10. management policies can be incorporated in an Automated Design Engineering System with the design logic, to insure that management policies on engineering are carried forward. Now that we have discussed the advantages of A. D. E., I would like to show you how we can assist you in developing and implementing an Automated Design Engineering system in your company. IBM has available tools, techniques, and methods to help you in developing your system. There are two steps in the development of an operating Automated Design Engineering system. These are the Survey and the Implementation. The survey represents step. th~ first Its purpose is to determine the applicability of the A.. D. E. system to your product lines. of your present system. It will provide you with a:n analysis It will determine the requirements of the new system and will give you a preliminary design of the new system as it can be applied specifically to your company. Lastly, upon com- pleting the survey, you will be able to measure the advantages that will accrue to your company through the use of Automated Design Engineering. The new tools have been developed to prepare an analysis of your present design engineering operation from four related but different standpoints: Those of time, cost, accuracy, and o operations. After the survey is completed you are ready for the next step, implementation. The purpose of the implementation phase 11. o is to develop a completely tested and operational A.utomated Design Engineering system. In order to do this, an analysis of your cus- tomer specifications, and an analysis of your product structure must be prepared. The techniques and tools which IBM will provide will also enab Ie you to capture the design logic of your product line onto decision tables, to perform the detailed systems design, the programming, the testing, the conversion and finally the initial operation of your Automated Design Engineering system. techniques and methods are tested and proven. in the form of printed material. These tools, They are available to you For example, the A. D. E. General Information Manual will introduce you to the survey and implementation phases of this new system. In addition, we have prepared a de- tailed reference manual to provide your engineers with the "How To Do It" information necessary to develop an A. D. E. system. I: I A PAYROLL AND LABOR DISTRIBUTION PROGRAM PACKAGE by ELIAS C. TONIAS and RICHARD C. DEVEREAUX December 1963 ERDMAN & ANTHONY 82 St. Paul Street Rochester 4, New York 0'· ' ,,'I' 70 o c - ''(-r"""--- iT ------ --- ---- -Y'n TITLE A Payroll and Labor Distribution Program Package AUTHORS Elias C. Tonias of Erdman & Anthony, Consulting Engineers, 82 St. Paul Street, Rochester 4, New York and Richard C. Devereaux of IBM Corporation, 540 Main Street East, Rochester 4, New York DIRECT INQUIRIES TO EliasC. Tonias ABSTRACT The objective of this paper is to demonstrate how some free 1620 time may be utilized in a relatively small scientific or engineering installation through the use of a package of commercial programs. The successful operation of such a program package since the first of this year (1963) has helped to justify the installation of a 1620 in another firm. This package, designed for use with the basic 20k 1620 Computer, with paper. tape 110 and without any peripheral equipment, produces the payroll report,. and complete labor cost distribution and breakdown reports. Written in FORTRAN for easy maintenance, the ideas from this program package might prove a worthwhile tool in justifying the installation, or in increasing the production ratio, of a small account. With a few or even no alterations, parts of the package may be used to handle other various time and cost distributions. o 11 o o mmwrr 0· ', TABLi£ OF CONTENTS ":, I. Introduction •••••••••••••••••••••••••••••••••••••••••••• 1 II. Systems Analysis and Design III. The Programs ••••••••••••••••••••••••••••• 2 .................................... 4 B. Payroll Register .................................... 5 C. Payroll Checks ...................................... D. Payroll Deductions Register ......................... 6 E. Labor Distribution Data Sorting ..................... 6 F. Labor Cost Distribution ............................. 8 G. Labor Cost Breakdown Bi weekly .................... 10 A. Data Preparation 6 H. Labor Cost Breakdown Accumulati ve IV. Timing Considerations and Limitations V. Swmnary VI. Appendix ................. 10 ................. . 11 ..................... .......................... . 12 ' •...•.........................•... Time and Expense Account Sheets ...•.......... Payroll Master ........•...................... A. General Flow Chart 14 B. Sample 16 C. Sample 18 D. Table of Addresses for the Data Preparation Program ••• 19 ............................. 20 F. Sample Payroll Check ................................ 21 G. Sample Payroll Deductions Register .................. 22 .............. 23 H. Sample Labor Cost Distribution I. Sample Labor Cost Breakdown Reports .................. 26 E. Sample Payroll Register ~eports o 7:1 - - - - - - - - _ .. __ ...•.. __ ._,-_.. -- ...._._..._... _----_--..---_ _._-_.._. ........ c o WE I o I. INTRODUCTION It is no secret that the 1620 is not designed for excessive data handling and the IBM Corporation definitely is not marketing it for commercial uses. In fact it is a small scientific computer and as its many users will attest, it performs with excellence in this role. Despite the machine's fine record in the field, its inability to handle excessive data (especially the basic 1620 with paper tape) produces one distinct problem area which can present itself during the selling phase or after the installation of the system. It is known as 1Il8chine utilization. Present day engineering or scientific establislunents do:have considerable, if 1 not abundant, amounts of commerical (accounting) type of work which could be automated and thus increase the utilization of a computer. This problem of the 1620 utilization is not as predominant in a large data processing oriented company as it most probably has a commercial computer system. In such a company the 1620 is being looked upon for the engineering or research and development departments. A large company would presumably not expect high utilization on a machine which would be operated in an open shop atmosphere. Their objectives would be more intangible, such as the release of engineers for more creative type work and the increase in speed of routine computations. On the other hand, in a small engineering account, where the rental cost of $1,600 per month represents a large investment, the subject of machine utilization is of great concern. Even though a two or three hour run a day can justify the installation of a 1620 in such an account there still is the feeling that more should be gotten for this kind of money. It is for such an establishment that this paper is primarily intended. The comPAny for ahich these programs were developed has a bi-weekly payroll for two offices, one of ninety and the other of forty employees. It takes about two days (16 hours) of the computer's time every other week to produce the complete payrolls and labor cost distri~ bution. The operation of these pregrams since January 1, 1963 has helped to increase the average utilization of the computer to 40 hours per week. As for economic value of these programs, it cannot, be measured in dollars and cents. It places valuable up-to-date information about the cost, progress and estimation of projects in the fingertips of management in a matter of a few hours. Neat reports may be produced in a moment's notice. It may be said that these programs in addition to increasing machine utilization do contribute in improving management operation. '0.-: '. I'" 73 __ - - - - . .. _------_.. _._ ..__....._.._--------_. o 2 II. o SYSTEMS ANALYSIS AND DESIGN In trying to analyze and successfully design a payroll and labor cost distribution program package the following must be taken into account: the memory capacity of the computer, the presence or absence of special features, the type of input-output and the method of operation of the shop including personnel availability. The concern where these Payroll and Labor Cost Distribution Program Packages were developed employed the basic 1620 with a 20k memory, without any special features, and with paper tape input-output hardware. This concern, a consulting engineering outfit, operates on a semi-open shop basis without a specialized machine operator. One cannot start any simpler than that nor can one have any more handicaps as far as the 1620 is concerned. Of the three major languages of the 1620, machine, SPS, and Fortran, the latter was chosen by the authors because the programs could very easily be maintained and revised according to the needs of the company. In addition, Fortran offers a lot easier means of programming and debugging and to the best of the author'S knowledge such a task would be a first. It must be noted here that all of the following programs (see exceptions later in this paper) were originally written in Fortran with Format and later changed to UTO Fortran. For those unfamiliar with UTO Fortran, it was developed by Mr. E. Stewart Lee and Mr. James A. Field of the University of Toronto, Ontario. UTO in general is similar to Fortran with Format except that it brings the program origin down to 06950, is slightly faster, the subroutines are about 10 per cent shorter, the use of EXECUTE PROCEDURE n statements (similar to CALL statements) facilitate programming and save memory when properly used, and has quite a flexible input fonnat (input data does not have to be in strict accordance with the fonnat statement). One question that had to be answered early in the planning stages was the availability of checks for the 1620 typewriter. Cardboard checks were out becaU5e they necessitated a typewriter RPQ. Paper checks were chosen and designed by IBM. A picture of a check appears in the appendix of this paper. The check is of the high-low design with the stub under the check proper. The perforation of the right .an be placed any where. When designing a pay check the following hints should be considered: (1) When designing the check, it is advisable to make the spacing between the printed lines a multiple of two or three. This means that at program time during the carriage control operations, the typewriter can be run in double or triple space mode and thereby save a considerable number of carriage returns • • 3 (2) Hint (1) may be accomplished with statements of the form 1000 FORt\1A T (f/) or if using UTO Fortran with the CONTROL 102 statement. In UTO the statement CONTROL +08 may be used for tabulation. (3) The check should be designed so that variable alphameric data appears only on the left edge. Note the example check where the two dates (XX/XX/XX) and the name are all left adjusted. This is very important in a situation such as the following. The employees name is part of the master record and must be the first data in that record. For example, the name and year to date gross, FICA, and withholding taxes would read in as: 0 ACCEPT TAPE 100, YGR, YFICA, YF\vT 100 FORMAT (ISH EMPWYEE NAME F9.2,F7.2,F8.2) The variable alphameric name is stored in the format statement itself. When saying PRINT 100, the name alone will be printed out. If the name were not left-adjusted in the master record and in the Format statement, it would be impossible to get it out alone. Another technic that could be used is to carry the name all by itself in a separate master record. The payroll system for which these program packages were designed is based on the following criteria: (a) Employees are to be paid bi-weekly with time being reported on two time sheets (the programs may easily be changed to any other pay system). (b) Reported time is to be reported by each employee by operation. (also referred as activity herein) for each project. (c) There are three types of paying systems (pay codes): straight salary (code 1), straight time (code 2) and time and a half for overtime (code 3). (d) There may not be more than 130 employees. This company uses 20 activities only. If need be this number may be increased at the expense of running time. (e) Three master records are to be kept:- (1 )the payroll master tape containing the name, number, rate, pay code, dependents, year-to-date gross, FICA, tax withholdings and expenses, quarter-to-date gross, FICA, tax withholdings, social security number, hospitalization, bonds, savings, life insurance, pension, and miscellaneous deductions; (2) the project master tape containing the project number and todate totals on direct labor, expenses, and indirect labor (sick leave and vacation); (3) the activity Master tape containing to-date totals for each activity for each project. I 1:,,1 I !l' I; IwivU"W' 1I"",. . . .t TWU"F7g tHh.-"fl"t"tt!!",,»"' !"Nil 4 o As mentioned in the introduction the biggest handicap in this opera~ion is that of handling data. Because of this, changes to any of the master tapes should be kept to a.minimum as the slightest revision requires the reproduction of the entire tape. Master tape edit programs may be written to facilitate tape revisions, especially massive ones such as the zeroing of the registers at the end of a quarter or the year. the data. II I. The other case where data handling enters the picture is in This is discussed in the following section. THE PROGRAMS Below is a general description of the programs in the Payroll and Labor Cost Distribution Programs. A. Data Preparation Considerable thought was given to the input data since this was the start of a chain of events which would produce all of the payroll and labor distribution reports. As part of this paper's appendix one will find the time sheet used for reporting work time and expenses. Since this data form the basis to both packages it is typed into the Data Preparation program. The sequence of this data is composed of a minimum of three records per employee. The first contains the employee's number and the remaining contain the activity, number of hours or personal expenses and the number of the project. A negative activity denotes the end of a time sheet. For a bi-weekly payroll two negative activities are required. The program immediately punches on tape the accepted data in a similar format. This is the Labor Distribution Data Tape. Then it proceeds to compute the regular and overtime hours and personal expenses on a weekly basis which is stored in memory for ea.ch employee. This information will l~lter produce the Payroll Data T~pe. To do this the Payroll Haster Tape containing the employees' pay codes is read into the program at the start of the run storing each employee'S number and pay code. It should be mentioned here that activities 18, 19 and 20 are reserved by the programs for vacation, sick leave and personal expenses. t •• t , I • I t I J r o One bit of logic that might be helpful in the prog~m is a technique for being able to go into memory and make corrections after all data is typed in and while the payroll dat~ tape is still stored. This could be accomplished by assigning a storage cell number to each employee as he is entered into memory for the first time. This storage cell nmnber would teen represent the position in all four arrays (man number, regular hours, overtime hours, expenses) where this particular employee's data could be found. The number, along with a memory map of the arrays would then give the exact location of any data in question. A portion of such a table is included in the appendix. If an error were made in the typing of data for the say ninth employee the table will show that this employee's number, regular hours, overtime hours and personal expenses would be found in addresses ,19750, 18450,17150, and 15850 respectively (see table in appendix). 5 It will be advantageous in the Data Preparation Program to have a dump of grand totals at the end showing regular hours, overtime hours and personal expenses. These totals should balance back to an adding machine tape which should have been taken when auditing the time and expense sheets. The totaling routine should be programmed so that it may be "branched to" after any corrections to the memory table have been made. At this point if everything balances, the Labor Distribution Data Tape may be removed from the punch and the punching routine of the Payroll Data Tape initialized. • o This marks the end of the combined operation between the Payroll and Labor Cost Distribution Programs. B. Payroll Register The first step in the Payroll Program Package is to obtain the payroll register showing all incomes and deductions for all employees of the current period. This, the Payroll Register Program, necessitates as input the Payroll Master Tape and the Payroll Data Tape as well as the starting check number and period ending date which are entered via the typewriter. The payroll data is stored in four arrays - employee number, regular hours, overtime hours and expenses. The Payroll Master Tape is then read in, the arrays are searched for a match, and the pertinent data from the master and detail are combined to calculate the net pay for each individual. Two lines of printed output appear for each employee· showing the standard payroll register items • Accumulative totals are kept for each item and at the end of the report these totals are crossfooted for a check against the accumulated net .pay. In this program as well as in the next there is the problem of half adjusting and truncating a product. Let it be. assumed that a. straight time employee waged at $4.375 per hour has wori{ed 48.50 hour$. Hence PAY • 4.375 x 48.50 • 212.18750 For obvious reasons the employee should be paid $212.19. Fortran however will carry it as 2121875003. When several such products are accumulated there is going to be a disagreement between the actual swns and those of Fortran. This may be remedied by the following statement: PAY • RATE x HOURS + 0.005 + 100000.0 - 100000.0 which will be executed PAY PAY PAY PAY • • • • 8S 4.3750000 212.18750 212.19250 100212.19 follows step by step x 48.500000 + 0.0050000 + 100000.00 - 100000.00 • • • • 212.18750 212.19250 100212.19 212.19000 When completed the payroll register should be spot shecked by the"personnel responsible to insure complete accuracy. There is no punched output to this program. I oj I 17 wm . -'JWoof 6 C. o Payroll Checks As soon as the pay~ll register is checked and accepted the Payroll Checks Program may be used to produce the paychecks. Basically this program is similar to the previous one using the same input (instead of the check number the check date is entered). In addition to the checks this program updates the Payroll Master Tape concurrently with the checks and at the end it produces an Adjusted Pay Rates tape. Such a rate is merely the employee's gross pay divided by the number of hours worked.. This will affect only the straight salary and time and a half employees. For example, a straight salary employee who is always paid for 40 hours and ha.s a rate of $4.375 has worked 48.50 hours. To give a more complete picture in the distribution reports all of the 48.50 hours should be considered. Thus this employee's adjusted pay rate is Adjusted Pay Rate • 4.375 x 40.00/48.50 • 3.6082474 Had the employee being paid time and a half for overtime then his adjusted pay rate would be Adjusted Pay Rate • (4.375 x 40.00 + (4.375 x 1.5) x 8.5)/48.50 • 4.7996134 This adjusted rate may later be used to distribute dollar increments of the 48.50 hours over the various projects on which the employee worked. This rate can be calculated at the same time as the employee's check and stored in memory. Since the Payroll Master tape is updated during the production of the checks the Adjusted Pay Rates will have to wait until the end of the program to be dumped on tape. These rates are later used with the labor cost distribution programs. D. Payroll Deductions Register Because of memory and output limitations not all deductions are itemized in the payroll register. These deductions, bonds, savings, hospitalization, etc. are itemized for each employee by the Payroll Deductions Register Program utilizing the just updated Payroll Master Tape. E. Labor Distribution Data Sorting Having completed the payroll phase of the operation we may proceed with the labor cost distribution phase. It was stated under the Data Preparation Program that data read directly into the program from the time sheets were divided ·into the Labor Distribution Data Tape and the Payroll Data Tape. Since it is possible for an employee to report time under several activities for several pro jects time data will have to be sorted in one way or another. Such a sort would be easy for a card machine. For a tape machine though it is a different story. Data could be sorted manually before the Data Preparation Program run but this would defeat payroll automation. As a matter of fact it would be a rather involved affair. Also mentioned earlier was the Project Haster Tape containing all projects of the company with their respective costs. o ---~---'-----------~"''"-'''''-'''"'--~'~=~=----=----=----=-----=--=---, 7 It is possible for the operator to type an erroneous project number that is not included in the Project Ha.ster Tape or that a new project had been added without the master being updated. For these reasons two programs were developed in machine language to utilize memory space. These are the Project Completeness Test and The Labor Distribution ~ Sorting programs. The Project Completeness Test program first reads the Project Master Tape and stores the project numbers only; then it reads the Labor Distribution Data Tape and compares each project number with the table previously stored in memory. If a match does not occur the project number on the Labor Distribution Data Tape is printed on the typewriter. At the end of the run the number of projects and number of records are typed. The number of records re fers to the records of the Labor Distribution Data Tape containing an activity, hours or expenses and a project number. When this test has been made the Labor Distribution Data Tape may be corrected if need be through a tape correction utility program. The Labor Distribution Data Sorting Program does a little more than just sort data. First it counts records to insure that no unauthorized records have been dropped or added while correcting the Labor Distribution Data Tape. Second it sorts data. The program inserts the employee number to each of his records, changes each record from an alphameric mode to numeric mode dropping all decimals and arranging them in the format shown below: XXXXX XXX XXXXXXX ... J \ whe re 1: 2: 3: 4: ______ 2:3 • ., 4 :t- pro ject numbe r employee number activity number hours or personal expenses Each record contains a total of 16 characters including the record mark. These records are stored in memory beginning in location 19999 and working on downwards. The sorting of the -records begins when all records have been stored. It is a simple replaceable sort routine which starts at the top of memory and keeps comparing two adjacent storage cells (records). If the upper is less than or equal to the lower, no replace takes place. If the upper is grea,ter than the lower, the two records .are interchanged, a switch is set and the program steps down one record to compare the next two. This continues until one complete pass on the records has been made. At this point, the switch is checked. If it is on (meaning at least one record out of sequence), it is turned off and the comparing routine is repeated. When the switch is finally off after a complete pass, the internal sort is considered terminated. It should 'be emphasized here that during the compression phase (changing data from alphameric to numeric) where data is edited in the input area and placed in a cell, the comparing field can be set up as one field and still get three breakdown levels - activity, within man, within project - providing the field is arranged as described in the previous paragraph. o 1 ~! II·'.,', ' II! 8 o Third, at-the completion of the sorting routine the program will commence reading the adjusted pay rates which were produced at the end of the payroll check production. The rates and employee numbers are converted into numeric characters and the memory cells are searched for an employee number match. When the match occurs the rate is multiplied by the corresponding hours (no multiplication is perfonned for activity 20 - personal expenses) and the hours are now replaced by the cost. Each employee number of the Adjusted Pay Rates tape searches the entire memory map as it is possible for one employee to appear in several records. Fourth, when all hours have been convered into costa the program will type the total direct and indirect labor cost which should agree with that of the payroll register. It is obvious that the adjusted pay rates of personnel other than those of straight time will not be the same with their original base pay. These adjusted pay rates may contain as many as eight significant digits. This problem is similar to that of the payroll register and payroll checks of rounding or truncating products. To compensate for this each product (rate times hours) is half adjusted to the nearest penny and all truncated parts are accumulated to form the truncation or rounding error. This which may be negative is assigned to project one, administration, to activity one, general, and to employee zero, fictitious. Hence the gross payroll figure plus the rounding error of this program should equal that of the payroll register. Fifth and last the fields of each record are changed back to alphameric and are punched on tape to produce the Sorted Labor Distribution Data Tape. F. Labor Cost Distribution This program is essentially a listing with some accumulating of the sorted labor data tape. This tape is now in project sequence. Within each project are the records of each man who worked on that project during the last pay period. Within each man are the operations which he performed on that project as well as the distributed money. There are actually three phases to this program which give three separate reports. The first report shows by employee number all the employees who worked on each project this current period. All the activities for each employee are accumulated and this accumulated money total is segregated into direct and indirect labor and personal expenses. The second report of this program is merely a listing of the totals lines for each project. It gives a neat, condensed, summary report of what the projects did this period. This is accomplishee by storing project number, total direct labor, total' expenses and total indirect labor for each project while producing the first report. 0 ",: , , --------.-----~--. 