ECONOMIC OPERATION OF POWER SYSTEMSGeneral Electric Series WRITTEN FOR THE ADVANCEMENT OF [ENGINEERNG PRACTICE ECONONIC OPERATION OF POWER SYSTEMS by toon K Kehmayer DIGITAL COMPUTER FROGRAMAING boy B.D, MeCrecken THE ART AND SCIENCE OF FROTECTVE FELAYNG by €. Russll Mason ‘APPLIED ELECTRICAL MEASUREMENT by hove F. Kinard AIRCRAFT GAS TURDINES by W. Sth [AN INTRODUCTION TO POWER SYSTEN ANALYSIS by Froderck 5. Rothe TRANSIENTS IN. POWER SYSTEMS by Harald A. Peterson [SERYOMECHANISMS AND REGULATING SYSTEM DESIGN, TWO VOLUMES: by Harold Chasnut end Robert W. Mayer TRANSFORMER ENGINEERING bythe lte LF luna, A. Boysen, G. Cell. C.tamor, §. inne end eM, Moninger, Second Editon CCRCUIT ANALYSIS OF A.C POWER SYSTEMS, TWO VOLUMES by Edih Care CcaPActTORS FoR NoUSTRY bby W. €.Bloomgut CR Crate, RM, Fa POWER SYSTEM STABLY ‘Volume |-Steody State Stitt; Yoline Tron Siabiiy, by Selden 8. Crary NATERALS AND PROCESSES bby 1. F Yeuna, Second Eaton len and RC Won ECONOMIC OPERATION OF POWER SYSTEMS LEON K. KIRCHMAYER Manager Sites Generation Aglicl Engineering ‘lecrie Unity Enginvory Nec ‘Genera Bcc Ce Sehonetedy, New York One of a series written by General Electric euthors for the advancement of snyinvering prstice JOHN WILEY & SONS, INC., ‘New York * Lonnon + Syoney WAL Nebo i98 Copyright © 1958 by General Electric Company Fproduction or translation of any part of this work beyond that pormities ty Sections 107 or 108 of te 1976 Unted States Copyright Act witout the permission of the copyright ewner {s unlawul, Requests fo: permission or further information should be adcrossed fo the Permissions Department, John Wiley & Sons, ine Library of Congress Catalog Cord Number: 5812710 Printed in the United Stotes of America ISBN 471 48180 7 To OLGAPREFACE ‘This book treats of new analytical and computing techniques which have resulted in signifieant direct annual savings in the production economy of electric utilities, ‘The emphasis is upon theoretical develop- ments and computer methods which supplement the practical skills of the electric utility engineer. ‘The methods discussed here have been widely applied by electric ulilities in both the United Statos and Canada, Improvements in fuel economy of approximately $50,000 per year per 1000 mw of installed capacity have been achieved in a number of systems by including the effects of transmission losses by means of the tansmission-loss formulas given, Several powerful tools aro deseribed which should greatly enhence the engineer's eapability of solving system problems. ‘Those axe 1. Matrix methods. 2. Analogue and digital computers Matrix metheds are used to derive and ealeulote transmissionloss formulas, ‘These matrix methods provide general methods of dealing, with systems in a simple manner. They also lead to orderly computa- tional procedures which are readily handled by digital computers, ‘The application of analogue and digital’ computers to the problems of cal- culating transmission-oss formulas and generation schedules is also treated. ‘The application of many of tho principles incorporated in this bock to aittomatic economic control of power ysterss is eovered in my book Economie Control of Interconnected Systems, now in preparation. ‘This companion work also considers energy accounting, control, and e¢onomic theories for economic operation of interconnected systems a8 well 2s the subject of evonomie operation of interconnected steam and hydro- electric aystems. ‘The present volume is based upon a course T have been giving sinee 1952 to participants in the General Electric Company's Power Systems avi PREFACE Ws Course, It is intonded that the material contained here be suitable for special senior courses or for use as a graduate course in power systems engineering. A knowledge of algebra, calculus, and elementary cireuit theory is presumed. Chapter 1, entitled Introduction, provides a brief historical review of developments leading to improvements in the economic operation. of power systems, To the reader who is new to power systems engineering it is suggested that this chapter be read last, since an appreciation of the historieal development will not be acquired until the reader has become familiar with the vocabulary of the power systems engineer. Our research in this aren of endeavor was undertaken with the en thusiastic support of 8. B. Crary, I wish to thank Mr. Crary sincerely for this encouragement, OF particular significance to the developments presented here is the fundamental work of G, Kron concerning the theory of transmission-loss formulas, I especially wish to acknowledge the contributions of A. T. Glimn, G. W. Stagg, C.D. Galloway, and R. Hsbermann, Jr., to the material of this book and to thank Professor J. J. Skiles for his helpful comments in reviewing the manuseript. I also wish to thank my seoretary Isabel Gedlewski for her excellent work in the preparation of the manuscript. Leon K. Kuncmsaven, Schenectady, New York Suly 1998 CONTENTS cmapTEn, eo onewe an Introduction Characteristics and Economic Operation of Steam Plants Development of Transmission Loss Formula ‘The Practical Calculation of Loss Formula Coefficients Coordination of Incrementsl Production Costs and Incremental ‘Transmission Losses for Optimum Eeonomy Practical Caleulation, Evaluation, and Application of Reonomic Scheduling of Generation ‘Transmission Losses as a Function of Voltage Phase Angle Desoription of Alternative Coordination Methods Evaluation of Energy Differences in the Eeonomic Comparison of Alternative Fecili Index BBO 186, 218 238, 252 2571 intropuction 1.1 SCOPE AND OBJECTIVES ‘The soope of this book, as indicated by the Table of Conteata, includes the theory and practical applications involved in devermining the ec nomic operation af a power system, Tt develops the nevessary eircui and mathematical techniques as required in addition to deseribing the important role that modern computers can play in improving power system performance. ‘These computing and analytical techniques will be of increasing value to the electric utility engineer, since systems have tended to become progressively complex as they expand. Many of the economic, mathematical, and computing concepts developed in this book are applicable to othor fields of endeavor 1.2. WHY CONSIDER TRANSMISSION LOSSES? Part of this book concems methods of calculating transmission losses through means of transmission-loss formulas. In the past much effort has boon expenced in the analysis of fuel casts and the thermnal perform ance of generating uniis at equal incremental fuel costs. However, with the development of integrated power systems and the interconnection of operating companies for purposes of economy interchange, it is neces- sary to consider not only the incremental fuel costs but also the incre- ‘mental tranemiseion losses for optimum economy. ‘The consideration of incremental transmission losses through means of transmission-loss for ‘mulas has resulted in fuel savings up to one hundred dollars per year por mogawvatt of installed peak capacity. “Another important probiem in the operation of interconnected systems is the determination of transmission losses for purposes of billing in vari- ‘ous interconnection transactions, The revenue to be gained by properly billing for losses involved during interconnection transactions may be a very large sum. For example, tho value of the losses incurred by one large operating company in supplying and wheeling power to certain atomic energy loads was of the order of magnitude of one million dollars per year.2 ECONOMIC OPERATION OF POWER SYSTEMS ‘Transmission-loss eonsiderstions have often proved to be important in the planning of future systems in particular regard to location af plants and the building of transmission lines, 1.3 BRIEF HISTORICAL REVIEW A transmission-loss formula expressing the total transmission losses ia terms of source powers was first presented by E, E. George * in 1943, ‘This formula was of tho following form: Pi, = total transmission losses = By Py? + BaaPa! + BasPa? +++++ BunPr? +t 2BiaPiPs + 2BysPLPa +++ + 2BegPoPs +++- 2BynPPn a) = LL PBanPa Py, Pm = source loadings Bmn = transmission-loss-formula coefficients ‘The determination of the Bn coefficients was based on a longhand pro- cedure which required two to three weeks’ work by two men fora system of eight to ten generators, Present methods require approximately five per cent or less of this time. The application of the network analyzer to determine a similar loss formala was developed later by Want, Baton, and Hale * of Purdue Uni- versity and published in 1950, However, for the number of generators normally encountered in system studies the arithmetie caleulations in- yolved were prohibitively Inrge. Abthe 1951 ATBE Summer Convention G. Kron, in conjunction with G. W. Stagg and L. K. Kirchmayer, presented companion papers 4* which described sn improved method of deriving a tolal transmission- Joss formula requiring considerably loss network-analyzer measurements and arithmetic calculations, Reference 4, in addition, evaluated the dioorepancica introduced by the assumptions made in obtaining a loss formule. ‘The application of automatic digital computers to caleulate a loss formula was presented by A.V. Glimn, R. Hebormann, Jy, L. K. Kirch- mayer, and G. W. Stogg in the summer of 1953. W. R. Brownlee * has indicated a method of expressing transmission losses in terms of generator voltages and sngles and the X/R ratios of the ‘transmission circuits. Also, loss formulas invalving linear terms and a INTRODUCTION 3 constant term, in addition to the quacduuic terms indicated by equation 1A, have beon recently described.