mcgrawnLcom/sites/O0 o 4g4gg4 7
s
New chapters added on ) ) ) ) )
i
Fr
'Ju;;
utilrl ruwEt
DaEir4iE
va
r9 rd
TI
t
Rm T r l r u t r t g r l , Anllueis ryrrr
€ co o' FT
il
fv =
A
rt
In simple language, book providesa modernintroduction power the to system operation,controland analysis. Key Features of the Third
sr o =, o o
o
Ir
€
.PowerSystemSecurity State Estimation Powersystem compensationincluding svs and FACTS Load Forecasting VoltageStability
Third Editior;
New appendices on : > MATLABand SIMULINKdemonstrating their use in problemsolving. ) Real time computercontrol of power systems.
L' r
Ff
llf
r\(J lltd,fl
J( ^r*I
^..!
lJ Nagrath
From
the
Reviewen.,
The book is very comprehensive, well organised,uptodateand (above all) lucid and easy to follow for selfstudy.lt is ampiy illustrated w1h solved examplesfor everyconceptand technique.
lMr
ll=i
iltililttJffiillililil
Modern PorroerSYstem
Third Edition
About the Authors
D P Kothari is vice chancellor, vIT University, vellore. Earlier,he was Professor, Centre for Energy Studies, and Depufy Director (Administration) Indian Instituteof Technology, Delhi. He has uiro t."n the Head of the centre for Energy Studies(199597)and Principal(l gg7g8),Visvesvaraya Regional Engineeringcollege, Nagpur. Earlier lflazs: and 19g9),he was a visiting fellow at RMIT, Melbourne, Australia. He obtained his BE, ME and phD degreesfrom BITS, Pilani. A fellow of the Institution Engineers (India), prof. Kothari has published/presented 450 papers in national and international journals/conferences.He has authored/coauthored more than 15 books, including Power system Engineering, Electric Machines, 2/e, power system Transients, Theory and problems of Electric Machines, 2/e., and. Basic Electrical Engineering. His researchinterestsinclude power system control, optimisation,reliability and energyconservation. I J Nagrath is Adjunct Professor,BITS Pilani and retired as professor of Electrical Engineeringand Deputy Director of Birla Institute of Technology and Science, Pilani. He obtained his BE in Electrical Engineering from the university of Rajasthanin 1951 and MS from the Unive.rity of Wi"sconsin in 1956' He has coauthored several successful books which include Electric Machines 2/e, Power system Engineering, signals and systems and.systems: Modelling and Analyns. He has also puulistred ,"rr.ui research papers in prestigiousnationaland international journats.
Modern Power System Analysis
Third Edition
D P Kothari Vice Chancellor VIT University Vellore Former DirectorIncharge, IIT Delhi Former Principal, VRCE, Nagpur
I J Nagrath Adjunct Professor, and Former Deputy Director, Birla Ins1i1y7"of Technologt and Science Pilani
Tata McGraw Hill Education private Limited
NEWDELHI
McGrawHill Offices
New Delhi Newyork St Louis San Francisco Auckland Bogot6 Caracas KualaLumpur Lisbon London Madrid Mexico city Milan Montreal San Juan Santiago Singapore Sydney Tokyo Toronto
Preface to the Third Edition
Sincethe appearance the secondedition in 1989,the overall energy situation of has changed considerably and this has generatedgreat interest in nonconventionaland renewableenergy sources,energyconservationand management, power reforms and restructuringand distributedarrddispersed generation. Chapter t has been therefore,enlargedand completely rewritten. In addition, the influences of environmentalconstraintsare also discussed. The present edition, like the earlier two, is designed for a twosemester course at the undergraduate level or for firstsemesterpostgraduate study. Modern power systemshave grown larger and spreadover larger geographical areawith many interconnections betweenneighbouringsystems.Optimal planning,operationand control of such largescale systemsrequire advanced computerbased techniques many of which are explainedin the studentoriented and readerfriendlymannerby meansof numericalexamplesthroughout this book. Electric utility engineers will also be benefittedby the book as it will preparethem more adequatelyto face the new challenges. The style of writing 'Ihe is amenable selfstudy. to wide rangeof topicsfacilitates versarile selection of chaptersand sectionsfbr completion in the semester time frame. Highlights of this edition are the five new chapters.Chapter 13 deals with power system security. Contingency analysis and sensitivity factors are described. analytical framework is developedto control bulk power systems An in sucha way that securityis enhanced. Everythingseems have a propensity to to fail. Power systemsare no exception.Power systemsecuritypracticestry to control and operatepower systemsin a defensivepostureso that the effects of theseinevitable failures are minimized. Chapter 14 is an introduction to the use of stateestimationin electric power systems.We have selectedLeast SquaresEstimationto give basic solution. External system equivalencing and treatment of bad data are also discussed. The economics of power transmissionhas always lured the planners to transmit as much power as possible through existing transmission lines. Difficulty of acquiring the right of way for new lines (the corridor crisis) has always motivated the power engineersto develop compensatorysystems. Therefore, Chapter 15 addresses compensationin power systems.Both series and shunt compensationof linqs have been thoroughly discussed. Concepts of SVS, STATCOM and FACTS havcbeen briefly introduced. Chapter 16 covers the important topic of load forecasting technique. Knowing load is absolutelyessentialfor solving any power systemproblem. Chapter 17 dealswith the important problem of voltagestability.Mathematical formulation, analysis, stateofart, future trends and challenges are discussed.