9 The third report is the pro ject sta'tus report showing all pro jects of the company whether work was done or not on all of them. For this the Project Master Tape is used. As the master is read in the project number from the tape searches the memory (see previous paragraph) for a match. If a match does occur the figures are combined and the master is updated. If no match occurs (no work done on this project this week) the figures remain the same. There may be printout options at this point. One which the user would probably want is a project-todate report showing three totals for each pro ject. Line 1 shows project totals to-last period, line 2 shows project totals of this period (zeroes may be present) and line 3 shows the updated project totals (sum of lines 1 and 2). ~~., I, ~ point of interest to this program as well as to the last of this series is that of the grand totals appearing at the end of the various reports. Fortran handles only eight significan figures thus limiting totals to $999,999.99. If the elements of a total added to more than a millien dollars the machine computed total would be a truncated figure sometimes several dollars off the actual sum. This problem may be eliminated using Fortran in the following manner. Let SUM represent the accumulative value of an item and COST the value of anyone element of the same item so that A SUM • L COST Before incrementing SUM by COST, COST is compared with the difference between 999,999.99 and SUM. If the cost is smaller than the difference the addition is performed. If not a carry-over factor is incremented by one and the sum reinitialized by the difference between COST and the first difference. Thus in UTO Fortran a procedure could be set up as follows: BEGIN PROCEDURE 100 DIF - 999999.99 - COST IF(DIF-COST) Ill, 110, 110 110 SUM - SUM + COST RETURN 100 III KARRY - KARRY + 1 SUM - COST - DIF - 0.01 END PROCEDURE 100 where KARRY indicates the millionth carry-over factor. Note the penny (0.01) in the last statement of the procedure to take care of the difference between a million and 999999.99. o 'hI I I'i Ii I \ "l 10 G. o Labor Cost Breakdown - Bi weekly This program merely rearranges data from the Sorted Labor Distribution Data Tape. That is, it reads in all the details for a project number, accumulates the amounts by activity nuulber, and punches out a new tape, the Labor Cost Breakdown Tape, which contains the project number followed by each of the operation numbers and their respective monies. This is done for the entire input tape, so that the output tape contains all the project numbers followed by all the activity numbers and amounts. This program generates a new activity code number for each project Activity 21 representing an accumulative total of all the money from all the other operations on that project. There is also an optional print-out to this program. Any or all projects may be printed showing money spent by each emPloyee under each activity for a project. Because of space limitations and the extend of the arrays only eight employees are to be included on a page. Thus several pages may be needed for one project. H. Labor Cost Breakdown - Accumulativ~ This is the last of the Labor Cost Distribution series of programs. As one may recall the third report of the Labor Cost Distribution Program updates a Project Master Tape showing labor breakdown by direct labor, personal expenses and indirect labor. This last program gives the labor breakdown by operation within project. Because there are 21 possible operation under each project, the Activity Master Tape is actually composed of three reels of tape. The first carries all the projects with activities 1 through 7. The second aarries all projects with activities 8 through 14 and the third carries all projects with activities 15 through 21. As a reminder activity 21 represents the total cost of activities 1 through 20 inclusive. This program is executed in three passes. At the beginning of each pass the Labor Cost Breakdown Tape is loaded and the program selects and stores activities 1 - 7, 8 - 14 and 15 - 21 depending whether it is pass one, two or three. Having stored this information the corresponding Activity Master Tape is passed printing the breakdown report and punching the updated Activity Master Tape. The report may take a similar form to~hat of the project status mentioned under the Labor Cost Distribution Program. The company that developed these programs however chose to eliminate from this report the top line showing project-to-last period totals. This final report is an extremely comprehensive breakdown of the allocation of funds on the various projects. When a project is completed, the final master records become an invaluable reference ,to be used in the future for estimating costs of similar jobs and in a way to predict or forecast job progress and man power alocation. 11 IV. TIMING CONSIDERATIONS AND LIMITATIONS Below is a list of limitations to these programs: 0 (a) The number of employees is limited to 130 in anyone run. Slight variations in paying systems and in state or local taxations may increase or decrease available memory and thus introduce new limitations on the number of employees. (b) In the Labor Distribution Data Sorting Program there is room for 1048 records. (c) An increase in the number of activities will increase the number of passes in the Labor Cost Breakdown Accumulative Program. (d) The following timings are based on 90 employees working on 35 pro jects at anyone time and reporting their time in about 500 records. The total number of projects being reported on the distribution and breakdown reports are 70. It should be kept in mind that these figures of time are for a bi-weekly payroll. 1. Time and expense sheets are turned in by Friday night. 2. Time and expense sheets are audited Monday morning. 3. Revisions and adjustments to the master tapes ••••••••••••••••••••••••••••••••••••••••• 1.00 hours 4. Data Preparation Program .••••..............•... 5. Payroll Register Program ....................... 0.75 2.50 " II 6. Payroll Checks Program ••••••••••••••••••••••••• 2.25 n 7. Payroll Deductions Register Program •••••••••••• 0.50 II ................ 0.75 ff Report #1 ....................................... 1.00 II Report #2 ...................................... 0.25 II 8. Labor Distribution Data. Sorting 9. Labor Cost Distribution Program Report #3 (project status) ••••••••••••••••••••• 0.75 " 10. Labor Cost Breakdown - Bi weekly without any print-out •••••••••••••••••••••••••• 0.50 (allow 5 minutes per project print-outs) " 11. Labor Cost Breakdown - Accumulative •••••••••••• 2.25 It 12. Average minimum total •••••••••••••••••••••••••• 12.50 It G ZUMTfi'IYW 12 o The secret to a smooth and fast run is to avoid last minute changes, and to train the employees to report their time and expenses properly thus eliminating corrections and revisions while a $1600 a month machine is running doing nothing but tape corrections. v. SUMMARY It was the purpose of this paper to present guide lines and documentation for implementing a payroll and labor distribution application on a basic paper tape 1620. It takes but a little imagination to develop additional utility programs to handle information developed by these programs to prepare special reports such as Employee Taxable Wage Reports, W-2 For.m Information (to the best of the author's knowledge there are no W-2 forms that will fit the basic 1620 typewriter), Sick Leave and Vacation Tally Reports and a number of others. With slight if any at all modification the distribution programs may be used to handle other distributions unrelated to the payroll. These programs have supplied their developers with up-tothe-minute information on the financial status of all their projects information which in the past was delinquent and incomplete. It is the authors' belief that such programs as these presented in this paper although not money making will provide a user with invaluable service. VI. APPEND IX Included in the appendix to this paper is a general flow chart of both program packages and input-output samples of their most important phas~s. The back side of the the packages. In the register represents the horizontal sums grand total is obtained twice; horizontal sums. time sheet is not used in either of of payroll deductions the last colwnn of the other columns. Note that the one from the vertical and once from the Two projects of the labor distribution report are shown on same page. Actually they would appear each on a separate sheet the carriage being controlled by the program. Totals of the project SUlllllary are obtained at the end but are not shown in the sample. The project status at the bottom shows a sub-total and a total. The developer of the programs wanted to divide the projects into two groups each having a subtotal. The subtotal shown is that of the second group of pro jects • t~ o 13 The bi-weekly Labor Cost Breakdown report shows a partial listing of project No. 61151:' There are only 16 of the 20 activities shown. The number acress the top indica,te the employee numbers. Had there been more than eight employees a second or even more sheets would have been needed. In such a case the column headed ACC would indicate the accumulative totals to that sheet. The last display shows a partial Accumulative Labor Cost Breakdown report. The numbers 15 - 21 across the top indicate the'activity numbers. There are actually two more sheets to this report for activities 1 - 7 and 8 - 14. Note that activity 21 is the sum of the first 20 activities. ~\ ~; o 'I" jiU "ii" r .. -- *i't :'@i'j"'rftli,U"i&",Ii' " "#!i W"-ifE f# -"""-W ""TEii8"fHi*'¥"""E"R"" ],"'TWI5---'"" "-" TrWKw""YW 2- I#- GENERAL FLOW CHART - o PAYROLL - lidding Time SlJeets MaclJil7~ 70~ of Hour,$ and EX,CJenses ~K -·-'L-.-/---II I.-- Expense 5heets L3olonce ;riflJ fhi~ ~,a!' lIoars O'~ £.xp~/'}se s I Totdls ChecK Labor f)istribufiol7 Oofa 7dpe Dofa 70 PreporQTiol7 Program Ad./'vsfed Poy Rare..s ?Ope LaborCotjf Ol:J fnol./tlol1 Proql'?:1m Pac~of1e Payroll 001'0 hpe Payroll Chec.ks Payroll Regisler Program Program Poyroll A1Q.s/~r Tape tlpdo/~d 'pqyro// MO.5f~r 7Ope. Use it? Pay- /Jed. Re9. ·0 t:JaIa/Jc~ OJeds -~ fo ...-- Pdyr()/I £e?isler Payroll Checks GENERAL FLOW CHART - Lfl80R. COST DISTR./I3UTION 0 Lobor '/)i6fr: ~-~ ";'1 . /Jato Sorf. ~----' Ad./v~f~d Pay Rofe.s ?Ope Program Lahor Di.5lrib. Dafa ?Ope j Pro.leef Masfer ?ape Sorfed Labor Oi6frlo"f/on Oofa 7bpe. Labor CO.$f fJreak.(8/JI'eeJ) Opfiol'Jd/ .Program Labor COst Olslribuflon f7roqrCln7 8reak. ,ee,oorf (8iJl'~ LaborCosf Breakdown Tope t/pdafe ProJect A1asler lOpe _-..1..---__ Acfillify LaborCosf Bred/:::. .(Ac~UIJ) A1a~f~r lOpes Prog/?::J/77 LoborCOsI P,sfn'b" Oi.slr;buT: SU/7Jl17Qry t/pdofed Acf/I//ty Mosf~r I I I \ Topes Prqjed .5faftls t ~ f}%nce • I Report- 701015 ,.J 70 Payroll Reg is fer ~7 o o a 'ERDMAN - LOCATION. DAILY TIME / HOURS W -- 8 3 8 /3 LOCATION; R~esfe;; /\I. Y NAMEJoha -- T - F TOTAL PROJECT HOURS NO. ! i i TITLE: WEEK ENDINGI OPER. J. ../oh/JS -# 1 Protl:smon /1-22-~3 ! - T M S DAILY TIME REPORT I ! //-/5-~3 --OPER. DAILY ERDMAN 8 ANTHONY I REPORT John ../. John s ~11 TITLE: fl. 1'1. '/0 'S/nQI? WEEK ENDING: S A'NTHO NY NAME: Roc/;e.sler; Il/. Y. NO. o DAILY HOURS NO. S .3 8 M S T TOTAL T W PROJECT NO. HOURS F I I ~- 8 8 8 Z - 8 2214. I 8 8 24- 22/4. I 24 2214. J 12 10 1114. 13 I I ~ 4- 4 8 8 220/. 8 //14. J I ~ 4 .4 /8 1 ~ J I ! 4- 4 /9 ! ! I i ,---- -- f - , - ---~ ~--- ---- _ _ _ _ .w I 1 - . _ - - - - ~'-------I I I ----1 ._- --- .... i 1 ---, I I . - - f-. 8 8 8 /0 8 s'otement #kr v this rl • () Employee( JSitr1tire ~A.A ,T'Y.£..J , ...ad II I 4Z 1 HEREBY CERTIFY that the ti"" reported above is a true and comp.et. fA myll~.lr period. I liz I Zr it fl'.:h. I' ---- 11 ... ! I HEREBY CERTIFY 'that the time report.d above ia a true cmd complet. lng p.rlod. sto(U' . 1/ Empoye, ;511 no(0/" '-' p-YV Agpd I II I I. I I l 2 ~ (J\TTACH RECEIPl"S WHEN ~BTAlNABL~ }'Olt ALL ITEMS) ==ii:=:;=='~=~~=;==;==~=::::::~==T=====\ii;=1 ====:.~~=. A~~ [~T ===:' (Please print) statement of K'£NETfI Date Itjtl;/63 /p.zj&3 - Brief statement of itinerary, etc., showing nature of expense items listed Day in columns right ~i S IIIi T Jab //l~?eCfiOI? ~1/5/. JOb /PI/51, 1115jlPch'oJ? ===: Mileage II 11M! II r and Transp. Lodg- 4.DO 6.50 ing ....-: Other Meals (Itemize) . 7,50 ~y ToW /8.00 4.f)O 4.00 I __..--~ ;./ = S I 11.1: -#:'7 ff~ JII/, !3ROWN Man~ Address ~= I~_~===Y_! W II T =====:::~ F -== -=== ==== --=-: t I S S l\{ IpJfo3 I~D#3 T w Job //J.5jJecfi olJ . 5VfJj'it'f's 4.00 60/5/. I 22-14. il , Zoo 3.00 Z,2B 2.28 /,2.tJO 0,50 10.50 2#2/3 8QB \I I'I T - F ~iEREBY CERTIFY that this 1s a true and c:)mple!c st.atement of authorized expenses incurred by me in connection with company b-:.lDineSSl for thls p1lY'rC!'9 period. J Employee • keMdlt (J). h1!!-o/ ~~~~~.----------------------- r? ------ 0 ----~- - --- 0.. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~j l ERDMAN AND ANTHONY, CONSULTING ENGINEERS ---'1 i2 2/63-------------------- PAYROLL MASJER.J'APE- - - - - - - - - -- ---- - --- -- - - - - - - --- - - -- - - - --1 ----------~-~----------------------------------------------------------------------------~ ~ ) kmpL Pay - ____4 ___A:. - - - --- - - - - - -- --- - --\ KENETH W. BROWN. 7705.92 174.00 1064.21 164.2~~L29 I 94 1 33 46 1 .00 5. 50 • 60 " i i ------------------------------------------------------------------------------------------, ______ 3__ 3.!~SQ_~__ _.. Q9 ___ .!1l1l __ ~... ~___ ...,oj_1_.l29j_.OO_____._O.0_j52_.J 0_. ___ 2J __.61 ____ .______________ J l MARY BENSON 126 34 9929 .00 • .00.60 6909.45 174.00 813.16 108.81 515.68 ____999 ___ .1009_ 0___ -,.00 ___ .. .0.0 ___ ... .0.0___ ...o.o_!)_____ .. DD__ __ ....0.0_______0.0__. __ ....00__________________ .: .2 _________________ .. _________________________________ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 1' _________________________________________________ ______________________________________ .__ ~ ~ ,,/9 TARtE OF o ADDRESSF~ The addresses shown in the table represent the address or the left most position of each Tariab1e. Em'RY NUMB. MAN REG. M. NUMB. 1 19830 HOORS 18530 HOURS 17230 2 19820 18520 3 19810 4 PERS. EIP. ENTRY MAN REG. NUMB. NUMB. mRS • OVT. HOUJG EXP. 17010 15nO 15930 • 23 19610 HOURS 18110 17220 15920 24 19600 18300 17000 15700 18510 17210 15910 25 19590 18290 16990 15690 19800 18500 17200 15900 26 19580 18280 16980 15680 5 19790 18490 1n90 15890 27 19570 18270 16970 15670 6 19780 18480 1n80 15880 28 19560 18260 16960 15660 7 19770 18470 11170 15870 29 19550 18250 16950 15650 8 19760 18460 ·11160 15860 30 19540 18240 16940 15640 9 19750 18450 In50 15850 31 19530 18230 16930 ]5630 10 19740 18440 11140 15840 32 19520 18220 16920 15620 11 19730 18430 11130 15830 J3 19510 18210 16910 15610 12 19720 18420 11120 15820 34 19500 18200 16900 1.5600 . 13 19nO 18410 17110 15810 3:5 19490 18190 16890 15590 14 19700 18400 InOO 158CIO 36 19480 18180 16880 15580 15 19690 18390 1.7090 15790 37 19470 18170 16870 15570 16 19680 18380 17080 15780 38 19460 18160 16860 15560 17 19670 18~70 17070 15770 39 19450 18150 16850 15550 18 19660 18360 17060 15760 40 19440 18140 16840 15540 19 19650 18350 17050 15750 41 19430 18130 16830 15530 20 19640 18340 17040 15740 42 19420 18120 16820 15520 21 19630 18330 17030 1.5730 43 19410 1811.0 16810 15510 22 19620 18320 17020 15720 44 19400 18100 16800 15500 · c· , r --l:)-------------------------------------------------------------------.---------------~ t ----------------~-----------------------------------------------------------------------_ . _---- ------------ . -ff722163----2~:----------------------------------------~--------------------------------- __________________________ EBQ~A~_AMQ_A~I~ONY,-CONSULIING_ENGJNEERS _______________________ ~ i _ _~REG.H OVT.H REG..P OVT.P GROSS EWJ FICA SWT DED EXP NET. __ i -JOHN-J:-JOHNS---11858~-------------4443~75-----------------------------------------------1 ____ 1 ___ 80 .00_10.00 __ 180 .00---33. 7S--213 .. 1S---1 .. 1S __ 38 .. ~8 ___ 5.29---- 25.60- - - __ .• 00_--136.63--., ~ -RAItY-SENSON----- f 1860: ---.- ----. -----7169-:45- .-- ----------.--------- --- ----- - -- -- --- -- - --- 3 80 !O _.28 .083.1 L ... 10 .01._ . 6.34_225.12_ I -.! ~ -------------------------------------------------------------------------·-----------------f ~(\\ 80&10.9-6 ____ ~ ___ ~~QQ_~?~QQ __ ~~?~Q~ _____ ~9Q __ 33~~Q~ ____ ~QQ __ ~§~ll ___ 1~J~____ ~~J9 ___ 3J~~~__ 3Q§~~j )U ~ I __ 1 """KENnA W. 81foWtCrf859. -D ~ ------~~~!~-OVT.H I REG.P GROSS "." .'--" OVT: 'P'" - - -.'. -.. -_. - _. _- Fl CA- - - _FWT ........ -- -- --- -- 'SWT-- - - - --"OED .."-- - -- . . -EXP--- - - .. -"NET ". -". _. I f' ~ ~ ----~-----------------------------------------------------------------I 1 _!9!! - - ~~ !QQ-IiO:OO- _1Z~ .!Q~---33:"5--.oo!)..!1~- --':75- 1H..!~3_ 56- - _!l1 __ 1L - -"37:62-- ~.6B.• 56__ -'5: ------------------------------------------------------------------------------------------1 ----------------------------------------------------------------__----1I ----------------------~-------------------------------------------------------------------\ _________________________________________ --______________________________________________ J I .2 __________________________________________________________________ - _______________________ I' 11 _________________________________________________________________________________________ _ 10 ____________________________________________________________________________________________________----------_______ 9 __________________________________________________________________________________________ _ l\) ~ ~. ERDMAN AND ANTHONY 223 ,/ CONSULTING ENGINEERS ,/' ROCHESTER, N. Y. DATE [i1-;~9/63 *4 CHECK No·118 • : : EAST AVE. BRANCH ~ •I ~ 136.63 /" Or,c/C/~L Y. 1:022 3 111 00 / DOLL.. r(S .CENTS /j,ef'llr TO GENESEE VALLEY UNION TRUST CO. ROCHESTER. N. t?~~// 1• JOHN J. JOHNS PAYROLL ACCOUNT II ,6~CI 58 /"-//0>- PAY TO THE ORDER OF // ~81: 5 ~0"'8 ~8b S/6/V'/J r(/£E.5 3 111 :iii' -0 (.ri STATEMENT OF EARNINGS AND DEDUCTIONS REGULAR HRS. OVERTIME HRS. 80.00 10.00 FED. TAX F.I.C.A. 38.48 7.75 REGULAR PAY RATE OVERTIME PAY 2.250 180 .00 STATE TAX 5.29 HOSP. INS. 33.75 BONDS .00 GROSS PAY EXPENSES 213.75 SAVINGS PENSION .00 25.00 .00 .00 N.Y.DISA. .60 MISC. .00 NET PAY 136.63 YEAR·TO-DATE TOTALS PER. END. DATE 11/22/63 GROSS PAY FED. TAX 4443.75 799.88 F.I.C.A. 161.20 STATE TAX PENSION 100.58 ~ ......... CJ (j c:>----------------------------------------------------------------------------------c=r--l -----------------------------------------------------------------------------------------~ I ----'----1 - - - r- - - - - - - - - - - - - - - - - - - -- -.- -- - - - - -. -- - _. - •. - - - - - - - - - - - - - - - - - - - - - - - - - -- -- - - - - - - - -- - - - - ERDMAN AND ANTHONY, CONSULTING ENGINEERS - - - - - - - - - - - -. - - --·--1 i --------------------------REGTsTER-OF-PAYRO[[-OEDucrIONS---------------------------------PER I 00 EtUfTNt;-FffrOAY ___ ~Qy~~~~~_~~L~12~3 _______ ~----------------_--------- ___________________________________ _ -- -- --Si)AOS- -- - --SAV1NGS ---AospTfA[:-----HTsc-------------------- GROUP- - - - -- --- -- --- .. - LIZATION DEDUCTIONS DISABILITY INSURANCE ---. -. ---' ----. ---I Il ---JOAN-J:-JOANS-------y:----------------------------------------------------------------~ ________ !. QQ _______ ~ ~ .! QQ _________ .! ---XE'NETH"W:-"BROWN-' 2 • ... ________ !.QQ _________ .!Q9_ -0 QQ __________~ 09 _. _ ..__ ."" • 60 ___________ __ 00 _________ 25.60. -----_. __ t ....... ___ .!9Q _____ ,...,. ___ .00 .• (>0 _ _ ____ .. _____ 5_: 50 ____.__________ 6.• _1 0 ______ _ +- -- -RARY- BENsON- -- --- --- 3~ ------;--4-0----------~; - - ~~ ___ .60. ________ .00_ .00 .00 ___ ~ _.c _ _ ._ _ _10.01 i --------:oo-------25:00--------9~40---------~01--------1~80--------5~50---------41:71----·· ----------------------------------------------------------------------------------------_.! 41.71 .~-------------------------------------~~------------- -------------------------------------11 _____________________________________________________________________ ~ ___________________ _ 10 _______________________________________________~----------------------------------------.-------------------------------------- ~ l\) 9 _______________________________ ---------~--------------------.----------------------------.· .•-.L. _ ...... .. •._... .... ... ... . .:..--.. - -. ------ ----- .......... - _--_ _-- ---------_ -._._--_ _---_.- -_._ _-_ ------_ - L1l8o£ Cos! .J);sr.el!3vTIO;tJ ----------------------------- -------------- 23. ------------------------~----- . . . " . __ ERDto1AN AND_. ANTHONY. CONSVLll NG_ .ENGJ N.E.ERS ________________________ .____ . . _________ ._ . . _ _ _ _ _---=-PRQJECT ~H) •.. _____t . ________________ _9127163 _______________________ . - .- -.. - ~ - - - - - - - - - - - - - - - - -. -- - _. - - - - - - - - - - - - - -~ ') C ...;..' . -' ______________~A13DJ~_____ .E~P.ENSE..s.______ ~ .. -.t_Y.• ________ JDJALS ________________ . o -.06 ._____ J __________ J .eO ~.oO_________ ... 00 _________ .. .00:.. '-"" ______ j .80 __ 00 ________________ . .____ 3.7__________ 3p{)..f.o.o_________ ..• 00_________ .. .00________ 3.60_.00________________ . _-,5~ . 1 4 7 . 2 5 5 ..•.70._ .____ 91-_ .:. _________ 39~.5.J ________ ..... ft .00 152~5~_ _ _ __ 00_________ .,.0.0_________ 39-,.5j ________________ . .____ 93__________ J31..,135_________ .• 0.0_________ .,'o'O_~ _______L31.,~.5________________ . _--.-J9~7:..._ _ _ _--=1~·1...£5..:...1.:..J.5!:-_____,.Q_Q.._ _ _~.~0~0_ _ _ _1!.....!l:.....1!5~._=_1~5_ _ _ _ _ __ __ _________ ____ _91!:t,l'O________ .5...,JO_ . . _______ -',0.0 ________ .9~.o_,~________ .9~O_,~~t_. __________ .E.ROM..l\.tt .l\.ND_ ANI.tl.ONY -1_. CCjNSU.LJJ N~._ .ENGJ N.E.ERS _____________________ ._____ . llJ. 4..._. __._________________ PROJECT ...NQ. _912.]!.93_____________________ ...._. . .__________________ !.A13.DB_____ .E~.P,ENSES........ _. _...S.. 'ty_. ________ J.OJ~_L.S _______________ .,.. _ _10=-_ _ _ _1.~9~..:::..0__- -....... ,. Q.O _........__. _----'<-.00~_ _--"l5.0_!_QO~_ _ _ _ __ 22 216.00 ,00 ---------~-------------~-. 00 216 00----------------._------------------------------, .. .____ 3_5___________ ~1_.,2._5________ ... ~O.O. . ________ .,.00 _________ ~_2.J.Z5 _____ ~ ___________ _ .____ ~.5___________ 1.5~'O'O__________•.9.0_________ ..,.0.0__________2.5_•.00____________. ____ . .____ §.9 __________ J9J..,..?.5_________ ~ .PJt ________ -,.9.9________ J .9] _, ~.5________________ . 80 8.14 ...... ,.Q.Q.......____. .00 8!~-=-1~4_ __ __ ____ ______ _____ ]]5..,~~_______ ._....•.0.0.. _. _. _____~]..,,O'O ________ ~.O_2_,_9.~______. _,_~O 2. 64-'0"'.' . - - - - - - - - - - - - _..._... '. __ _-----------------_._----_ _- _ .. .. ... ... _ ... \ --- ~ --- --- _ o .. _. L~l3oe CCJS T aS7e/8tJ770AJ SU/J1/Jf/l.ey . _... ...... ". ... -..... -- -- - - - - - - - - - - - - - - - - - - -Z4-- -- - - - -- - - - -- - - - - ~- __ ._ _ - - - - - .... - --_ ..... _. _.._--- .. _9J1-.7J93____________________ . .__________________LA~.o~_____..EXPEN.SES..._. ___ JJ,!"Y... _________TOTALS ______________ _ 1• .......... ~~. 70 __ . __.. 974. 70 !.Qo 9.~Q ..._~Q _ _ _ _. .____ _l.~ _________ ~~.lJ~.9J _____ ...... _.• 00_ ". _______ ~99 ________ ~.4.4.• PJ _____________ _ .__ JJ J ~~ _________ Jl5.-!'§_~______ ...... ,.00. ... _.______ ~]~.OQ ________ j3P2.•.6~_ ._.___________ _ _ L1J 5. 117 .Q.O_ _...... ..00. ............. _._ .Q.O ttI._Q.o~_ _ _ __ .__ JJ .5l___ _______ J~9§-!'~.!3____ ._ ... 4.5.J 4........... _____ ~.9.0 _______ J~.5.~~.O_2_____________ _ .__ 1-.?.9J________ ___ J]~~.!3~_____ ...... 7?-..~.3 7_ ...... ___ J~9~.9.9________ .9_~J ~J .9_____________ _ ~.1.02. _ 140 .02 _....... _..__ .~OO __ ._..._ .00 1~_!.P.2:....-_ _ _ __ .__ 1-l9~___________ 59J~5.9______._..__...~.OO __ ..: ___ _1.19__ ]3 ________ ~_6.~.,_?3 _____________ _ __l.Z.9.5-!' ______ __ J93.!3-!'.3~______........__~.po __._____ 1]~9.9 _______ JJJ5_.3.9_____________ _ _ ~2.06. 29.5---2.5 ...........__.......00_. .0_0 295___25_ _ _ __ _ __Z.2.0J3,J1 _________ 3~5_•.o.0______ ... _. __•.00_______ 5~_.'O'O________ 3J;.9,J1.0_0_____________ _ __ 1_2J1._.__________ 1..3_.1.5________ J_._86_________ ,JI_O!t ________ .~.5,J1_9J _____ .- ______ ... _ _ ~2J3 • l49.4.~_Q4 S3-L~5 _._ ....._._ .•. 00 1.345 .J~9._ _ _ __ _ _1.2-' _4_. ________ .3.3.1.8_, jt8______ .5J _9__~ J.7_________ _•.0.0_______ 3_83.4 •.2 5... _". ___ .. ___ :- ___ _ _ _..2.2-'_5_. _________ .257_.3.8_____ .__ ""' _.... 00 _______ .21-•.00- _______ .284. 38....... - ....... . __2.218_. ...• 00 9 ..55 . 0 0 - 9. 55-----.. __ .2221... _________ J.65.J 9- _- - _ - - - 5 •.55- - __ - _- ____ .00_ - - - - -. - -.170. 74 ............. - -.-.- -- ___ 2222... _________ J.80...}3_________...._0.0_________ ....0.0________ J.80... 73-. _. -_... _. - -.- - - -_2223_.. 32 .•.55 •. 00 .00 . --32.55 ....----.----------- __. .3102.___________ .21.•. 25 _________ ..•.0.0_______ 5.0..•. 63- ________ 71.88 ...-.. . 8 _._ -, :.. ,,; 310.1-._____ ,- __ _ J29 __ J8 ______ ..1.0.2..•.59.. ______ Mt•. 63_____ - -- .873.00 ..... _. -...... - .... - - - -310.8._ ... _.4_.25.. ___•. 00 .00 _. __ .. 4.25 ...____ . _____ .___.__ J.1 .40.__________._0.0.______ .... 368.56 ... - 6 _ 5~ 105._________ 357.16.. ______ r: . 60107 •.. ". __ . 111.80 _______ ........00.. _______. _....0._0.______ . . 111.80 4 2 ;::. •. . -.- - -- 60'51 ....._._ ... ___ 1666.80. _____._.. 61...39 ___.__10. .• 0.0. __ ..._._. __ .. 1738.19 _61.1 0.2.•_____. ____ ..1556 •.7.6. _.. . ....0.0 . . ___ J.6_4.__.0_0_ .. " qb - - - - - - - - - - - - - - - -- - - -- - - - - .- -. ..' "- - -. - - - - - - - - - -- -_. - ~- ... ... - 1720.76 - - - p,eOJECT _._._---_._----,-, ST~r(JS 25 ---.---.-------, ---- ERDMAN AND ANTHONY; CONSULT I NG- ENG'! NEERS----·--·-· 10/11/63 LABOR 2203. .00 ",00 • 00 538.20 .. 00 .00 ~206 0 13 .00 0 13 4293.42 5iJ.·1.:·.oO .00 wDO 3101. -----.. ---.-----"'-,--, . -313.31 .00-' 44. L~5 ._~. ~ 3104. ----........ ---...". . ._.~ •• ••• _____ ..v··· __ ..w.· _ _ _ ··_~ _ _ _ ·•· 61.88 • _ •• _._ ..•.• -·~··oo 76.00.--- .. 198 . 1.1-0 .00 198.l.j·O 760.47 3106. ,- -----"---. -.. --- ....... " -- - ....11838.01 -.. 000·----11838.01 .--.-.. .. ----' - . 2915.55 .' -3107. ---. -.. ,,-.- - . --------.. -- ---,746.28 50.38- -- . --, 102.59 , ~ ___ ~5 7 .. 00 4.00 .00 24. 1 5, 48152.42 1 575022.51 .,48897.95 ___ • _0. _ •• _ _ _ _ _ • _ _ _ • ____ • _ _ _ • _____ ••••• _ ••.• ~ _0 _ ____ • . •. __ • " _ _ _ ._ •••.• _ ." •• - 55.93 1014.80 55.93 1014.80 0_0, ••••• " -.-~ 000 ----~OO ----~-OO"- .... 4.25 -..---..,-. -.00 ~1.8 _______ ?771 • 7~. .~Jl:t~_z.L.r 58380.27 1 625612027 .09 58944. 3~_____t_ . 94L~589 036 564-~-ff------:--·---18977 60548~30 1 664724.58 564. 1'-------- 20308.29 7Lt 5. 53 •• -.... -.. -. ---'- 214.74 ---------2723.79 - . • a0- ____. ____ . _1 5380 1 9 • 9547625 "03 _ .' •••• 61.88 - - -- .. -- - -----..-- - .-. - --1 .-- -- _..520352Q50 17667 "L} 5 --.. --- --4687~.50 745 • 53 18998.65-- ------ ----~OO"-------- _. . ,_____ ... Q5L ___________ .. _~ •. 25 .00 1 556023.86 . .00 ·'·-----'-.. c . _, . 61.88 • 00 0 2LL.15 32. 90_. _____.. . 889050 , .~ 00----·----.. · 50.38 40.63 939 88 102 ~ 59 --"62204. ,- .... -.----.------ ----- 2484.90 48.00---"---" " 40.63 .00 .25 ---~- ~OO 2915.55 ~ .. 112.54 14866. 10 --'----;00---------.... -- - .-----.. -----, 112.54 14866.10 . .... .. ..... --.-.-..-. - ......... _ .. . . -----.• ,---_ _-------_ .. _- -~-.---.-----~-.-'--.- 3108. ---._------------.. -..-----------.4.25 ."- . 00 -----,-.--.. ---.. -- 1330.22 -.-~.----~- ...... -.'00-------'-' -000 ------.._------------- .00 796.66 ...... --~OO----------- .00 ~_ _____}} O_~~___. __ .. __.. , _____ 760 • L.~ 7 .00 - Lt 357076 76.00 .. ... --------.--- - 1330.22 •00 .00 .00 --~ .00 5~~ 63 _ . .__ .________ 3686.63 7'4.33 .. ....... ---"-' 61.88 --, 00------ .-, .00 5018.31 ~ .00 __ • 180.89 50.63 2264.82 9 00·---'-·-------·-1421 .81 --.. -..-----.-.. -.. --- .... _----.. - 7 L}. 33 1179.89 w. _ _ _ • _ _ _ _ _ . _ . · . _ _ 50~8.31 .. -.. -, ·--·-··---:00---------··· -. ".00' .00 357.76 02 • 2053 • 95 60 .' 2. L;. .-----,-3-...1---.---------.---00----'--1384.89-----·-.. -· . ·, 127,,92 3438.-84 188" 16 1179.89 0 .00 l{~L!-. l{.S -. ---- - -.• 00 538 20 ~80"G9 .00 _.__ ._. _ , . _._. _______ ,_______ .._ 313.31 .00 _. 128.50 ______ 2507.57 ... 00 .00 128.50_. _____._____ 2507 .57 .. 00 5L:-4.00 , 538.20 ..CiO .00 ______..... ____ ._____ ,__ ._____ ~.2_93 _.42. 3103. TOTALS 538 .. 20 _§?'_?_Q? ~______.___________ 2206 _____ .___ . _______ s.-:-\~ 61112" ~J. _____ 1 685032 87 0 97 1 625612.27 18977.09 6445890 1 664724.58 .~ 20308.29 1 685032.87 • • .'I~~ o o ----------------------------------------------------------------------------------------. ERDMAN AND ANTHONY CONSULTING ENGINEERS ---------------------------------~----~---------------------------------------------_.---- ----- - - ------------------- ____ f.RP)J:.cJ_.NP___ f>JJ5J _________________________ ---- ----- - - - - - -. - ! --I! ' I I ---------------------~----~-------------------------------------------------------------1 _~/21~~ ________________________________________________~-------- i 3 6 13' 21 56 0 0 0 TOT ACC I -------------------------------~--------------------------------------------------------~ ___~_____ ~p..!~~_____ ~.9.9___ ~.§..!J..?