*? The form of the loss equation is th ™ Pu = XL PnBanPs + X BusPn + Boo (+9) ‘This form of loss formula allows more flexibility in the assumptions réat- ing to the manner in which each individual load varies with the total load. ‘The first major step in the development: of a method of coordinating {neromental fuel costs and ineremental transmission losses wes presented in 1949 by E. E, George, H. W. Pege, and J. B. Wan ® in their use of the network analyzer to prepare predicted plant loading schedules for w lange power system, At tho same time the clectrival engineering stal? of the American Gas and Blectris Service Corporation, also with the aid of the network analyzer, developed a methed of modifying the incre- ‘mental fuel costs of the various plants on an incremental slide rule in onler to account for transmission losses. Next, the American Gas and Blectrie Service Corporation, in cooperation with the General Elec trie Company, successfully employed transmission-loss formulas and punched-card machines for the preparation of penalty-factor charts to be used in the economic scheduling of gonoration.' ‘The incremental production cost of a given plant multiplied by the penalty factor for that plant gives the incremental cost of power delivered to the system load from that plant. Optimum economy with the effect. of transmission losses considered is then obiained when the ineremental cost of delivered powor is tho same from ell scute Tn 1952 a paper entitled “Evalustion of Methods of Coordinating Incremental Fuel Costs and Theremental Transmission Losses” * pre- sented 1. A mathematical analysis of various methods of coordinating inere- ‘mental fuel costs and incremental transmission losses, 2, An evaluation of the errors introduced in optimum system opera tion by assumptions invelved in determining. « loss formula, 3. An evaluation of tho savings to be obtained by coordinating incre- ‘mental fuel costs and inetemental transmission losses, Progress in the analysis af the economic operation of # combined thermal and hydroelectric power system was reported by the Hydro- Electric Power Commission of Ontario and the General Electrio Com- pany in the paper “Short-Range Beonomie Operation of a Combined ‘Thermal and Hydroslcetrie Powor System,” ™ which was preseated at the 1953 ATEE Pacific General Meeting.4 ECONOMIC OPERATION OF POWER SYSTEMS An iterative method of calculating generation schedules suitable for the use of a high-speeil automatic digital computer has been deseribed. in the paper entitled “Automatic Digital Computer Applied to Genera tion Schecluling,” by A. F. Glimn, R. Habermann, Jr. L. K. Kirchmayer, and BR. W. Thomas. For given total load, the computer calculates and tabulates incremental cost of received power, total transmission lees, total fuel input, penalty factors, and received load, along with the allocation and summation of generation. Since the program of cal= culation is general, a single routine is maintained in the computer library which will permit scheduling of any size system in the most efficient manner. ‘The American Gas and Electric Service Corporation early in 1955 installed an ineremental transmission loss computer in their Columbus Production and Coordination Office specifically for the use of the system load digpatcher.'* ‘This computer ealculates incremental transmission Josses and penalty factors for varions system operating conditions. The coordinated operation of this computer and an incremental cost slide rule furnishes a flexible and accurate method of taking into account the various and rapidly changing system conditions in the plant and on the transmission system, Other computer developments include anslogue dispatching computers which inecrporate both plant ineremental cost rep- resentation and penalty-factor computation within the compater."™! Rapid progress has been made by the industry in developing esonomie automation schemes by means of which system frequency, net inter- change, and economic allocation of generation for a given area are simul- tancously and automatically maintained.!#%%38225263128 These de- ‘vices offer important opportunities for savings by 1. Improvement in fuel economy by closer adherence to the optimum schedule than would be possible by manual operation, 2, Possible saving in man-hows by elimination of coriain manual procedures. For operation with the foregoing systems, the net interchange out of the ares is set manually and is determined by contracts and bargaining with neighboring areas. Reference 29 deseribes means of automatically determining and controlling the most economic interchange between With respect to interconnected systems energy accounting, meth- od #t44221224.5 regently developed permit determination of loss formulas for each separate company and analytical interconnestion of the loss formulas of these companies for study of lasses in interconnected opera- tion. Loss formulas ean be derived that express losses in a given area in INTRODUCTION 5 terms of the source und interconnection loaslings of that ares or in terms of source loadings in all areas and the scheduled interchange between areas, These formulas can also be used to express change in losses from. 1 given condition in terms of the changes in souree loadings and sched- uled flows from » given condition. 1.4 TRENDS IN GROWTH OF POWER SYSTEMS. Because of the following trends in the growth of power systems it has become progressively important to give inereasing attention to economic system operation, 1, In many cases economic factors and the availability of primary estentials, such 28 coal, water, ete., dictate that new generating plants be located at greater distances from the load centers 2. The installation of larger blocks of power has resulted in the neces- sity of transmitting power out of a given area until the load in that area is equal to the new block of installed capacity. 3. Power eyatems are interconnecting for purposes of economy inter change and reduction af reserve capacity. 4. In a number of areas of the country the cost of fuel is rapidly inereasing. References 1. Heononie Loading of Sioam Power Plonts and Bledric Syitems, M. J. Steinberg, ‘T. H.Smith. John Wiley and Sons, New York, 1948 2, Intinsystem Transmission Losses, E. B. George, ATEE Trans., Val. 62, March 1948, pp, 158-158. 8 Total and Inoremontal Losses in Power Tranamission Networks, J, B. Ward, J.R, Waton, H.W, Halo, ATER Tene, Vel 69, Part 1, 1950, pp. 626-631. 4, Analysis of Total and Incremental Losses in Transmission Systems, L, K, Kircl- mayer, ©. W. Stagg, ATER Trans, Yo. 70, Part T, 1961, pp. 1197-1208. 5, Tensorial Analysis of Tategrated Tvansinissicn Systems—Part I: The Six Basie Reference Frames, G. Keon. ATK Trans, Vol.70, Past, 1951, pp. 1280-1218, 6, Loss Formulas Made Easy, A. F. Ginn, R. Habormana, Js, L, X, Kirehmayer, G.W, Stagg, ATEE Travs,, Vol. 2, Part III, 1958, pp. 790-735, " 7 Coordination of Incremental Puel Costs and Inoremental Transmission Losses bby Functions of Voltage Phise Angles, W. R. Brownlee, ASE Trans, Vol. 73, Part ITT, 1954, pp. 529-841, 8. A General Transmission Lose Kquation, B. D, Bary, RB Watson, Q, Ts Smith, AIKE Teavs., Vol. 74, Part TI, 1985, pp. 610-616, 9. A New Method of Determining Constants for the General ‘Transmission Loss Equation, E. D. Farly, R. B, Watson, AEE Tvass., Vol. 74, Part II, 1955, pp. MAIT-1431, 10, Coordination of Fwol Cost and ‘Tranemiasion Lass by Use of the Network Ana lyzer to Determine Plant Loading Schedules, B. 3, George, H. W. Page, J. B, Ward. AYEE Trans., Vol. 68, Pars IE, 1949, pp. 1152-1160,uL 2, a4 3B i. Bw Bw. a. ECONOMIC OPERATION OF POWER SYSTEMS ‘Transmission Loses and Heonomie Louling of Power Systems, Ly K. Kirchmayer, G.H. MeDsniel. Gener Blectric Review, Schenectady, New York, Vol. 54, No, 10, Ootaber 191, pp. 1152-1168, yalaation of Methods of Coordinating Tueremontal Fuel Costs and Incremental ‘Tiaranission Losses, L. K. Kichmayer, G. W. Stags, AIBE Trans, Vol. 71, Part II, 1952, pp. 513-521 Shori-Range Hoonomie Operation of a Combined Thermal and Hydrooleetsic Power System, W. G. Chandler, P. TE, Dandono, A. F. Glinn, Ts I, Kirohmayer, AIEE Prana, Vol. 72, Part 11 1953, pp. 1057-1065. ‘Automatic Digital Computer Applied to Goperation Scheduling, A. P. Glimn, R. Habermann, Jr, L. K. Kirehmayer, B. W. Thoms. ATER Trans, Vol. 73, Part II-B, 1954, pp. 1287-1275. A Transmission Loa Penalty Factor Computer, ©. A. Imburria, L. 1. Kirehe mayer, GW. Stagg, ALEK Trans, Vol. 78, Part IIEA, 1954, pp. 637-570. Design and Application of Penalty Factor Compater, C. A. Imbargia, L. I. Kircamayer, G, W, Stagg, K. R. Geiser. Proszelings of the American Power Conference, Vel. XVI, 1955, pp. 087-007. ‘An Toeremental Cost of Power-Daivared Computer, E. D. Barly, W. I. Philips, W. Shreve. AZEE Trons,, Vol. 74, Part TI, 1665, pp. 520-584. A Computer for Feosomie Scheduling and Control of Power Systems, C.D, Morrill, J. A. Blake, ATBB Trane., Vol. 74, Part TI, 1986, pp. 1196-1141, ‘Lost Bralustion—Part V: Beononie Dispaich Computer Design, R. B. Squires, H.W. Calbors, R. T. Byesly, W. R. Hanailion, JBI Prant., Vel 75, Patt TH, 1956, pp. 719-727. Load Sohodaling Goes Automatic, D. H. Cameron, W. April 18, 1956, pp. 125-128 A'New ‘Type Automatie Dispatching System at Kansas City, D. 1. Cameron, BL. Mueller, ASME Paper 50-A-215. Automatic Losd-requeney Control Systom for Ceateal Station Power, Gustave Bhrenberg, ALE Trans, Vol. 74, Part TIL, 1955, pp. 787-705 5 Burt, Bledrisa! World, 1 A. Now Aittomatio Dispatching Systom for Blocixis Power Sysioms, K. N. Bar nett, D. W. Halfhill, B.R, Shepard. AZBH Trens, Vel. 5, Part IIT, 1956, pp, 1049-1054 How Medeen Tools Tmprove Power System Operations, L. K. Kicchmayer General Bede Keuiw, Sohenectadly, New York, Soptember 1956. "The “Rasly Bied” Goss Automatic, BJ. Kempast. Contral Bnpinecring, Decem- ber 1956, pp. 77-88. Beonomie Aspects af General Electric Automatie Dispatching System at Kansas City, D. H. Cameron, K.L. Mueller. ASE Conference Paper 57-14 presented at the Winter General Mesting, January 1087, ‘Automatic Heonomic Dispatching and Lond Control, R. Hi. ‘ravers, AEE ‘Tears., Vol. 7h, Pact II, 1987, pp. 291-801 Diltercatitl Analyser Ale Délgn of Hectic Vuluy Automatic Dispatching ‘System, L. K. Kirchmayer. ATER Conference Paper presented ut the Feedback Conival Conference, Comnputars in Contra, at Atlantie City, New Jersoy, October 1087. ‘Automatic Operation of Tateroonnected Aroas, A. F. Ginn, LK. Kirchmayer, HL W. Chamberlain. AZEE Conference Paper presented ab the Summer and Pacific General Meeting, San Franciseo, California, June 1950. |. Analysis of Lasse in Interconnectod Systems, A. I Giimn, T, X. Kirehmayer, GoW. Stagg, AEP Trave,, Vol 71, Part Ill, 1953, pp. 706-808. 