___ J~~..?2-----~.92-- _____~.9______•.9.9_____ -'9J>___LLO_•.31____ JJJ)_,.3.7 .. _.: I 2 .OQ_._~90 _____•.00 _ _ .OQ__._7___ ~2_ _ _ ~Qg___~.Q_'~ .., ____ .00 . . ___1.~ . ~.? ____..... ,. 1. ~2_._: 3 .00 ~ ""\/"'l .'.JlJ .00 () ~ '\ ~ ~ ~ ~ 4.85... ._ ._ ... .l•• 85 ....!.9.9__._. __ ._ ~ .9.9.. ____ . __ .!..9.9___ .... _.~.9_0_. _____.__.9_°. _ .__ .__ • Qq. __.._,__ !t.Q9________ .•.00.. _ 5 ... ______~,~_.~_~ ...____. . ~ _ ._.Q9 ___ .. _ . .~_ !~.5..____ . __9 !.5.0 .~_oo _._ .. ___._~QQ..._ . . _~Q____ ..•. QO._._4tt.1.3. ..._... _. __.28. 73. __ § ____ ___ ~.9.9____ 2..!99____ ___ 99_. ____ ~9.9_____ ~9J>______._~LO___ _:- ~-9-0---- ... -,Oo. __. __ .5.,_0_0______ 5 •.00 __ : 4 -0 ..4.• .?5.___ .... _. ~ . O_~.. .,.... -. ..• .00__. _, _ .....9.0._ ". _......~.OO .00 • i .00 ! ~ __ 1____ ~9~..! }~___ 1.5~.9.9___ ~J__ 5)___ 51~.9.9___ 55 __ J~_____ JJtO_____ ~1tO_____ ...JtO__ }jl.9,,-Z5___ -1l7.9 .•.25_. : "~ i ~. 8 .00 5.00 .00 .00 19.25 .00 .00 .00 24.25 -.29.. 25.__ ~ I 9 .00 .00 131.08 128. 25_ J_~J.!§J ________O_O______~JLO_____ JJtO__ ~.2_2_ •.91L- ____~2..2,.94. ~ -----------------------------------I 10 - - .-- - 37.50 - - -- - - -- -- - --- - .00 - - - - - - - - .00 - - - - - _. - - - - - ~ f'.. ~ ___ ..9.9_____ ..!..9.9_______ ..9J>_______..9.9_____ -'..9_0___ .31.•.5_0_____ 37. .5.0.. ~· I I .Qo .00 4 ..Z5. _ _~.• 15_~ _J1________ ~..9_9_______ .9..9_______ .9..9_______.9.9_____ -.!.9Jl______•..9.9_____ .....9..9_____ ,,_o_Q_____ "JlQ_________ .• _OO_.~ 11 .00 .00 .00 1t..75 .00 .00 I _J_3.._._______ ~.9_9____.5__ .9..9_____ ~..9..9_____ -,..9..9___ Jlt,~.!L ____ J_o_o______.J>Jt _____._O_Q____19.• 4!.L __ .__ ....1.9. ~~4.__, I __ J~_ . ______5.~.~_?~. __ ._.___2..Q_Q_. .0c) 14 •..l5 ,00 .00___.._Q.o .00 1-0.51_______ 7.0.51 ___ _ .. 15 ..... _. _ . , .. ~.9.9. _________~.9.9_. _____~.9..9___ .;- _."..9.9_____ -,..9_0_ ~ ___ ._•.99______•..9Jt ____ J_o~t ____ .•. OQ. _________ .•_0_0__: _J_~_. _. _____ ~.9.9_. ____ ~9..9_____ ...!.9_0;.. ____ ."..9..9__ ~ __ ."..9..9______•..9.9________O..9_____ J_O"'O______ ".QO ___. ____ "J),O_ ~ l\) (;\ ______________________ -------- - -- 11/08/6) - - - - - - -- - '" ._ .-- __ ,.., . . _ ..... __ .....,. ~.,. .-.. ~ _ _ ....... .-- _____ ----------.---.. . . . . , .. '" ... -----.--~.-.-.--~.-.- , ~ ___ • ..--. .. - - _ _ _. , _ .. _. ...,...CONSULTING -- - _.ERDMAN -- .. - -_. AND - .... _..ANTHONY .. .............. . .... 4 -.. ~_ ... _ *. ___ ... _ ........__ .... " _ ENGINEERS .. .. . . ... - . ,.. ...... ... "".". ~ .. _ • _ _ _ ..... .-. ". " ... _~4 _ _ •• -.- .. _ _. . . . . . . _ _ . . . . . . -. ... . __ • _ _ ~ 9"'4 • ___ • ~ _ ••• ...... . -- .....-- ..........--.- .__ ........•• -.................................-._.__ ................. -- ......._..... . _P8-Q~ g~! _________ ! 2____________ 1§ ______. _. _____ 1..7__,. _________.1 ~ _..__ .. _____ J 9.___ .. _'__ .. _....20 _.. ____ .. _.. _..... 21 ~ - - - - - -": - - - - - - - -: 00 - - - - - - - - 60 - - - -- - - - 00 -- - -- - - - 00 - - - - - - - - 00 - _. ---- -- -. _..: 00 --- --- -. 92S'. 13 -- --.. ._ _ _ _..,---.:.00 • 90______ ~.Q.Q__.___.-l3.~lt..•_.~4._~. ___ ._11Jt•.!J.6 __________ ..1263 .67.. . ._. 31 588.98 _ .. . ~ ------2~--------~OO----- - - - - ~ ~ ~ I - - - - - - - - - - -- -- - - - - - - - -- - _ .• -- - - - - - - - - - _. _ .• _._ - _ .. _- - _._ - __ - __ ...... ___ . .. .00 .00 4.50 .00 3.25 . . ....... __ .1 722.00 ____________ .... ___ ~QQ ___. _____ .!QQ _________ .___.. P(L. ___ .__. !f35.• !fo _____ 19!t. 25..... _., ...____15.75 ._. __ ..11919.71 ~. -D .-, 1s.' () b"".." - ,,00 .00 ." 00. .". _. .11 00... •0 0 .00 .00 .....• 00 .......... " ...• 00 000 1100 . •. oQ ......... 2241 .20 _._.__._..__~_~.!l.!.9_~_ . _._~._. ___ ~.f_.J. Q"...... ___._.... -~oo- -- .00 40.00 13'~'27 1862.66 3437. 16 -----~ 00'·-- - --.. _... _. -~oo--·- --. ----- -~O(>"- --- - _. . _. _. ___________ ~ ZZ§ 1!. 13. _________ .!QQ ________ .!QQ _____. ___ -~QQ .. ___ .. _.._....... 00 ..... _, ~oo 1463.52 133755.23 -"-'' ' _............ 00'" ..... ~ __ ._..• 00..... _._ 3 7761 • 13 _. ~ -'S8rO-' -.. -··,,· . -·. -.. . -·-· . ·o--. .. .0'0 ____________ 6~Z3 z~. ~3 ____ 3~~I7.! 1~__ ._.______"QQ ___. ___ .___.QQ~_ .. _._._ ....... oP_ .......__ ,.,_,_ ,.00,. ,____ .269471 .35 .--··~·-.·oo-· --:OO-·~--~~--··-·_:_O·O-·-----~-"--~:-ho ---0·0 ·~oo -···~·oo·· N ~ ------- - - -, -" -: -" -S8 1-0lf~ - --- -----~60 _. --'"" --- - - - ~ 06 --- - - ---~oo-- ----- --~oo 00--···- -.- _.. - -~ 00·. ·- - -- -.. -. _. -... ~.()()..... . _ _ _ _1.93. ~.3 . ~_Q9_~_. __~Q.1.9..J,.15 .._._._·. .=~_"'....... ,P.O.~~~ .__~.~.J_4..oo~.. _. __ . _....____ .• 00 . ___._~__.. ___... 00 ...... _._._881 58. 71. __ . -'-5§1-65~--- - - -. - - •••••• ' .- .... -- _. -- -- .- .... -- • • • ., ...." -_.- ... 0'. -- .- •••••• -" . , .. ~ • • • • • • , .• ~ ....... , •• , -- ..........- _ ••.•. '.. .... -.. ..- _ ... _ ••• - - •• _ • -- -- ••••.• -- • • - ' .......... , ... • • •. •• • •.• _ .• ~ :. .00 .00 .00 .00 .00 . 0 0 , 235.29 . .. ___. _________ ~~Q~ .•. Q!,L.._. _.__.117 .• 75 ______ .2~!t.• 50 __ '"_ .. .1500 .• 36 .... _.... 1513 .14, .. o. _ -..31 0 • 76 .... J 22213.57 ....... ---59r(f3--:----· . ~ l.i\ 21014.22 -s:J - - - 5701': - - - - - - --:00 - - - - -- - ~ ~ .. ' 58153 ...·····...... ··. ······ .. ".'Cjo ................-..... .. 00 __ ._. _______. _"" ~ 1Q1~ . . ~ f _... _. _____.. _. _........ QO ... . - - 60 t 5 r.· ~ . '" -_.. ~ . 00 ........ " ·. ·---~OO"·......---.. --·- 0----__.________ ~7. ~.~?------.-- .• QQ _________ .• Q!:L._____ ...• 00 __ ......... - .00 .. -- ... -. - 5.85 - -- -- -- 3d 7 -=·b(r----~--·--:-()O--..--~~~---~':oo--··------~~ ·:·o·(j"---·-··-···-_o.~.~ --·-·-~·O'o ..._,.. . ~ I~ ~ ~ "l • o SPIRE (System for Personnel Information REtrieval) Prepared by: G. J. Reed Albuquerque Division ACF Industries, Inc. December 9, 1963 o /trD • c I ! 0' , I • I ~. ~ ~ I I • l. o INTRODUCTION The management of ACF Industries, Incorporated real ized that manual methods for accessing personnel information were inadequate both with respect to speed and accuracy. Many requests for summarized information either could not be answered or became a costly venture in man hours alone. As an attempt to remedy th is problem, a punched card system was designed to hold personnel information in coded form. Th is card system proved inadequate in speed, type and amount of information stored, and updating capabil ity. With th is system as a background, SPI RE was developed. SPIRE is an acronym for "System for Personnel Information REtrieval. II Th is system was designed for and implemented on a magnetic tape 60K IBM 1620 and has been operational since February, 1963. Since that date, no revisions have been necessary in the type or format of the information. 11. CONTENT The informat ion content of SPI RE was spec ifica Ily designed for ACF's use. Personnel information was carefully analyzed and organized to determine what information had been recorded by the previous system and what part of this information had been useful. By categorizing the useful information, it was possible to devise coding techniques wh ich mad~ each item unique in the system and thus recoverable. The information in SPIRE is broken into the following five categories: 1. Personal - name, date of birth, se,x, date of hire, marital status, socia1 security number, etc. 2. Previous Experience v ision, etc. 3. ACF Experience - complete file on all pertinent happenings at ACF with the employees status (job title, grade, salary, department, etc.) recorded at the time of the action. 4. Education - includes all completed levels of education and the extent, ma jor, minor, and univ.ersity, if applicable, 'for each level. 5. Mil itary Experience - contains the same information as the previous experience category with the branch of service added. , job code, position title, time on job, extent of super- I ,t I t ,I l 0 i I I t To obta in th is information in coded form for all employees, superv isors interviewed their subordinates and coded their education and experience backgrounds. The ACF experience involved so much detail that it was necessary to code each part of the employee's experience using the permanent personnel records. As new employees are h ired, the ir backgrounds are coded and added to the master file. I • ~ . I /DI III. APPLICATIONS Although SPIRE was developed as a general purpose information system that could be used for the generation of virtually any report involving personnel information, one appl ication appears to be the most notable, that of in-plant recruiting. Prior to the development of SPIRE, to fill a vacant position from within required the personnel in Salary Administration to have an intimate knowledge of the background of all salaried personnel. Th is, of course, became a virtually impossible task as the number of employees increased. Qual ified people could easily have been overlooked, and the preparation of readable resumes for the managers became a monumental task. SPI RE now assures that no qual ified individual is overlooked and automatically prepares the ir resumes Th is resume, as printed, contains virtually no coded information, thus providing management with a concise, complete, easily read and understood document on any salaried employee c' 0 0 Of course, qual ification is a matter of definition. A professional analyst takes the vacant position and determines the background necessary to qual ify an individual. The analyst specifies acceptable work experience and the minimum years of experience in related fields. He also specifies the acceptable areas of education and the minimum acceptable level. From the years of experience and the education level spec ified, a minimum number of "points" are computed by the system which will qual ify on individual for the vacant position. Individuals may accrue the necessary points in any combination of related experience and education. Thus, if an individual had a masters degree in an appl icable field and only a bachelors was required, he could qual ify for the position with less experience than that specified The equating of education and experience used in the SPIRE system is based on the ACF position rating plan. The analyst also spec ifies the acceptabl e sal ary grades for the search. Th is keeps sen ior men from "dropping out" for junior jobs for which they would be qualified. In addition to the preciseness with wh ich th is search for qual ified employees is conducted i a considerable savings in both time and money exists. 0 The normal time to complete a search for any given position varies between six and eight minutes. This same job accomplished by manual methods would take about six man hours. Another appl ication which has saved a considerable number of man days is the generation of the quarterly Merit Rev iew Report. Th is report is by department and contains pertinent .information concerning past salary increases for all employees who are to be reviewed during the next quarter . The normal time to complete th is run is about 15 to 20 minutes as opposed to approximately 32 man hours 0 Other current appl ications of SPIRE include: creation of summary reports on salary increases for any time period, skills inventories, projections for salary budgeting, preparation of data for salary surveys, current salary status reports, salary distribution studies, and the generation of resumes for specific purposes. 10 J.. o '~i""'tF"jj£#ttjRriii" • '""["TV [Tf ...... _······*jiiBihN ,Silli9W·jijWijpq---J· s----mY' W W"i'WTWT' . 'TW"WfTTwqrnoo • IV. o COMPUTER TECHNIQUES SPIRE is centered about a master information tape. The information stored on this tape is first punched into cards. Each card contains all the pertinent information about a single happen ing in the employees past. Every card entering the system is identified by the category number and the employee IS permanent number. The remainder of the card contains identifying codes for the data as well as alphameric descriptions. The master tape is blocked by employee. All the information about any employee forms a single tape record. Within th is record the information is held in card images sorted by category and date where appl icable. Th is means that the maximum number of records to be read or bypassed will be the number of employees in the system. The employeels current salary grade is carried as part of the first card image so that for many appl icatioris a decision can be made as to whether or not a particular record is of interest to the search before actually performing a resume scan. This is particularly useful in the recruitment appl ication. The accuracy of the information contained in the master file is, of course, quite critical. To provide this accuracy, a great deal of time and effort has been placed in the program used to update the file. A total of 47 different errors are uniquely detectable in the updating data. This program allows total resume replacement or deleting, individual card replacement or deletion with in any resume, and the addition of any new information to any given resume. In general the customary way the master file is used in supplying information, other than complete resumes, is to scan the master tape and write a summary tape of the requ ired information. The information on th is summary tape may then be sorted or interrogated in any prescribed manner. Several of these codes that prepare summary tapes have been written In the ma jority of spec ial requests, one or more of these summary tapes can provide the necessary data. In these instances it is usually possible to write a program using ACF's magnetic tape FORCOM system which will interrogate the summary tape and prepare the necessary report. This provides a quick, economical way to satisfy demands made on the system. 0 V. CONCLUSIONS SPIRE has proved a complete and useful system. It has resulted in large cost savings and has been well accepted by management. Th is acceptance is born out by the constant increase in requests for information from th is system. To date there has been no request that could not be satisfied because the data had not been included in the system. o SPIRE was designed specifically for ACFls needs by ACF personnel and, as such, it may or may not be appl icable to any other organization. Th is system is operational and in fact is far exceeding the original expectations in speed and versatil ity. 103 • •••• ACF PRIVATE ••••• ACF PRIVATE ••••• ACF PRIVATE •••• ACF PRIVATE ••••• ACF PRIVATE .••••• ACF PRIVATE ••••• ACF PRIVATE SECTY-STENO 4N HS GRAD PLUS SHORT SPECIALIZED TRNG 1-3YRS EXPR fEMALE - MARRIED 61205 BROWN S A *.... PREVIOUS 1 YR. ••••• JOB TITLE DATE 0 08/01161 04/01/62 07/16/62 09/0l/62 EXPERIENCE **.*. STENOGRAPHER TYPIST 1-3 YEARS ...c- DATE OF REPORT - 01/01/63 JOB TITLE Tlt-1E CODE MO~.- 11 DATE OF HIRE - 08/01/61 OATE OF BIRTH - 08/16/31 3 TOTAL POINTS SEX-MARITAL STATUS PERM. NO. NAME TYPIST STENOGRAPHER STENOGRAPHER STENOGRAPHER EQ. - 1 YR. COL. HIGH SCHOOL YR.COMP. 1961 1955 SALARY DEPT. GRADE 284 314 325 345 108 108 108 108 2N 3N 3N 3N ••••• EXTENT CODE EXPERIENCE ACF EDUCATION MAJOR TITLE SHORTHAND COMMERCIAL ••••• PERS.ACT. NH RP SA MI RATING V. GOOD PROM. SHIFT 6MO-1YR 1 1 1 1 NEXT REV. 1-6 MONTH 5-ANNUAL ••••• MINOR TITLE UNIV. CODE o o or c o • .. iN iij·rlriM --W EH¥SS"" i¥tit6·-tHf···Wf "i& ··_·",·tiFY±i &6£+·" riH£Saii5ri5¥ttliHliiW--.-••#irii5Rfij f itFT. -iri6"f"i'tRiiif#iiW··... ··.I·irFFittiWiHitiRH±, iF... · -0--- +Hfri·. iiii!i;i _··-•• %6 -55fEcEFiR*hf"iri·_··b.. ··-iicwHlM.*rifHilFFriWtiWfi--PHrW;iwdftiFHH·"jf--\-wHdH¥i6iAf""tWzPdHtlSF9iitiiiii*MFH.J!j'-··fffBf¥wS"f. n POSITION REUUEST ·FOR EDP ~L-314 POSITION REQUESTED BY 0 HAS TH'S POSIT.ION BEEN .-EQU·ESTEDPREV I OUSl Y? POSITION rvl YES ~NO l'F" YES, ARE PHEVIOUS ~~.~I~MEN~=-~':QUA TE 0 ° TITLE AND CODE 1,3 ,7 , 1 ,. ,4 , 1 1 C~ 0 YES NO IF" NO, SHOW REQUIREMENTS OR ADO IT' ONAl REQU I REMENTS BELOW REQUIREMENTS a. ,4, N, , H. S, G R ,A D ,p, L ,U , S, IS, H ,0, R , T , . X, ,P R, , , • E D T , R I NG 1 - I 3 I , Y , RL.S E S P E ,C I I ,A ,L,I Z " I , 1 , I I ~ I CIT I Y I - IS, T , E , N, I I NO. I I , E:XP CODES ,0 ,8 EXPERIENCE REQUIREMENTS COOES I f TITLE NO. EO EO CODEI EXT ,2,5 ° ,4,4 CODES 1 ,3, 7, 1,0 I 41 1, C STENOGRAPHER 1 ,3, 7, 1,0 1 4, 1, C SECTY - STENO ,,,, ,1 I I , , , I , I 1 1 , , I , 1 ,,,, 1I , I 1 , 1 1 1 ,,,I 1 , ,,,,11 1 1 13, 3, 0,1 I 41 2 I C SECTY DEPT HEAD 1 1 11 3 17 ,3,2,0,0,0 1 ,3, 7 I 1,2 1 ° I ° 1 ° 1 ,3 17, 1,0 , 0, 0, ° 1 13 13 1°,1,0,0,0 1 13 I 71 3 12 I 4, 1, C I , , , , , I , , ,, , TYPI ST (PREV.) STE NOGRAPH ER (PREV.) SECTY - STENO (PREV.) SECTY DEPT HEAD (PC-EV.) TYPIST ,,, ,, , , I CODES 1 1 , ,,, ,,,, I , CODES T'TL.E ,,, ,,, ACADEMIC HIGH SCHOOL I SECRETARIAL SHORTHAND J 1 13,7, 3,2 TYPING I , 1 , I I 1 I 1 , 1 ,,, 1 I , , , ,, , 1 , ," , ,,,, I 1 , , I I 1 , , , I I o I I 1 I I I ,, I , ,,, ,,,, I f I , I I 1 , ,0 ,2,9 TITLE I , , ,, I 1 1 1 I MI N GRD t ° , 1 , 1 , 0, ° 1 ,3,3, 0, 1 1,3,7,1,2 I EXP ,0 , 11 1 , , I YRI I TITLE CC 11-80 1 MAX GRO ,0 ,3 j II DEPARiMENT NAME AND I NU_'_\B:~ I REVIEW PERIOD DATE GRADES I PERM. ,. NN,\~ NO_ I 'I J' JOB T,LE ~'I DATE DATE "",m" '" ,m MIN =0325 MIOPT =0 13 ITOR ·0404 '3" 1°'>102159 61206 - PRES. M" ...... ' i ! S HNIlGKAPHf.R SMITH M F I i 1~:ITIAl I"~ PROM I Gi{AOE =03N MERIT REVIEW ____ .Al-44_9:.I.f J..=1-.sz,63,u1_ _ __ A C F INDUSTRIES ~ TYPE Of I ~,. HCOM. INCIIfASI! """": I I rr T MAX .04JO IfftCfM "~ . . . . , . . . - --- .. :~lO I AN:~UAL - -.-~.--.- 10. I 25 101115/62 25 07101162 08/01/62 30 ! GKADE =04N [BRowr~ - S A 61205 wYL::::!~ MIN =0360 MIOPT =~~13 !4;~ SEC TY-STENO ;ro'{ =0450 :08101161 MJ\X =04 dEn rill. 10 5 j') 1°1116/63 1315 6-MO. 1439 ANNUAL -.-~.-- .... --.---. -- ! SfCTY-ST!:':-':U J K _4N 01101/57 05/15/6J 30 09/01/61 i I 61207 2'5 I Dc/OlN,' I 25 --+ --------- --+-------------I Gi\ I\f)E CLOUGH G ..... =J5~-J MI;\l =0400 MI')PT =0 60 SECTY DEPARTMENT HEAD ~ 7120R 5N TO .... =C412 i 01/06/59 \03/01162 /;IH iJ':PT =()5 0 [Ve. It) 450 40 6-MO. I o 0"'---------+------------ -~.--- - ------ --- ---- --------- -- ....... --------........ -- -.---------+- ---------4-----_--- I DEPARTMENT MA~jAGEI! - o DATE I II II ADMINISTRATIVE MANAGER--- - - - - - - - - , DATE -+--~-- I , ! APPROVAlS ------+ I I GENERAL MGR.ALwooiROOe- DIVisicit- -+--+-----1'--'---+-----I 1f-c-O;;:DA-:-;T:;-E~-nU-ANAL::::::;--:m:;:--;:---'-----'-------'-----,,---::'fI=:;lIid~;=&jft;or.;"'--~__iiiiiiiiiMii-=-'----'--:.....----- ~ ." o A COMPUTER PROGRAM FOR THE CALCULATION OF PRIME IMPLICANTS FROM A LIST OF BOOLEAN MINTERMS THOMAS R. HOFFMAN UNION COLLEGE SCHENECTADY, NEW YORK o ID7 p' , . A Computer Program for the Calculation of o Prime Implicants from a List of Boolean Minterms I Introduction Boolean algebra has proven to be a most useful tool in the logical design of digital circuits. In general, the method employed involves four phases: 1- definition of the problem (verbal statements) 2- translation of verbal statements into Boolean equations 3- simplification of the equations 4- implementation of the simplified equations Step 3 - Boolean simplification - can be done in a number of ways. ·1 method is attributed to Quine.. One The program to be discussed here enables the 1620 computer to simplify Boolean functions in a manner very similar to that proposed by Quine, obtaining as a result all the prime implicants of the original function. II The Quine Method 2 To simplify a Boolean function by the Quine method, it is necessary to express the function as a sum of minterms. The N-letter minterms are then systematically compared with each other two at a time, and all pairs differing in the state of only one letter are combined according to the identity: XA+XA=X where A represents the letter eliminated and X is all other letters in the minterms involved. After all possible comparisons have been made, there is in general a list of terms having N-I letters. If any minterm all during this process, it is a prime implicant. did not combine at Minterms are checked as they are found to combine, to facilitate spotting of "non-combiners". Comparisons continue among the N-I letter terms, with the restriction that only terms containing the same letters can possibly combine. o The N-l letter terms fall in as many as N different sets, and further combination is possible only within these sets. Successful combination of N-I letter terms produces N-2 letter terms. Again, uncombinable terms are prime implicants. This process is continued until no further combinations are possible. Uncombinable terms at any level are prime implicants. The Quine method then goes on to select the simplest set of prime implicants that is logically equivalent to the original function. This last phase is not handled by the program to be discussed - its output is simply a list of all prime implicants. The reader is referred to Reference 2 for a much more detailed description of the Quine method, along with numerous examples. III The Program - General The program to be discussed was written by Mr. H. Huhta of the General Electric Co. as part of his M.S. thesis project at Union College. The required inputs are: 1: KV - the number of variables (cannot exceed 9) 2: KN the number of minterms in the function (cannot exceed 60) 3: the list of minterms, coded in octal form (see IV A) The resulting outputs are the prime implicants, coded as described in IV D. The FORTRAN-O language was used. The following paragraphs point up various aspects of the program organization. A detailed example is included, together with Veitch diagram verification of its correctness. The Union 1620 has only 20 K memory. program may not fit. If both KV and KN are large, the To insure the ability of the machine to cope with any 9-variable problem, a second program was written to perform a single level of reduction. Thus the output, with 9-1etter minterms, would be 8-letter terms which would then be reentered to produce 7-letter- terms, etc. The second program will not be discussed further in this paper. The first program has been shown to handle an 8-variable, 24-minterm problem successfully. The exact limits are not yet known. IV Program Details (A) Input Coding Input minterms are coded as fixed-point octal numbers. octal~digit Each specifies the state (i.e. - complemented or not) of three variables, hence three octal places will suffice for a 9-variable IO~ o , \"j""(jmj""')iifli"' 'MY' """,i",mi,"ijmii"j " o problem. For example the 5-letter minterm ABC D E would be entered as 25, using the conventional scheme of representing an uncomp1emented letter by 1 and a complemented letter by 0 - but converting to octal (25) rather than decimal (21). (B) Comblnabi1ity Terms at any level are combinable if: 1- they contain the same letters 2- all letters but one are in the same state Satisfaction of requirement 2 is detected by a two-step process. First, terms must differ in only one octal place, and second, the differing octal digits must differ in only one binary place. The fact that the second of these steps can be carried out at the octal level follows directly from the relationship between the binary and octal number systems. A table L (I) contains the necessary information. This 43-entry table relates Reference Octal Digit M· and Compared Octal Digit N. It is addressed by the number N + SM, and contents of each address is either 4,2,1 or 0 - to indicate the weight of the differing binary place. 0 means: "These two occa1 digits do !!2!. differ in just one binary place, hence they are not combinable." Terms at any level in the process are always arranged in ascending order prior to comparison. Thus if M is larger than N, it means that the terms being compared differ in at least one other octal position hence they do not combine. The entries for all cases of M) N are there- fore 0 in the table. (C) Set Identification As mentioned in II, comparisons at all levels beyond the minterm level take place only within~. The program must therefore generate a set identification number ( SIN) to identify each combined pair. The SIN increment generated by any particular comparison is simply the octal weight of the letter eliminated. As terms pregress through successive levels, an up-to-date SIN is maintained by cumulative addition of the weight of each eliminated variable. o (D) Example - to determine the prime imp1icants of the function: ItO j', (1) List minterms in ascending order in octal systemm3 .... 0003 m 4 -+ 0004 m5 ~ 0005 m7 .-. 0007 m9 .... 0011 m lO ~ 0012 m ~0014 12 m 13 -+ 0015 (4 digits are shown, because this is the way FORTRAN 0 represents fixed point numbers internally. Leading O's need not be typed.) (2) Investigate all possible combinations. Start with 3, and compare each term with all higher-numbered ones. ex: 3 = 011 differences.~ three binary not combinable 4 = 100 3 5 3 7 t = 011 = 101 two binary differences, .not combinable = 011 = 111 combinable; the most significant binary place is the one that combines The decision as to whether or not a pair of octal digits is combinable is made by referring to the table L (I), as discussed in IV-B. In the 3-7 comparison just described, the most significant place is the one that combines, so address 22 (7 + 5 X 3 = 22) of the table contains the number 4 - the octal weight of the eliminated variable. the SIN of the combined term. 4 now becomes It is attached in the 5th place of the lower-valued of the two combined minterms, and the result stored to await processing at the next level (5 places are now possible because the combined term is given a floating point FORTRAN name). This process continues until all possible combinations are tried. Minterms involved in one or more comparisons are tagged by putting a 9 in the 4th (most significant) place. If any minterms have not entered into any comparisons when the processing of level N is complete, they are printed out. They are prime implicants. III o "!"'''i'iI'' '-WWWRZU • The complete level 4 results appear below: o Initial Level 4 Final Level 4 Transferred to Level 3 Combination of min terms 0003 9003 40003 0003 and 0007 0004 9004 10004 0004 and 0005 0005 9005 100004 0004 and 0014 0007 9007 20005 0005 and 0007 0011 9011 100005 0005 and 0015 0012 0012 40011 0011 and 0015 0014 0015 9014 10014 0014 and 0015 9015 At this point, 0012 (minterm ten) is outputted - it is a prime implicant. Output coding will be discussed in paragraph (3) Processing of level 3 starts by arranging the transferred terms into sets having identical S I Nls, and putting the terms in each set into ascending order. 10004~ SIN The result is: =1 10014--' 20005 S I N- 2 40003} S I N 40011 =4 10000')- SIN = 10 100005 Digit-by-digit comparisons are now made within each set, just as before. Since the S I Nls of two terms being compared are the same, comparison need be made only of the four least significant digits. (The SIN is actually dropped before comparison, by returning to use of fixed-point variables.) The terms with SIN =1 combine, since they differ only in the second least significant octal place, and 0 and 1 combine (Table L (I» to eliminate the variable of weight 1. o The new SIN is therefore 10 (generated at level 3) + 1 (generated at level 4) = 11. transferred is therefore 110004. The term Note that the SIN increment of 10 indicates that the variable eliminated had weight eight. __ ~" ••_~".;.;" ..•• ~:•.• ::.. -:. .o- f, f There is only one term having S I N at the conclusion of level 3 processing. with SIN =4 = 2, hence 20005 is outputted Similarily, the two terms do not combine, hence both 40003 and 40011 are outputted. 0 . All three of these terms are prime implicants. The terms with SIN = 10 combine, since 4 and 5 combine to eliminate the variable of weight 1. The new SIN would be 11, and the transferred term 110004 - identical to that derived from the SIN =1 set. The program avoids this duplication by a means that will be discussed in paragraph 6. Since only one term was transferred to level 2, it is of course a prime implicant. The problem is now complete, except for the decoding of the computer output. There are five prime imp1icants, so far identified as follows: At level 4: minterm 00012 (note that the SIN of all minterms is 0) At level 3: terms 20005 40003 40011 At level 2: term 110004 (3) Output Format Three fixed point numbers are printed out for each prime implicant. They are: 1- the minterm from which the P.l. was derived; 2- the final SIN, to tell which letters have been eliminated; 3- the level, to tell how many letters will appear in the P.l. (this is a check on the others; it contains no new information) If two P.I. 's come from the same set, the SIN and the level will be printed only once. . ., (4) Output Decoding o The output format of (3) can be translated to literal form as follows (using the first prime implicant to illustrate): 1- write letters for variables: ABC D 2- translate octal minterm to binary: 3- translate SIN to binary: 12 0 -+ 1010 ~ 0000 Now each 0 in binary SIN indicates a letter present in the prime implicant (in this case, all four). The binary minterm now indicates complemented variables by 0, true variables by 1. is therefore notation. - ABC D, The prime implicant which is minterm ten, in the usual decimal The output level -4- verifies the fact that the term should contain four letters, which it does. Similarly, the other prime imp1icants decode as follows: ABC D 010 o0 1 5 2 3 o0 3 4 3 1 010 11 100 4 010 3 010 4 11 100 2 ~t ~J-+ ~~ ~~ ABD AC D AC D BC (5) Check of Results The solution of this problem by Veitch diagram is shown below. It is readily verified that the computer solution is correct. Prime imp1icants: ABC D (m A I .. J '3 I o • I If IS ,t I,. c. I • 7 3 I t- O , ) B C (m4 + mS + m12 + m13) I 'I Is 10 A B D (m A C D (m D S 3 A C D (m 9 + ID7) +m ) 7 + m13 ) r :>1 I I (6) Elimination of Duplication It is inherent in the Quine system that identical combined terms will arise from different series of comparisons. For example, a term A C could arise by eliminating B from ABC and ABC, or by eliminating D from A C D and A C D. The program eliminates such duplication at all levels by the following procedure: When a combinable pair is found, the 9ld SIN and the weight of the variable just eliminated are compared. The term is transferred to the next level of comparison only if the old SIN is less than the weight of the variable just eliminated. For example, in the problem just illustrated, the two terms with SIN =1 combined to eliminate the variable of weight 10 (decimal eight). Since 1 <10, the term was transferred. Later on, the two terms with SIN = 10 combined to eliminate the variable of weight 1 (which would have produced the same term). This term was not transferred because 10) 1. References (1) "The Problem of Simplifying Truth Functions" by W.V. Quine (American mathematical monthly, p. 521, 1952) (2) "Logical Design of Digital Computers" by M. Phister· (Wiley, 1958) (p. 68-75) /y/pm~ /f. A'# ~ Thomas R•.Hoffman Prof. of Elect. Eng. Union Co lIe ge o • o CR I T ICAL SPEED. STRE SS, AND [3El~\R ItJG REACTION CALCULATIONS FOR A GENERAL SHAFT US I tJG NUt1ER I CAL INTEGRA T I Ot~ 8Y RALPH B. BATES COtJTENTS SUMMARY------------------------------------PAGE 2 INTRODUCTION-------------------------------PAGE 3 T~~ EOR '( -------------------·------------------P AG E 3 NUMERICAL INTEGRATION----------------------PAGE 6 t~ETHOD-------------------------------------PAGE 8 REt1ARKS------------------------------------PAGE 12 REFERENCES---------------------------------PAGE 15 116 ... ,. ii' h·W·jfb······tj"··ij'·,······¥·· ·ff·¥t-Ef ·W·T§··5 '·T -'p"rrn'rp 1t'""feTFPI2M?'TllirUg-prJ'yr' "W"U"j-j'j f n ... _. 'wwwnw • StlOlAlY o Critical apeed,atre •• ea, and bearing reaction. can be calculated on a digital computer for a general cwo bearing .haft, which caD have each bearing located anywhere on the ahaft, any number of croe •••ctions, variable loading, and any length. B•• ide. elLminatlng the tediou. labor of the calculations, the computer provide. flexibility. A number of calculations may be rapidly made to opttmiae design or to check out application variations on a standard deaign. Bae18 of the Calculations In ganeral, the critical speed calculation i8 baaed on "yleigh'. method using the following formula where Hc i. the first critical speed. rpm W. is the load at the I po.ition excluding external loads without ma•• , such as belt pull, pounds. Ys i. the Itatic deflection of the load at the a poeition due to the loads W. with the loads reversed between bearings. The .tatic deflectiona ud bending moMnt. are found by u.aing the same principlea o~ mechanica which are the baai, of graphical determination of be. . deflectiona. Numerical integration ia u•• d to relate load, .hear. bending moment, alope. and deflection. Thi. integration, althouah numerical) gives exact areas &8 lODl a. the loads and momenta of tnertia are con.tant over the incrementa of ahaft length. The bearing rea-ctlons are found by taking momenta about each beariug equal to zero in accordance with statics. .cr•••• and combined shear .tre•• are calculated according to tne usual equatlona of strength of material •• The ahaft bending tenaile stre.l, tor.ional shear Method The method i. illustrated in detail by an example calculation. Briefly, it con.iata of dividing the shaft into increment. of length, determining the 10a4 aDO abaft momenta of inertia in each increment anel the cOliputer calculate. critical speed, .tr••••• and bearing reactions. If there ia only .haft weight in a number of increment., the O.D. and I.D. of the ahaft are sufficient input data and the computer will calculate the load and momenta of inertia. o Other Factors in Critical Spe.d There are other factors affecting critical speed, such as, .bear deflection, gyroscopic effects, bearing length, and bearing flexibility. Bearing support and oil film deflection may be accounted for by adding the bearing deflection to the deflections used in calculating critical speed. JI7 Page Z Critical .peed. .tr••••• , ad baar1ng,~ r ••ct1en. for a leD.ral .haft MY be calculated on a d1altal computer by utilizing numerical tDtegratloa. The '''.1'.1 .haft baa each of two b••rin&' located anywhere on tbe .haft, any nu.ber of cro•••action., and any numb.r of load•• rf--"'\. ~) Be.ide. altainatina th. tedi0U8 labor of the calcalatlona, the ~t.r It allow. the calculation of • number of variatiaa. in order to opt~l.e dea1gn or evaluate application variatton. of .taadard offer. flexibility. de·Lan•• THIqIlY Ipi'lIb '. l . .rpKetho •... (4) ud (1)1 byleip'. energy method may be ueed to determine lova.t natural fr. .uency of a 'Yltam. The _tbod eoa.i.t. of a configuration for the .yat. which vill approximate the max~ amplitude of the fUDdamental mode 01" lowe.t Datwal frequency. • ...4 on thi. conftl.atloD. the aaaxiaum potential "'.iDa energy 18 calculated at maximua dl.placement (zero velocity) ad "uatad to the ..x~ kinetic eaerlY at the .y.t. . . .ul11b~lum po.it1oD (max~UD veloTbl. r.latlen.blp may then be lolved for the lowe.t natural frequency. city). The dynamic deflection curve at max~ .lspl.c.... t fer lateral dl.,laceaeat of • b... may be ...... 4 to be very cl••• to the .tatic dafleetloa curve. When the confiaurat1ea due to .tatic lo.41a" .f • be. . with lta CNn low 1a ue.d for calculattRa the fundamental fr ..~cy by "yletab l . . .tbod. the freqUeRCY 1a wLthto the accuracy ....treG fer ... t . . . iD.e~tD.~cal~ul.tlon•• SiDee lAyle1ah'. method conailta ....ati.1ty of ....t1n& inertia for c•• of the rotatinl _ •••• to the re.torina forc •• of the ahaft. 'at.nal forcea which do not have " 8 . auch .. belt p.ll t . . . . t Ht be i.acladed. a&yle1&h foad that the .yetem vibrate. 1D a IMDQeI' wblch make. tbe freq.mcy a a1DiIaua. Tnia _ _, that tile inertia forc •••f the ...... aut act i.a 4irectl.a ca.aing max~ defleetton.. ID. two b.ar1D& abaft the .tatic load. between the b••riD,.... t be rever.ed te produce lI&Xiaaa deflection and millllaua f't'eq\18llcy. I.&yl.ip·. methed 1. tile work clone on the be. . 1n moviog ita load. from the equilibrium po.1t~ to the 4yQaalc dafl.eti_ curve .t .aiJDua deflect1oa. ('rbi. i •••• u.4 the .ame .. the .tatlc The potential eoerlY iAvolved 1a _flection cutve without. external loada and with loa" raver.ed between be.rial·.) '£hue l , .1. • 1/2 ~ Iw.y.1 The uxiaum velocity of the weigbt W. wlth amplituda ,. in har1aonle motion 1. pya • Bence. the max~ kinetic en.ray of tbe .,.tem i. ~.I. 1. • ..L ~ tv. (p1s)21 o 2,'" Humber. in parenth•• 1A cteatpate refer_c•• at .ct of papel'. 'age 3 II~ -------------------~--~--------- ..----------- ----- .----~~-~ The .ymhol. in the previoua equation. are: 0 '.1. 1, potential energy. W. 18 the load at the • poattion, Ys i. the deflection of tbe load at the a position. It.!. ie the kinetic energy. g is the acceleration of gravity. p ia the natural circular frequency. 8 • 1,2,3,---. The la. of conservation of enerey require. that the potential energy ....1 the kinetic enerlY in tmclamped .yatema. Solv1Da fen: p2 giv•• the fol1ow1aa re.ult: 1: IW.~sl ~lw.y.21 The flr.t critical speed of a abaft may be conaldered to occur when the circular frequency (&) of shaft rotation e,Wil. tba natural frequency of lateral vib~ation of the rotors (considered aa a beam). The rotatina abaft 1. deflected by the ...11 unbalance force. in the rotating .y.t... wbieh cannot be elia1a.ated entirely 111 actual roto.... Th••• exciting lerce. have a frequency 1n a lateral plane equal to the eircular frequency of rotation. When the exciting forces have the same frequency as the natural frequency of lateral vibrat1on, r.~anc. occur•• whlch uaually caua•• UDde.tr ••bl. exc•• aive vibration of the macbine. Thua2 W 2 Henee: • 187.7 IV,Y11+IW2YaJ+lv,Yl l+ •• IW1)11' IW2yz2!+ 1·'Y3" where 1a the fir.t critical .peed, rpa. i. the load at the • poaitioa _clGd1n& external 1••d. without mae., 8uch . . belt p.ll.~, pouDda. Ys 1. the .tatic defl.ettoa of the load at the I po.lt~ (W.> with the l ••da 'l"ner.ed between tu bearl.... 1Ilcbu. Deflection The It.tic defl.etl~ of the abaft .y.t. . may be deter.iDed by utl1lataa the prlncipl •• of . .chanic. which ar. the bull of araphlcal det.rainatiOD of beam deflectl. . . . .beND in riaure 1.(13). (10), and (3). o B'Artpl ...ctlona The b••ring r.actions are foun4 by taktng moment• •bout each beartna .,..1 to aero in accordance with atatics. 'f ~ Page 4 ... Simply Support.Jtd Beam 1!11111!'l l1B the lO3.d per un1t length- v r. 1s the bea.r1ng reaction at the left and ,',- converted to loa.d per un1 t lengt.h. r z 1s the beari'llg react10n at the right end oonverted to load per un! t length. If ) Loa.d D1asram r.l 1s the load per un1 t length including the be~r1ng react1ons. l V V 1s the shear. V• Sw dx :: Area. under the W ourve; and ~ :: W ~ --1f-dx ~ .. x M Moment Diagram M=SV dx+externa.1 moments; and E Shear Dla.gr8.!ll L b M 1s the bending moment. 1-1 '- x fx = V = Area under the V aurve 1- Mext ( Mo st shafts have zero external bending '!loments.)' . -..f t--dx • x is the modulus of elc.sli1city. Ia 1s ·the moment of 1nertin of the beaa cross-sectional area ~.,n th respect to the axis in the neutral plane. . D.x ~~4x -,-,--~------------------------------:-..:.:.:--- -& is the area under the (!~/EI~) curve. S + Cs· S(M/EI a) dx =e-; ruld.~:~ SZ 1&:1.S 1s the slope and as 1s the consta.nt. 5=&-Os; S dS - ~(M ({X-If; and. -u 1nte~rat1on b is the area under the & curve. ,.+0,.:: SS dx ~S(e--CI) dx=~-Osx; 1=~'-CsX-Oy ; and ~#-.4L..c,; ax u l' 1s the derlectlorl~ and 0., is 80 t.~e 1 a.ni ~,.e!-OB ~.& ax consts.nt. '·0' . FIGURE 1 - BASIS OF 3-RAPHICAL DETER·!INAT10N OF DEFLECTIONS Pa.ge 5 .. ~ ) F 'ttMIU o .he_ The .baft badiq teuil• • tre•• , tor.loaal .bur .tr. . . . . . cOllb1Ded .tr••• are ealculatett aecordlnl to the uaual equation. of .tt_8tb of JUterlal. (13). Sba • K.; I Su • T. ~I Where all value. are at .ectlon • of the .baftl 1. beDding teD.il • • tr •••• p.l to the outer .haft .urlac. d1yf.da. by the IIIOIII8Ilt of iIlertia .f the ero•••eetlollal area with r •• peet to radial axl., 1. the rad1u. Ln.-'. 1. the .baft tor.lonel u the torque in the lu..r laut, .tre•• , pa1 lD. - lhe. ta tbe radlua to tbe outer ahaft .urface divided by the aoaeot of tnertla of the crol. ..cti~ With r ..pect to the Icmgltud1Dal &Xu, in. ar.a 1. tile .haft aaxiaum cOllblae4 abe. . . tre•• , p.i '.ur8 The iDtearatlona .bowD in 1 ...y be ,...for.d by ll.-rical . .an. with & COIIIputft. lhauical 1ntear&tloD teebDi,• • IDA)' be approxiute, (8) and (IS). Iowever, tbi. numerical 1ntearattoD 1. exact if tbe load and the IIIOIMIlt of inertia are COD.tant OYC the lncre.ent of x, ( a x) .ad the follow1na formul.. are u.ed. Zaylol'" r....l. (8) rr.. calculua Tayler'. fermala with rema1nder fo11"" f'Ca) (x • .) + • • • • + f (a> ,. ( ) (x • .)Il· + I (~\ x ·f() . + --r!t i Ix _ .)" + 1 f( ) 1. Ia ria_e 2 •.• Yz a. 1f x • b tbea (x - a>. • ~ n+ ! _4 11 x • c thee (x • .) • 6. x + ~~ Ax i. the f.acr_t of l ..... th aloq the be_• • 1. the .tatioo liDe 111 tAe center of II x. Sub.cript I de.1in.taa the value at the 'tation lin•• Sub.cript • • ·Yz. de.ignatea the value at the left of Ax. Sublcript • + 'I" cIe.1.pate. the value at the right of Ax. rIGta! 2 I~J ' •• 6 -- - -- - - .-------~---~--- Shear formulas Vi • dV • di, v • (constant over A x). ws; V" - 0 C l ... \, ,; Bending Moment Formulas M' • dM + y~. Hs • dX M't V; Ms .. Y2,. - + V8 • V' r::I _ w • Y2,. (l::t. x) + .L w. ( A x) 2 2_ Slope ForlUu.laa dS. M ~ and since S i-C.; • BIz where EI 9' Z8 is constant over 6 dS • ~ 4 a x .2i • dx - Deflection Formulas ~ ds • d~ di'"'" S S + Ca • c.(3) eJ • y and since 9 a ft • i $ S - c. therefore x - Cy ~' !I. • dx • !.i dx - C • • S· i ,(S) V 0·0 BIz. JJ. '" Page '1 Ii 1 o Therefore, the only .rror introduced in the integration 1. due to the approximation which i. necessary to keep the loading and moment of inertia constant over the increment of x, ( l1 x). If desired, the amount of error can be determined by approximating the load with higb value. and the moment of tnertia with low value. for one calculation. A second calculation i. then made by approximating the load with low value. and the moment of inertia witb high valu... The total etTOr will be lea8 than the difference between result. of the two calculation.. The error may be reduced by maktng further calculation. with amaller increments of x. Th1a 1. not require' for moat engineering calculations with the u.ual .afety factor •• a.- The calculation method 1s i1lu.trated by uaing the ahaft ~ in ,1gure 3a. This is a compre •• or rotor mounted on a stepped sbaft with two bearing.. cause the loads need to be regarded differently, the calculation. are done ln two parts. Example Caleu1.tion .' Part 1 Firat, the bearing reactions and'.tr ..... are determine. incluciin& external load. without maaa and with load. in their normal, direction (cIowmIard) ae tabulated 10 Table 1. The input data for 'art 1 of the calculation. i. de.cribed belowl Colum Description Sll!!bol bx The~ increment of length. • Statioa numbers of bearing location•• 1 • Station l10e number ira the center of l1 x. 2 J.s Load per unit l.oatil includina external 10ao without mus in noraa! dlrect1oD.. Coneentrated load. BlUst be distributed over the lncremeat6x. Actual .haft and rotor load. are more often distributed than concentrated. 9 C/l~a Se. Page 6. 11 s•• Page 6. 12 s•• Page 6. The calculation. made for Part 1 are de.crlbed below! o ColWDll Sl!!ol 3 "'2 4 De.crlRtton / MomenU about ba.ulng 2 station Une aa follow. 1Il2 • .£ 8 (s(of bearing 2) • a) calculated at .ach station line, lb/tn x atation number. / Momenta about bearing 1 .tatioo 11ne aa follovl: 1111 • 1s [.(of bearing 1) - . ] calculated at each station line., 1b/in x station number. PageS 1 2 3 4 5 6 9 8 7 10 a. Rotors 0 -100 is r-r-~~--;r=*=*=M=-_=_*,==-=.~,~~=*~-----.~~--,-r-. ! I: 1 -x I b. Loads, Part 1 LJ -200 500 ! : '-~" . I , . , , ; -1,000 r' 5,000 2,500 0 'x Sh.ear. Part 1 -500 se I ' ,-'-- 0 V . I ; '~.~.t---_.....-- fj :\ · ~- .~ . I· •. : 4---~------ .. . I _------+-- - : - - '- x d.Comblned Shear Stress, Part 1 1 0 I -.,_--.i-_-I- - - - ;- - + - -.....-/~=-~~~"""'"""_.-"""'-;-- I---=--' __ -"-" .... "-. -5,000 f·r -10,000 ..'" ,)8 ",",,-// .~,.,.. ~ I X ' ' . : e. SendiDg Moment, Part 1 o -~OOl t . . Deflection, P[\rt 2 o FIGURE 3 - B:{A.7·!PLE CALCULATION DIAGRA.!-1:S Pa.ge 9 Sol,. 0 • 5 ,..1 "ac~lRt1on ........ n"t1oft I, lba/lD "1 ~l II ~ It w) 'S!f~ liM} • at ll.... lq 1 •• at --iDi • • 11 leal't.q reactt. 1, lbl. 1'2 "'~1q ~2 + • 7 8 • 1'1 ( ~ x) z::.(ate_G!!!"" ,ytM U.~ LDa 2 •• at --Iai 1) 1.U'iIIa I'MCtiaD 2, IN. W, LeU aZ. 1'2 ( 6 x) ,.1' UDit l_tla iacl"iDa be_tna I'uctt..i L• ...pt . , b--iDa 1, w. • w, • aM at burial Z, 6 1 react1ea 2, lba/ia &2 w. · 1 2) '. +'1, II. + 'II. . . . . . . . ' . . 7. X. _ i. + ~1 i. + 1"2 ........ 6 • • 7. ..... 6 . . 7. 10 Sb. see ,,,. 6 13 Sta s.. .... 6 14 SCI S.. 'a,8 6 a iI'lIl! Calc!I!S"BI • lIES i.... S.cad J tile c~ltieal .,... :La .. tenlMd .-:1-.4" utuaal 1. . . vitlMtut . . . uti With the 1.... bettle_ b• • reYer••a talMtlate4 (Q ~abl. 2. lot. the .hr1ak. fit of the ntel' ad tIw . . , l l q Ilub MY be e.- t.,.....) .. COIIpI'.". l1de"l. to tacr.... the abaft dl... ~ wh.a calealatiaa ••1tical _peed, (14). !be tap., ••ta for .art 2 of the St1- - o 11"& b.x c.lc.l.t~ 1a delcrlbed bel. . : blttl,S", !b. 1.cr....t .r 1...,11" • • Statioa .WIIb.r. of b__ 1q locatiou • 1 • StatloG liM . . . .r. 10 tbe .entel' of 2 1. II x. Load pel' _It leqtll uclu.1a1 atenal 10'" v1tbeut • •'. _ _ u-ace4 10... _ t be 41•• tr1butecl over the :Lncr-.at A x. Actual abaft . 4 I'otor 10." .... _tte .ftaa cU.. ~lb.ted than eone_teat.d. 12 IX. 1I,w. at at_tieD I. S. . . . . . 5 • ••.,.10 The calc"latiolUl .... for Put 2 u. "acriNd Ml... Ctl,.. S_l 3 ~ 4 ~ S. . • r1 See hl't 1. • Jr S.. Part 1. S ". ,"SlpS1ft C' s.. , .. t 1. 2 , , , ~7 Puc 1. I S_ 'art 1. !he loe1.. _ of --iDa rMCtlMaa .., be que.tl_ed .laGe c.u, .........1 1"- JIIMrI.. .1dane ..... it .., .. . . . . that tIM iDelua1ea. . , beu1q wuctlou 1a .......,. 1» _ _ tIM' iatepoat1oa , • 'eet al-. the l-.tJa .f tbe abaft. f ... .... 6 LNd, . .ludLaa a.Canal lea_ at .tatia . . . . ., lb.. V. • 1. ( A x) • 7 s.. , .... 8 See ..... 6 aftd 7. 10 I. S.. . . . . . S,6_47. + 'II. b. 11 6 ael 7. s....... 12 S, 6, aM 7. See ' . l l. .taa. Deflect£on. 11 14 ,w.t, , Ab.elute value .f 001_ 6 ·x ..1.. 12 at '•.,.2, Abaol"te .tatloa I. "al_ of _1.- 13 x .001_ 12 atatia '. 'irat Crf.~lCa1 " ' • Sea .... 3. Ie • 187. 7 ,W' }Ii • £,. .~~ _atarat.at!_ of Y' It ... ahewD <'aa. ' 'OIl b. 5) that y Since u :It b· • tbe c:llatan¢e a101&1 C~c • y tu __ = x • • (A x) Let '. • the atatiDG number at befttBC 1 sb • the .tatioa Dumoer at beartDa 2 , • the ~ .t b. .ciDa 1 •• ~b the • ~. at beari.q 2 Since the deflection y • 0 at bear1n& 1 an. bearina 2 than ~a • C.s a < A x) + Cy and bb • Cas b ( A x) '~b + Cy ••• u ct' Solytaa thea. two a ..ut1OD1 fO'l' C. aad C, aDd ••batitutlq iDto tb. . . . .t1.a for y thell ' •• S•• o Thla 18 the . _ 15 • (abb • • I. cS b ) - (gb· s.,> I .1 bI - ~• (~b -~.) - ( • b- . I. s,)Ax )4x ( •• lleace y i. the cU.ff. . .ce 1D value Dew.. tb8 value of b. ad the .t..alallt liD. '1 which 10" throvah each beuina . . .hIND 1n rtpre It. rlCUll 4 .AUS U.e of • CO!p!tec Beeauae of the ted£OYa laboc 10901••• 18 th... laner.1 abaft calcu1.tlODe, the uee of • cOliputer 1a ......U')'. The input data for: the co.put.r coaputat1eD.a differ_ al1sbtly fn. that -..cr1bed iD the . . , 1 . , whicb i.11_tI"'tec1 the .thod. The 1Dput dat. to the c.,utu ia •• follow I o 1. ldentific.tion 2. A. x. 3. • ...iD. location at.tlOD DWIb.... 4. bten.1 10ade vlttaout .... and loc.tion Itation number•• 5. Ior. .,..,.r tran8lDltted with .tart and end atat101l numb••• 6. rpm 7. Modulua of elalticlty. J~ 1 'ace 12 In additi011 either of the 8. l.i1_iDa, Load. .xcl.dIDS Ihaft _lpt aDd extera.l 10••• vlt~ " ..... I. Shaft O.D. witb .cut ad ... atatieD aUllb.I. 10. Shaft 1.D. vitia .tut _cI ... atatlora .u.b.... Le... excl_iaa utenal 1.... wf.thMac _ I •• 10! 11! z,. C. wbleh 18 liZ ahaft O.D. 't .. Val• • •uch .1 5. 4 10 vIllell .., be the .... f • • 8 . . . . . of .tat1ou ... e.. ter to put iDto tIM COIIputer . . . '1a&1. val_ f ... ltart .ucla a.ber to end It.CiAm 0\18)_. 1xten&1 1..... 4, • • ptat 1Dto tbe COIIpUt. . . ..,...tel1 eo that they . .y be _e4 COl: ule.la's... lia11w to • • c 1 of the .....1. ad . excluded fro. calc.l.' ...... 1la111a to , . . t 2 of the ..... If tM 'baft baa an \IOuual .eeel. 110 that e, 1a aad 1. camaot 1M c.lcuate4 froa abaft 0.1 • • 1.1. tile" 8; ,~ 10~ ad 11· ar. the f.Dfut uta of that It.tieD laItucI of 8, , _41 18. 1.. The caleulatl... . . . by tM coaplItn' an aWl_ to tbN. 1D tU .....1. acept to .laaft "labt), I''.., CI Ip.. I . . . llu •• calculat•• , .... Itn,ut data S, 6, 7. 8 . 4 ,. !be COIIpUUI' uti. . aldlel' laput ••ta 4 + 8 + ~ 1. (..... to ,Daft wt) 01' 1aput clata 4 + 8' AI J.. ill caleu1etlou aWI_ to , ..t 1. ID ca1culatloD' ....11&1' to 'ut 2, the .....tc .... 1ft,.t uta 8 +1, (. . to Daft wt) 01" input data 8', with the 1."SH-" beRt.aaa ""....H, .. 1. •• ...1,('''. ~lvgt.'. !be p~lnclpal 1. 2. .....tae•••f th••• leGeral abaft calculatteDa awe lilted be~1 .,,1, to any two beartna ahaft witll each beRtaa l ..aced aaytIbere 011 the .uft wlth • ., v_iul. cre•• aactlaa ael loadiq aM with .., l-atla. fte calc:ulatl.a ..UIlcluy ...aitl• • •a .valutad 18 the u1 ••1atloG. by ..._ the fUlt lHCI at 1 ael -lal un cleflectiAtnat ucla be-1a&. .t.,.... 3. De.ip• .., b• .,tia1sed by al••lattal • Daber of • • latleraa by Cllllputer _41 ••1ectiq the Mat. 4. Vulatwa. ill ltaadad d•• ll1l ••\IOb . . . D. . . .r of belt ,..J..1I, .., b. cIa••ked by Mkia& a ...... • , aalovlaciGu by cOIIpU". Mcveey 'l'he accUZ'acy of the caln1&tioDa an affected by tbe f~11ow1D& J 1. lavina the 1. . . . . . taut ovel' tM i.Der.-nt of lalllth (Ax). 2. 1&9101 the ahaft cn••••oti_ ...-u of la.rtia COG.tant ta. increment of l-.th ( A x) • 3. ...aumiD.g the atatic deflection curve 1a· the dyJaam1c def,.cit1oa 18 O'N'I' caleulattna eritical apeed. '.ae 13 o 0·'· .,..d. 4. CODe.ntr.ttas the l.ad 1n the tDcreaent of lenlth 1D calculattDs s. U.ing Rayl-lIbr-'. _tbed to calcualate critical .peed. critical I, Tbe error d_ to 1 ad 2 precedf.q .., be evaluated .. lDcl1catect _ ' ' ' . 8. n. en.. clue to 4 .., be rdaiai. .cI by Uk"" _11 1acr_t. .f !be error. due to 3 _d 5 .., be raduced ud... the iaanla f . . . . . : - , (4) V •a ~ • 11 Y .1 11 ( c1 , iii~ i_til ( 6. .) • OJ ..eo.lo\ll.t .... tM cl'ltlo&1 ....4 1... 2 V•• -aa Ax calculattDa • lecoad critLeal ..... value .02 • 187.7 £,-,2'12' 1:"1 ' .:1 Sub.cript 1 r.feca to va1uea 1. the flr.t calou1atlea ..d luhacw1ft 2 ~.f.... ee val... ill the ••cond calcutatin. Icl liDee tU Ida.tlc --ayla • itaetl. of .... whil. potatial __., la • fuaotioD .f the _ .. Cia foI'oe. lIota that -.1 11 ...eI ill tile calculacieD of ttW: 'HCD' a. l! Crit&ctl . . I" ... !hue are OeMI' factor. affeot1q c.. ltloal .,... heaI: de'lecCia. (3) eel (5). &y1'. .coplo .ffeete, (2). (3). act (6). __ lq 1_tIl (3) . . . . -'laa f1albil1C7 (3). ---iDa ...port _cI .11 fila flulblllC)' "y, be u,*-ced , . by ad4taa the beal'1Di 4ef1eotioa to Cba 1. to ~ calculat~ .f c~ltlca1 .p..d. o .... 11t 1. ADder.au, R. A. "'laual Vlbratlo1l8 of UraifOl'll ..... Accord1la& to tbe Tt.• •Make !heory". Jetil'. Ap,l. llech, v. 20. A. 4. (Dec. 1953) 2. Bert, Chari•• V.....f1ectlou :La Ste,,.. SUfte·. 24, 1960 ". 121-131. .~ IMhLp, ...", 3. Cburcb, A. B. "Celltrlbaa1 ..... _41lANa'." . . . . I.kt J.... Wiley. 1t44. pp. 291·296 4. Church, A. B. !!J.cha'"l "'t 1957, pp. 5. Cw1e, Al_der 'lI.t..,. •.. .20 • York, .,.. Vl1e)' & _ , t • • Tabular Ketlled f_ Calc,datiDa Defleocw.. ef Step,.. ad Tap.ed Shaft.". !If!b1ae I!.ya t. 1916) eheuko, S. and D, H. YoUlll. V1b!:.t12M Ir'bl~ 1n haW...m, 3rd edt J Nev Y.rk: D. Van ... tr_COIIpany, 1934;,. 40. 15. tJaaar, Eric C. "Numerical IDtearatiOl1 Sillp11fiea CorIplex .... ip". Product 1y1n..... W (Harc.b 16 J ~9S9) I pp. 50 .. 54, S. lr, 1 13 () 'aa. 1S 1 1',1 . . TABLE 1 .. EXAHlLE CALCULA'rION • PAllt 1 (InclUding external loads without maaa and wlth normal direction of loacla) o Ax • 1 3 m 2 2 • 1. '/In 1..2 1.S 2.0 ..... , ,; -70 I 1. inches • Bearing No 1 at I 4 -,.'t·;,· :.. ' -490 I 144 r864 2,5 t i I .20 -4 l..Q. f 3.51 ( j ~ -4 ,-16 ~ 4.51 I ) f J 5,0 I -4 ~ ! 5.51 6,0 6.5 i I I ~ 7,01 -4 .- -2 0 I t 1 t ! l~ i ~ j, I 9.0 iI 10,0 I I I I i t .~ i ! i I i 480 I -4 i ! I f Ij I I I , 464 -70401 () 0 1-41361 0.16 • 3100 t I 365,2'6300 ~ 2781.6300 , ~ j 91 I 0.40 -112 i \ I. I i 26300 0.20 fS260 5260 ; 11 126300 0.40 J 0.20 +----+". -.~-I 1 L- r 1 • + 346 I/ln. r 2 •• 94 I/in. R, • 1384 lb. &2. -376 lb •• 13 / t. 5260 I ! ! i l f !! ! I I t it __.L. . . . ___ . .J.. _. . _._.L___.. . ---~. -_l * .. 10.20 15260' 5260 t } II l I ~ -28 i I t 1120 1122 1 1 ! i 0.40 f 10 • 08 r i Is l i 0.08 1120 1128 1 f I 1 'i I ~ -672 ) I .}-2280 ,i 0.16 ·352 0.08 \1120 11132 I I 1 ~ -228 L I I f I, I! ; ! I \ I 10 •51 I ( ! 0 I 0 1 I ·.···----r-.·-·\---t·.--·.·--t--.---.- . -.. --.-...-i-----+--t;._-376: 0.,08 1110 1138 t 661 126300 I -2 L 1175 26300 1 -5088! J. -14 0.08 1110 1130 /-61661 0.16 I 9-86126300 - I f 730126300 I -96 56 +84 0.20 5260 5261 563 26300 I I l ! TotAL -252; .1~84 in-lb, I 448 I -137-61 ·+10 1 Su 1Il-. P_1..l p.l I I '-7348 0.16 I I 1 I +28 r: S CI I i j -14 14 13 I -8721 -6288 64 I 1'.1 -2832 I 9.5 I ~:' 1-4552 0.16 J ~. '. +8 -2 '1 +2 Sba _l~ I I I I I • i i I 8.5 12 • -1408, 0.40 496 i 1+12 11 T I -560 -280 ~ I 342 -140 I I• i ~ I' i! !l ~ 7• 5 t I i ~ -4 t ~ H. ! 0 I -856 i +8 i 8.0 Ii I -4 j i 1-4 0 Ii i+4 ,i ! I 0 t J -4 t -8 i l i 1-12 j 10 9 ......s..- iI II ~ ~ i.!1 I -70 1 !-288 1~144 It -4 8 in -lbr;. . \n;Jl.;s.· I J ~ I I I l i 7 Bearing No 2 at • • I ". c:o .+~a. Illn !!. .. 8 • K.+ y& V w. 0 , 6 5 lilt 111'0 }'lln I iQ. 0.5 !t Page 16 1-I .__.<_ TABU 2 - IXAMJ'LI CAJ.ClJLATlmi - .AaT 2 (Ixcluding external loads witbout .... and with lo.de rev.r ••d between bearing_) 2 1 • ! 1. v. Illn '/ln I , Ilin 'lln , in-lb. o 0.5 -210 -490 -70 1.5 -864 I -20 -16 12 I-144 I -5761 I ,-856 I -4 -4! -16l -288 4 330 r t t I t -4 f -81 I -17 -1154 3240 3740 1 -75 -560 1000 -2832 3650 2320 -677 -12 I 38 9 -519 0 o 0 -966 105 17- 2 -1497 126 20 3 -2086 89 14 1 108 -2726 1 o 0 108 -3318 -99 8 l -4030 -200 28 56 -5717' -6288 360 4 I 16 1 ! I -108 1 -2 I 464 -5312 t 1-208871 t 16 360 1 I 48.0 I -3424 f -260671 -1472 -8 64 -28832 -352 -8 -112 .56 -32525 360 o 10.5 I 360 56 -14 1 I 16 4 13937 - 360 I i 4 -7136 ! 9.5 84 -235 ~ 448 .. I 10 -202 360 -16 ~ ! I 1 28 I -872 Q! 8 10.0 -14 , 0 f i 496 :1 o t i 4 ! 1000 -280 r 8 -28°1 -70 1I la.2. rad .. I 00 !Jl D.I o -32567 . . . .---1--- .--'-'+----t----4---.+----+---.--+-----I----+----+---.......-----I---1336 -424 totAt~288 7015 ----,--~--~--.--~--~----~._ _- - _ L_ __ _ 1"1 • 334 111n • r He • 187.7 613V~, --106 Ilin ~ IJ~ I I,~, Page 17 I ! , • o~ . ...... CRITICAL SPEED. STRESS. AND 8EARING REACTION C _ _ _._ .•._". __ , .... "." ............. ~N·_··· CA~CULATIONS ........_ ..... .. FOR A GENERAL SHAFT USING NUMERICAL INTEGRATION C c ********** C PROGRAM DOES NOT INCLUDE A TRACE ROUTINE ... ---.-.... ...-.. -- .. .......-.------...----.-._. __ .,..-NO SWITCH SETTINGS --.--.-.~,------~ C . -~. -"~ ***** C ~.- ----~- PART A-1 ***** **** DIMENSION STATEMENT C DIMENSION A(102).b(102).C(lQ2) C **** INITIALIZE ARRAYS AND SUMS LOCATIONS 50 DO 51 K=1.102 A(K)=O.O - - - - - .. _. __._ •. _...._.. _--_._--_..._.. _.....---_._ ... -_ .. B(K)=O.O .... --.--..-. -" 51 ._ ....- .. -.... -.-.----.- ...------....- . - - - . - - - - ClK)=O.O SUMA=O.O --------_.._-_.._.__ __._--_.._._.• ---_. __ ......... ---_.._-_. . SUM8=O.O - - - - - - -.. _---_.._ SUMCcO.O .. .........-.. "'~------------------ C ... --_... -- c** READ IN EXTERNAL LOADS __._ _ _._---_._... .. ----- ._-------.._-_ ........ _.... ,lOt W1 ,lIt W.?' I 2. W3 , I 3 , W4 , 14 K=I1+1 A(K)=W1, K=I2+1 A(K)=w2 __......._.......-..........._._------- -_._-_.-._---- ........ ... K=I3+1 A(K)=W3 K= 14.+ 1 A(K)=W4 o IF (10)60,60,100 c **** READ NO.1 INPUT DATA 100 READ PUN CH ,IS I ZE,. I CAL, I TYP ,M , S 1 , S 2 ,H P t RPM, DX 133 C **** CALCULATE .IORQUE •• T•••THE TURNING MOMENT T=63025. *HP IRP.r~ M=M+l 51=51+1. 52=52 +1. C **** READ =2 INPUT DATA 110 READ J=J+l L=L+l A(l)=O.O IF (OD) 114.114.115 C **** TO DET.IF IX.W ARE GIVEN 115 IF C 4 4 t 4 (W)111,111.11~ **** CALCULATE SHAFT WEIGHT W=PI/4(OD**2-ID**2)*0.285*DX • __ • • w"O. _. ..