5 al. e INTRODUCTION 7 ‘Tongorial Analysis of Integrated "Transmission Systoms—Patt TIT: The Primitive Division, G. Kron. AZBE Trens,, Vol, 71, Pat: IL, 1952, pp. 814-822. . Tensorial Analysis of Integrated ‘Transmission Systoms—Part IV: Tho Inter- connection of Transmission Systems, 0. Kron, AEB Trane., Vol. 72, Part I, 1853, pp. 827-888, . Analysis of Lossos in Loop-Intereonnceted Systams, A. F. Glin, L. K. Kirche mayor, G. W. Stagg. ALEE Trane, Vol. 72, Part TT, 1953, pp. 944-968, Power Losges in Intereonnosted Transmission Networks, H.W. Hale. ATER ‘Trans, Vol. 71, Part TH, 1952, pp. 993-998. Improved Mood of Intereonnoeting Tronsmission Toss Formulas, A. F. Climn, LK. Kirebmayer, J.J. Skil, ATES Paper prossnied at Great Lakes Distviet Meeting, Hast Lansing, Michigan, May 1958,2 quaracteristics AND ECONOMIC OPERATION OF STEAM PLANTS 2.1 DEFINITION OF TERMS To define tho terms involved, simplified performance curves of a given turbine-generator boiler unit are drawn in Figures 2.1, 2.2, and 23. Im ‘Figure 2.1 the fuel input in Bt per hour is plotted asa function of out LA 1200) 3 8 8 ] (va input in Buy) 2 10-8 8 T 8 T { if | { 7008) 8 1 Power cutput ve ‘igure 2.1. Inputoutput curve, ut in megawatts. ‘The corresponding heat mite, which is obtained by dividing the input by the corresponding output, is given in Figure 2.2 It is to be noted that the units associated with heat rate aro Btu per ‘kw-hr. The ineremental fuel rate is given by the following definition Ainput incremental fuel rate = — PS" 1) Souipat In other words, the incremental fuel rate is equal to a small change in the input divided by the corresponding small change in the output, As a ECONOMIC OPERATION OF STEAMA PLANTS. ° the 4 quantities become progressively smaller, it is seen that the a (input) incremental fue rate = OEP ey d (output) ao - pages Sere | sae 2 so & 14,000 i _| fom Si 0.0 + }— | 80095—2g 4050 100, fo uote ewe 22, Hatt chars 180 Sasa] 3 saan | | 0, the fuel input to the system is inereased by inereasing the generation of source 1; and in this case it would not be desirable to increase the output of source 1. In fact it may prove to be desirable to decrease the ontput of unit 1. 3. If AP; = 0, no improvement is obtained by taking up a small inere- ment of generation on source 1 ‘Theoretically, any deviation from the optimum loading would result in an increase in fuel input in dollars per hour. However, for practical purposes, the total cost varies slowly with changes from the minimum, cast point, and criterion 3 may be used to a precision within the size of the increment AP;, This criterion becomes previsely correct as AP; approaches zero. From criterion 3 1+ AF AF, = AF, 44) ECONOMIC OPERATION OF STEAM PLANTS 13 Dividing through by APs, ah ate 5 Fram equation 2-1 ‘Then oc nth Bing a ey for optimum economy. 2.4 INCREMENTAL FUEL RATE SLIDE RULE ‘Tho scheduling of generation for operation at equal incremental pro- duction costs may be efficiently accomplished by the use of an incremen- tal-cost slide rule, ‘Figure 24 schematically illustrates an ineremental- cost: slide rule for a two-generator system. lobe sins. 000} — 25000] — 20000) — wm} fiw F - wo]-—] 10} 3 18,00} — 3 g 100-4 ; o|é 10,000 | «4 12 Buw-he mw ow | baa Calibration scale Generator 1 Gonerater 2S 0 Figure 24, Ineremental-cont side rule,14 ECONOMIC OPERATION OF POWER SYSTEMS ‘This slide rule consists essentially of a logarithmetic calibration seale, a movable strip for ench generator unit, and a fuel-cost adjustment seale. ‘The calibration seale is graduated in Btu per kw-hr to a logarithmic seale. Each movable generator sivip is calibrated in megewatts and indicates ‘the relation between the ineremental fuol rate and output of a given gen ‘erator unit, Differences in fuel cosis may be accounted for by displacing given generator strip s0 2s to line up the bottom of the stzip to @ posi ‘tion on the [uel-cost adjustment scale corresponding to the ratio of fuel costs. For a given incremental cost of received power, corresponding gonorator outputs con then be read directly from the strips. If the fuel casts for both generators are assumed equal, it is seen from Figure 24 ‘that fora total load of 250 mw, 150 should be scheduled from geninator 1 and 100 from generator 2. 2.5 APPROXIMATION OF INCREMENTAL PRODUCTION COSTS Practices vary greatly among companies in the representation of the ineremental production costs. Some companies have refined the curve in Figure 23 to inelude discontinuities due to valves, as indicated in Figure 2.5, whereas others use a block representation, also indicated in Figure 2.5. sec E ca|— Sy a sathoske dec § 12000 | i Pg 5 col af fo ic 4 i «or oa 60 —«B 0 Power outpt-= ma igre 2.5, Repmsentalion of inotomentol fuel rate characteristics, ‘Various control echemes for automatic economic allocation require that: methods be devised of representing incremental cost data by an alogues. In the interests of obtaining an economic design it is necessary to discover the simplest straight-line approximation of incremental cost data that may be made without incurring appreciable loss in operating eronomy. Digital calculations may also be simplified through use of straight-line approximations ECONOMIC OPERATION OF STEAM PLANTS 15 ‘The cost increase resulting from straight-line approximations of inere- mental cost data are evaluated in this section by consideration of the incremental rate characteristics of Figures 2.6 and 2.7. The cost of fuel is assumed to be the same for both units, 14.000 Unt J 2M gen) — | == os pian 3 Dect aon 4 & r2400 c r & v Ene z “sm Eo iz a a D et gration m) igure 26, nore fol rte for unit 1 1300 — 2 2000 cad ra é in worn A — z <= cin peti / 12000 2 in 2 Se - 3 = 2 1000|— t _ 2 + | a "Net generation iv) igure 2:7, Ineramontal fuel rate for unit 2 COMPARISON 1 ‘The exact incremental rate curve for unit 1 and the two-line approxi- mation thereof (Figure 2.6) is used here. ‘The approximation was made16 ECONOMIC OPERATION OF POWER SYSTEMS so 2s to make the area under each curve the same when integrated from P, minimum (50 mw) to P; maximum (107.11) and attempting insofar as possible to keep the difference in ordinates to a minimum. ‘Two units, 14 and 1B, were then postulated. Unit 14 is to have an incremental cost curve corresponding to the exact incremental rate curve of unit 1; unit 1B, sn incremental cost curve corresponding to the two- line approximate incremental rate curve of unit 1. Units 14 and 1B ‘were then considered to be located at » common bus, so that scheduling could he done on an equal incremental production-cost basis. ‘This schedule, identified as goncration schedule 1, is shown as Vigure 2.8. 0 Unit generation (nw) g 80 io 080 ‘Total generation (may igure 2.8. Schedule 1. ‘This schedule is then compared with the schedule that would result if ‘unit LB were to have the samo incremental eost curve as unit 14 or, in other words, if the incremental cost carves of both units were to be exact. Tn this case both units would supply one half the total generation, #0 that no plot of this schedule, which is called generation schedule 2, is required, ECONOMIC OPERATION OF STEAM PLANTS 7 Tt ean now be seen that, for a given total generation, unit generation will depend on whether schedule 1 or schedule 2 is followed. This dif- ference in schedules is the result of using an approximation for ineremen- tal cost in the ease of unit 1B. Schedule 2 represents the most economi allocation of genemtion. ‘Therefore, tho use of schedule 1 will, in genoral, specify an allocation of generation resulting in sn hourly fuel input in ‘excess of that required if schedule 2 were used. ‘The example which fol- lows illustrates how the resulting hourly increase in fuel input may be caleulated, ‘Assume a total generation of 149 mw. Schedulo 1 specified that for this condition unit LA shall supply 78 mw and unit 1B shall supply 71 mw (78 +71 = 149), Schedule 2, however, specifies that the total generation shall be divided equally #0 that unit 14 and unit 1B shall each provide 74.5 mw (74.5 + 74.5 = 149) ‘Thus the use of schedule 1 rather than schedule 2 eauses the generation at unit 14 to be increased from 74.5 to 78 mw, whereasat unit. 1B gonoration is deereased from 74.5 to TL mw, ‘The inerease in Btu per hour resulting from the increase of unit 14 generation is then (78.0 — 74.5) times (Btu/mw-hr) where (Btu/mw-hr) is taken to be the average incremental Btu per my-hr in the interval from Pi4 = 74.5 mw to Pra = 78.0 mw. Referring to Fig- ‘ure26, (Bta/mwv-hz) can be found from the exact incremental rete carve to be equal to 10.280 X 10%. Thus, 4 (Btu/he), due to the increase in unit 14 generation, is 10,280 X 10" x 3.5 = +35.980 X 10°. Simi- larly, the 4 (Btu/hr), duo to the docrease in unit 1B genoration, ean be found to be (71 — 74.5)(10,254)(108) = —35.889 x 10°, The nes in- creuse in Btu per hour eaused by the use of schedule 1 rather than ached- ule2 is then A Btu/hr due to change in unit 14 generation + A Btu/hr ue to change in unit 1B generation = (A Btu/he)La + (A Buu/tn)ip = (85.980) (10°) — (85.880)(10°} = 0.091 X 10° = 91,000 Btu/hr. ‘Ta- ble 2.1 shows this procedure in tabular form for various values of total generation. Note that at those points where schedules 1 and 2 are iden- tical, there is, of courso, no resultant cost inereaso. ‘An assimed load-duration curve is given by Figure 2.9. By use of this curve we may plot Beu per hour as a function of the per cent time that the load assoeiated with a particular Biu per hour increase ocvurs. ‘Thus a plot is made up of Btu per hour vs. per cent time. Such a curve is chown in Figure 2.10. ‘The average of this curve now represents the average increase in Btu per hour caused by the use of an approximation fr incremental fuel east. Knowing this, we may assign a dollar value to Btu's, say thirty conts er million, and evaluate annual savings. ‘This ealealation is show Figure 2.10.Percent tine Figure 29, Loudduration eurve 1 8 ECONOMIC OPERATION OF POWER SYSTEAS ECONOIAIC OPERATION OF STEAM, PLANTS 9 Tan 21 rat fue Pn ti wro(t)a| ane cor (S) | (28) 4 | te Be at) BEE | catty | Co) ote) © On} 2) | iy wen, eer. | vat ?12)| Thr, Saou, ie) ie oe ts) | Cae) a fein) | art Shuto” [ORM] MGR to™ | Cats ata) | Mele (ei, |Get) ae) ety] ay | So oss ms sai Basar tot njescine | "Btyte™ | Faewe 26) | am “ . a i des | at i | gt the ee ‘ = | i is | ims = | ie = | ae veer | ea So | ie 28) 2 ) 3s ae " 5 oom | me wr | — das so | a fe | cam ae 2 og | : 7" as 2 of |e he sata a Se || ie @ sant = = 2, | 2 A : 3 som] B oe ber ¢ 2. 2 3 ie o 20 200 _ 100] = a om T_T t T_T z ner ea 510 200,000} ‘et run a | i 1 vad SEO ne Zim 2 oso HOY - oe SI aL SSE ee we ee a Figure 2.10, Tevaluation of ineroasod annual gost of sohedule 1 over schedule 2.20 ECONONIC OPERATION OF POWER SYSTEMS COMPARISON 2 ‘This comparison differs from comparison 1 only in that. single straight, lino is used to approximate the incremental fuel rato of unit 1 rathor than the two-straight-line approximation used previously, Again two units, 14 and 1B, were postulated, with the incremental fuel rate of unit 1 being taken as the exaet curve of Figure 2.6and that of unit: 1B being considered as the one-straight-line approximation also shown by Figure 2.6. ‘The resulting allocation of generation is given as Figure 2.11, schedule 3, This schedule is then compared with schedule no 109| 8 Unit generation (mv) io) 204010 1e0 0 ‘Teta gneraton (mw) Figure 211, Schedule 8 2, which, as mentioned cartier, is based on using an exuct representation for the incremental fuel cost of both units, ‘The two schedules, 2 and 3, are then compared ina manner similar to ‘Fable 2.1, and the annual increase in cost resulting from the use of the one-straight-lino approximation of ineremental rate is then evaluated. ‘The plot of Biu's per hour vs. per cent time is given as Figure 2.12, which shows the evaluation of the annual cost increase involved. vr ECONOMIC OPERATION OF STEAM PLANTS a 3900.00: —} |_| 1) toy ob 4 = 13 tonnes 2900,010}— tujhe 1,900,000 10 20° 304080 Forcont tina Figure 212, Kvabuaticn of increased annual cost of schedule 8 over schedule 2 COMPARISON 3 In both comparison 1 and comparison 2 the two its being scheduled ‘vere chosen to be identical, although in some instances approximate characteristics were used rather than tho oxact anes. Now, for com- parison 3 we shall consider the scheduling of two different, units, first using the exact incremental fuel rate for each unit, and then a two-line approximation for each unit, Figure 2.6 gives the exact and approximate incremental-rate data for unit 1, and Figure 2.7 gives the same information for unit 2. Figure 2.13 shows two schedules, 4 and 5. Schedule 4 shows the allocation of genera- tion beiween units 1 and 2 when scheduling is dane on the basis of equal ineromental cost with the exact incremental-cost data used for both units. Schedule 5 shows how generation would be divided between units 1 and 2 when equal incremental-cost: scheduling is followed but with the incre- mental-rate data for each plant being represented by a two-straight-line approximation. ‘The difference in cost resulting from the two schedules is then com- puted in a manner similar to that outlined for comparisons 1 and 2. Since unit 2 has a maximum rating greater than that of unit 1, a new oad-duration curve is assumed, ‘This curve, lond-duration curve 2, is shown as Figure 2.14,ECONOMIC OPERATION OF POWER SYSTEMS ECONOMIC OPERATION OF STEAM PLANTS 23 Figure 2.15 shows the determinstion of annual cost inezease due to the scheduling of units 1 and 2 by means of two-line approximations of the no incremental produetion costs of each unit 1,400,000 100 | io 7 Ve a - 120000 }— + E V7 é L g 80 000 L 2 1000, (1 wos oa ven r= 10800 2, Une {2 conparmiton gts = 929 Ar it et = 105000 4760 4 ouon m6 © Sa tne tard? — | ae ] {rcstnepponraion ot 5 Easy z sol 4 00,0] 10 120d i680 200 Total gereraten (rw) 400,00) Figure 2.18, Schedules 4 snd & 200 | 220 | 2entoo 200 a Ero i °9 700 160 er ant tin x0 | COMPARISON 4 100) r ‘This calculation is similar to comparison 3, except that when units 1 and 2 are scheduled on an approximate basis a one-line approximation of ineremental fuel rate is assumed rather than the two-line approximation used before. 10a B18 00 igure 214. Load-duration eurve 224 ECONOIAIC OPERATION OF POWER SYSTEMS Bxack incremental-rate data and the one-line approximation of this data is given by Figures 2.6 and 2.7 for units 1 and 2, respeetively. Schedules based on equal incremental-production-cost scheduling. are then drawa up. Figure 2.16 shows the two schedules obtained. Sched- 119} — f 2 Or fF a eB" a, S J fal Z d oe so ie 10014618007 Total generation (rw) Figurs 2.16, Sohodulea 4 nad 6. ule 4 results from the use of exact incremental-rate daia; schedule 6 from the use of the one-line approximations. "The difforence in sehodules is than snslyzed, Again by the use of load- duration curve 2 (Figure 2.14), an annual cost inerease resulting from the use of approximations for incremental fuel cost is evaluated. ‘This evalustion ia shown by Figure 2.17. From the results on page 26, it appears that a one-line approximation for incremental-fucl-cost curves of the type of Figure 2.6 is not sufi cciontly accurate for purposes of economic scheduling. Far this partien- lar problem, the use of a two-line approximation seems to be sufficiently accurate for practical purposes. 3,600.00 2,000,000] 1,000.00} ECONOMIC OPERATION OF STEAM PLANTS 25 fe (a) Avrag ou nea in = 212000 (2) Cot pr milon Sire = $030 Anmloot weease m 212000 030 Gro oo! a be67 stigure 2.17 70 70 60 on 100 [Bvaluation of inorsased aan east of schedule 6 over schedule 426 ECONOMIC OPERATION OF POWER SYSTEMS ‘he resus obtied bythe preceding comparisons are irda below (comparison 1 14 1B Unit Schedule 2 ‘Schedule T costs $101 per year more than Sched 2 ‘Comparison 2 Unt ‘ust Treenatae | vs (mene a oe Sehedulo’3 Schedule 2 ‘Schedule 3 cass $1,280 per year ror ta Sehedule 2 ‘Compatison 3 Seheae 5 ‘shade 5 cosis $275 per year mere Comparison 4 schecule “Sehedule 6 cass $OB7 per year more then Schedule 4, |. ll ECONONIC OPERATION OF STEAM PLANTS 7 2.6 OTHER METHODS OF SCHEDULING TURBINE-GENERATORS ‘Although the criterion of equal incremental production costs will result in the optimum eccnomic scheduling of genezation, the following methods of scheduling are sometimes still found in use? 1, Base Loading lo Capacity. ‘The turbine-gonerators are successively loaded to eapacity in order of their efficiencies. 2. Base Loading to Mest Efficient Load. "The tarbine-generator units fare successively loaded, in ascending ordor of their heat rates, to their most efficient loads, When all units are operating at their most efficient loads they are londed to eapacity in the same order: 3. Proportional to Capacity. ‘The loads on the units are scheduled in proportion to their rated capacity. 2.7 DETERMINATION OF CAPACITY TO BE OPERATED ‘The discussion thus far has considered the optimum allocation of gen- eration fora given connected capacity. A problem which is not answered by inspection of incremental-cost data is the determination of the capae- yy to be operated for a given totel load. Determination of this capacity ia based upon such considerations as 1. Beonom 2. Reserve requirements, 3. Stobility limitations. 