__ _ • ". _."._ ~~ .~_." ¥ _ _ ." .. _ _ _ _ _ ~,_ .. _..._ •• _ ". _ IIi W=O.7853982*(OD**2-XID**2)*O.285*DX 112 IF (XIX)l13,113,120 C C **** CALCULATE THE MOMENT OF INERTIA OF THE CROSS SECTIONAL AREA.IX WITH RESPECT TO THE RADIAL AXIS--IX=(OD**4-ID**4)*PI/64 .... . _--- , . t 113 XIX=(OD**4-XID**4)*0.049087385 C C 4 **** CALCULATE THE MOMENT OF INERTIA OF THE CROSS SECTIONAL AREA,IP 1 WITH RESPECT TO THE LONGITUDINAL AXIS--IX=IY,SO IP=IX*IY=2*lX XIP=2.*XIX C C **** CALCULATE C/IX(S),THE RADIUS TO THE OUTE~ SHAFT SURFACE DIVIDED BY THE MOMENT OF INERTIA OF THE CROSS SECTIONAL AREA W.R.TO c 120 CDIX=(OD/2.)/XIX I C **** CALCULATE C/IP(S),THE RADIUS TO THE OTHER SHAFT SURFACE DIVIDED C BY THE MOMENT Of INERTIA OF THE CROSS SECTIONAL AREA W.R. TO C THE LONGITUDINAL AXIS.IN**-3 1, f I CDIP=(OD/2.)/XIP GO TO 116 C C **** CALCULATE LS(ACK)+LS)=LOAD PER INCREMENT INCLUDING EXTERNAL LOADS WITHOUT MASS IN NORMAL DIRECTION. CONCENTRATED LOADS MUST l311 o c aE DISTRIbUTED OVER THE INCREMENT DELTA X. ACTUAL SHAFT AND C ROTOR LUADS ARE MORE OFTEN JISTRI8UTED THAN CONCENTRATED C **** DIR~CTION OF LOAD, DOWN = - 114 CJIX=O.O CDIP=O.O 116 DO 130 I=J,L PUNCH,CDIX,OO,CuIP,XID A(I)=-(A(I)+w) C **** CALCULATE M2 (B(J»=MOMENTS ABOUT BEARING 2 STATION LINE C M2=LS*«5 OF dEARING 2)-S)---WHERE S=STATION NO. SO THAT M2 IS C CALCULATED AT EACH STATION LINE. S=I 8(I)=A(I)*(S2-S) C **** CALCULATE M1 (C(J»)=MOMENTS ABOUT BEARING 1 STATION LINE C M1=LS*«S OF BEARING l)-S)---WHERE S=STATION NO. SO THAT M1 IS C CALCULATED AT EACH STATION LINE C(I)=A(I)*(S1-S) SUiY\C=SUMC+C ( I ) SUMb=SUMd+t:)(I) 130 SUMA=SUMA+A(I) If (L-M)110,140,140 140 I=S1 A(I)=A(I)+SUMd/(Sl-S2) SUMA=SUMA+SUM8/(Sl-S2) 1=52 A(I)=A(I)+SUMC/(S2-S1) SUMA=SUMA+SUMC/(S2-S1) C **** PUNCH OUTPUT DATA---S,W,M2,M1 DO 150 I=1,M 0,",:, S1=1-1 ,,'. 150 PUNCH,S1,A(I),d(I),C(I) C **** PUNCH TOTALS OF M2,M1,W GO TO 50 END ***** C C PART A-2 **** **** DIMENSION STATEMENT DIMENSION A(102),P(102),XC102) C **** INITIALIZE ARRAYS AND SUMS LOCATIONS 100 DO 101 K=1,102 P(K)=u.O 101 X(K)=O.O SUMV=O.O SUMD=O.O SUME=O.O READ ,ISIZE,ICAL,ITYP,M READ ,Sl,S2,HP,RPM READ ,UX PUNCH,ISIZE,ICAL,ITYP,M,Sl,S2,HP.RPM,DX C **** STORE C/IX IN X( I ) ARRAY C **** STORE ClIP IN PC I ) ARRAY M=f'.1+ 1 DO 102 I=1,M READ ,CDIX,OD,CDIP,XID X(!)=CDIX 102 PCI)=CDIP DO 103 I=l,M 103 READ READ ,S,A(I} ,Z,ZU ,T,SUMd,SUMC,SUMA 5=0.0 V1=O.O D1=0.0 DO 120 I=l,M C **** CALCULATE SHEAR A STATION (S+1/2}---V(S+1/2}=V(S-1/Z)+WS*DX V2=Vl+A(I)*DX C **** CALCULATE bENDING MOMENT AT STATION (5+1/2)---- o • ---M(S+1/2)=M(S-1/2)+V(S-1/2)*OX+1/2*WS*OX**2 ( D2=D1+0.5*DX*(V1+V2) o ( **** (ALCULATE dENDING MOMENT AT STATION (S)---- M(S)=E(I) ---MCS)=M(S-1/2)+1/2*VCS-1/2)*OX+1/6*WS*UX**2 ( ( **** (ALCULATE bEN0ING STRESS,PSI---SBS=MS*C/IX T~NSILE S8=E1*X(I) 104 SB=-SB c **** CALCULAT~ SH~FT TORSIONAL SHEAR STRESS,PSI---ST5=T*C/IP 105 ST=T*P(I) C **** THE ShAFT MAX. (OMDIN~D SHEAR STR~SS,PSl--- ---S(5=(ST**2+(5dS/2)**2)**.5 ( ( CALCJLAT~ **** PUNCH OUTPUT DATA---V~5+1/2),M(S+1/2),MS,C/IX,SBS,T,C/IP,STS,SC 5=1-1 IF (X(I»110,110.111 110 Q=O. 111 PUNCH, S, £1 ,x ( I ) ,S6 ,Q, P ( I ) ,S T ,SC Q=T 5=5+.5 PUNCH,S,V2,D2 V1=V2 120 D1=D2 GO TO 100 END C REARRANGE THE OUTPUT DATA OF 62-001SA1,2 INTO TABLE D I ME NS ION A ( 1 02 ) , 0 ( 1 0 2 ) ,C ( 1 02 ) , T ( 1 u 2 ), V( 1 02 ) ,D ( 1 0 2 ) 100 READ,ISIZE,ICAL,ITYP,M o READ,Sl,S2,rlP,RPM READ,DX PUNCH,15IZ[,lCAL,ITYP,M,Sl,S2,HP,RPM,DX K=M+l 137 C **** R~Au,C/IX,OJ,C/IP.lu • 00 110 I=l.K 110 READ,Rl.R2,R3,R~ C **** READ S,wS,M2,M1 DO 120 I=ltK READ,S ,R1,R2,R3 PUNCH,S ,R1.R2,R3 S=S+.5 120 PUNCH.S S=S-s READ,R4.5UMd.SUMC,SUMA PUNCH,S ,SUMA,SUMb,SUMC R L=SUrvtBI ( S 1-52 ) R2=$UMC/(S2-S1) READ,Sl.S2,HP.RPM READ,DX C **** A(I)=C/IP,o(I)=STS,C(I)=SCS DO 130 I=l,K READ,Sl,5M,CDIX,SB READ,5, V( I ) ,D ( I ) 130 PUNCH,Sl,SM,CDIX,Sb DO 131 1=1tK S=1 5=S-.5 131 PUNCH,S,V(I),D(I) S=S-s PUNCH.S.R1,R2 DO 140 I=l,K GO TO o 100 eND I' &Z » C o ***** C ***** PART B-1 **** DIMENSION STATEMENT DIMENSION A(10Z),X(lU2),Y(102) C **** INITIALIZE WuRING AREAS 90 DO 91 1=1,102 A(I)=O.O X(I)=O.O 91 Y(I)=O.O Vl=O.O Dl=O.O Fl=O.O Gl=O.O SUMb=0. SUMC=O. C **** READ INPUT DATA NO.1 READ PUNCH,I5IZE,ICAL,ITYP,M,Sl,S2,rlP,RPM,DX C **** READ C/IX «(vIX) M=M+l Sl=Sl+l. 52=52+1. DO 101 I=l,M READ ,X(I),Ou,CP,CX IF (X(I»101,101,100 100 X(I)=(OD/2.)/X(I) 101 CONTINUE C **** READ STATION,LS,M2,Ml DO 105 I=l,M READ o S=S+l. C **** CALCULATE LS=LOAO PER INCREMENT EXCLUDING MASS AND wITH C ~ITHOUT C CONCENTRATED LOA~S LOAO~ MUST d~ REVERbED EXT~RNAL d~TwtEN LOADS bEARINGS. DISTRIo0Tt0 OVLR THE INCREMENT DX. C ACTUAL ShAFT C CONCt:NTRATEO. AN~ RuTOR LOAD~ ARE I~OR£ • 0FTiN OlSTRIoUT£D THAN IF (5-52)102,104,104 o 102 IF (S-51)104,104,103 103 A(l)=-A(I) b=-8 C=-C 104 SUMB=SUMb+b PuNCH,I ,A( I) 105 SuMC=SUMC+C READ,T,tj,C,SA K=51 b=A(K)-B/(S1-52) A(K)=B+SUMB/(Sl-S2) PUNCH,K,A(K) K=S2 C=A(K)-C/(S2-S1) A(K)=C+5UMC/(S2-S1) PUNCH,K,A(K) DO 170 K=l,M C **** CALCULATE SrlEAR AT STATION (S+1/2)---V(S+1/2)=V(S-1/2)+WS*DX V2=V1+A(K)*DX C **** CALCULATt: C 0EN01~~ MOMENT AT STATION (5+112)--- ---M(S+1/2)=M(S-1/2)+V(S-1/2)*OX+I/Z*WS*DX**Z D2=Dl+O.5*DX*(Vl+Vl) C **** CALCULATE THt AREA UNDER THE (M/EIZ) CURVE AT STATION (S+1/2)-- C --F(S+1/2)=F(S-1/2)+(2M(S-1/2+1/2V(S-1/2)+M(S+1/2»*DX 1(3*EIX) IF (X(K»110,110,120 110 F2=Fl o G2=G1+F1*DX Y(K)=G1+.5*DX*F1 GO TO 130 120 F2=Fl+(2.*Dl+0.5*Vl+D2)*DX/(3.*X(K)*30.E06) l'io C **** THe AREA UNDER THE F CURVE AT STATION (S+1/2)--- G(S+1/Z)=G(S-1/2)+DX*(FCS-I/Z)+(12M(S-1/2)+DX(3V(S-1/2) C o CALCULAT~ +V(S+112) »*DX/24EI) " C G2=Gl+DX*(Fl+(12.*Dl+DX*C3.*Vl+V2»*DX/(24.*X(K)*30.0E06» C **** CALCULATE THc AREA UNDER THE F CURVE AT STATION S--- C ---G =G(S-1/2)+1/2DX(F1+CM(S-1/2)+1/48*DX(7VCS-l/2)+V(S+1/2) Y(K)=Gl+.5*DX*(Fl+(Dl+.02083333*DX*{7.*Vl+V2»*DX/C4.*X(K)*30.E06» C ****CALCULATE DEFLECTION AT STATION S---GS1=Gl.G=GS2 130 PUNCH,Vl,V2,Dl,D2,Fl,F2,G1tG2,Y(K) V1=V2 D1=D2 F1=F2 G1=G2 170 CONTINUE K=S1 G1=Y(K) A(K)=B K=S2 G2=Y(K) A(K)=C C **** CALCULATE ASS. SUM OF WS*YS--WS=WS*DX SUMP=O.O SUMK.=O.O DO 240 K=l,M S=K Y(K)=Y(K)-Gl-(G2-Gl)*(S-S1)/(S2-S1) SUMP=SUMP+(ACK)*0X)*YCK) SUMK=SUMK+CACK)*DX)*Y(K)*Y(K) PUNCH,SUMP,SUMK,Y(K) o IF (SUMP)210,220,220 210 SUMP=-Surv1p 220 IF (SUMK)23C,24Q,240 230 SUMK=-SU~1K ___ \ .. !t. • •••• ff •• ~ ".l: ~_'i-!_l!._=-: __ ""'--"" - .. Q .. --------------... - •.. - - . - - - - 240 CONTINUE C **** CALCULATE THE FIRST ~RITICAL NC=187.7*(SUM OF WY/SUM OF WY**2)**.5 C XNC=187.7*(SUMP/SUMK)**.5 C SPEED,RPM **** PUNCH JUTPUT--YS AND NC Rl=SUMB/(Sl-SZ) R2=SUMC/(S2-S1) PUNCH,Rl,R2,XNC DO 250 I=l,M K=I-l 250 PUNCH,K,Y(I) GO TO 90 END o THREE DIMENSIONAL SURFACE FIT M-151 December 9, 1963 D. G. Kitzinger /43 c • o TABLE OF CONTENTS Page ............................ 2 ANAL YS IS· .... • • ... • • .. • • • .. • • • . . . . . . . . . . . . 2 FIGURE 1 .••.••..•...••.....••••••.......... 3 SURFACE FITTING TECHNIQUE·.············· 5 FIGURE 2 ...•...•.•.•..••.•.••.........•.... 6 TRANSFORMATION OF VARIABLES..... ..•.... 6 I NTRODUCTI ON .................................... .................................... 8 I NPUT PARAMETERS ......... · • .. • . • . . • . . . . . .. 10 I NPUT TECH NIQUE .................•...•.... 12 TABLE 3 ...................................... 13 SAMPLE PROBLEM ......•.•.........•.....•.• 14 TABLE 4 ..................•....... • ... • ... • .. 14 TABLE 5· ............... • .......... • .. · • . . . .. 15 TABLE 6· . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . .. 16 FLOW CHART ............................... 16 TABLE 1 TABLE 2 o 7 .• " ... " ,,~' INTRODUCTION An ever present problem in engineering is the need to represent the three dimensional array of data by a mathematical equation. Many physical quantities can be described only in three-space. Com.puter storage limitations prevent table look-up of a large number of these quantities, expecially if their range of values is extensive. M-1Sl uses nth order multiple interpolation together with extensive transformation of variables to describe most three dimensional functions. The method used has the advantage of separating the three dimensional characteristics of the array into components described in two dimensions, hence easy to define by the user. Since many functions (exponentials, sinusoidals, etc.) are not best described by the polynomials used in interpolation, the first approximation to the function might be sadly deficient. Better successive approximations to the function can be easily made using the appropriate output option accompanied by transformation of variables. An error analysis assists the programmer in choosing the best fit of a function, and in the case of second and th ird order interpolation, the best fit is selected from a number of possible fits. ANALYSIS Let us represent a three dimensional array by x = f [Y(y), Z(zll, where Y(y) and Z(z) are polynomials in y and z, respectively. Let us I imit the order of these polynom ials to nand m so that Y(y) =a +ay+···+ay o 1 n n .+b z m n L = a.y i=0 m (1 ) i I m L = (2) • i= 0 At the ith value of y, m Z. = I b .. z i E i=0 II Define: n x = f(Y, Z.) I = L i=O n m L L i = 0 i= 0 b .. z iy i (3) • II m Then, c. = I L (4) • i= 0 -2- Note that x is a smooth function of y and z with (n + 1)(m + 1) coefficients and hence requiring the same number of known members of the array. In general, an (n + l)(m + 1) order matrix would have to be solved to obtain values for the unknown coefficients. Referring to equations (3) and (4), note that both a. and x vary in the same way with changes I in z. Thus, by studying the behavior of x with changing z, the behavior of a. with changing I z is also known. By holding y constant, a. is a function of z alone. If values of a. are I I known at each of (m + 1) values of z, the b .. values in equation (4) are defined by the I' appropriate matrix solution. If z is held constant at z. and if (n + 1) values for x = g(y, Z4) I I are known, (n + 1) a •. coefficients are determined. Define a restriction that allows easy II . determination of the a .. and b .. coefficients: y takes on, at most, (n + 1) values; z takes on, II II at most, (m + 1) values; and the value of x is defined for all allowed values of y and z. At particular Z.I I n (5), x.(y, z.) = a i + a 1iY + . . • + an iY O I I wh ich is a form that allows solution for (n + 1) values of a .. , when the (n + 1) values of y are used. . II x X m+ l=a O, m+ 1+a 1, m+ 1Y . +.. ·+a n,m+1 y n I / / I / / / / I / / ~ ____ / ~/~ zl / / __ ________________ __________________________ z ~/ z2 ~/ z m+1 FIGURE 1 -3- t III .Choose the (m + 1) values of z successively, then an array of a.(z.} values results: I I i = 0, 1, 2, . . . , ni i = 0, 1, 2, . . . , m. Fit the (m + 1) values of a. by I ' a.(z} = boo + b .z + . . . + b .z 1I I I ml m (6), with the result that the b .. coefficients in equation (3) are evaluated. EquQtion (3) defines II a function of y and z that passes through (n + l)(m + 1) data points. Furthermore, the function of y and z is smooth, and if planes parallel to the y or z axis and perpendicular to the y z plane are chosen, they intersect the x = g(y, z} surface in I ines described by polynom ials of order nand m, respectively. For practical purposes, nand m may be selected by choosing their values dependent on the study of the z = constant plane intersections and the y = constant plane intersections with the surface x = g(y, z}. Note that experimental data are often determined most easily by holding all variables constant except one. This procedure guarantees compl iance with the restrictions imposed regarding values of x at each combi nation of allowable y and z. The usefulness of the assumed fit given by equation (3) is I imited when the surface to be fit is best described using exponential, trigonometric, fractional exponent and hyperbolic functions. AI though it is true that most usefu I eng ineering functions can be described by infinite series approximated by the form given in equation (3), it would be better to use the appropriate functional transformation directlf" For instance, assume that a surface is known to be of the form x = f(y) = [sin(7y+3)]la +a y ]. Let xl = f(Yl)andx = f(Y2) and define o 1 2 g(y) = sin(7y + 3), x; = x /g(y), x = xig(y) and, in general, Xl . = x/g(y). Then fit 1 Xl = a + a y by a first order polynomial using x; and x as data points. O 1 2 2 Having determined the values of a and a l' x is defined by xlg(y). The expression for x is O valid not only at Yl and Y2' but over the entire region in which g(y) is a suitable transformation. It should be noted that the data points xl and x could have been fitted by 2 x = a + aly without considering g(y). However, only the two data points and the points O 1 where (7y + 3) = sin- (l) w~ld be satisfied by x = a '+ 0l Y' O Theoretically, provided the matrix solution is exact, the polynomial fit given by equation (3) shou Id fit exactly the (n + l){m + 1) primary data points. If the function to be described is of the form x = g (y ) c n m i=O L j=O L b .. z iy i, a study of the primary data points will not give II information concern ing the su itabil ity of g(y) as the appropriate transformation of variables, . since all primary data points are fitted, independent of the form of g{y). Hence, secondary data points not used in the determination of the b .. coefficients must be studied to arrive at II g{y). M-151 allows secondary data points for either arpitrary y or z, but not both arbitrary. Either y or z must be an lIaliowed li value, in the sense that defines primary data points. The other coordinate, z or y respectively, may be arbitrary. J47 -4- To help decrease the importance of the user in deciding on allowed values of y and z for primary data points, over-definition of the problem is allowed in the cases of second and th ird order fitting. As many as seven allowed values may be chosen and the program wi II use all combinations of seven points taken three and four at a time, respectively for second and th ird order fits. . SURFACE FITTING TECHNIQUE 1. Select the dependent variable, x 2. Plot famil ies of curves a. y = constant pi otted on the x y plane b. z = constant plotted on the xy plane n 3. m b.z i i=O j=O I on the xz and xy planes, respectively. If there is uncertainty concerning the best value of nand m, all possible values should be. used in separate cases, using the average error and maximum error features of the code to decide between cases. It is not crucial that the very best case be used, if transformation of variables can be used to force a best fit. For instance, x = sin(7y + 3) is fit best in certain regions Assume nand m for .fitting y and z such that x = L ajyi and x = L 2 by x = L a.yi without transformation of variables, but with transformation of I i =0 variables x/sin(7y + 3) = x' = 1 is fit exactly by a zero order fit: x' = aO = 1. In the interest of developing a technique independent of curve fitting experience, it 2 L a.yi, provided that the region of definition will remain I i = 0 small. However, any curve fitting experience at the user's disposal should be used is better to use x = 0 4. To effect more accurate fits, use the output option in which transformed data is output ready for further transformations. Th is option prints the transformed values of x (for example, x/g{y) is printed if the transformation is Xl. = x/g{y) and x is printed if there is no transformation), the fit of the transformed variable, the difference including sign between Xl and its fit, and the ratio xi/(fit of x'). Referring to figure 2: x = f(y), the fit of x = g(y), and h(y) = x - {fit of x)o -5- g(y) x / /' lZJ / f(y) / I l1J I o primary points El secondary points fit of points, g(y) function to be fit, f(y) y FIGURE 2 The transformation that effects an exact solution is f(y) = h(y) + g(y). Since h(y) can be approximated by a sinusoidal with appropriate ampl itude and period, the sinusoidal transformation is made on all data points using the program to make the change: Xl = x - h(y). Upon rerunning this case with the transformation included, the printout will include Xl, the fit of Xl, Xl - fit of Xl, and the ratio xl/(fit of Xl). Once again, hl(y) = Xl - fit of Xl = fl(y) - gl(y). To determine our fit, Xl = hl(y) + gl(y) and x = Xl + h(y) = h(y) + h'(y) + gl(y). Correspondingly, 'the ratio x/(fit of x) may be used by plotting th is ratio and describing it just as h (y) was described. Code capabil ity allows as many as ten transformations to be made. It is necessary, in order to ach ieve an improvement in the fit, that the transformation of variables be such that the original data points will not be seriously sh ifted by the transformation. As an example, consider the translation without distortion of the fit of the array to more closely approximate secondary data points in the array. If this translation is fit once again, the fit of the translated array will be parallel to the original fit with no improvement over the original fit in regard to secondary data points in the array. TRANSFORMATION OF VARIABLES o The util ity of transformation of variables was demonstrated in the section on surface fitting technique. Basically, there are two types of transformations available in this code: functional transformations that deal with the man ipulation of complete generated functions of x, y and Zj and transformations of individual variables x, y and Z to form the functions used in functional transformation. Functional transformations available are I isted in Table 1 and transformation of variables are I isted in Table 2. -6- T TABLE 1 ~-( -~. FUNCTIONAL TRANSFORMATIONS = NMODE = 1 NMODE None None 2 f(x) f(x) 3 xg(y) xh(z) 4 x + g(y) x + h(z) 5 x/g(y} x/h(z) 6 x[g(y) + k(y)] x[h(z) + I (z)] 7 x/[g(y) + k(y)] x/[h(z) + I(z)] 8 xg(y)h (z) x/[g(y) h (z)] 9 x [ 9 (y) + h (z)] x/[g(y) + h(z)] xg(y)/h(z) xh(z)/g(y) NVT 10 ISo 2 -7- • TABLE 2 TRANSFORMATION OF VARIABLES Let v take on the value x, y or z. v shall be transformed. TRANSFORMATION INVERSE TRANSFORMATION None NT ::: None NT ::: 2 0+ be cv + d NT ::: 3 0+ bv c NT ::: 4 a sin-1 ( (a + b sin ~CV + d) - a] ) _ d + b sin (cv + d) NT ::: 5 c In [(a + b~ dv) - a] d In c NT ::: 6 exp ([a + b In b(cV + d)J - a) - d c a + b In (cv + d) lSI -8- (Table 2, Continued) ([a + b I\{v + d)] - a) NT = 7 - d c a+bln (v+d) c NT = 8 c a+bln v+d - d ce a + b (v + c) ~ a+ d . NT = 10 (In [11 + ([a+bSinh{~V-l:d)J - a )2(2 + [a + b sinh{~v + d)] -a cos- 1 ( [a + b NT = 11 a + b cos(cv + d) ] _ d)! co~{cv + d)].- a ) _d c tan (la + b tan- ~ (cv + d)J - a ) _ d NT = 12 a + b tan -1 (cv + d) c = 13 ( ± In [I - a + b cosh(cv + d) + [a + b with the restriction i-hat (cv + d) ~ b - c -a (v + c)d a + b sinh (cv + d) NT lid b NT = 9 1 + ( [a + b cosh~cv + d)l - a c~sh (cv + d ~l a ) 211/2 ] _d Y > 1. /s~ -9- • Cl I NPUT PARAMETERS AT, BT, CT, DT - These are the coefficients a, b, c, and d defined in Table 2. Referring to Table 1, these coefficients apply to each single function transformation, but apply only to g(y) if NMODE = 1 or h(z) if NMODE = 2, when a transformation is used that involves two d ist inct changes of variables. (An examp Ie is NVT = 6 and NMODE = 1 from Table 1, where g(y) and k(y) are distinct changes of variables having distinct characteristic coefficients.) AT2, BT2, CT2, DT2 - These are the coefficients a, b, c, and d defined in Table 2. Referring to Table 2, these coefficients apply only when a two function transformation is used. When NMODE = 1, only k(y) or h(z) is described; and when NMODE = 2, only I (z) or g(y) is described. DC - x values defined as primary (see analysis). Values of x are allowed only for (YC, ZC) coordinate pairs. To fill input cards, nonvalid values of DC are set equal to 1.OOOE-9. The order in wh ich DC values are placed on cards is defined by the order in wh ich ZC values are read. YC is held constant for any DC input card. DY - x values defined as secondary (see analysis). Arbitrary values of z are allowed for each YC value. To fill input cards, nonvalid values of DY are set equal to 1.000E-9o The order in wh ich DY values are placed on cards is defined by the order in wh ich Z values are read. YC is held constant for any DY card and packets of DY and Z cards are read in the same order as YC values were read. DZ - x values defined as secondary (see analysis). Arbitrary values of yare allowed for each ZC value. To fill input cards, nonvalid values of DZ are set equal to 1.OOOE-9. The order in which DZ values are placed on cards is defined by the order in which Y values are read. ZC is held constant for any DZ card and packets of DZ and Y cards are read in the same order as ZC values were read. KO - output selector: KO = 1 - only the b .. coefficients are printed; KO = 2 - b .. coli II efficients, x, y, z, fit of x, x - (fit of x), average per centage error, and maximum per centage error are printed; KO = 3 - same as KO = 2 with the addition of the transformed x = x I, fit of x I, Xl - (fit of x I), and xl/(fit of x I} printed. Table 3 shows a fourth order function fitted with a first order approximation using K 0 = 1, 2, and 30 MO - order -of polynomial Z defined in equation (2), where MO = m. MO may be as large as ten, but calculational error is such that it is suggested that MO be kept less than seven. MC -(MO + 1) except for second and third order fits where MC may be as large os seven. There ,are MC values of ZC. MC ~ 12 c NC - (NO + 1) except for second and th ird order fits where NC may be as large as seven. There are NC values of YC. NC~ 12 153 -10- T NC2 - the number of secondary values of x to be read in for each valu'e of YC. There are NC2 arbitrary z values. NC2 is::; 18. NC3 - the number of secondary values of x to be read in for each value of ZC. There are NC3 arbitrary y values. NC3 is ~ 18. NE - operator that governs the type of error analysis us~ to distingu ish between possible fits when there are multiple fits allowed. Consequently, NE is useful only for second and th ird order fits where over-definition is possible (see analysis). NE TYPE FIT numerical average 2 product 3 numerical average X product 4 /. average2maximum error 5 product with minimum error stripped out 6 average with maximum stripped out 7 product with maximum stripped out 8 product with maximum and min imum stripped out NIv\ODE - an operator, together with NVT, that is used to define transformations in Table 1 . NO - order of polynomial Y defined in equation (1), where NO = n. NO may be as large as ten, but calculational error is such that it is suggested that NO be kept less than seven. NPASS - number of transformations allowed sequencially. Referring to the section on techniques, NPASS takes on the value 2 if x" transformations are made 0 NT - defined in Table 2 and treated similar to coefficients AT( BT, CT and DTo NT2 - defined in the same way as NT but treated sim ilar to coefficients AT2, BT2, CT2 and DT2. NVT - an operator, together with NMODE, that is used to define transformations in Table 1. Y - an arbitrary value of y which, together with a ZC, defines a secondary point, x. YC - the value of y used to define primary data points in the DC array and secondary data points in the DY array. 154 -11- ;e' . , __ oC o Z - an arbitrary value of z which, together with a YC, defines a secondary point, x. ZC - the value of z used to define primary data points in the DC array and secondary data points in the DZ array. INPUT TECHNIQUE Cards 3 and 4 below are read in repetitively in packages consisting of one card 3 followed by one card 4 prov ided MC ~ 6 or by two card 4s when MC > 6. NC of these packages are read in. Card 5 is repeated MC times. Cards 6, 7 and 8 form a package that is read in NC times. If NC2 = 0, cards seven and eight are not included in the package. If 0 < NC2 ~ 6, just one card 7 and card 8 follow NC2. If 6 < NC2 ~ 12, the sequence is 6, 7, 8, 7, 8. If 12 < NC2 ~ 18, the sequence is 6,7,8,7,8,7,8. Cards 9, 10and 11 form a package that is read in MC times. This package is handled the same as the 6, 7 and 8 package. Cards 13, 14, 15 and 16 fo"rm a package that is read in NPASS times. If NVT (I) < 6, cards 14 and 15 are omitted. 55 symbol comments card 55H 2 NO, MO, NC, MC, NE, KO 614 3 YC(I) EO.4 4 DC(I,J), DC(I, J + 1), DC(I, J + 2), DC(I, J+ 3), DC(I, J + 4), DC(I, J + 5) 6E8.4 5 ZC(J) EB.4 6 NC2 14 7 DY(I, J), DY(I, J + 1), DY(I, J + 2), DY(I, J + 3), DY(I, J + 4), DY(I, J + 5) 6EO.4 () Z(I, J), Z(I, J + 1), Z(I, J + 2), Z(I, J + 3), Z(I, J + 4), Z(I,J+5) 6E8.4 NC3 14 10 DZ(J, I), DZ(J +1,1), DZ(J + 2,1), DZ(J + 3,1), DZ(J+4,1), DZ(J+5,1) 6E8.4 11 Y(J, I), Y(J + 1, I), Y(J + 2, I), Y(J + 3, I), Y(J + 4, I), Y(J+5,1) 6E8.4 12 NPASS 14 13 NVT(I), NMODE(I), NT(I) 314 14 NT2(1 ) 14 15 AT2(1), BT2(1), CT2(1), DT2(1) 4E8.4 16 AT(I), BT(I), CT(I), DT(I) 4EB.4 () 9 0 Format Description Card 166 -12- KO XIHk£E OINENSIONAl SURFACf FIT caUE, "-151 "-151 DEMONSTRATION PRO~lEN. C~fffICltNTS POwER OF T~E nROER-I~1 SERlf~ :JI AI I, II I, 11 - .'0.15 71E-Ol IO.S032t:-04 BI 01 <'. 1I i. 1'.dOb41::-04 -t.:>.}496t-08 1.1 3 KG -= , IN XTHkEE DI"f~SIONAl SU~FlC[ FIT CODE. M-l~1 C) "-1~1 DEMONSTRATION pqO~l(N, URDtA-IXl ~F CUEfFICIENTS rHE P~.fR StRIE~ II I, 21 -10.'1571t-01 l , II 131 i. 21 1'.1j0t.1t,,-)1t -tHo }496t-\)8 All. 81 I~ Z 10.d011~-01t 131 THI) OPflON OUTPUTS T'iE F[,Ltrl .. I·~G Ol 4LTt:p.IUn lINF')-- tlRIl>I~Al nATA T A'llIl y XTH~ll OIM~~~IO~AL \U~F~tE FIT LuOF, TA~UlA'fU . 50.00001:-00 30.0000{t02 3).0000E.OI d7.11001::-0282.II00f-01 .0000t.-)) }i.)OO~E-OI 32.1191f-OI lO.0000E-JH 81.ll00t-02 82.IC?1E-01 10.0000(-08 17.81001::-01 11.8H99l-01 10.COOOl-OM 10.?100t-02 11.~~30E-01-72.b104f-1l 10.1400l-01 14.4q~OE-01-43.§~08~-n; so. OOOlil -Or) n.ClOOOf.Ol 81.IIOdt-OI 82.1100E-02 :'O.OOOO(,tOI 32.30001--01 3l.199'1E-01 10.OOOlll+01 27.0000ftOl 82.1100£.-01 82.10,}9E-02 ~0.00001:+01 17.89001--01 11.8899£-01 SO.OOOtil-DO 30.0000h02 H.0000£t02 31.0000t't01 41.0000t+0} 47.0000£t02 49.0000E.OI 10.00001:+01 30.0000E+Ol H. OO'OOE. 02 37.0000Et02 41.0000£t02 1t7.0000£+02 49.0000H02 AYERAGE PERCENTAGE EAROA 1\7.11001:-01 .0000£.-91 H. JOOOE-Ol 10.00001:-08 82.11 OOf-Ol 10.OOOOE-OS 17.11<)00£-01 10.0000E-08 90.9100E-02 10. hOOt-OI 11.90001'-01 14.HOOf-01 13.80001:-01 1').0900(-01 81.1100E-OI .~lInG~-?~ 10.00001'-01 1l.299h--01 1.1. )'")OL.t-08 <)9.?99 H'-Ul 82.109JE-~1 11.00001:-08 <)<).9999E-0£ 17.8b99f-01 10.0000t-08 ,)9.999'E-01 A9.~ltOOE-Ol ~4.7141t-Ol-~1.9478f-OJ 11.1~301::-~I-~2.6204t-01 1".4950F-OI-").SSC~E-OI le.h8~4E-01-b7.d44'e-Oi 22.813QE-01-81.63R v t-O. 29.1519f-Ol-S).57?~f-O.l 31.2S26f-01-ll.62bjE-O} 91.2700£-02 10.7359E-OI-IO.089~(-Ol 10.1000E-Ol 12 •• 192E-Ol-1S.1926E-Ol 11.• "HOOE-Ol 14.1015f.-01-ll.}2~b£-01 15.56001:-01 16.627~E-Ol-10.61~2E-Ol 11.0ltOOI::-01 il.~691f-01-~2.1111E-O~ • 12221486Et02 .36110770f.Ol 31.0000f.O.l 41.0000E.02 47.0000E.02 49.0000e.02 10. OOOOFt 02 10.0000e.02 3).0000E.02 37.00001.02 41.0000e.02 47.0000(.02 49.0000E.02 11. JOOOE-Ol 18. 6844E -01-61. 844 7E-OI 14.11DOe-Ot 23.8000t-01 21.0900E-OI Ie). 5400£-02 97.2700E-02 10.9000E-Ol Il.3700E-01 15.5600t-01 lr.0400E-01 .l2.~73AE-01-81.638bE-02 29.157JE-01-53.571Sf-02 31.151t.[-01-21 •• 265£-02 ')4.7 )41E -02-51. 9418E-03 10.7J51£-01-10.0895E-02 12.4192E-01-15.1926£-02 14.1015E-Ol-11.)256£-01 1 •• 627~E-Ol-l0 •• 752E-02 17.4691E-01-42 •• 172£-0) ~'I I TABLE .1lZZlu,e.Ol • UtlOnOf·Ol 'rJ 6 3 (, , '-/ -13- o SAMPLE PROBLEM I A practical problem involving the fit of hydrogen thermal conductivity as a function of temperature and pressure was met with success using M-151. The technique of regression analysis proved unwieldy in solving the same problem; lack of success be ing attributed to inexperience in picking suitable form for the fit plusthe inherent I imitations of regression analysis, such as inaccuracies resulting from the need to solve for large numbers of coefficients simultaneously. Some of the data points used are presented in Table 4. Referring to the section on Surface Fitting,Je,chnique, thermal conductivity was made the independent variable, pressure = y a~d femper(]t~Je= i:. Th'e'families:of curves were not plotted since it was decided to choose many combinations of n andm and study the. resulting error analysis. : Table.5 presents input us~dwhen n =:= m = 4. Step 4 in the,technique section was not used extensively since good results were ach ieved with the 4 x 4 order fit. Because of the vot0meofoutput wi-th the 4 x 4'case, only the error 'analysis w;ilJ:be, presented together with th~, ~rrors fro"", other case,s in Table 6. Table 3 presents examples of output available.' , ,. TABLE 4 THERMAL CONDUCTIVITY OF HYDROGEN 6 (X10 ) ;. ~ 50 . TEMPERATURE, OR .' , .' 