4, Voltage limitations. '. Ability to pick up load quickly. ‘Very frequently, and in particular in widespread systems, conditions 2 to 5 overrule condition 1 ‘The dotorminstion of the most economic combination of eapacity to be operated at a given time is accomplished by inspection of the total fuel input to the system for various assumed combinations of eapacity. Of course, for any assumed capacity in operation the economic alloeation of generation is given by equal incremental-cost loading. {In general, in given station the units are placed in service in asecnd- ing order of their heat rates assuming the cost per Btu to be the same ‘To determine the most economic combination of units for & given station load it is necessury to plot total station heat-rate curves of successive combinations snd to note the combination providing the lowest heat rate for a given station lead. Anether problem of importance is to determine the economies advis- ability of taking units off the line for relatively short periods of time, sch a between the morning and evening poaks. "This determination is ¢ evaluation.28 ECONOMIC OPERATION OF POWER SYSTEMS based upon ealeuinting the total fuel input in dollars to the system dur- ing this period of time with the units in question both on and off the line ‘This calculation should include cost of restoring the units under con- sideration back in service and losses involved in banking the boilers 2.8 BRIEF DESCRIPTION OF A TYPICAL MANUAL PROCEDURE IN LOAD DISPATCHER'S OFFICE ‘Load forecasts for the expected hourly lowdings of a given day are usually prepared on tho praceding day. ‘The capacity available is de- pendent upon the maintenance sthedlule for tie system as well as any forced outages that may have occurred. ‘The choice of units to be oper * ated from this available eapacity is diseussed in Section 2.7. Based upon the forecasted hourly Ieads and the eapocity that is to be operated, the load dispatcher then determines hourly generation sched- les for the following day’s operation by the use of the incremental slide mule, The plant operators are informed of their respective plant sched- ules and manually follow that particular schedule, In case of variations from predieted conditions the loacl dispatcher is called upon to revise his prevalculated generation schedule. Recent developments in computers and automatic dispatching systems have removed the need for many of thes: manual steps. Several dis- patching compatars are described in Chapter 6, Various automatic methods of obtaining economic allocation of generation, together with the maintenance of frequency and net interchange, are discussed in Baononiic Control of Interconnected Systems by La. Kirchmayer (in preparation). 2.9 EFFECTS OF ERRORS IN ECONOMIC DISPATCHING OF POWER ‘SYSTEMS # An anelysia of the effects of errors in the economic dispatching of sys- tems is important in understanding and ehoosing the accuracy require- ments of the components of a dispatching system. Deviations from the most economie schedule are obtained if, L. The representation of the incremental-produetion-cosb eurve is i 2. The servemechanism loop which matches the desired generation h the actual genemtion is inaccurate. In the ease of manual opera- tion the station operator represents the servomechanism loop. ‘The two types of error indicated may occur in an eutomatic dispatching system as well ns in the manual dispatching of » power system, ECONOMIC OPERATION OF STEAM PLANTS 29 SYSTEM STUDIED ‘Phe system considered consists of two 100-mw units, as shown in Figure 2.18, Incremental cost data is givon by a APs per mowv-he = 24 0.0127; (216) rh igure 218, Sehe- {incremental production cost of unit 2 in dollars mae represents tion of sytem snd. ied @4 ineremental production cost of unit 1 indollars Ps por mw-he = 15+ 0.015P, (217) Assume that 1, Both units are connected and operating. 2. Minimum load of each unit is 10 mw. 3. Meximum eapability of each unit is 100 mw. 4. The total load varies from 50 to 200 mw, as shown in Figure 2.19, Figure 2.19, Loacldurstion eurve, ERROR IN REPRESENTATION OF INCREWENTAL:COST DATA Assume that di’;/d?; is incomectly represented as being high by a factor e((al,)/(aP,)) and that difz/dl?, is represented as being «{(@F2)/ (@P.)} lower than the correct value, Consequently, seledules that sre different from the most economic schedule are obtsined when eis other than zero.30 ECONOMIC OPERATION OF POWER SYSTEMS Figure 2.20 shows the correct scheslule and modified schedules for units 1 and 2 for various values of ‘The inerease in production cost in so -- Pee De yy sx dis MW PST = Zina | ii Lr, bof) | YP, comet JAF he t i 7 fn Ly i ZA | “ys \ i E40 le 5 a » Pie LAG, 20 PLP, «= 00 (Anesti Fi ee °o 0a 6080100 7201401604820 Seen eo Figure 2.90, Genecation schedules for two-unit sytem, dollars per hour for the different schedules is plotted in Figures 2.21 and 2.22 as a function of system load for various values of , ‘The method of analysis involved in the calculation of Figures 2.20, 2.21, and 2.22 is given in Appendix 2.2. To establish an order of may ude of the devin~ tions in economy involved for various values of enote the deviations in dollars per hour for a total load of 125 my give 0.00 ont 0.03 0.05 o ‘Table 22: TABLE 2.2 Toss in Beonomy in Dollars per Hour 0.00 0.05 046 128, 51s a Figure 2.21, ECONONIC OPERATION OF STEAM PLANTS a ss Baal a & : Loos am i | 8 oso $ a a sites ae vinci nnsce erat va es = Z x oe ie a [sso 1 g se a system ioad—mw ‘Figure 2.22 Hxotse conte incurred in deviation from optimum sohodules, Ui the multipliers (1 + «) and (1 — ©) are interchanged, then the loss in economy becomes as shown in Table 2.3 oot 0.08 0.05 OL TABLE 2.3 Lass in Boonomy in Dellars per Hour 0 0.05 045 1.35 49432 ECONOMIC OPERATION OF POWER SYSTEMS ‘Tho intograted annual loss in oconomy as a function of « assigned in the two manners discussed is illustrated in Figure 2.28, It is to be noted 49 ' wa-o = wl Bare 8 8 £320 | i ag “i £" 10 bt Gare _| ee |_| a-0_| a 14 002 008 B06 068 O10 Figure 2.23, Annual less in economy as a function of « that the annual loss of economy is approximately a paraholie function of ¢. ‘The difference in the tavo curves of Figure 2.23 is due mainly to the fact that minimum-maximum unit-generetion constraints are encoun- tered at different total loads, In Figure 2.24 the deviation in economy is plotted as the ratio of on = i Bano 3 oo Baro : 5 ome | 2 ocoa|—} | z a0 ao a5 a0 a0 Figure 224, Por unit deviation in annual production cost as funetion of & annual loss in economy 10 the total fuel consumption when operated most economically. ‘The fuel consumption was obtained by integrating . 6 ; i ECONONIC OPERATION OF STEAM PLANTS 33 ‘the proper incremental-cost curves and adding the assumed no-load production-cost interespts of twenty dollars per hour and forty collars per hour, respectively, for units 1 and 2. As noted in Figure 2.24, the ‘per unit deviation in fuel economy is much smaller than the correspond ing deviation assumed in the incremental-production-cost data. ‘To as- sist in obtaining an understanding of this reletionship Figure 2.25 plots Bano Faro = |, Zo Baro aa-o~ | a a IC) Figure2.25, Ratioof per unit deviation in annual produotion eostto as function of « the ratio of the per unit deviation in fuel economy to the deviation in incremental-cost data as a function of the deviation in incremental-cost data Appendix 2.3 presents an analytical determination of the loss in oper ating economy resulting from an inerementel-cost displacement multi- plier error (1+ ) for one of two identical units on the same bus and (1 ~ © for the other unit. ‘The hourly loss in economy is given by the ‘oxpression ar, (218) where AF, ~ dollars por hour lose of opernting economy, 2 = incremental-cost level for two identical units on the same bus as the land slope of ineremental-sost charaeteristie of exch unit «= deviation in inoremental-cost representation,34 ECONOMIC OPERATION OF POWER SYSTEMS Ti is seen from equation 2-18 that the deviation of fuel economy varies as the square of the error e ‘The loss in operating economy may also be expressed in terms in the total generation as AF = ¢ Pet ») = (219) where b = intercept of the incremental-cost characteristic and the other quantities are as previously defined. Both equations 2-18 and 2-19 are applicable only when both units are uneonstrained by min imum or maximum generation limitations. ‘These expressions may be used as approximate tools to apply to sys- ‘tems in which the unit chumcteristies are not identical. For example, consider the twe-unit system previously discussed and a load of 125 mw. ‘The value of ) involved for the load of 125 mw is $2.61 per mw-hr. For 0012 + 0.015 the value of te slope, se the average 2212 0015 _ gop, (2.61)? ‘Then P= COD Go 0135 505(0" From the above equation we obtain the following results: TABLE 2.4 Lass in Beonowy ‘ in Dallars per Hour 0 0 0.01 0.05 0.08 045 0.05 126 Ou 505 ‘By comparison of Table 2.4 with Tables 2.2 and 2.3 it will be soon that close agreoment ie obtained betwaan the resulte of the approximate formula, as given in Table 2.4, ond the results presented in Tables 2.2 and 23. EFFECT OF ERROR IN HOLDING GENERATION AT DISRED VALUE Although the incremental-production-cost data may be represented accurately, an error in the unit losding may occur because of errors in = ECONOMIC OPERATION OF STEAM PLANTS 35 the automotio or manual exccution of the schedule. Wor purposes of this disoussion assume that the error is 2 fixed amount of the maximum ‘unit load but that the unit minimums and maximums will still be held at 10 and 100 my, respectively. A per unit error of 0.1 in loading by this definition corresponds to 0.1 X 100, ora 10-mw shift in loading; ie., ‘unit 1 loading would then be low by 10 my and unit 2 loading, high by 10 mw. The loss in economy as 9 function of various per unit errors is plotted in Figure 2.26 for the two-unit system described. Figure 2.27 2 evens Wy AP [ Petes duces by SP Peincened iy ae Anus lssin economy in thousands of dot D02__—B6F 008 “Gor O10 er uni or in loading (AP) Figure 2.20, Annual Joss in eeonomy as a function of error in Touding. presents the same results in terms of the ratio of the annual loss in evon- omy to the produetion est for optimum leading as a function of the Joading error. Figure 2.28 gives the ratio of the per unit loss in economy to the per unit loading error as a function of tho deviation in loading. Appeadix 23 presents the derivation of the following expression for the loss of economy resulting from errors in maintaining generation ab desired values for the special ease of two identical unite: AR, = APs 220) where AP = displacement in loading on one of two identical units 4 = slope of incremental-cost curve of either unit,36 ECONOMIC OPERATION OF POWER SYSTEMS 008 _ § ooo g 3 ? Freese by 2 & onoe Pico a 5 oon Pinca ae & ‘0002 0s G05 —-os a0 er unit ero in loading (AP) Figure 2.27, Devintion in annual production cost asa funtion of error in loading, a” 3 nemo wa . 2 emu oy a sl ‘ema Bina 9 a g 3 a Bo 3a“ 908 To For unt eror is fading (AP) igure 2.28, Ratio of per unit deviation ia ennual production cost to errr in Joading, 1s funotion of error in louding. ‘The leas in ceenomy varies aa the square of the error in loading. Ap- plying this expression to the small system of two dissimilar units previ- ously discusced, we have AF, = AP*(0.0135) where @ = 00135 = average slope ‘Thus, for a simultaneous deviation of +10 mw of one unit and —10 mw ECONOMIC OPERATION OF STEAM PLANTS. 37 on the other ar, (10)7(0.0135) = $1.35 per hour CONCLUSIONS 1, From the study of the effect of simultaneously displacing the inere- mental cost of one source by (I+ «) and the ineremental cost of another source by (1 ~ ¢) the following conclusions may be drawn for the two- source system investigated: (q) The loss in hourly economy varies as the square of the per unit error ¢ in the representation of the incremental production cost. (b) For a given valu of « the lose of hourly economy varies directly as the square of the ineremental-cost level and inversely as the average slope of the incremental-cost characteristics of the sources in question, 2, From the study presented of the elfect of displacing the output of ‘one souree by ++ AP mw and the outpat of another source by —AP saw from the desired eeonomie value it is noted that the loss of hourly econ- ‘omy variesas the square of the AP mw deviation from optimum sehedule, 2.10 SYSTENS WITH TRANSIAISSION LOSSES Jn the general case all sourees of genoration are not located at the same bus but are connected by means of a transmission network to the various loads. Some plants will be more favorably located with respect to the load than others. Also, if the criterion of equal incremental production «costs is applied, there will, in general, be transmission of power from lov-cost to high-cost areas, It will be necessary, of eourso, for optimum ‘economic operation to recognize that transmission losses occur in this operation and to modify the incremental production costs of all plants to take these line losses into aecount. A simple system * that is used as an example is given in Figure 2.29. 20 mn less for 100 mi tafe igh-s0e ree ead Pigue 220, Sohomatie representation of system.38 ECONOMIC OPERATION OF POWER SYSTEMS ‘This system is a simple representation of the American Gas and Electric System (1950) and illustrates the relatively high-cost generation in the Indiana Division as compared to tho low-cost generation in the Ohio Division available for transfer west. ‘These two areas are interconnected, by a 132-kw tmnsmission system and are separated by a distance of ap~ proximately 250 miles. ‘The meximum practical transfer over the trans- mission system coresponds to approximately 170 mw, 40 109 200) 300 100 ‘Nw utout ofeach area Figure 2.40, Tneremontal ful cost dasa ‘The approximate incremental production costs for plants 1 and 2 are shown in Figure 2.30 and may be expressed by aly Sp, 7 ineremental procustion cost of plant 1 in dlls per mw-hr = Pui th (2-21) = 0.002P) + 1.7 as Sp, 7 msremental produetion eost of plat 2 in dllars per mw-br = FaaPo the 004? -+ 20 If the effect of transmission losses is neglected in the scheduling of gen- eration, then optimum loading is given by PuPi this 2-28) FaPat f= (2-22) ECONOMIC OPERATION OF STEAM PLANTS. 39 ‘Yarious values of total generation are obtained by varying the value of 2. It will be noted that if the two areas are operated at the same incre- mental production costs 150 mw will be scheduled from the low-cost area before generation is scheduled from the high-cost area. ‘The division of generation as a function of total generation is given in Vigure 231. ‘The i b ‘Pant No. 1 i 100] Plant No.2 0 Tor 200300400500 Total geraraton—mw Figure 281, Generation schedule. fuel input as a function of reesived load with generation scheduled by ‘equal incremental produotion costs is given in Figure 2.82. However, if sai a op eae ohana 9 Sane : =p 7 oo» x - j Z ja Bol Ae “200 Q {oo — 305 an Receved lead In mw ‘Figare 282, Fuel inputas function of received loud,40 ECONOMIC OPERATION OF POWER SYSTEMS the effect of incremental transmission losses had been included in the scheduling of generation, the fuel input would be reduced, as shown in Figure 232. In order to be able to coordinate incremental production costs and incremental transmission losses properly, we shall next study the theory involved in the determination of a trensmission-loss formula. 2.11 SUMMARY With the effect of transmission losses neglected, the minimum dollars- per-hour input to the system for a givon total load is obtained when all units are operating at the same incremental production eost, "The inere- mental production cost of a given unit is equal to the slope of the dollars per-hour input vs. megawatt outpatcurve, ‘The incremental production cost is expressed in dollars per mw-hr or mills per kew-hr, APPENDIX 2.1, OPTIMUM SCHEDULING OF GENERATION (Transmission Losses Neglecteel) ‘This derivation follows directly from the mothod of Lagrangian multi- pliers described by Courant.® Let F¢ = total input to eystem in dollars per hour =IFs input to plant 2 in dollars per hour where Py It is desired to minimize the total inpat (F,) in dollare per hour for a given reesived load (Pp). Let Pp = given received load By application of the method of Lagrangian multipliers, the equation of constraint: ia given by (P,, Ps, Pps P,) = Pa — Pp 0 (2-24) ‘Then minimum fuel input for a given received load is obtained when, ae (25) ay where F= Fay (2-26) A = Lagrangian type of multiplier as am. oY wo Mg oz ap.” ap, ae, ¥ ECONOMIC OPERATION OF STEAM PLANTS 4a arr. a ‘Then a P.-P, 7 oP. Pe [= * are St a - a] =0 aa Pe ore ay 2-28) wy (=m) or, ( r Be @-29) pat a, Py ‘hen equation 2-28 becomes, mad (2-30) a, e Equation 2-80 is identical to equation 2-5 previously presented. APPENDIX 2.2. METHOD OF CALCULATION OF LOSS OF ECONOMY 1. Caleviation of Schedules Assume two units whos lowing equations: icremental costs are represented by the fol- 7. sneremental cost of unit 1 in dollars per mo-br aPy where Py 2 seroma tnt Sn lars permet ay ooPa +o oad on unit 2 ‘The most economic schedule for the case of no transmission losses occurs when the incremental plant costs are equal. For this ease the most evonomie schedule is given by where Pa ar = athi=s (2-31) eS Pith (2-31) d's oe PaaPa the = (2-82) wp, taPat has (2-82) where > = incremental cost of reesived power2 ECONOMIC OPERATION OF POWER SYSTEMS From equations 2-31 and 2-82 we obiain PoP tht Py = EM thf 88) Pan Defining, Pa as the total load on both units, Pa= Pa Pj (a4) Busting 2-53 and 2-34 and solving for Ps, we obtain Fabs —fith py tcf (2-35) iu + Poa Values for P snd Py for various values of Pe may be obtained from ‘equations 2-34 and 2-85, respectively. Assume that d?,/dP is multiplied by (1+ ) and that dP a/dP, is multiplied by (I — ¢). Por these modified ineremental casts tho expres- sion for Py then becomes _faPa Sit fe~ fi + fot PnP) fq oe 2-86) a + Fon + (Fi ~ Faale es) ‘The value for Po is, of course, given by P= Pe-Py (234) 2. Coleulotion of AF; for Verious Values of Pr Tet Figure 2.33 represent the ecact incremental cost data, For a given ae OF Figue 233, Representation of incremental east data. ‘ineremental cost ) and the corresponding value of Pr let the value of P; and P, obtainod with the exact incremental data be denoted hy P' and Py. ECONOMIC OPERATION OF STEAM PLANTS 43 If the modified ineremental-cost data is used, different values of Pr and Ps will be obtained for the same value of Pr. Denote these vslues of Py and Pa by Py" and P2" Jn going trom Ps! to P” the fuel input to unit 2is increased by [OC where AP2 = change in goneration on unit 2 = Pp” — Pa! (#2) a <2) = for Py = Pr!” IPs) ~ dP; (2-87) (dF) dF, 2) = 1 4 1] Aas | Zaks + Zaks [—] 2] Zar | Z2oks + Zaks 3] Zs: | Zaks + Zaks where Zig = Ziaks + ZasKs a = Kolar + Kean (3-31) Zag = KaBoaKy + KaaaKy + KyZaaky + Kaeaks Tn terms of numbers we have ok 7 in Tr = —| (3-33) DEVELOPMENT OF TRANSMISSION LOSS FORMULA 67 whore the voltages are given by equation 3-30 and the impedances by equation 3-31 Ta terms of numbers we have 08-03%, 03K; igure 3.16. Modified eireuit after transformation, 3,6 TRANSFORMATION TO REFERENCE FRAME 2 It is desired to eliminate the individual equivalent load currents as variables, sinee the final result should involve only generator powers, As will be recalled, the equivalent load ewrrent at a bus is defined as the ‘sum of the line-charging, synchronous condonser, and load current at that bus, The first assumption involved in the development: of a loss formula is the following: Tt is assumed thet each equivalent loud eurrent remains a constant complex fraction of the total equivalent load current. Define i= Diy Ga) By the above assumption xs = bin (3-25) For the system given by Figure 8.10 and equation 3-12 we may write ies = toa tu = hie, tne = bin ‘Tho preceding wlation may bewrilten in terms ofa matrix of transforma tion,68. ECONOIMIC OPERATION OF c2 63 POWER SYSTEMS L 1} in, 62) ive 30) 63) ica tla ‘Thus the currents of reference frame 1 (1h) are related to the currents of reference frame 2 (Is) by a matrix of tran GL G2 usformation C3! where co @-37) The symbol Gy’ is used to indieats the frume or step j to reference frame or step ‘The new impedance matrix is given as transformation from reference k indicated by equation 3-24 by GZ.iC. Performing the Z1€ operation firsi, we have a @ @ nm @ aa ob Fara Leno DEVELOPMENT OF TRANSIAISSION LOSS FORMULA 69 @ G38) a a as L Zay-als + Brae a @ @L G39) Bor—nil + Barta = Zea-nilh + Zea Boa—tabr + Zortalr by = bisa + be*Ziaor (40) by = Zeon + LL r2—a by = Biro + be*Zi2e9 we hrc +h". where a ale tala + e*Zia—wils + bb*Zra—rale79 ECONOMIC