2700 3100 3500' 3900 4300 ~82f1 '.9425 1<094 1.31'2 1.687 .', 2.166 , ~ 100 .8211 .9352 1.073 200 '.82 t'1 :927"1 1 .04,1 1.262 1.571 1.212: . ', 1.451 ' 4600 14800 5000 '2.626 3.230 ,. , ",. 1. 947 ' 2.301 2.760 J .721 1,.965 2.273 , " I/) 0.. ... 300 L.. :> I/) I/) Q) L.. 0.. .821,1 " Q) .9264 1.047 '; 1.203 i 1.430 ! 1.682 ::8211 .9250 1,.043 " 1.194: " 1.409 :1 .643.;', 500 ,,' ;.0211 .9237 1.0~9 1 .184, , 1.3Sa < J..604, 400 1.907 2 191 l~850 2" 109 .793 "'1~:: 678 2.026 1.627 10789 v ',: --, 700 .'ri2 1"1 ~. 9210 1'~ 031 1. 166: " 1.34$' ""1.526' 1000 .8211 .9204 1.029 1 . 159 1.327 1.491 >"J , 1.862 c -14- T TABLE 5 M-151 4 50. .8211 100. .8211 300. .8211 500. .8211 1000. .8211 2700. 3500. 4300. 4800. 5000. DEHONSTRATIONPROBLEM, ORDER=4X4 4 5 5 2 3 1.094 1.687 2.626 3.23 1.OOOE-9 1.073 1.571 2.301 2. 76 l.OOOE-9 1.047 1..43 1.907 2. 191 1.000E-9 1 .039 1.388 1.793 2.026 1.000E-9 1.029 1.327 1.621 1.789 1.000E-9 1.014 3300. 1. 19 3700. 1.471 4100. (t 2.38 700. 2.909 4900. 1 .0 3300. 1 • 158 3700. 1.396 4100. 2. 112 4700. 2.510 49QC. .9838 .3300. 1 • 119 3700. 1.304 4100. 1.788 4700. 2.041 4900. .9789 3300. 1.107 3700. 1.276 4100. 1.694 4700. 1.903 4900. .895"4 .9127 3000. 0 0 3300. 1.09 3700. 1.237 4100. 1.556 4700. 1.704 4900. 6 .9091 3000. 6 .9047 3000. 6 .8992 3000. 6 .8974 3000. 6 0 0 0 0 -15- 0 ' do. TABLE 6 ERROR ANALYSIS Order (n x m) Maximum Average % Error % 1x 1 12.22 36.31 1x 2 7.755 18.20 2x2 8.681 18.82 2x3 2.645 13.25 3x2 9.439 21.38 3x3 2.653 9.839 4x4 .04456 .3445 5x5 .07827 .7006 6x6 2.201 38.67 7x7 51.29 1211. FLOW Error CHART START LOAD M- IS I Se:...T PUNCH NrAPE. WR\TE WR",EA> = I", ,,",COC'{O .. 't../ ~OP =I HEAO ING OA> TAP£. OArA LOAO ON TAPE: M- J5Z-.. 'AJP VI -16- I LOAD "'1- 15L.- I t?EA D .,- GOEF"F" I SE.T fA) THE I c., £ ~T·~ MATR' ')( I I APE:.. I I GAUSS/~A,) 2- NTAPC. ELltVt INATIOAJ fv1,AT R \ )(' f? 0 l>T , AJ E.. ~ ~ I ~O~O~( ~ 0" 1.1 z. SU 6«0 Uti Ne:. I Z 3 /VCORO ~ -i ~4 IOUTf'VT ---I LOAD I M- 1.53 a.. M- LOAO i.j IS4-1 t k.. (.t G E'AJE RATE 0 == "3 I J W~tTE. DATA OAJ T'Af'E.J I LOAO [AJ M-153 IREAD OR M-/rcr f TAPEI I Rf:CALL R 0 UTII\)F"- DATA I SCT , 'SELEGTS' PO [.A...)TS I fA..) COE F'':-IC,I E:NTS I GAU~SrAtV MATR\X THE.. 1 I ELIMIA..lATION Z. I NGoRD MAT~\)( ROVT .,,-.,,--- 3 /bD I i I ~GORO M- .5"1 , L.OAD GENE~ATe:D -17- o 1 I S,.-A"ST(GAL (3E:5T $E.T EAc"H CONSr.ANT..,. OF E.~~O~ THE. SELe:.CTS ROUTlA.lE.. A 1..0 A) ~ 0..- - (t 4:,./.N£.. Of' EI!.A-rO R:", STATISTICAL SCo. LE:C;'-S USIAJu ,~ EAC.~ N£. 6EST T!-Ie:. k_. OF" ROUT. /VE~e.. T TI-lA-r FI TS a... .. ( I A~e:. ALL of POI AJ,-.S" OF VALU~.s OF Me.. c,oM eJ IAJAi/O/US TI-fE... "} GolV~ 10 C( £C' ~ INO [AJ A~e:. ALL LIA.,)f:S :} 'YES 100 1 A~E: a....C- [AJ A OF 8E..,ST f" SET OF (N 0 C ~OSE..A.j ARE. VALUeS FOR VAL-VC:.s A«E. THIS , Nc,oRO = I VALUES ALL • Se:T OF INO YE.S eL._ (1- ~ I NTAf'£ = "L I I 9.ATA OJ.) TAPE. I I I I ILoAD OATA OF [AJ r WRITE. 3 VALUES' FIT ~ SAVE-D. I I WRtTE. I) -I- ALL M-.~5~1 ON TAPE. , I LOAD M-15L fbI -18- 1 41.. ,1 1 • MAX IlVfUlvl LIKELIHOOD RESOLUTION OF TWO () MIXED NORMAL DISTRIBU'rrONS Reimut Wette Biomathematics Unit, Division of Hesearch, The University of Texas M.D. Anderson Hospitat and Tumor Institute, Houston, Texas Abstract: Samptes exhibiting particutar deviations from an assumed normat distribution can, in certain situations, be interpreted on the assumption that the parent distribution consists of two normatty distributed subpoputations, with different means and/or variances, and mixed in a finite proportion. The statisticat probtem is, then, to estimate the respective parameters. In the generat case, where no simptifying assumptions (e.g. equatity of variances) can be made, five parameters have to be estimated, viz. two means, two variances (or standard deviations), and the proportion of mixture. The author's estimation method of choice was the maximum tiketihood scoring system. The derived estimator is of the iterative type, and modified insofar as the information matrix is estimated as the dyadic square of the score vector. Numericat execution of this iterative estimation procedure requires initiat estimates of the five parameters; these are, at present, obtained ~anuatty from a normit graph of the data. A generatized distance from origin of the score vector (i.e. the quadratic form of the estimated variancecovariance matrix) is used as a criterion to exit from the iteration cyctes. rr'he pro cedure was programmed for the IJ3~\1 1 620 (with 1 311) in FOR1 HAN II-D. The main program consists of two parts: 1.- Routines 1 for controt transfer set-up, standard form data input or transfer to either of two optionat non-standard data input subprograms. 2.- The resoLution procedure proper, routines for optionat intermediate and other output ~~lnd transfer to two optiona t subprograms. / bJ. The program ?rovides, at present, the fottowing I/O (punched cards) and test opt ions (under parameter card contro t) : Data input: (1) Ungrouped or grouped data not exceeding 200 variates or ctasseso (2) Grouping of ungrouped raw dataJ ctass interval. and grouping range computed from initial. estimates. (3) Grouping of fixed-interval. vatued ungrouped data. Output and test options: (1) Output of grouped data frequency diStribution. (2) Output of expected frequency and cumutative frequency distribution together with observed distribution (grouped data onty). (3) Chi-squared test for normatity against skewness and kurtosis (cumutant test), with programmed bypass of resotution peocedure when not significant. (4) Intermediate iteration output of estimated information and variance-covariance matrix, parameter estimates and corrections, and convergence criterion. Input options 2 and 3 and output options 2 and 3 are subprograms catted as LOCALs on a 40k machine. A data traiter card controts exit to MONeAL or return "Linkage to part 1 of the main progra.m. Maximum core storage used is 35,604 cores (Monitor I, modif. 2, att-in-core subroutines), program package (2 main/"Link, 4 subprograms) disk storage is 435 sectors. Interation cycte running time is about 2.7 seconds per point (variate or frequency ctass). Convergence in this iterative estimation procedure de,ends on the goodness of the initial estimates, on sampte size and structure. Experience gather'ed so far indicates that convereence, if attainabte, is comparativety fast (tess than 10 iterations). Convergence coutd not be attained in a few instances of s~att and itt-behaved samptes, of size around 20 and l.ess, from \v11ich rel.ia"bl.e initiat estimates coul.d not be obtained. I I b'3 1: o The author gratefu1ty acknow1edees the extensive assistance of Mr. Lawrence E. :rewton, Jr., in ~rop:rammine and testing different versions of the method in the Computer Science Laboratory of the hl~D. Anderson Hospital" and Tumor Institute. Note: Methoc1ol"ogical" and e1sewhere. proEramminr~ deta.i 1.s wi 1.1. be pub1.ished o EVALUATION or TWO METHODS or FINDING SIGNIFICANT CONTRIBUTORS IN MULTIPLE REGRESSION ANALYSIS 1/ Mo Jo Garber and Richard H. Hill- With the advent of computers in recent years tremendous strides have been made in the ability to reduce to manageable numbers vast accumulations of data. In evaluation of experimental results the major emphasis is now on punching and proofing the observed values with the assurance that a generalized computer program is probably on the shelf ready for use. In the majority of cases this is true, but thoughts and concepts have expanded along with the increase in ability to do arithmetic quickly and accurately. We are, of course, now find- ing problems even large scale computers cannot solve in a reasonable period of time. One such problem is mUltiple regression involving many independent variabIes. A number of years ago the first of approximately 60 such problems was handed to the senior author o One phase of the experiment dealt with growth in a greenhouse of citrus seedlings in 102 nonfumigated old citrus soils. of 35 measurements was made on each. A total These included chemical, physical, and biological properties of the soil, plant composition, and relative growthe Locating statistically the best single contributor and the best set of 33 of the 34 independent variables is relatively a trivial operation. The best single contributor is the variable with the highest coefficient of determination (r~oy)e ~ The best set of 33 (or ndl) contributors is easily found by entering all n vari= ~ I I I ables into the regression equation, and deleting the variable with the smallest 1/ University of California, Riverside, and Informatics, Inc. 2 absolute partial correlation (Ir·. . YXj • x.;x • .J =,,.,~ k I ') ~ ' Mre Hill (then at the University of California. Los Angeles> wrote the MISLE program for the 709 The best set of two~ 0 the best set of n ... 2 9 and all best sets between these extremes present a very different problem. The best set of two will not neces- sarily contain the best single contributorS) and the best set of 1'1-2 could contain the variable not included in the best set of n-10 analysis all best sets should be calculated. problem. For a statistically efficient But this presents a formidable For example. the final regression equation for the above problem con- tained 10 independent variables. Finding the best set of 10 among 34 variables can be quite a chorel! even for a computer. is 34!/(lO!24!) or 131,128,14011 The number of sets to be evaluated Assuming that a computer can solve these at the rate of one per second, over 36,424 hours are required for finding the best set of 10. A number of questions immediately come to mind: (1) Is the best set of 10 significantly better than the one selected by this procedure? (2) How does the best set of 10 compare statistically with the best set of lIt or any other best set? (3) As the correlation coefficients are statisticsf) and not parameters, is the additional labor worth doing? After some soul searching it was decided that initially the following threephase procedure would be used: (1) MISLE Program (presently UCRBL 0052~ 1620 Library File No. 6.0.37) All independent variables are entered into the regression equation, the inverse matrix af sums of squares and cross products being calculated in the process. 2 All b./e .. are calculated, where bo is the partial regression coef1 11 1 ficient of y on the ith independent variable, and e i i is the major diagonal 1 166 l~ I:) -.----~---- ~- --- -------- 1'1 3 o element for that variable in the inverse matrixo As Fisher (1) has shown b~/cii is the variance (equivalent to a sum of squares with one degree of freedom) which would drop into the error sum of squares of y if the variable were deleted. 2 The variable with the smallest b /c is deleted. and the remaining n-l independent variables are entered into the regression equation. This procedure is continued until only one variable remains. An example of such analysis is illustrated in Table 1 in terms of the squares of partial correlation coefficients. Note that. in general. the simple correlation coefficients, the squares of which are shown in the second line of the table, do not reflect in any way the magnitudes of the squares of the partial correlation coefficients as variables are deleted. The leftmost 10 variables were found to be statistically significant contributors (0.05 level of probability), and it should be noted that four of these showed nonsignificant simple correlation with the dependent variable. The magnitude of R2 was 0.5387. Variable 34 has the largest simple correlation. other variables its contribution is relatively small. Yet. in combination with On the other hand. Variable 31 which was not significantly directly associated with the dependent variable is relatively a large contributor in the presence of other variables. (2) Stepwise Regression (UCRBL 0018, 1620 Library File No. 6.0.031, a modification of the BIMD 9 developed at UCLA by the Biostatistics group) Here the procedure is to begin with the variable most highly correlated with y co From the remaining variables the one with the highest (absolute) partial correlation is selected for entry (assuming that it meets the F criterion), and the magnitudes of the two contributions are evaluated. Either both are retained or one is deleted / b7 .. --------------,-.~,,- .. ----"" ~,--, .. --- ------- ----------- 4 (a parameter card entered with the data contains the minimum F value required to enter a variable into the regression equation and the minimum F value required for retention of a variable) 0 This process of entering, evaluat ing. and deleting var i- abIes continues until the criteria can no longer be met. For the problem being considered here nine variables were found to be 2 statistically' significant contributors (R = .5205) when grouped, Eight of these variables were also selected, by the HISLE program. (3) MISLE Runs (a) The eleven variables selected by either or both of the above programs were fed into the computer for another HISLE evaluation. Ten of the vari- abIes were found to be significant contributors, and the first variable of the eleven to be deleted was the one unique to the Program IB run. There was thus no significant gain. (b) It was then decided to make another run, this time including the eleven above and certain additional variables, (~ Examination of the F values 1) led to the choice of the two variables last deleted by Program 52 and the succeeding three which would have been entered by Program lB. Two of these were common to both runs. Of the fourteen variables entered for the final run 10 were found to be significant. Seven of these were among the variables found significant by both programs, two common, but in the noncritical area, and the tenth being a signifi-cant contributor found by Program lB. The magnitude of R2 was .5433. The 10 variables statistically significant in this run were the ones coneluded to be real contributors (paper by Martin, Harding, and Garber (2» The four curves are shown in Figure I. t. The curve of multiple R2 calculated by Program 52 (labeled 521) is almost smooth. running essentially horizontally o 5 until approximately six varidbles have been deleted o steep as addit lonal variables are deletedo The drop off becomes quite It should be noted that the last remain= lng variable is not the one most: highly correlated with yo The la'tter Sl Variable 349 was dropped after the 26th run. The curve of mUltiple R2 calculated by Program 18 begins by entering Variable 34 then rising in a somewhat oscillatory manner to essentially ma.tch the 521 curve after 9 variables have been entered o The short spur labeled 521Ia is the graphic result of the MISLE run with 11 variables& The short section·of curve (52I1b) exhibits a sharper break than either of the first two. The remaining two curves in the figure are from the 521 rune. The curve of 2 "Cumulated b /e" shows the addition to the residual sum of squares (all variables initially entered into the regression equation) as variables are deleted.. The curve of "Residual Mean Squares" falls off from its initial value i then rises t) Recalling that the residual mean square is the quotient of the residual sum of squares by the residual degrees of freedom it is seen in this example that in the initial phases of variable deletion the denominator is increasing relatively more rapidly than the numeratoro. Only after approximately half the number of variables has been deleted does the curve begin to rise o A summary of the evaluation of 55 experiments· is given in Table 20 left side of the table the r criterion was set at 1 00 0 0 For the For the right side of the table the 0 05 level of probability was the criterion o 41 In eight of the 55 experiments Program U programs., No~ 18 was the better of the two In four experiments Program NO e 52 gave better results o In one experI- ment each of the programs provided information both significant and unique u • 6 In one experiment. the order of entering variables indicated the advisability of a third run. The outcome was the set of four variables selected by both programs. For the remaining 41 experiments both programs gave the same results in the initial runs. The third run was unnecessary. In summary, there is no clear cut indication of the superiority of one method over the other. For our purposes at Riverside we will continue to evaluate large problems by both the deletion and the stepwise methods. LITERATURE CITED 1. Fisher, Sir R. A. 10th Editione 2. (1948). Statistical Methods for Research Workers. Hafner Publishing Company. Martin, J, P., R.. Bo Harding, anq M, J. Garber. (196l). Re'lation of Soil Properties and Plant Composition to Growth of Citrus Seedlings in 100 Nonfumigated and Fumigated Old Citrus Soils. 17D- Soil Science 91(5):317-323. '0 ............... ~- - o o • '-. -.......1 ......... 6184(3 6184 6184 6184 6182 6179 6175 6171 6164 6146 6131 6120 6104 6067 6035 6018 5967 5948 5876 5817 5714 5693 5644 5489 5387 5171 5006 4682 4426 4194 3866 3434 2565 1154 1(1) 32 12(2)11 8(4) 13 8 13 8 13 8 13 8 14 9 14 9 14 9 14 9 15 10 15 11 15 11 15 11 15 10 15 10 15 9 15 10 17 10 17 15 16 14 15 12 18 12 18 11 17 14 17 13 17 12 17 16 24 12 25 14 22 28 19 26 15 26 16 16 16 12 31 1 10 10 10 10 10 10 10 10 10 12 13 13 13 13 13 13 12 11 12 11 11 13 14 14 14 17 21 19 16 17 13 12 11 7 0 0 3 4 7 7 7 7 7 9 8 8 9 8 8 8 7 7 7 8 8 7 6 10 10 11 12 9 7 6 7 6 0 3 8 5 2 3 8 8 8 3 3 5 6 5 8 5 8 8 7 7 10 7 9 7 9 7 9 9 8 8 10 10 9 10 9 8 12 8 12 7 11 8 11 10 9 10 9 10 8 10 8 11 9 11 9 7 9 9 6 4 6 5 2 29 34 24 20 27 22 16 17 23 19 9 15 5 4 30 18 28 33 13 7 26 3 21 12 14 10 25 4 o 16 0 1 12 11 0 0 0 4 3 10 4 2 3 1 1 0 6 11 0 1 0 2 1 6 3 4 4 5 5 7 1 3 1 1 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 '2 4 4 1 5 7 6 1 1 1 ../ 2 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 4 4 7 1 5 6 1 3 1 :!. 0 2 2 0 0 0 0 1 0 0 1 0 0 1 1 1 4 5 8 1 6 5 1 3 0 0 2 2 0 1 0 1 1 0 0 0 0 1 1 1 1 4 0 3 5 7 0 6 0 0 2 2 1 5 1 3 0 0 1 0 1 1 1 1 1 0 3 5 4 0 0 5 8 3 2 6 2 1 3 0 0 1 1 1 1 1 1 1 3 8 4 0 5 5 6 2 2 3 2 2 0 3 1 0 1 0 1 1 1 1 3 5 8 4 0 5 6 2 2 3 0 0 3 2 2 1 1 1 1 1 1 3 9 6 5 6 5 2 4 2 3 2 2 0 0 2 1 1 1 1 1 4 5 6 8 4 6 6 2 1 3 2 2 0 0 2 1 1 1 1 5 6 6 8 5 6 2 3 1 3 2 2 0 0 2 1 1 1 5 7 6 6 8 4 2 5 2 1 2 2 0 2 1 1 1 5 8 8 5 6 5 2 1 3 2 2 2 1 2 1 1 5 7 8 4 6 6 5 2 3 2 2 2 2 1 1 5 7 7 8 5 4 3 6 1 2 2 1 0 1 2 1 5 4 3 1 8 8 5 2 6 6 2 7 5 4 1 0 7 6 8 2 3 3 2 9 8 6 6 5 4 2 6 3 2 2 7 7 8 8 5 4 3 2 3 1 8 8 9 5 6 3 3 2 2 4 8 7 7 5 '3 0 2 7 7 7 5 4 3 1 7 4 4 6 6 3 8 7 6 5 2 LEGEND 8 5 4 6 7 3 1. Variable number (Variables were deleted from right to left) 6 7 6 2 (Two decimal digits) 5 2. SiJap Ie rYXi Table 1. 3. 2 Multiple Ry.ijk--- (Four decimal digits) 4. 2 Partial ryi.jk--- (Two decimal digits) Computer re.ults (34 independent variables, 102 sets of observation.). Table 20 No. of Indep. Var. 34 32 32 31 31 30 - - ..J >U Degrees of Freedom Summary of evaluation of 55 experLments. No. of Variables No. of Selected by Var. in Common 0018 0052 101 44 44 44 44 12 44 7 4 2 7 9 5 4 12 3 10 9 6 4 5 8 3 2 2 6 3 2 1 29 9 44 15 31 31 10 44 44 3 2 6 2 6 72 72 5 1 2 2 33 101 6 6 4 32 44 4 5 4 4 Final Selection No. of Variables No. Unigue to No. of No. Unique to 0018 0052 Variables Common 0052 0018 10 2 9 9 6 4 8 3 3 3- 3 1 6 2 1 3 1 3 2 3 2 6 2 2 1 1 2 2 2 2 2 5 3 1 4 4 2 2 7 3 1 3 2 1 2 2 5 1 1 1 2 1 3 2 2 3 1 2 3 4 2 2 2 1 1 SU1IID&ry of 41 additional experLments: No. of independent variables: 6 to 32 No. of degrees of freedom: 24 to 152 No. of significant contributors selected in common by 0052 and 0018: 0 to 5 • ~ " ' 1 'i (,! <." C' II'IWI"z' 'lUI! !' ! ' i j kl b tetHr'W+ '+h we +lrl. u,,+i"+# • 0 • CD rn 3. ~y~ • ® [4] .~~ h\hos ~ lIJ .1 -'d, + ~..iL ® ~ -- ®- 1 7 ACNODE Bl N(1 20)2 G=.25 2 C=.l B2 N (2,0), R = 5.0 B3 N (0, 1), L = 2. 0, E = 1/0 B4 N (1, 2), L Ml B (3. 4). M = 1. - = 3. 0 ° FREQUENCY = 1. 59 EXECUTE o -.. T :0 1 2 3 .. 561 8 910'1 2 3 4 5 6 1 8 9 i O 1 2 3 .. 5 6 7 8 9 , I \ i 123456789 i , 1 1 J 4 567 8 °10 1 2 345 6 7 8 9 1 2 3 .. :5 6 1 8 9;0 1 2 3 .. 5 6 7 8 I I i i i AbNOOE I NI~ ( 1 ,0) ,G,= • ~ '5, C =• 1 I B1 I '\2 N~(2,O).R=51. 114 NI:: ( 1, 2 ) • L =3,1. ~1 ~l ~~(O,1),l=2~,E=1./O. B~(3,4),~=1. F ~ E CUE N ~ V = Ep ., ... ~~ ..~,." ....- - - - - - - -_ _- - _ _ _ t - - - - Fig" 6a LCR Voltage Regulator - Schematic Diagram .2 mhy ---- VI V2 ~~. 84 Bl .In 2,uf + 15v. + > (lM,OS) 1 V3 -...- t-L B5 TV4 > +> 86 ~ 111 ~ F!g. 6b LCR Voltage Regulator ... Equivalent Circuit 1 tl t~ <.....,. B7 6n B8 < J±. 1 I (12.0,lJ.·· o . o ... 1 2 3 4 5 6 7 P cO' 2 , o 4 , 6 ! rl 9 0 I 2 1 4' ~ /:1 7 8 ., 0 2; 4 " 6 1 q ., 0 ! .. " 4 5 6 7 8 'l 0 i 2 { 4 ,i 6 ", H Q 0 , 2 :) 4 5 6 7 B 9 0 1 2 :, 4 1 2 1 ~) 6 7 8 9 0 I 2' 4 TRANSIENT c c C LCR VCLTAGE REGULATOR C C 81 62 N=(O,l),R=(.l,l.E6),E=(lS.,15.) N=(O,ll,R=(1.E6,.S) N=(l,Z),R==.l N=(2,3),l=.2E-3 N=(3,4',C=.2E-5 N=(O,4),R=1. N=(O,3),R=6. N=(3,O),R=(1.E6,1.E6),E=(-12.,-11.7) 6=8,(1,2,8J,OFF OElTA=1.0E-6 63 84 65 66 B7 B8 Sl ~AJOR=20 1ERROR=1.OE-2 FINISH=200.0E-6 EXECUTE EXECUTE PROGRAt-1. TR CALLED. T = O. V 1 2 3 V it O.1500E 02 O.lSOCE 02 0.1167E-04 O.1155E-04 V V 4 56! 8 9 G 1 2 ~ T ::: 0.2000E-04 1 2 ::: ::: O.1487E 0.1414E O.4436E O.SS18E 02 02 01 00 V 3 ::: 4 .: I I I I I 1 = O.1291E 01 2 = -O.1481E-04 3 ::: O.1297E 01 4 :: 0.1297E 01 5 == O.5578E 00 6 ::: -O.5578E 00 1 == -O.7394E 00 S 8 ~ 0 '"'i Q. 7 S 6 <4 It.".Q § 2 I ,1 O.1188E-04 V 1 :' = 1 ::: -0.1945E-OS 8 ::: -O.1200E-04 V 1 ::: 2 == -0.lSOOE-04 3 == o. 4 == O.1500E-05 5 :: O. 115 5E-'04 6 ::: -O.1155E-04 "V o :I 1 = I I I I I I I I 0 :: 1 .'i 6 _~ : L 5 6 : 8 9 0 1 2 10 3 4 5 6 1 8 9' 0 1 2 J 4 ., 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 4 .'i 6 7 8 9' r) 1 ? 1 4 OJ? i ,J \ 6 .' R ~) ,) i) V V V 1 = 2 ~ 3:: V 4:: I I I I I I I I 1 2 3 " 4 'j 6 7 1" B 9 0 4 " g , ", (, 2', 4 O.1480E 02 0.14SgE 02 0.g586E 01 0.4516E-OO = 0.2049E 01 -0.1480E-04 0.204gE 01 = :: :: = = 1 2 = O.S318E-04 0.1411E 02 0.1454E 02 O.l2eCE 02 0.3031E-OO V V V V 4 = I I I I I I I I 1 = 0.2303E 01 2 ::: -0.1417E-04 3 0.2303E 01 4 :: 0.2303E 01 5 = 0.3031E-OO 6 ::: -0.3031E-OO 7 = -O.2000E 01 8 = -0.4399E-10 3 :: ~ SWITCH T ::: 1 IS ON 1 ::: -0.IIS2E 01 2 = -O.1382E 01 3 = 0.l200E 02 4 :: O.3026E-OO I I I I I I I I 1 = 2 ::: 3 4 5 6 7 8 4 O.5378E-04 V V V V 0.1615E-04 0.2303E 01 = 0.2303E 01 = 0.2303E 01 = O.3026E-OO = -O.3026E-OO = -0.200CE 01 = 0.302SE-06 T::: 'V'61G~ V 7 4 0.2049E 01 5 = 0.4516E-00 6 = -0.4S16E-OO 7 = -0.ISg8E 01 8 = -O.2414E-OS T 1~4~,6-'890 : 0 O.6COOE-04 ~Q~94S5i,QQ2" 2:: -0.1135E 01 '0 1 8 9 0 ' 2 3 4 5 6 ; 890 , 4 ') 6 8'; 0 ' 4 6' 8 'j 0 1 2 1 4 ' .', " 8 , 0 ' :> " ,,! I i • • (\ 3 4 V V " .. 4 \ ") 7 . 11 ') ,. ," C :: :: f' ~ 1 O.lIS3E 02 -0.803SE-Ol . , ;:~ - " 'I O.lS95E-04 O.18Q1E 01 O.1891E 01 4 .: O.1891E 01 S ::; -O.8035E-01 6 == O.8035E-01 1 :: -0.1971E 01 8 .: 0.1279E-06 I I I ~ 2 == 3 .: I I I I [ T .: O.6116E-04 -O.9080E 00 -O.lOqOE 01 3 .: 0.1110E 02 4 = -0.1340E-OO V V V V I I I I I I 1 .: 2 :: 1 2 :: 3 4 :: = ::I 5 .: 6 1 I I e .: ~ :: S~ITCH T == O.lSqlE-04 0.l816E 01 O.lSl6E 01 O.1816E 01 -0.1340E-OO 0.l340E-OO -O.lqSOE 01 0.3191E-IO 1 IS UFF 0.6116E-04 V V V V 2 3 =! 0.1482E 02 O.1464E 02 :::Ii O.1170E 02 4 == -0.1338E-OO I I I I I I I I 1 :: 1 == O.1816E 01 2 = -O.1482E-04 O.1816E 01 O.lS16E 01 5 - -O.1338E-OO 6 :: O.133BE-OO 7 = -O.1950E 01 8 :: -0.3011E-'06 3 ::I 4 ::I T :: 0.1991E-04 1 2 3 4 .: =i -, O.147ge 02 0.1458E 02 0.1200E 02 O.8145E-Ol D V V V V o 2 3 4 ~ 6 ; 8 4 f) I 2 ; - I " 1 R ~ () , Q. 2 Q e 7 it .,Q 1 2 I 2 = -O.1479E-04 f) ~/~ i 4 , 6 ;' 8 9 U : 2 3 4 5 6 7 8 ~ 0 I 2 :i 4 0' 6 ;- 8 y 0 1 2 3 4 5 6 7 8 9 0 1 2 '1 4 5 6 7 8 9 0 1 2 : ~ I I 0123456789 0.2081E 01 0.2081E 01 3 = 4 = 12 3 { 5 6 58 '=10 I 6 I 1 I 8 1 =l 3456789 123456789 123456789 123456789 123456789 1234 (,c-", -O.2000E 101 0.1895El10 1 I I sWITe1 T f)~e'1t5E,,:()12 ~ -O.8145E~01 1"-..,,'/, I I 1 liON 1 o.7991El04 4 v -O.1044E 101 1 2 =1 -O.1252E 101 3 ~ O.1200EJ02 4 == 0.87301: 101 V V V I I 1 = 2 = I I I I I I I I 3:oJ 4.J O.1604E~04 101 0.2087E 0.2087E 101 0.2081E !01 5:J O.8130E~01 6 J -0.S13ce-iol 1 -0. 2000E 101 8 0.3001E-I06 1, i I I I I T:d f 0.aOOOE1 04 i 1 ~ -O.1041E 101 2 ~ -0.1249E !01 3:j 0 • 1200 E !O 2 4 =j O.8142E-:Ol V V V V 1 I 1 1 =rJ 2 ~ 3 .:; I I I I I I 6 I 1 O.1604E-j04 O.2081E i01 0.2081E :01 :i 0.2081E i01 i =I O.8142E- Ol = -0.8142E-;01 ~ -0.200CE j01 =1 O.2981E-i06 1 4 5 a I 1 ! i T ~ I 012345678 o 1 2 3 O.8386E- 04 i I J V V V V 1 -O.9145E 100 2 :: -0.1091E iOl 3 :;1 0.1110E i02 4 ~ -0.1209E- OO I I I I I O.1591E-04 0.1829E iOI 3 =! 0.1829E 01 4 -, 0.1829E 01 5 :: -0.1209E-'OO 6 = .0.1209E-00 "18 :;: 0"" Q • 19 S Cit 9 Q 1 2 a = -0.3254E-I0 I I I r 1 =!i 2 5 6 =1 -:J 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 n '/:3 0 1 2 3' 4 5 6 7 8 9: 0 1 2 3 4 5 6 7 8 9 !, 123456789012141 , il ,I' , • • 0123456789 123S,h'llCt;0123i567tsoOfF456789 0 T =: 1 =: 0.1482E 02 2 =i O.1463E :02 3 l O.1170E i02 4 :; -O.1207E-100 1 1 ~ 2 =; 123456789 1234 1 i ! :3 I I I I I 4 0.1829E i01 -O.1482E~04 5 6 1 8 0.1829E :01 O.l829E :01 -0.1207E--00 == 0.1201E-,OO = -0.1950E ,01 :; -0.30llE-06 T =I O.lOOOE-403 i O.1419E i02 O.1459E :02 0.1192E :02 O. 162 3E-~O 1 I i ~ :; = V V V 1 =i 2 ... 3 ~ V 4 =4 I I I I 1 I I 1 I I 123456789 O.8386E"104 V V V V I 1234567890123456789 I 1 ~ 0.2063E '01 2 ~ -O.1479E-i04 :3 ~ 0.20b3E :01 4 ~ O.2063E ,01 5 ~ 0.1623E-'Ol 6 =1 -0.1623E-iol 7 ~ -O.1987E 101 I 8 :d -0.7665E-i07 I i! i T =1i V V V V iI 1 ~ 2 =: 3 ~ 4 =1 i 0.1016E-(03 0.1479E :02 0.1458E 02 0.1200E 02 O.8494E~01 I I 0 I I I I 1 I 1 I 1 2 3 4 5 =! O.20S5E 101 i =1 -0.1479E-I04 ~ O.2085E [01 O.2085E iOl C.8494E-:01 6 =1 -0. 8494E-iO 1 7 =! -0.2000E :01 8 == -0.2325E-:10 i =i =1 I i I : SY4ITCH . 1 IS ON :0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9f0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 S ~l& (1: 0 1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9! 0 1 2 1 4 T , 0 1 2 3 4 5 6 7 8 9 0 123 "i5 = 0.1016E-03 6 i8 '=l0-O.lCit2E9!V12 3 4 56789 V 2 ~ -0.1251E \01 V V 3 =1 4 123456789[0123456789 123456789 123456789 1234 'I 0.1200E 1 O.8419EI01 ;02 I I 1 ~ 2 ~ I I I I I 1 I I 0. 1604E-j04 O.208SE 101 O.2085E 1101 "~ O.208SE ]01 5:j O.8419E~01 6 -0. 8419E-j0 1 1 -O.2000E 101 8 O.3001E-106 3:d 1 V i ! I : T 1 1 ::I 0.1056E~03 ! !oo -0.9136E -O.1096E 101 3 =i 0.1110E :02 4 ~ -O.1228E~OO 2 V V V J I ~ i I i I 1 I. I I I I I I 2 3 =I " =1~ I O.1591E~04 O.1827E 101 0.1827E '01 0.1821E 101 5 -0.1228E~00 6 ~ 0.1228E~00 1 =1 -0.1950E \01 8 ~, 0.5I62E-IIO 'I I, ! SWITC~ 1 I I I T J I V V V V 1 2 3 4 I S! OFF ~ O.1056E~03 I I O.1482E !,02 0.1463E [02 O.l170E 102 -O.1226E-ioo ~ j I I I 1 =1 2 =1 I I I I I I I I 3::d, 0 • 182 7 E '10 1 4 ~ O.1821E tOl 5 d -0.1226E-100 6 ~ 0.1226E-1 oo 1 ~ - o. 1950 E 10 1 8 :::j -0.3010E-!06 1 . i T =4 V I °1 2 3 4 5 6 7 8 9: °1 2 V 3 V5 V O.1821E iOI -0.1482E~04 6 I I O. 1200E-!03 i 1 ~ O.1480E02 4 = O.6492E-Ol 2 ~ 0.1459E ;02 1 8 ~ 0 1 Q. 11 tH. E 9 :Q 2 2 3 4 5 6 7 8 9 J. 12345678 234567890123456789 123456789012341 Ii] , o 1 2 3 4 5 6 7 8 9 0 1 2 3 15 6 18 '=10 1 I 2 1 3:i I 4 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 123456789 123456789 123456789 1234 123456789 123456789 12345678 1234567801234 0.2039E tOl O.2039E 101 O.6492E~01 i 5 ~ 1 0 ~! 03 9E 9:012 ~ -0.1480E~04 6 ~ -O.6492E~01 1 ~ -0.1974E :01 I I 8 ~ -0.1576E~06 I I I I T ~: Oo1235E~03 I V V 1 ~ 2:j 3 ~ V It:l v i 0.14 79E \02 O.1458E !02 0.1200E !02 O.S534E-j01 I i 1 I I I 1 =j 0.2085E 101 2 1 -0.1479E104 3 ~ O.2085E 101 O.2085E j01 5 0.8534E101 6 -0.S534E101 7 -0.2000E 101 8 -O.2384El12 ".1 I I I I I 1 SWITC~ 1 I I T V V V V 1 2 1 I I I I I I I =i ~ O.8518E~Ol 2 j 1 J ! I I I O.1275E-103 ! I ! : ' 1 =1 -0.9136E iOO 2 ~ -O.1096E 101 3 ~ 0.ll10E 102 4 1-0.1228E-!OO V V V V 1 2 3 1 4 T o 1 ==l I I ,1 0 12345678 , I 1 Oo1604EJ04 O.20a5E 101 3 ~ 0.2085E !01 4:j O.20a5E 101 5 =l O.8519E-!01 6 -O.8518E-lol 1 ~ -0.2000E iOI 8 O.3001E-lo6 I o I O.1235E- 03 -0.1043E 01 -0.1251E 101 0.1200E ,02 3, It I 1 ~ I~ ON O.1591E-~4 I 1. I I 2 =1 O.1821E :01 3 8 ~ 0 1 Q. 18 21 E 9 iQ 1 2 0.1827E 01 I 5 6 ,,= 3 4 5 6 7 8 9 5 = -0.1228E-00 6 = 0.1228E-OO { 5 6 " 8 =0 ..... O. 1950 Eco . V12 I 8 ~ 0.5484E~10 I • I ',0 ! 