OPERATION OF POWER SYSTEMS From equation 3-23 it is sven that our new voltages of reference frame 2 are given by C;*en1g as indicated by Ga nm a @ @s oun) n 12 ‘The calculation for H;, — Hy is indicated in detail below. LB — Be) + "(Bis — Be) = W*Bis + "Bia ~ (G* + hE Deline By = bE + lata Since t+h4) =1 te) + "(Eno ~ Bx) wehave 4 *(Bx — in En From equations 3-85, 9-39, and 341 we have Zoo Foo: Is will be noted that the effect of each load current has been replaced by a single total Load exrrent ‘The preceding steps accomplished by the transformation matrix Cy! may be thought of in terms of a number of algebraic steps. Consider the reference frame 1 equations for this example as repeated below: DEVELOPMENT OF TRANSIAISSION LOSS FORMULA n aa As before, let fn = hit 43) ins = bie Substituting equation 3-48 into equation 3-12, we obtain a Louth + Zor-aale a ah + Zora @ oxta + Hoxmaals | C4) a 1 ‘The impedances in equation 3-44 correspond to ZaaC. Define a hypothetical load voltege J, such that the power loss con- tributed by (Bui — Beliaa* + (za — En)ina* remains invariant. ‘Thus (Bn — Bain* = (ia — Ea)ina* + (Eta ~ Brdiva® = (Bis — Hayhitic* + (Bia — Exdatin* (48)72 ECONOMIC OPERATION OF POWER SYSTEMS Dividing equation 345 by tr (Gs, ~ Ba) = Bia — Ends + (Bia — EnJa* — G-46) Performing the operation indicated by equation 3-46 upon equation 3-4, we obtain equation 3-42 as befare. From equation 8-42 we note that. the mutual impedaness between genertors and loads are not symmetrical. ‘That is, ay ¥ ty, da x boy and ¢3 > b,, For example, a1 = Zor—tili + Zor—tale = Rorna + iXo1—ra)th’ + jh") + (Raita + JX e1—z2(h! + jh”) = Rox—tah! + Rott’ — Xoruah” — Xer—saby! + i(Xertahi! + Xoxsah! + Rorarh" + Rer—rale”) (9-47) boar =U! = HW Rira +5X 1-0) +! = le Binor + 5X00) 1 Ramos + W’Riaa. +h" X1.—01 + be! Xea. +HU'X tor + BX a—e1 — hh" Bia—ar — b!"Re2—ai) (8-48) by = Zia Consider next the more general ease in wiieh the number of sourees = m, nand the number of loads = j, & and for which reference frame 1 ‘equations are given by equation 3-6 which is repeated below: ‘The matrix of transformation C,! is given by (3-49) va [| [ie DEVELOPMENT OF TRANSMISSION LOSS FORMULA 73 Cn 7 (3-50) ‘The matrix Lk isa column matrix with the number of elements quel to k, the number of load curzents. ‘This follows by inspestion of tho matrix, since the number of eolumns correspond to L and the num ber of rows to Lk. Since there is orly one hypothetical loud eurrent, thors is only one column, ‘The transpose of C,' is given by Gm ij Gm (Ge is a row matrix with the same elements as the ect tis 4 iY th sets won oo stead of a column. ‘The resulting voltages, impedances, and currents, are given by Gn L fon B51) whore Om = Zan-tile bn = YZ 100 3-52) w= UPZrtle ‘ By = UB ay74 ECONOMIC OPERATION OF POWER SYSTEMS By means of the above transformation the eireuit of Bigure 3.7 has heen changed to the citeuit given in Figure 8.17. ‘The load point L does Reference point. Hypa son ‘ed a Bon (=) {> Figure 3.12. Reference frame 2 not exist in the actual network, and consequently it is referred to as a hypothetical load point. As previously noted, the mutual impedances between the generators and loads are not equal. As noted by equation 3-51 and Figure 3.17, the component of the voltage dzop om — Bp due to load current iz is given by Gyiz. Similarly, the component of the voltage drop By, ~ Bp due to current én is given by bnigg. ‘The impel- ance w isthe self impedance existing between the hypothetical load point and the reference bus, 3.7 TRANSFORMATION TO REFERENCE FRAME 3 As will be noted in equations 3-42 and 3-51 and also Figure 3.17, the individual load currents have been eliminated as variables and replaced by the total load current zz. ‘The next step in our analysis involves eliminating the total load current iy, as a variable, We may accomplish this by the relationship that the surnmation of the souree currents must be equal and opposite to the summation of the load euents, ‘Thus Lin = in For the system of Figure 3.10 and equation 3-42 we may write fa = tax fates a8 ios = for tn = —(ar + tee + tas) DEVELOPMENT OF TRANSMISSION LOSS FORMULA, 75 ‘The above relation may be written in terms of a matrix of transforma- tion as indicated below: GLI tor G2) igo | B54) 63| tos ‘Thus the currents of reference frame 2 (J) are related to the currents of reference frame 3 (Js) by 2 matrix of transformation ,? where Cl G2 G37 ECONOMIC OPERATION OF POWER SYSTEMS DEVELOPMENT OF TRANSMISSION LOSS FORMULA 7 ‘The new impedance matrix, us indicated by equation 3-24, is given by. From equations 3-54, 2-56, and 3-38, the reference frame 3 currents, CALoaC. Periorming the C*Zs1a operation first, we have impedances, and voltages are given by It will be noted irom equation 3-59 that only the generator eurrents appeer as variables. ‘Also Cl G2 63 Zorn —b % | ct] Be - 1] tor Zea-ox 0 Zama +0 where Zp Zon—an ~ Om — byw (61) Zoro —) ~0; +0] 2 Zoras~y~ m+ jane ‘The results of performing the operations indicated by equations 3-22, Zoro — bat +0 —-8-28, and 3-24 with transformation C;° may be visualized by a number of algebraic steps. Consider the reference frame 2 equations for this 6-38) example as shown in equation 3-42:78 ECONOMIC OPERATION OF POWER SYSTEMS. Subtracting ion + bain + baton + iw from each of the previous equations in 3-42 we have Zao Zor-aa~ bs Zas-as be (62) ‘The voltages in equation 3-62 correspond to C,*esia, as indicated by equation 3-56. Also, the impedanees indicated by equation 3-82 cor respond to G,*Zaia, as indicated by equation 3-37 Substituting t= — (ar + tee + ies) into equation 3-42, we obtain equation 3-59 as before, From equations 3-59, 3-60, and 3-61 we note, as in reference frame 2, that the mutual impedances are not symmetrical, DEVELOPMENT OF TRANSMISSION LOSS FORMULA, 79 Thus Zy2 = Zai-c2— 4 —~ by Fw= Ria tsXi-a (3-63) Fa = Toray ~ tb, + we Rea t+iXe1 (8-64) ‘From equations 3-40, 3-47, and 3-48, Ria = Raror — (Rar—tih! + Rer-tal! — Xe. = Xo1_aala") ~ (Rasash! + Reaoah! + Xr—ach” + Xia-aah!) + w! (3-05) Mia = Koren — Sor-nil’ + Xer—iale! + Rornah" + Ras_tah”) — (X11odh! + Xtaoal! — Rasach”” = Rizal") + 0" (9-03) Raat = Roa-or ~ Rar—tah’ + Roz-tale! — Xara” — Xea-tale") — Wy Rinar + h'Ria-o1 "Xuron +h" X12-a1) + (3-67) Xa = Kei — (Kora! + Xor—tala! + Rernli” + Ros-tale”) — (b’X sas + ba X aan — "Rison = be" Praai) + w" (3-03) where = ul + ju" (3-69) It will be noted that Ria — Raa = 2(Xer—na” + Xer—sa"” — Xir—aah!” — X—aile") (9-70) Xia Xe =2(—Roriah” — Ror—uaby” + Ria aaah" + Rraasls") (o-71) ‘The asymmetry in the real part of Z».4 results from terms involving the products of imaginary load currents and mutual reactances between gon- emtors and loads. ‘The asymmetry in the imaginary part of Zp.n results from terms involving the products of imaginary load currents and mu- tual resistances between generators and load. ‘The reference frame 2 equations for the general case are given by equation 3-51)80 ECONOMIC OPERATION OF POWER SYSTEMS emmy os ‘The matrix of transformation C;? is given by o Cn en} i a (3-72) Ld where dy = —1 for all values of n. By application of equations 3-22, 3-23, and 3-24, we obtain ‘The cireuit of reference frame 2, given by Figure 3.17, has been modi- fied as indicated by equations 3-59 and 3-73 to that given in Figure3.18, Hypotvetiet eae pont Figure 318. Referonoe framo 3. As noted by 3-70 and 3-71, the mutual impedances are not sym- metrical. Consequently, it is not possible to represent this equivalent cireuit on the network analyzer through the use of static circuit ele- DEVELOPMENT OF TRANSMISSION LOSS FORMULA al ments. ‘The losses in the equivalent cincuit of Figure 3.18 correspond to the losses in the transmission linos of tho original eireuit, 3.8 CALCULATION OF LOSSES ‘The real losses in the equivalent circuit of Figure 8.18 and equation 3-78 may be calculated as follows: i Py = Gly*hs (3-74) GUs*Zshy (3-75) where Bs, Is, andl Zj denote reference ftame 3 quantities and the symbol Gi denotes the real part of Is" Lot us define the real and imaginary components of tge by den and gu, respectively, ‘Thus tan = fen + Stan (3-76) ‘For the system of equation 3-60 we have a @ ea [zo — | Kas — em 3 | Bas ~ B, = % where Zan = Zom—on — Gy by +0 ‘Then (Ry —rian + Byatt + Risin — Xvriyn — Kista — Xisigs) + i(R1ater + Ri_aiga + isis + Xiattan + Xt-viae + X1stea) Basia +-Re-aita + Ra-ain — Xia — Xsiea ~ Xz-aiea) + Jlasign + Rasta + Resins + Xacvian + Xa-aima + X2-sia) (Raariar + Romain + Rosias — Xasign ~ Xesice ~ Xs-sies) + jBa—aigt + Basin + Rasta + Xacti + Xs—siea + Kastan)22 ECONOMIC OFERATION OF POWER SYSTEMS and BDZ alg = tas Rsster + ta Rviea + taxRy-ateg = tax Xa igs ~ tag X sign — tasX1—aign + tor aiat + fghsaiga + igi sige + ta X aia fgg Xmaien + fen Xi ates + tials + fealha_oien + faa-aiea — teaXatie, ~ teaXa-aiea — faaXa-sioa + tealessiat + tealta sign + igaRaiea + iaXosia, Fig Xaasin + feaXaatea + tialtat + feaRaaiaa + iags-aisa — tia Xatigt ~ taXa-aiea — tasXa—vigs + fealesik (3-78) + igXa-aian + faXs-aiva ‘Combining terms, we have Ma Bg (EBD yg (Beat Me afte ie sien + Big (“2 PE) ig ie din Byat fant) . (f= + Bs 2 faa + Bia (= Fieve + Bin ( hae (eh) igs + igo(Raa)ica 2 ii (B= = Hs se ig (PB), em It will be noted in the foregoing that only the symmetric part of the frame 3 resistances and skew-symmetric part of the frame 8 reactances contribute to a real loss. DEVELOPMENT OF TRANSMISSION LOSS FORMULA 83 A matrix is ealled symmetric if the sume components eccur on both sides of the main diagonal lino. ‘Tho main diagonal line is a line drawn from the upper left-hand corner of « matrix to the lower right hand eor- ner, ‘The following is an example of a symmetrie matrix: reali Pa] « a2 | a fs, Tt will be recalled from the definition of transpose given in Seotion 3.4 thet the transpose of a symmetric matrix is identical with the original matrix, ‘A matrix is denoted as skew-symmetric if the components along the ‘main diagonsl are zero and if the same components oceur on both sides of the main diagonal line but with opposite signs. ‘The following is an example of a skew-symmetric matrix: 12 3 af o. | +a [47 In general, a given matrix A may be divided into a symmetric matrix Band a skew-symmetric matrix C such that A=B+O where the symmetric part is given by Ata > (6-80) and the skew symmetric part is given by A-A e @-81) Tn terms of index notation we write Ap = But Op where the symmetric part is given by Ant Ay B, 284 ECONOMIC OPERATION OF POWER SYSTEMS and the skew-symmetric part is given by Thus, if thon DEVELOPMENT OF TRANSNISSION LOSS FORMULA 85 Tn the ease of a quadratic form, such as PAP or P;AsuPa, in which the clemonts of P and A are real mamhers, the matrix A may always be re- placed by the symmetric part (4 + 4,)/2, siuee the components result~ ing from the skew-symmetzic part reduce to zero. ‘Vhat is (4“)p-0 2 Vor example, consider the expression PjljuPs for j, k= 1, 23° PyA;Pu = PiAuPe + Pedals + PaAsiPe = PiAuPs + PiAaPo + PrArsPs + Poda:Py + PadaaPs + PrAnals + PrdaP + PrAwsP2 + PrAgaPs ‘The result obtained using the symmetric part of Aye is PBaPe= ‘Au + Aw ‘Aja + Aay (Ais + Agi (22) oa (A nn (At oy +, (“ <4) Prt Ps (Aai4e + A) b+ Ps Cae ff *2) Py 2 ‘Ay + Aus) ‘Aaa + A: 5 +r( * ) m+ (2) y+ p, (AAP) py 2 2 2 It is noted that Pilg Pe = PiBiPe Next consider the skew-symmetrie part of Aj. We then have =) an(test ann (tn an PrtPy ¢ tn aay Prt Ps @486 ECONOMIC OPERATION OF POWER SYSTEMS Returning to the general case in which the number of sources = m, n we have Yay = Zn —nltin + Sign) = Ronan — Xm—nfgn) + G(Rn—nign + Xan) RGdea — Figm)Zn—n{teiw + Sign) amb ntin ~ itm Xm OIs*Zals an + fgnfm rian + gn %m—nin GamBen tan + tanPennign ~ tinXmanian + in Xnantion Halas + Hamlin (82) + ian 2]. (3-83) ut m,n = 1, 2, 3, equation 3-83 reduces to equation 3-79. From equa- tion 3-73 Zim = BomOn ~ On ~ by + ‘The real symmetrical part of Zp. is given by Rann + Rare Bot ien GBZan n+ Fund = BE Zeman ~ Om — by + w+ Zan—om ~ Oy — bm + 10] (BI Zam an + Zena ~ (lu + Bn) — (Ay + by) + 20] = Ronan — dm — dy + (3-84) where a= a(t) (@-85) w' = aw (3-86) ‘The calculation of dm is first discussed in terms of the system of equation 3-60. It will be recalled from equations 3-40, 3-47, and 3-48 that 41 = Ze.tal + Zar—tale by = Zrsoa* + Zig-oale* DEVELOPMENT OF TRANSMISSION LOSS FORMULA 87 Then ant by = Zorarals + Zaa-ah* + Zartala + Zia—orb® ‘but Zora = Zrr-0r Zeit = Zia—01 Hence attr = Zia-ovlls + U*) + Zaaala +") Let We Wt a= hth!" Then 2a or(h!) + 2na-calld) and = dy = Bisel! + Ria-osle! @-s7 ‘Similarly, ty = 0 Ra cah! + Raacold 6-88) * 2 ‘Thus Psat a Bo Ratner dh = btw = Resa: — Ris—oily! — Rra—cale! = Ria-aal’ ~ Ria—oah! +" (3-89) ‘This result may also be obtained by (aking one half the sum of equations 3-65 and 3-07. — For our gencral system it will be recalled from equation 3-52 that ay = Zom—tile by = W*Zi5—cn Since the indices n, m and j, # are dummy indices, we may write by = be Zre—am = Zr—anli™ Then Oy + Oy = Zom—sile + Zee—ombi™ Since Lanta = Zie-0m we have ig +n = Ferm omlla + le) = 2ir—col! where Wy = Oly In = Racal? G90)88 ECONOMIC OPERATION OF POWER SYSTENS ‘rom equation 3-52, it will be notod that w= *Z15—24le ‘Then w= WU i5 tale (91) It is desirable to derive an altemative method of obtaining w’ which would not require the measurement of all the Zzj_n1 impedances. For a system of 60 loads the number of Z,;-14 impedances would be 3600. ‘The calculation of w’ fora system of 60 louds would thus bo prohibitively lengthy. A much shorter method of calculating w’ is discussed later. ‘The imaginary skew-symmetric port of Zp is given by -x, where # denotes imaginary part of quantity indieated. From equation B52 Om = Zan—vile = Zon sa(le! + dle) Om = Zae-anls* = Zom—caile” = Zim—talla! ~ jh") Similarly, on = Zan—talli’ + jl!) bn = Banta! — gle!) ‘Then Om — bm = eam alle) (Gm — be) = Wonca” Similarly, Han ~ a) = 2Ran—ile” the Amen ane [Sas Monn) 2 2 = [-Rom—tale!” + Ren—tade'") = [Hin + fal (3-2) where Jn = Rentals” (8-93) Su = Ran—enhe” DEVELOPMENT OF TRANSIAISSION LOSS FORMULA 39 In terms of the system of equation 3-59 we have Xia X = (Reta! + Ror—tal;") +(Rartah” + Ras—tale’’) As previously noted, the skew-symmetry in the reactanees results from terms involving the products of imaginary oad currents and mutual re- sistances between generators and loads. ‘In summary we note that the transmission losses are given by Pi = tinRm—sttn + tga ton + Phanlfm — Jolign (8-94) hero hereafter Ry» Will be understood to be the frame 3 symmetric resistance given by Ruan = Raman ~ dy — dy + 10 (4) where dy = Ram—tale! (90) Also Sn = Ramus!” (03) 3.9 TRANSFORMATION OF GENERATOR CURRENTS TO GENERATOR POWERS Equation 3-£4 expresses the transmission losses in terms of generator currents, The load dispatcher customarily works in terms of powers. Hence it is necessary for us to transform equation 3-94 to generator powers so that our expression for losses will be the most useful. ‘The steps involved in proceeding from generator currents to generator powers are described as reference frames 4, 5, and 6 by Kron. Tn this section ‘ve shall proceed directly from reference frame 8 to reference frame 6. Denote by @ the angle botwveen the reference axis and the voliage of generator m, The reference axis is the common axis upon whieh all Yoltages and currents have been projected in all our previous work. Leb Py, = real power output: of generator m. Qn = reactive power output of generator m: Yn Jbaolute valuo of tho voltage of generator m90 ECONOMIC OPERATION OF POWER SYSTEMS From Figure 3.19, it will be seen that 1 fam = (Pm 008m + On tn (3-95) -1 fom = ve [Pr $10 Bn + Qn 008 On} (8-96) Figure 3.19. Vector diagram for trigeometsie projections. To eliminate Qq as a variable, i is assumed (hat the ratio of Qn/P, will remain a constant value 8. Onin ‘Thus, equations 3-95 and 3-96 may be written 1 fam = F108 Ba + Sn 8 OP (3-97) -1 fam = T= (— sim Oy + 8 008 0g]? (8-98) Substituting equations 3-97 and 3-98 into equation (3-04), we have DEVELOPMENT OF TRANSNISSION LOSS FORMULA 91 1 P= Pa[ qe 080m + su tiny] Rane Fe Oy + tn Bit oo]. x yoo ny BIN By) a re[ tet toveund + Pm y,! ‘80 Om + fm 008 Im) | Rn x [cana + event | Fr lrsints + yet | Ps ~ 2P a (cos a+ ssn) | fe Fl x [Resins treet) [Pa (3-09) Pala “Ly + Sn Sit Oy 008 Ip SS sit On si 60a] BS (C08 Hy, C050 + $y £08 Oy i i , +l ee (in Oy sin Oy — 8 Sim Dy 608 2 ~ Sy 0 yy Dy Fy £08 Oe 0s )en-n] a 1 VnVn ~ $n 8in yy 8in By + Buby BD By €08 OF -10| Pa (8-100) Recall that 08 jn = C08 (x — On) = £08 Oy, 608 Oy -f Si By Sin Oy, +101) - 2 { C05 Oy itt By, + Sy COB By C08 Py sin Oan = #i (Gy — 04) = Sin q C08 0, ~ sin By COS Oe @-102) Substituting equations 3-102 snd 3-101 into 3-100, we have Py = PaKunkm—aPn — 2PylanPa (108) 1 where Kan [CL + Sfx) €08 Bymn + (Sm = 84) Si Ona] (3-104)92 ECONOMIC OPERATION OF POWER SYSTEMS (08 0p Sin Oy “+ 85 COS lx COS Oy —~ Sm Si Om Sin Oy + Sp Sit yy C08 8p) Fu — Jn) (8-105) ‘The calculation for ~2P lyn?» may be simplified by revalling that only the symmetrical part of Fn is required since —2P,FngPq is a quad~ ratic form, as disoussed in Soction 8.8. ‘Then Pon + Pom [ q sin Oy ++ 8 C080, C08 Oy 2 2VnVn L~ in Gy Sin Gy + 884 Sin Om CO8O,) = (= 008 Oy SiN Om + 8p C08 By C08 Oy, = 8p SiN Oy SiN By “FSS SiN By COB Oy) Substituting equations 3-101 and 3-102 into 3-106, Fant Pon 1 Um fn) (3-106) [CL + tmnt) sit Opin + (8 — 892) €08 Oran] (fon — Sn) 2 BWV = Mania fu) 107) where Ha = >—[(l + S48) St Gyn + (Be — fm) 608 na) 108) Va¥n ‘Then oquation 3-108 becomes + Pom Pi PnKankennPn — 2p "5 —" Py (8-109) = PuKrnBm—nPn ~ Pallan(fn —In)Pn @-110) = PaBunPa where Brn = KrauBmn — Hrn(fa — fn) (111) Ti the term Hran(fa — fx) is neglected, the expression for Buy reduces to Burn = Kone G12) ‘The conditions for which Hnn(fm — fn) may be negligibly small are dis- euated in Seotion 4.3 of Chapter 4, ‘The equivalent circuit corresponding to equation 3-110 is given in Figure 3.2. We now have impressed generator powers instead of genera tor currents. ‘The Byn represent an equivalent loss notworic through which the genentor powers flow in supplying the over.all system load. Since Bun = Bryn, the number of loss-formula coefficients to be calcu- lated for a Toss formula with n sourees is [n(n + 1))/2. noel ey DEVELOPMENT OF TRANSMISSION LOSS FORMULA 9 The relative magnitudes of the Bra terms ean be estimated from a physical knowledge of the transmission system. For example, those sources which are the greatest distance from the system load will have the langest valve of the self Byq terms. Sourees close to each other usually have positive mutual terms, and sources on opposite ends of the system usually have negative mutusls, In any given row or column of coefficients the self term is always positive and generally the largest positive number in that row or column, 3.10 DETERMINATION OF w’ From equations 8-111 and 8-84 we may write Bun = Krunln—n — Hons m = fa) = Kun(Rom—an — du — da + 0!) ~ Half ~ fn) = Kaan(Ram—an — dn ~ dy) + Kennto! — Hran(Sin ~ Sa) = Aan + Knut’ — Hina (Sm — fn) (@-113) where Ann = Knn(Rom—on — dn — dn) (3-114) ‘The losses from the loss formula may be equated to the base case losses obtained by summing the #22; losses in all the lines where ‘ig = sealar current in line k Ry = resistance of line k: Dw Re = PaBxPn PeAnmPn + PaKnntt!Py — PnBaa loa — Solving for w’, we obtain (3-115) Neglecting the effect of Hax(Su — fx), the expression for 1! becomes: DaPRe — PaAnnPo PuKnaPa Calculation of w! in this manner, as given by equation 8-115 or 8-116, eliminates the wi @-116)