2 3 4 5 6 7 8 9·0 i 2 3 T V V V V =I 0.1275E-j03 1 =i 2 ~ :3 .::Ii 0.1482E :02 0.1463E :02 0.1170E :02 ! It :II -0.1227E~OO I 1 ~ 0.1827E :01 I I I I 4 5 .:: 1 I 7 =i -0.1950E 01 8 ::I -0.3010E~06 1 1 : 2 3 4 5 6 7 8 9; 0 1 2 3 .4 5 6 7 8 9 iO I 1 2 3 4 5 6 7 8 9 10 1 2 3 4 I 2 == -0.1482E~04 3 -: 0.1821E ;01 6 ~ = 0.1821£ 01 -0. 1221E""00 0.1221E~00 T =i 0.1400E-;03 1 :j O.1480E '02 0.1460E 02 0.1171E 102 0.5081E-10 1 V V V 2 ::Ii 3 =9 V 4 :i I 1 :i 1 2 ~ , 3 ::i '+ ~ , 0.2013E '01 -0. 1480E.... 04 0.2013E ,01 O.2013EOl 5 =1 O.5081E~Ol I I I I I I 1 2 3 4 5 6 7 8 9 0 IS OFF 1 SwITCI1 3 4 5 6 7 8 9 i 6 =!I -V.508lf-Ol 7 =I -0.1962E :01 8 .::I -0. 2296E~06 T O.1454E-03 ::I i V V V V 1 =I 2 =I 3 :1, 4 0.1419E ,02 0.1458E ;02 0.1200E i02 0.8534E-:Ol =1 I I I I I I I I I o 1234 5 6 7 8 9 () 1 .=! 0.2085E iO I 2 -0. 1479E-:04 3 =I 0.2085E :01 - 4 ::I 5 6 ': :: 1 -; 8 - 0.208SE 01 0.8534E-Ol -0.8534E-'Ol -0.2000E 01 0.1550E-:}0 , '1 3 456 .: 8 9 0 ; 2:1 4 5 6 7 d l7 S 9 U ',2 :J " 5 6 7 89[0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9iO 1 2 3 4 567 8 9 0 1 2 3 4 567 8 9 0 1 2 1 4 SwI1CH o 1. ~ 3 4 5 6 7 8 9,0 I 2 3 4"; 6 .7 8 9,0 0, T , 2 J 2 3 4 6 7 8 9,0 I ? = = 7 B 9,0 I 2 J 4 5 6 7 8 9 0 ! 2 3 4 56? 8 9 0 1 2 3 4 5 6 7 8 9 ~ io 1 2 3 4 1 2 3 4 5 6 1 8 9 i0 1 2 'l 4 I 2 3 4 6 7 8 9 -O.1251E 01 0.l200E 02 O.852CE~Ol 1 8 T = 5 6 0.1493E-03 1 == -0.9131E 00 2 = -0.1096E :01 3 ..., O. lI70E ,02 4 ~ -0.1226E~OO V .. V V 0.1591E .... 04 O.1827E 01 3 =! 0.1827EOl 4 == 0.1827E '01 5 ::I -O.1226E-iOO 6 == O.1226E-:00 1 :j -0.1950E 01 1 -= I I I I I I I I 2 =I 8 :;! 0.1252E~lO SWITCH 1 IS OFF T =: 0.1493E-'03 :; 0.1482E 02 0.l463E 02 0.l170E 02 -O.1224E-OO V 1 V V V 2 :: 4 ~ I I I I I I I I 1 =: 0.1827E 01 2 =i -0. l482E-'04 3 =1 0.1821EOl 4 =: 0.1827E 01 5 == -0.1225E-00 6 -, 0.l224E-OO 7 =i -0.l950E 01 e == -0.30 l1E-'06 I 2 3 4 5 J 4 5 6 0.1454E-'03 ~ O.1604E-04 = O.2085E 01 O.2085EOl = O.208SE 01 = O.8519E-01 = -O.8520E-01 = -O.2000E 01 = O.3007E-06 4 V 1 2 3 4 5 6 7 8 9 0 :) 1 .:: 2 3 -= I I I I I I I I :0 4 1 = -0.1043E '01 V V V V 0 IS ON 1 3 ~ J? =: ~ () 1 Q. 16 CQ E~ Q] 2 dl''b .1 4 5 6 7 8 0. 0 1 2 3 4 5 6 l 8 Q 0 : I 2 1 4 5 6 7 8 " 0 i 2 3 4 5 6 7 8 9: 0 • v v 1 := 2 -= 0.1480E 02 0.1460E 02 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9:0 1 2 3 4 5 6 7 8 9 [0 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9:0 1 2 3 4 I(~, i/"I\..J! I I I 1 ~ 2 ~ 3 :4 1 4 ~ I I 5 =1 0.1986E 101 -0.1480E~04 0.1986E '01 O.1986E 101 O.3425E-iOl -0. 3425E"";0 1 -0.1952E 101 ~ =i 1 6 1 I 8 =i -O.2891E"';06 V V V V I I I I I I I I T ~ 0.1612E~03 1 := 0.1419E :02 0.1458E ;02 0.1200E '02 2 = 3 ~ 4 -i 0.852aE~Ol -: i O.20a5E 101 -O.1419E..,04 I 3 =9 0.208SE lOI 4 ~ 0.20a5E :01 5 =iI 0.8528E~OI 6 ~ -O.8528E-i01 1 ~ -0.2COOE 101 8 9 O.1152EllO I 1 ::::j 2 ::::j i i I SWITCt1 1 lSi ON I i T V V V V ~ 0.1612E-!03 ~ -0.1043E 101 -0.1251E 01 3 1-i O.1.200E !02 4 ;; 0.8513E~01 1 2 ::j 1 i I I I I I I I I 1 ~ 2 =, 0.1604E~04 3 ;; 4 :i O.2085E '01 0.208SE ;01 0.2085E ;01 5 =i 0.8512E~01 6 =1 -0.8513E~01 1 :; -0.2COOE :01 8 0.3001E106 ie 4 T :9 I I I , ; .j !I I 0.1112E-I03 I I V I ; ; -0.914{)E iOO V 2:9 -0.1097E :01 V 3 ~ 0.1170E 02 V 4 ~ -O.1219E-,OO 012345678 1 2 3 4 5 6 7 8 9 0 1 2 ~ 45 (, 78>' 0 1 2 3 4 5 6 ' 7 I i 8j!O', ;}d(J ' . ! 2 3 4 : ; 6 7 8 9 \ 0 1 2 1 4 5 6 7 8 9 ; 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9!0 1 2 1 4 • o I 2 3 4 5 6 7 8 9 0 i '2 ~ • , I I i 1 2 5 6 38 = C.1591E-04 0 •. 1828E 01 = i:I 0 1 () I =\ ~ 1 82 e E 9 ! V 1 2 I It I 5 =\ -0.1219E~00 6 ~ 0.1219E~00 7 ~ -0.1950E 101 8 ~ 0.1371E~10 I I I 1 T ~ I I I I I 1 I I 1 2 3 =I 4 ::I :I :I 1 2 ::I 3 =I 4 =1 5 6 .~ 6 7 8 '7 I~ i a ' 0.1828E ;01 SwlTC~ V V V V J 4 : • I 2 3 4 5 6 7 8 9: 0 1 2 3 4 5 6 ' 8 9 1 2 3 " 5 6 7 8 9 0 I 1 2 3 4 5 6 7 8 910 1 2 3 4 ! . OFF 0.1712E"""403 0.1482E ·02 0.1463E 02 0.1170E 02 -0.1217E-00 0.1828E 01 -0.1482E-04 0.1828E 01 0.1828E01 =\ -0.1217E~00 0.1217E..,00 ~ 7 ..., -0.1950E i01 -' 8 =: -0.3010E-!06 V V V V T d 0.1800E-i 03 1 ::I 2 ~ 3 ::I 0.1480E :02' 0.1461E :02 0.1161E ,02 I 4 , 4 0.1547E~01 i r I 1 1 2 0.1960E iOl ~ -0. 1480E-i04 0.1960E :01 0.1960£ iOl 0.1541E..j,01 I 3 ~ 1 4 I I 1 I ~ :1 6 =1 -0. 1541E-;01 7 =; -0.1944E iOl 5 8 =1 -0.3344E-i06 T ~ 0.1890E-:03 i I I 0 V V V V 1 2 3 ~ 4 d I 1 ~ 0.2085E i01 2::1 -0.1419E 1 04 3 = 0.2085E01 4 ~ 0.2085E01 0.1479E i02 0.1458E \02 0.1200E !02 0.8502E-:01 I I I I 012345678 o 1 2 3 I I 5 6 $ 8 ~ 0 L Q .. 8 § Q 3 E~Q 1 2 6 = -0.8502E-01 :i 4 5 6 7 8 9! 0 ~).G>: I 2 3 " 5 6 7 8 910 1 2 3 " 5 6 7 8 9 i0 1 2 3 4 5 6 7 8 9 12345678901214 I , 0123456789 I I 7 = -O.20COE 01 8 = -O.2265E-I0 1 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 9 io 1 2 3 4 5 6 7 8 9 ., 1 2 3 4 5 6 7 8 9 1 2 3 45 6 7 8 9 1 2 3 4 5 6 7 8 9 123456789 1234 I I I 1 SkITC 1 " ON .. ~ I T O.1890E10 3 1 2 V 3 V 4 -O.1043E 101 -O.1251E 101 0.1200E 102 O.8494E~Ol 1 1 0.1604E~04 V V I I I I I 1 I I O.20aSE 101 C.208SE 101 4 O.2085EJ'01 5 0.8491E 01 6 I -0. 8494E~O 1 1 ~ -O.2000E 101 8 0.3001EI06 2 3 1 T V 1 V 2 V 4 V 3 I I I I I I I I J 0.1929E~03 ~ -0.9141E 100 ~' -O.1098E I 2 O.l829E 01 O.1829E 01 4 O.1829E 01 5 -O.1201E 00 6 ~ O.l207E-OO 1 , -0.1950E 01 0.311lE 10 3 81 1 I T V I 0.1929e-!03 O.1482EJ02 O.l463E 02 0.l170E 02 -O.ll05E 100 V V 4 I O.1829E -O.1482El04 3 O.l829E 01 It ] O.1829E pI 5 1 -O.120SE1 00 I I I I I 01 OFF 2 3 I 012345678 1 ~ ~" 'I I SWITC, V 101 O.l170E 02 -0.l201E 00 1 1 J ~ 101 2 1 6 7 4 O.1205E-~O =i -0.1950E :01 2315698~o~Q.]QIQE~G'23456789 12345678901234567890123456789 12345678901234 . ~. 10 1 2 3 4 5 6 7 B 9 I I~·;, 1'-'" I II 1 23456'8~O'O~2000E103"""" V 1 V 2 V 3:4 V 4 I ·1 I I I I I 123456789 123456789 I 2 3 4 1 ~ -O.485SE~02 i I i 1 123456789 O.1481E 1102 0.1461E 02 O. 1164E 0 2 I ! 123456789 1, O.1935E iOI 2 =1 -O.1481E-j04 3 =1 4 =l 5 :d 6::: 7 ~ 8 =9 0.1935E iOI O.1935E \01 -0.4858E-j02 0.4858E~02 -O.1940E 101 -0.3623E-i06 o o 1 2 3 456 7 8 0 1 2 3 4 5 6 7 8 9 0 1 2 3 456 7 8 9:0 1 2 3 4 567 8 9 0 1 2 3 4 5 6 7 8 9 I 2 3 4 5 6 7 890 I 2 3 4 5 6 7 8 9 1 2 3 456 7 8 9 0 1 2 1 4 I ,I i " • () • • If'ttt .. '.: ~V' U. S. ARMY ENGINEER DISTRICT, WALLA WALLA WALLA WALLA t WASHINGTON AN INTEGRATED SPS EARTHWORK SYSTEM FOR THE IBM 1620 A Paper Presented at the IBi-1 1620 Users Group Western Region Conference at Tempe, Arizona 12 December 1963 o Prepared by Cecil L. Ashley ADP Coordinator, NPW November 1963 t. • rid! S"HM"{bHi4¥"j"'-'" i!I'iier--"Uffi8P?T'"'%WtpW"" um'M'w i1 I) CONTENTS Page ABSTRACT • PART I PART II PART III - ...... .. . ......... • • • • • • ii • • • • • • • • • • • • • • • 1 Terrain • • • • • • • • • • • • • • • • • • • Variety of Criteria • • • • • • • • • • • • • 1 APPROACH TO . .. ... . . ..... 2 Acquiring Data ••••••••••••• • • Data Translation and Preparation • • • • • • • 2 2 INTEGRATED PROCESSING 4 GENERAL PROBLEM PROBLm~ . .... ....... Alignment and Realignment • • • • • • • • Template Generation ••• • • • • • • • • Quantity Computation • • • • • • • • • • • Design for Contract • • • • • • • • Special Options • • • • • • • • • • • • • Pay Quantities Computation • • • • • • • • PART IV PART V 4 4 4 4 5 5 ... ..... . .... .. . ... ABSTRACTS OF PROGRAMS . . . . . • • • • • 7 1. 2. 3. 7 8 9 5. 6. 7. 8. 9. 10. 11. 12. 13. o • • • • • • RESULTS 4. FIGURE 1 - • • • • • • 1 HORIZONTAL ALIGNMENT • • • • • • • • • •• GEOMETRIC COMPUTATIONS • • • • • • • • •• EARTHWORK DATA CHECK • • • • • • • • • •• PROFILE GRADE AND TEMPLATE GENERATION •• SLOPE SELECTION - TEMPLATE GENERATION •• EARTHWORK QUANTITY CO~~PUTATIONS •• • •• EARTHWORK LINE SHIFT •••••••••• EARTHWORK TEMPLATE SHIFT • • • • • • • •• EARTHWORK ALIGNMENT OFFSET • • • • • • •• EARTHWORK DATA PLOT REDUCTION • • • • •• CENTERLINE DATA PLOT REDUCTION • • • • •• STATIOn INTERVAL QUANTITY SUMMARY • • • • PLANE COORDINATE CONVERSION PHOTOGRAMMETRY LEAST SQUARES METHOD •• SYSTEI-1 FLOW CHARTS • • • • • • • • • • • • PREPARATION OF X-SECTION CARDS COMPUTATION OF CUT AND FILL QUANTITIES i ·. 6 9 9 10 10 10 11 11 11 11 11 12 .--~----- - ...- - -- --~~~ _...... .. , ' o ABSTRACT The purpose of this paper is to review the group of roadway design programs being utilized and developed by the Uo S. Army Engineer Division, North Pacific, in cooperation with State Highway Departments of Washington, Oregon, a.nd California, U. S. Bureau of Public Roads, Uo S. Forest Service and IBM. Programs discussed include horizontal and vertical control, data check, template design, alignment offset, quantity computation and several related utility programs, all of which are coded in SPS and make up an integrated system which is convenient as well as fast. The specific use of these programs by the Corps of Engineers is in conjunction with relocation of highways, railroads, county roads and municipalities involved in the design and construction of large multi-purpose hydroelectric projects. The "System" of programs has been developed through a combination of initial development, modification of others' programs and adoption of others' programs with no modification. Coordination of programming efforts is accomplished largely through facilities of HEEP. The abstracts of programs at the end of this paper are brief and general in nature. Writeups are complete but not necessarily in compliance with USERS GROUP Standards. Interested parties should address inquiries to: Division Engineer Uo S. Army Engineer Division, North Pacific 210 Custom House Portland 9, Oregon o ii o PART I GENERAL PROBLEM Terrain. The nature of the earthwork problems being encountered by the Walla Walla District may not be entirely unique, but they are certainly extreme at times. The Columbia and Snake Rivers have both cut deeply into Miocene basalt flows that form the regional bedrock. For 45 miles above the John Day Lock and Dam project, the Columbia River flows in a relatively deep valley with precipitous walls rising to a height of 3,000 feet above the general land surface on the north side, and up to 1,000 feet on the south side. The canyon walls consist of a series of stepped basalt cliffs interposed with steep talus slopes and terminating at the top in relatively flat plateau areas. Because of the constricted nature of the canyon, there is very little choice for , the alignments; however, where the alignment is benched into the basalt edges, a minor shift in alignment can mean a major difference in cost because of the rock excavation and high fills which may occur. Because of the type and shape of the terrain, alignment is quite critical and many trial alignments may be required in order to arrive at what is considered an optimum one. Variety of Criteria. Another feature that requires considerable attention is the number of agencies to be dealt with. In one project area there may be up to three State Highway Commissions, four railroads, and a dozen counties involved; all of which have from minor to major differences in criteria. This condition is emphasized when two railroads and a highway are aligned adjacently and there is hardly room for one of them in the canyon. Something has to move up the hill. The John Day Lock and Dam presents a good example of the magnitude of the work being done in the North Pacific Division. This will be a concrete gravity structure with an ultimate installation of 2,700,000 Kl1 produced by twenty 135 ,ago KWT units driven by Kaplan turbines measuring 305 inches throat'diameter. The total excavation for the project will be approximately 80,000,000 cubic yards, half of which will be hard rock. This total excavation figure is exceeded only by Owyhee, Fort Randall, Fort Peck, and Garrison project; all of which are earth or rockfill dams. As far as we know, the 40,000,000 cubic yards of hard rock excavation will comprise a world record. c The relocation at this project involves 80 miles of SP&S Railroad, 57 miles of UP Railroad, 40 miles of Washington State Highway, 32 miles of Oregon Interstate Highway, and the town of Arlington with a population of 633, 1960 census. When the alignments involved are required to be revised up to 8 or 10 times in order to reach a compromise with economics .and the agencies concerned, the workload approaches staggering proportions. 1 .. PART II APPROACH TO PROBLEM Acquiring Data. In order to plan and design such a project and its accompanying relocation work, surface data must be translated into punched cards representing ground cross sections along the alignment centerline. First. of course, an approximate alignment must be determined On paper. The horizontal alignment and geometric computation programs are utilized at this stage for horizontal control. Survey parties then either "flag" that alignment in the field, or establish control points. Helicopters are utilized to transport the party members to inaccessible locations. Aerial photographers are then hired to "fly" the line and the resulting film is converted to glass plates. Sufficient area is covered to enable the designer to shift the alignment or position several roadway alignments such as railroad mainline and shoofly. and highway and detour. Data Translation and Preparation. With the pictures taken and plates made, the next problem is to translate these to punched cards which are input media to the computer. This is accomplished by one of two ways, with emphasis on minimum manual work. For planning and design stages, if contour maps are available, alignments may be laid out on them and cross section cards punched by means of the Benson-Lehner Digital Scale. If contour maps are unavailable. or if for other reasons. it is desirable to do so. the cross section cards are prepared on a WILD A-7 Autograph Stereo Plotter connected to an IBM 026 Key Punch through a WILD EK-5. A hard copy is prepared concurrently on an electric typewriter. An optional feature of this system permits printing out on typewriter the X, Y. and Z coordinates at points selected. For initial phases. cross sections are taken covering sufficient area to allow for alignment shifts and multiple alignments. Once the cross section cards are punched. they are checked for "detectable bulls"by means of a computer pass using the "Data Check" program which senses such errors as overhangs, double or no centerline. excessive horizontal or vertical difference between rod readings, etc. as may be logically determined. At this stage, cross sections can be plotted by the Benson-Lehner Electroplotter. To many designers, cross sections are considered obsolete but the terrain and materials with which we have to contend make cross sections desirable and often necessary. Another means by which segments of cross sections are sometimes prepared is the Benson-Lehner Oscar Trace Record Reader. This machine 2 I"~ I i ' c uses fathometer graphs as input and produces punched cards in the same format produced by the Digital Scale or WILD A-7. This method is useful where the toe of proposed fill slopes extend below present water surfaces and in navigation channel and powerhouse excavation applications. A program is currently being developed by the Walla Walla District which will update cross section data cards by extending on either side. by inserting additional readings or replacing corrected readings. c 3 - - - . - -••- ...- - - - - -.••. ,... -- ········ .... ".~_.,c,c=_ ..... __ PART III INTEGRATED PROCESSING Alignment and Realignment. With the cross section cards prepared and checked, the question arises, "Has the alignment been revised during the process of preparing the data cards?" If so, the new alignment computation data is combined with the original or previous alignment data and processed through the Earthwork Offset Alignment Reference program which computes the offset and skew angle between the two alignments and new stationing. If the skew angle is not too great and the original cross section cards cover sufficient distance, the output cards of this computer pass serve as header cards which are combined with the cross section cards and processed through the Earthwork Alignment Shift program to compute new cross section cards referenced to the new alignment centerline. Template Generation. Roadway templates to be used for computing cut and fill quantities can be prepared on the computer at this stage through what is at present a two-pass process. The first pass establishes profile grade and the coordinates of the basic roadway points reflecting superelevation for horizontal curves. This basic roadway consists of two to four planes defined by from three to five points. The second pass completes the templates by adding ditches where required, and slope readings. This pass also serves as an additional data check by indicating slopes that will not catch, etc. The complexity of road design involving berms on fill slopes for embankment protection, and benches on cut slopes for rock fall protection or due to material classification change, has prompted a recent modification to this program which enables the designer to specify these features on the terminal slopes of the template. Templates can be produced with either slope readings or catch point coordinates at extremities. Quantity Computation. For computing quantities, the cross section and template cards are merged and processed through the Earthwork Quantity Computation program. The output of the pass provides cut and fill areas. accumulated cut and fill quantities and mass ordinate. This program also has optional features for dredging and levee applications. in which case. directed slopes and outside catch points are used. It will also make quantity adjustments for curvature correction. Design for Contract. Listings of this output are provided the designer who then determines if and where alignment needs to be revised vertically or horizontally. This is repeated usually for various segments of the line, then when the alignment is satisfactory, the complete line is recomputed for quantities. 4 I i o Special Options. If two alignments are adjacent and the embankments sometime overlap, the Stage Development program is used to combine the cross section and template cards to produce new cross section cards which can then be line-shifted and used to compute berms for riprap which is to be computed separately from basic quantities. The answer cards from the quantity run can be processed through the Station Interval Quantity summary to produce listings of quantity summations at selected intervals. Cross section and template cards are processed through the Earthwork Data Plot Reduction program to replace template slope readings with catch point coordinates and reference all vertical coordinates to a common elevation so the cross sections and templates can be plotted. Pay Quantities Computation. During and after construction when pay quantities are to be computed, the line is re-flown and "asbuilt" templates prepared. These cards differ from the previously mentioned templates in that they do not have slope readings but extend from catch point to catch point. These templates are merged with cross section cards of the final alignment and pay quantities are computed through the Earthwork Quantity Computation program. The "as-built" template cards are point-plotted over the "asdesigned" cross sections and used to check for over- or under-built conditions. ',,·, 0. " 5 PART IV RESULTS Opinions of the users of these programs are varied. so in order to avoid giving an overly optimistic view, this example was taken from the most "conservative" user. Sixty miles of railroad alignment on the Little Goose Lock and Dam project was processed through eight aligrunents before arriving at the final choice. By thoroughly analyzing the first seven alignments. five million yards of rock excavation were saved between the seventh and eighth alignments. If hand methods had been used, there would not have been time for more than four "shotgun" alignments and the savings resulting from computer utilization would not have been possible. The cost involved in design stages may be equal to or even greater than that incurred by "shotgun" manual methods, but this includes increased analyzation of alignments, making possible savings such as cited in the example. Even greater savings are being realized with a computer on-site in the Walla Walla District so that even more alignment trials can be made in certain areas where the alignment is most problematical. One installation has combined template design and quantity computation programs enabling them to obtain quantities in one pass without producing intermediate templates. This version also incorporates line shifting as an integral part of the single pass. It requires a different format of data cards than that being used by our District and 60 K. The type of processing usually encountered in the Walla Walla District require some insertion of hand-prepared templates before the final pass, so we have not seriously considered their degree of combination. The two passes of template design might be combined to advantage, providing header card storage did not become too restricted. 6 J31 o PART V ABSTRACTS OF PROGRAMS 1. HORIZONTAL ALIGNMENT This program computes horizontal alignment from the following basic input alignment data: a. Beginning station and coordinates (RP) b. Coordinates of PI's c. Degree or radius of curves d. Spiral lengths (highways) or chord lengths and number (UPRR) Output information includes: a. Bearing and length of course between PI's b. Stations and coordinates of all curve points c. Deflection angle d. Degree. length and radius of circular curves e. Deflection angle (DE). x and y. u and v of spirals (~) and semi-tangents Curve definition can be by arc or chord. Distances are rounded to nearest .01 foot. coordinates are carried to nearest .001 foot. "I '" 7 2. GEOMETRIC COMPUTATION The present geometric program combines the features of several related programs which were used on the IBM 650. The following types of computations can be performed singly or in combination in one machine pass. a. Survey traverse using bearings with azimuth or deflection b. Compass or transit rule traverse adjustment. I c. Areas bounded by traverses which may contain circular f angles. t segments. i d. In a traverse between known points, a combination of two unknowns (bearings, 2 distances or a distance and a bearing) may be solved for and stored for recall in subsequent problems in same pass with interdependency of geometric figures. I e. Location of points on tangents, circular curves, spirals or offset spirals and intersections of any two of these. f. The intersection of tangents or circular curves with spirals or offset spirals. g. to another. Convert selected coordinates from anyone plane system The ability to store data pertaining to any course for use in problems which follow makes it possible for the engineer to begin with a minimum of data and to solve complex office and field layout problems in one pass on the computer. Problems are arranged in logical sequence to develop intermediate answers as would be done by "people-computers" following traditional methods. A closed traverse can be adjusted in one problem, for instance, and an adjusted loop can be run from it to close on any point for which coordinates are known or have been computed. Two alignments can be intersected and the bearing and distance to land corners or other references readily determined. The area of gusset plates can be determined and a host of other geometric problems can be solved. As a matter of fact, some users say"it can do anything~" o 8 I .'" " • c 3. EARTHWORK DATA CHECK This program checks earthwork input data in format used by U. S. Army Engineer Division, North Pacific, and its various Districts. Errors which would result in erroneous results or machine halts are detected and error messages typed or punched identifying the station and specifying the type of error detected. 4. PROFILE GRADE AND TEMPLATE GENERATION This program converted from Washington State Highway Department's Five-Point Profile Grade Program. produces the following results: a. Generates basic roadbed templates for input to Slope Selection - Template Generation Program. b. Produces pavement elevation cards supplying elevation of five points at each station to .01 foot accuracy. c. Computes profile grade at selected or incremental stations. d. Reproduces templates inserting new profile grade elevation in the reference elevation field. e. Punches· a summary of vertical control data at end of each run. 5. SLOPE SELECTION - TEMPLATE GENERATION This program requires ground line cross section cards plus the template cards produced by the Profile Grade Program as basic data. Other input consists of slope specification cards and bench-berm data cards. Three fill slopes and four cut slopes may be specified for appropriate depths of fill or cut. Ditches can be specified with V or flat bottom with depth and slope optional. Terminal slopes may include bench (cut) or berm (fill). The templates produced may include catch points. or the terminal rod readings may be replaced with slope readings in the form xxx.xx horizontal to 1 foot vertical. These templates when collated with ground line cross section cards are the input to the Earthwork Quantity Computation Program. o 9 • 6. EARTHWORK QUANTITY COMPUTATIONS a. The purpose of this program is to compute quantities of earthwork to be excavated and placed on a given Job. The program is patterned after the I~~ 650 Cut and Fill Program (H-84l) as modified by Bureau of Public Roads, Vancouver, Washington. The computations are performed in the conventional manner by taking the original ground topography and the design of the completed section, as specified by the engineer, and computing the cut and fill end areas. These are then used to compute the volumes between sections by the Average-EndArea method. This program will apply a volume correction due to curvature if desired. Swelling or compaction can be applied to either the cut or the fill and these quantities are accumulated throughout a project to produce a mass-ordinate. The total accumulated quantities of unadjusted cut and fill are also available for each section. '- b. As a by-product of these computations, some design information is available at each section to aid the engineer in his appraisal of the design that has just been computed. These include the cut or fill at the pivot point, at the toe of both the right and left slope, and at centerline. Also, the rod and distance to the catch point of the template slope and the original ground line, if these were not part of the input data. 7. EARTHWORK LINE SHIFT, The purpose of this program is to shift topog card data (type "0") to a new or offset centerline. This enables the engineer the choice of running multiple trial alignments while using the original topog data but shifting the topog data to coincide with the alignment. The program can also be used for station adjustment and for the reverse line shifting of topog data. It is desirable in some phases of earthwork to shift the topog data from the original line to an offset line, generate templates at the offset line, stage the ground line, and then shift the staged ground back to the original line. By proper program control in the Job Control Card, the original line shift headers (used to shift original topog to an offset line) and the staged ground at the offset line can be used to shift the staged ground back to the original line. 8. EARTHWORK TEMPLATE SHIFT This program is designed to take the Type "1" templates that were prepared or generated at an 'offset centerline and punch new templates having original centerline stationing and shot distances referenced to the original centerline. The new Type ttl" templates can then be collated with the original terrain cards (Type "0") and a plotter deck is obtained which has the original centerline as a base for plotting of both terrain ~ 10 Ii, I;, I: c data and Type "1" templates. Use of this program will facilitate plotting operations when more than one roadway section is planned and is to be plotted on the same cross section of terrain. 9. EARTHWORK ALIGHMENT OFFSET This program is designed to compute offset distances. skevT angles, and centerline station coordinates of a base line and an offset line. Output also includes line shift headers which enables the Earthwork Line Shift Program to produce a new set of ground line cards referenced to the offset line. 10. EARTH\'/ORK DATA PLOT REDUCTION This program reproduces ground line and template cards adding plotter signs where necessary and references both type of cards to a common reference elevation. It will also provide demand origin offset codes and insert blank cards where necessary. 11. CENTEHLINE DA'rA PLOT REDUCTION This program compiles station numbers and centerline elevations of ground line cross sections or template cards in a format for plotting of centerline profile. 12. STATION IUTERVAL QUAnTITY Sill1l·1ARY The purpose of this program is to produce a summary of quantities of earthwork to be excavated and/or placed on a given job. Two types of summary options are available: Station or Interval. The listings are used by the engineer for design analyses and as tabulated quantity listings for plan and profile drawings for design memorandums and/or contract drawings. 13. PLANE COORDInATE CONVERSION - PHOTOGRAMMETRY LEAST SQUARES r-1ETHOD The purpose of this program is primarily to convert the machine coordinates of a universal stereo-plotter to a local coordinate system. applying the least squares method of conversion (Helmert). This conversion is mathematically expressed as follows: o x =p + k X cos a - k Y sin a y =q + k Y cos a + k X sin a The quantities p and q are ground coordinates of the origin of -the machine coordinate system, a is the angle of rotation between the two systems, and k is the distance scale factor. 11 PREPARATION OF X-SECTION CARDS PUNCH X-SECTION CARDS VIA WILD A-7 AUTOGRAPH STEREO PLOTTER COMPUTATION OF CUT AND FILL QUANTITIES A PROCESS THROUGH PROFILE GRADE PROGRAM TO COMPUTE BASIC ROADBED PROCESS THROUGH ALIGNMENT SHIFT TO PRODUCE NEW X-SECTION CARDS PROCESS THROUGH ......._ _ _ _ _-L:SLOPE SELECTION OGRAM TO COMPLETE TEMPLATES MERGE X-SECTION & TEMPLATE CARDS-INSERT HAND PREPARED TEMPlATES IF ANY o REVISE HORIZONTAL AND/OR VERTICAL !....io----t--=;:~T~7__j.-...J ALIGNMENT 12 ;;37 FIG. 1 c HYDRO SYSTEM DAILY OPERATION ANALYSIS PROGRAM CHARLES R. HEBBLE U.S. ARMY ENGINEER DISTRICT, WALLA WALLA WALLA WALLA, WASHINGTON c • ( I '" I.~ ' I \ I c HYDRO SYSTEM DAILY OPERATION ANALYSIS PROGRAM by Charles R. Hebble 11 INTRODUCTION The computer program described in this paper, along with certain fixed input data, constitutes a detailed mathematical model of a system of hydroelectric projects. It is a general program applicable to any river system .3.nd scheme of development. Natural lake and channel storage may be synthesized in addition to reservoir storage and backwater effects. The program is capable of accurately simulating the hour-by-hour power loading and water regulation of a hydro system. Through use of the program it is possible to determine the operating characteristics of planned future projects for various-sized power installations. Backwater encroachment on upstream reservoirs, pondage requirements, and the effects of peaking discharges on downstream river stages and reservoirs may be evaluated. The program may also be used to compare alternative distributions of system load among an existing group of hydraulically and electrically integrated projects. It is thus possible, through a trial and error approach, to determine optimum load distribution. The pxngram has been used in investigating existing system development, along with future planned developments of the lower Columbi~ and lower Snake Rivers. Here a series of run-of-river projects, below Grand Coulee Dam, will develop almost all of the available head (see Figures 1 and 2). It is anticipated the program will be used to analyse future operating problems in regards to power loading and water regulation. DESCRIPTION OF EQUIPMENT Two versions of the program exist: one for an IBM 1620 having a 40,000 digit memory with an" IBM 1622 Card READ-PUNCH for input-output; the other for an IBM 1920 System (combined IBM 1620/1401) having a 60,000 digit memory in the IEM 1620 and 4,000 digits in the IBM 1401. This latter system has both an IBM 1402 Card READ-PUNCH and an IBM 1403 on-line printer for input-output. Both systems have the following optional features required for program execution: Transfer Numeric Strip, Transfer Numeric Fill, Move Flag, and Indirect Addressing. o 1/ Civil Engineer, Automatic Data Processing Section U. S. Army Engineer District, Walla Walla Peripheral equipment required is as follows: card keypunch, card verifier, and card sorter. In addition, for the system with the IBM 1622, a card tabulating machine with a special wired panel is required. ANALYSIS AND METHOD OF SOLUTION General. - The program is intended for the detailed hour-by-hour analysis of hydro system operations; basic input to the program and output from the program is hourly. However, periods greater or less than one hour may be used. There is no limit to the number of realtime operating intervals which can be simulated by the computer. The program continues to run as long as there are input data in-the READ hopper. Ordinarily, however, real-time periods of more than a week are not analyzed because of the excessive length of the computer runs for all but the smallest systems. As a rough estimate, the program requires two minutes to compute one day's output for each hydro project in the system. The number of reaches and projects that can be analyzed in a given system is dependent on the computer storage capacity. The program instructions require approximately 25,000 digits of memory. A rule-of-thumb estimate for a given hydro system, including open river reaches, is 1,000 digits of memory for each hydro station. Variation of memory requirement is due to optional lengths of tables used to define system parameters. Program Operating Modes. - Discharges from the various hydro projects are specified, either directly or indirectly as input to the program for each real-time period. Four operating modes exist for this purpose: (1) station power loading given; (2) system power load plus generation allocation (breakpoint) settings given; (3) total project discharge given; and (4) project forebay elevations specified. Any of the foregoing, in combination with other fixed and variable data, determines the individual project discharges. Different projects may be operated under different operating modes at the same time, and the operating mode of any project may be changed at any time. Variable Input. - Each hydro station obtains operating data for each real-time period from coded variable input cards. Variable input data include project power loads (or discharges), system loads with project breakpoint settings for load-frequency control, local inflow, optional spill, and number of generating units synchronized on the line and available. These data may be varied as desired during the course of a run. ~ration Sequence. - The program begins with the upstream project of the series and proceeds downstream, routing flows through each open channel reach or reservoir. Local inflow between projects is added to the routed flows. The outflow from the projects is computed as the sum 2 c of power discharge, spill, and average fixed release. (The average fixed release, which must remain constant throughout the run, includes losses due to lockage, leakage, and useage.) The entire process is repeated for successive IS-minute computation (routing) periods. Project and desired reach answers are obtained upon completion of the fourth routing period. The variable data for the next hour is then read in and the entire process repeated. Routing and output intervals may be varied by a special job definition routine. The routing interval may be different for different projects or river reaches but must be a multiple of the basic routing interval. The computer program has two fundamental parts. One is the method of simulating a hydro-power station, the other is the flow routing procedure. Each part will be considered separately and in some detail. Hydro-Power Station Simulation. - Each generating station synthesized in the program is represented by individual unit characteristics as illustrated in Figure 3. This is in contrast to the method used in other power programs for coarser increments of time which consider only generating station characteristics as a whole. Such programs assume some operating efficiency between best efficiency and, say, full-gate efficiency. This assumption, while adequate for studies having a basic time period of day, week, or month, during which there are many swings in 'station power loadings, are inadequate for the detailed hourly computations required in the study of peaking operations. The number of generating units synchronized on the line may be specified as variable input or may be automatically selected by the program. The program has the ability to automatically add units as required to meet the load up to the maximum number available. When a power load in excess of the maximum capability of the installation is specified, generation will automatically be adjusted to equal plant capability as limited by head. The output is coded to indicate this alteration by the computer program. Under automatic selection of units the program computes the number of units for best efficiency operation. The total station load (or discharge) is divided by the unit best-efficiency loading (or discharge) for the particular head existing on the station to determine the desired number of units. Fractional numbers of units are truncated; thus unit loadings are at or higher than the best efficiency point. The number of units selected is limited to the maximum available. The program automatically causes all excess water above maximum pool elevation to be spilled. This is termed "mandatory" spill. Power discharge, fixed release, optional spill and mandatory spill are summed to arrive at the total project discharge. The alarm section of the output is coded calling attention to such mandatory spill. o Should a project attempt to draft below its m1n~um pool elevation as a result of releases exceeding inflow; the power discharge and hence, 3 generation is automatically reduced to prevent the overdraft, irrespective of the power demand or discharge called for by the input data. Here again the alarm section of the output is coded. The computation of power, discharge, head, etc. are accomplished by means of table look-up and interpolation. For existing projects, unit performance tables may be prepared from observed data. In the case of planned projects, the basic data must first be computed. A separate computer program has been developed which can quickly compute performance characteristics. The program is based on the turbine performance characteristics of a unit machine, i.e., a machine runner diameter of-one foot operating under a net head of one foot. These data are summarized in a unit performance hill in which power output is plotted against the peripheral speed coefficient. It is thus possible to quickly evaluate the performance of different sized units and units having different characteristics. Centra lized Load- Frequency Control. - A feature of the program is a provision for simulating centralized control of the station power loadings in a manner similar to that of centralized load-frequency control equipment. Such equipment is presently installed at the major Federal hydro stations on the Columbia River. The power loadings of these projects are centrally controlled from the system load dispatcher's office in Portland, Oregon. When the load-frequency control feature of the program is used, participating settings for each of the hydro stations are then included as input data and a single system load rather than individual station loadings are specified. The system load is apportioned among the hydro stations in accordance with station participation values. Other than this, the program operates in the same manner as when individual loadings are prescribed for the several hydro stations. This centralized, automatic dispatching of power loadings affords the opportunity for efficient coordinated system operation of the Federal projects. Use of the load-frequency control feature of the computer program will allow planning· studies and scheduling of power operations in conjunction with the equipment, in addition to it~ other uses. It is envisioned that ultimately the load dispatching of these hydro stations will be directly and automatically controlled by computer. The present program could well serve as a basis for such a future program. Flow Routing. - The total discharge from a project (power discharge, spill, and fixed release) is routed downstream by a method of flow routing known as incremental storage routing. This approach consists of subdividing reservoirs or river channels into incremental reaches. Each incremental reach is represented in the program by two tables: one giving 4 o o the relationship between storage and elevation for the particular reach; the other giving the relationship between discharge and elevat.ion. By proper selection of routing interval and choice of the number of reaches, actual hydraulic conditions may be accurately approximated. The discharge from any reach is computed as a function of the reach elevation, or where backwater effects must be considered, as a function of the next downstream reach as well. The change in storage content of the incremental reservoirs is related to the difference between inflow and outflow. The combined effect of the storage-elevation and discharge-elevation relationships mentioned in the preceding paragraph causes a time constant to be introduced into the routing. This is known a'S "time of storage" and is defined as the change in storage per unit change in discharge. Ts ~ S/ ~ Z Il Q/ Il Z where, Ts = time of storage S storage volume Q = discharge rate Z = elevation Dimensionally, using the units adopted for program use, Ts (min) = ~ ~ S (kc fs-min)/ ~ Z (feet) Q(kcfs)/ ~ Z(feet) The routing procedure used in the program does not explicitly consider time of storage; however, the routing interval used in the program must be shorter than the minimum time of storage for any given reach. Failure to observe this criterion will result in oscillations of the reach elevation and discharge. These oscillations tend to increase in amplitude until the range of tables is exceeded. A reservoir which is encroached upon by backwater from a downstream reservoir may have an extremely short time of storage. The time-of-storage concept may be visualized by assuming a constant rate of increase in inflow to an incremental reservoir. Outflow from the reservoir will eventually reach an equilibrium condition where it is increasing at the same rate as the inflow; at this time, the outflow hydrograph will be displaced from the inflow hydrograph by the time of storage. Figure 4 illustrates a discharge hydrograph routed through six identical incremental reservoirs. Storage-elevation and dischargeelevation curves are shown on this same sheet. c 5 Overlapping reservoirs are normally subdivided into two reaches: a forebay reach and a tailwater reach for the next upstream project. The sum of the storages of the individual reaches must equal the actual reservoir storage. In subdividing open channel reaches into incremental reaches, no definite rules can be given for the number of increments. In general, the greater the number of reaches, the greater the translation of the discharge hydrograph with minimum attenuation. Discharge hydrographs can be modified, to a lesser degree, by choice of routing interval. Here again, selection of routing interval beyond that required to prevent oscillation, is subject to rules established for particular problem definition. For river systems where all tailwater reaches have extremely short times of storages, an alternate routing method is available. This method permits use of relatively long routing intervals, up to the answer output .interval if desired. In effect it considers the time of storage to be zero. Tailwater elevation is first determined as a function of inflow (and next downstream reach elevation if necessary); the storage corresponding to the new tailwater elevation is then computed as is the change in storage during the routing interval; this change in storage is converted to an average rate of flow during the routing interval and added (or subtracted) to tailwater inflow to arrive at the outflow. The algebraic sum added to the inflow is dependent on change of tailwater elevation during the routing period, i.e., a rising tailwater will cause outflow to be less than inflow with the opposite occurring on a declining tailwater condition. The flow and storage-routing procedure used in the program has a sound theoretical basis which permits analysis of future reservoirs as well as existing reservoirs and river reaches. The computation of discharge, elevation, storage content, etc. are accomplished by means of table look-up and interpolation. Volume-elevation relationships for open river reaches and reservoirs are generally obtained from survey data. Water surface profiles are computed for various forebay elevations and rates of flow by independent means. (Here again separate computer programs are available for determining water surface profiles, and volume relationship, in reservoirs or open channel reaches.) From these profiles, the discbarge-elevation-volume relationships are derived. PROGRAM OUTPUT Output Data. - Program output consists of the following data for each project: Date and time project name Abb~eviated 6 c ,\O"i"ihiw' . St··· . "tiff -..¥j,iiNIIij'i"'wbcwJ["lW·\·l·\!!!·- .. JIIIIIl ". i.' End-of-period forebay elevation End-of-period tai1water elevation Average head Forebay storage content Storage change during period Local inflow Routed inflow Power discharge Spill Total discharge Generation Nu~ber of units operating Alarm codes Operating mode number Project number O~tput may be obtained for selected reaches, in addition to project data, at users option. Reach output consists of the following data: Date and time Abbreviated reach name End-of-period elevation End-of-period total discharge Reach number The 1920 System program output consists of both on-line printing and punched cards. The same data are given by both forms of output. Choice of output mode is optional with user. The 1620 System output is punched cards. The 1920 System uses an auxiliary listing program to list the punched cards for each project or optional reach and system Summary Cards. Output data consists of ~roject Cards, selected Reach Cards and Summary Cards containing end-of-period system generation totals and system potential energy remaining in storage. The sequence of the punched card output and on-line printout is as follows: An initial SUImnary Card gives potential energy in the system based on starting profile data. This is followed by data for each hydro project and for optional selected reaches for the initial time period. A Smnnary Card concludes the data for the initial time period. Project, Reach and Summary data are then punched out for the next subsequent output interval. Output intervals may be varied by spec:ia1 control. In the absence of special control, the output interval is one hour. Sample output formats of a typical study are shown on Figure 5 for the IBM 1620 System and on Figure 6 for the IBM 1920 System. 7 ACKNOWLEDGEMENTS The original program was developed under the direction of Mr. C. E. Hildebrand, Water Control Branch, U. S. Army Engineer Division, North Pacific, Portland, Oregon. Michael A. C. Mann, Consulting Engineer, assisted in the analysis and programmed the basic program and supplementary routines for the IBM 650. The author wishes to acknowledge the work of Mr. Lyle A. Dunstan of H. Zinder and Associates, who programmed the present version. This work was financed by the U. S. Bureau of Reclamation and the program used by them in their planning studies. Additional logic, and program refinement was accomplished as a joint effort by U. S. Army Engineer Division, North Pacific and Walla Walla District, Corps of Engineers. The assistance of C. E. Hildebrand and Robert D. Moffitt of the U. S. Army Engineer Division, North Pacific in accomplishing the latte~ and in the preparation of this paper is great fully acknowledged. Requests for program details should be directed to: Division Engineer, U. S. Army Engineer Division, North Pacific, 210 Custom House, Portland, Oregon, 97209. 8 e o • ,. CANADA UNITED STATES I I O~ %4 4t- oZ -0 Is I ROCK ISLAND J LEGEND ) Existing projects ~ Authorized projects -... SCALE IN MILES 25 0 25 50 ' EXISTING AND FUTURE PROJECTS OF THE LOWER COLUMBIA AND SNAKE RIVERS Fig. 1 ELEV ATION I N FEET ABOVE M. S. L. 8 I OUTER END OF JETTY BONNEVILLE EL. 74 < m THE DALLES _ _III EL. 160 ;cI ;a ~_.JOHN rm .en ;cI r- < m ~ m ~ 0 '" n 0r- ;cI rm en ~ ~ 0 en :J: C < m »'" ~ » z c0 0 -t Z »?\ m '" < m '" en 1 • DAY ~ EL. 265 0. " CHINA GARDENS I EL. 910 =~ ASOTIN EL. 842.5 _ _LOWER GRANIT EL. 735 ':.-:ds I ;cI - 8 \ 100 1..;1 LITTLE GOOSE '1" EL. 638 i ~....... .AII-. LOWER MONUMENTAL EL. 540 ~_. ICE HARBOR EL. 440 "':., BEN FRANKLI N I EL. 385 I PRIEST RAPIDS EL. 486.5 ~-..WANAPUM EL. 570 ~_ ROCK ISLAND EL. 606.5 ~~ ROCKY REACH EL. 707 ~-=T WELLS : EL. 775 550 CHIEF JOSEPH EL. 946 ----1 GRAND COULEE . . 600 . . . 00 8 -8 o o ELEVATION IN FEET ABOVE M. S. L. i.: 0, :", i 25~------~------~--------~---------------- 20 ....'"u FULL-GATE LIMIT ~ = f(H)~ ... w C> 15 0::: < ::I: U V) 0 w Z 10 c:c ~ ~ E :::> to- BEST EFF. LOADING 5 71 ~ w ~I 0 10 20 30 40 50 60 GENERA TOR OUTPUT, megawatts o GENERATING UNIT CHARACTERISTICS F!.a. 3 INFLOW 150 III ~ U ~ ... w ~ 100 ~ :r: u V) 0 50 o o 36 24 12 TIME, hours DISCHARGE , K cfs 10_O_ _ 410 0r---_ _ _-'T Q) Q) ~ w 200· 300 400 400 ~ .. ~ ~ u V) :r: 390 0 380 ~~~_~ o 10 _ _ _---L_ _ _ _- ' -_ _ _........ 20 30 STORAGE, M cfs - min, INCREMENTAL RESERVOIR ROUTING EXAMPLE WITH STORAGE ELEVATION AND DISCHARGE - ELEVATION CURVES fig. 4 II • e o ~ ~~ ~ ~.;::: rn ~~ :~ '~ ~ DATE-TIME il ~ ~ oil ~ ~~ ~ "ii ~ ~ -~ ~ 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 ·15 15 1 1 1 1 1 1 i i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 85 85 85 85 85 85 85 85 85 85 85 85 85 85 85 85 85 85 85 85 85 8S 85 85 100 200 300 .. 00 500 600 700 800 900 1000 1100 120C 1300 1400 1500 1600 1700 1800 1900 2000 2100 22CO 23CO 2400 PROJ GCl GCl GCl GCl Gel GCl Gel GCl Gel. GCl Gel GCl Gel Gel Gel Gel Gel Gel GCl Gel Gel Gel Gel Gel FORE8AY TAllWATER FF.ET FEET 1262.08 1262.16 1262.24 1262.33 1262.40 1262.44 1262.41 1262.32 1262.20 1262.09 1261.97 1261.86 1261.76 1261.68 1261.61 1261.53 1261.41 1261.29 1261.18 1261.07 1260.99 1260.96 1260.97 1261.02 945.41 944.47 944.05 943.89 944.48 945.99 950.20 953.79 955.99 957.01 957.57 304.1 304.7 305.4 305.7 305.1 304.1 303.5 303.3 303.7 305.1 307.9 311.1 04.0~2 05.485 06.991 08.497 09.765 10.496 09.903 08.301 06.310 04.295 02.268 00.233 198.556 197.192 195.930 194.470 192.41+8 190.1+13 188.371 186.491 185.117 184.555 18 ... 797· 185.524 7415.34 22866.42 , 315.3 317.1 317.9 318.2 318.0 317.0 314.4 310.6 307.4 305.7 304.7 957.89 301+.2 957.38 956.50 955.83 955.84 956.91 957.52 957.85 957.72 956.69 954.43 951.15 947 ••6 30281.97 ,Ii AVG FORE8AY STOR HEAD SlOR CHANGE 760.474 1.313 1,"23 1.506 1.506 1.268 731 594- lOCAL INFLOW ROUTED INFLOW '7~7'O 1.9912.0152.0262.0351.6771.3641.2621.460"; 2.0222.0342.042·1.8801.37..-· 562242 728 '6.700 36.650 36.650 36.620 36.650 36.620 ·'6.650 36.650 36.620 36,650 36.620 36.650 36.650 36.620 36.650 36.620 36.650 36.650 36.620 S6.650 36.620 36.650 36.650 17.223- 880.510 1.602- POWER DISCH SPIll DISCH TOTAL DISCH 5.740 2.050 GEM MO. MW UNITS C 6.240 2.550 500 500 6.190 19.100 50.870 75.090 84,430 84.990 85.280 85.470 76.880 69.390 66.920 71.690 85.150 85,470 85.660 81.740 69.630 50.110 30.840 19.190 ,.690 1•• 600 50.370 74.590 .'.930 .4.490 84.780 84.970 76~3'0 68.890 66 ... 20 71.190 '''.650 .... 970 85.160 81.240 69.130 49.610 30.340 18.690 133 45 133 439 1.197 1.749 1.94~ 1.944 1.944 1.944 1,749 1.580 1.527 1.638 1.944 1.944 1,944 1.853 1.~80 1.139 704 439 2.0 2 1.0 2 2 2 2.0 2 4.0 2 13.0 2 20.0 2 22.7 2 23.0 2 23.0 2 23.0 2 21.0 2 1e.5 2 le.o 2 19.0 2 23.0 2 23.0 2 23.0 2 22.0 2 19.0 2 13.0 Z 8.0 2 5.0 2 29,513 1.281,880 1,293.880 346.2 '11 . ;! ~ • va ........ ~, l>-I~' ,"'" ... I , : ,-,,' ~- ,.;. 11 ~i < ".' t~~;~~ -'<-~ . ~ .. ~ ~• f , DR. JOHN MANfOn:s COMPUTER TECHNOLOGY DEPT. PURDUE UNIVERSITY CALUMET CAMPUS HYDRO SYSTIM DAILY OPERATION ANALYSIS U. S. ARMY ENGINEER DIVISION, NOIT" 'AClfIC C.-l UN1T.S IFI . HAMMOND. IN 46323 PAGE <' . - . 4 • ... GRAND COULEE ELtV. ~ND PERIOD FOttEISA., TAIL"'" FEET FElT DATE AND TIMt. 1~ 1~ 15 1~ l~ l' I) 15 lS 15 1, 15 1~ 1~ 1~ 1") 1, IS 1) 1, 1~ 15 I') 1~ 01 01 01 01 01 01 Ol 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 01 85 85 85 85 85 85 85 85 85 85 8S 85 85 85 85 85 85 85 85 85 85 85 85 85 0100 1262.08 0200 12e»l.16 0300 1262.21t 0400 1262.33 0,00 12b2.40 0600 12bl.44 0100 12u2.41 0800 1262 •.Jl 0900 12~1.20 1000 - 12~2.09 1100 1261.91 1200 1261.86 1300 12b1.1b 1400 1261.68 1500 1261.61 1600 126'1.53 1100 1261.41 1800 1261.29 1900 12bl.18 2000 1261.01 2100 1260.99 2200 12b(,.96 2300 12bO.91 2400 12bl.02 945.41 91t4.ltl 941t.()' 91t3.89 91t4.48 94'.99 950.20 9.,3.19 955.99 951.01 951.51 951.89 951.38 9S6.!iO 955.83 955.81t 956.91 951.~2 951.85 951.12 956.b9 954.43 951.15 941.86 AVG. HEAD FEEl FOREeAY STORAGE CFS-DYS DELT' STORAGE CFS-DYS 315.30 311.19 204.062 1,311 205,lt85 1.1t23 311.91t 206,991 318.21 208.1t91 1.506 1.506 1,268 131 5941,6021,9912.0152.0262,035- 318.02 209,165 311.05 210.496 314.43 209,903 310.61 208,301 J01.45 206,310 30S.13 .201t,295 304.19 202,268 304.21 200,2)') JOIt.13 198.556 301t.l1 191,192 J05.43 195,9)0 305.14 194.410 305.18 192,448 301t.19 190,1t13 303.,8 . 188.111 303.33 186,491 303. lit 185,111 305.11 184,555 )01.96 184.191 )11.19 185,524 1,b111,3641,2621,4602,0222.0)42,042- . 1,8801.314562242 121 ••••••••• INFLON •••••••• TOTAL ROUIEO LOCAL CFS CFS CfS 0 0 0 0 0 0 0 .0 0 0 ° 0 0 0 0 0 0 0 0 0 0 0 0 0 31,150 36.100 16.650 36,6,0 36,620 36,650 36.620 36,650 )6,650 36,620 36,650 36,620 36.6,0 16.650 16.620 16.650 36,620 36,650 36.6,0 36.620 )6,650 36,620 36,6§0 36.650 31,150 36,100 36.650 36,650 36,620 36,650 36,620 36.650 36.650 )6,~20 36.650 36,620 36,650 16,650 36.620 16,650 36,620 )6,650 36,650 36,620 )6,650 36,620 36,650 36,650 ••••••• DISCHARGE ••••••• TOTAL POWER SPILL CFS CFS CFS 5.140 2,050 0 0 5.690 18,600 50,310 14,590 83.930 81t,lt90 84,180 84,910 16.380 68,890 66,,.20 11,190 84.650 84,910 85,160 81.240 69,130 49.610 30,340 18,690 0 0 .0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6,240 2,550 500 500 6,190 19,100 50,810 15.090 84,430 84,990 85,280 85.410 16,880 69,390 66,920 11,690 85,150 85,410 85,660 81,140 69.630 50,110 30,840 1~,190 AVERAGES 1Zbl.l, 952.,11 308.'11 198,353 133 itS 0 0 133 439 1.191 1.149 1.944 1.944 1,944 1, 9it~ 1,149 1,)80 1,521 1,638 1,944 1.944 1,944 1,853 1,580 1,139 101t 439 AVERAGE NUMBER UNITS 2.0 1.0 0.0 0.0 2.0 4.0 13.0 20.0 22.1 23.0 23.0 l3.0 ~1.0 18.5 18.0 19.0 23.0 23.0 23.0 22.0 19.0 13.0 8.0 5.0 OP, CD • 22 2 2 1. 1 it I ~ " 1 2 l.. , 1 1 2 2 2 2 1 2 2 2 29,513 11,223- TOTALS AVG. GEN. MW 0 36,690 36,.90 53,410 0 53,910 1.23C 11t.4 ;2 1 .0\ • F~ l,! ) \) / , i
Source Exif Data:
File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.3 Linearized : No XMP Toolkit : Adobe XMP Core 4.2.1-c043 52.372728, 2009/01/18-15:56:37 Create Date : 2014:03:05 15:46:58-08:00 Modify Date : 2014:03:05 16:05:52-08:00 Metadata Date : 2014:03:05 16:05:52-08:00 Producer : Adobe Acrobat 9.55 Paper Capture Plug-in Format : application/pdf Document ID : uuid:e2867b8e-b6e4-834c-8a28-51ea633ff873 Instance ID : uuid:0dcaf427-ac20-6c49-9fc7-480eb2c6c8a8 Page Layout : SinglePage Page Mode : UseNone Page Count : 281EXIF Metadata provided by EXIF.tools