Information contained in this work has been obtained by Tata McGrawHill, from sources believed to be reliable. However, neither Tata McGrawHill nor its authors guaranteethe accuracy or completenessof any information published hereiir, and neittier Tata McGrawHill nor its authors shall be responsiblefor any errors, omissions, or damages arising out of use of this information. This work is publi'shedwith the that Tata McGrawHill and its authorsare supplying information but are understanding not attempting to render enginecring or other professionalservices. If such seryicesare required, the assistanceof an appropriate professional should be sought
TataMcGrawHill
Private Iimited TataMcGrtrwI{ill Education O 2003,1989,1980, reprint2009 Sixteenth RCXCRRBFRARBQ in can No part of this publication be reproduced any form or by any "un, of withoutthe prior writtenpermission the publishers from India only by the publishers, This edition can be exported PrivateLimited TataMcGraw Hill Education ISBN13: 9780070494893 4 ISBN 10: 007049489 PrivateLimited, Published TataMcGrawHill Education by in 7 WestPatelNagat New Delhi I l0 008, typeset TimesRomanby Script Makers, at Vihar, New Delhi ll0 063 andprinted Paschim Al8, ShopNo. 19, DDA Markct, Delhi ll0 053 Enterprises, Gopaljee Coverprinter:SDR Printcrs
Wl
Preracero rne lhlrd Edrtion
are analysis for MATLAB andSIMULINK, idealprograms powersystem examples illustrating alongwith 18solved in included thisbookasan appendix
tem problems. The help rendered tive theiruse in solvin by Shri Sunil Bhat of VNIT, Nagpur in writing this appendix is thankfully acknowledged. Tata McGrawHill and the authors would like to thank the following reviewers of this edition: Prof. J.D. Sharma,IIT Roorkee; Prof. S.N. Tiwari, MNNIT Allahabad; Dr. M.R. Mohan, Anna University, Chennai; Prof. M.K. Prof.P.R. Bijwe PEC, Chandigarh; BITS, Pilani; Dr. H.R. Seedhar, Deshmukh, Dr. SanjayRoy, IIT Delhi. and While revising the text, we have had the benefit of valuable advice and and practising engineerswho used from many professors,students suggestions the earlier editions of this book. All these individuals have influenced this to our edition.We express thanks and appreciation them. We hope this support/ would continue in the future also. response D P Kors[m I J Nlcn+rn
Preface to the First
Mathematical modellingand solutionon digital computers the only practical is approach to systems analysis and planning studies for a modern day power system with its large size, complex and integrated nature. The stage has, therefore,been reachedwhere an undergraduate must be trained in the latest techniques analysisof largescale of power systems. similar needalso exists A in the industry wherea practisingpower systemengineeris constantly facedwith the challengeof the rapidly advancingfield. This book has bedndesignedto fulfil this need by integratingthe basic principlesof power systemanalysisillustrated through the simplestsystemstructurewith analysistechniques practical size for systems. this book largescale In systemanalysisfollows as a naturalextension of the basicprinciples.The form and level of someof the wellknown techniques are presented in such a manner that undergraduates can easily grasp and appreciatethem. The book is designedfor a twosemester course at the undergraduate level. With a judicious choice of advancedtopics, some institutionsmay also frnd it useful for a first course for postgraduates. The readeris expectedto have a prior grounding in circuit theory and electrical machines. He should also have been exposed to Laplace transform, linear differential equations, optimisation techniquesand a first course in control theory. Matrix analysisis applied throughoutthe book. However, a knowledge of simple matrix operations would suffice and these are summarisedin an appendixfbr quick reference. The digital computerbeing an indispensable tool for power systemanalysis, computationalalgorithms for various systemstudiessuch as load flow, fault level analysis,stability, etc. have been included at appropriateplacesin the book. It is suggested that where computerfacilities exist, students shouldbe encouraged to build computer programs for these studies using the algorithms provided. Further, the students can be asked to pool the various programs for more advancedand sophisticated studies,e.g. optimal scheduling. importantnovel An featureof the book is the inclusion of the latestand practicallyuseful topics like unit commitment, generation reliability, optimal thermal scheduling,optimal hydrothermalschedulingand decoupledload flow in a text which is primarily meantfor undergraduates. The introductory chapter contains a discussion on various methods of electricalenergygenerationand their technoeconomic comparison. glimpse is A given into the future of electricalenergy.The readeris alsoexposed the Indian to power scenariowith facts and figures. Chapters2 and3 give the transmission line parameters theseare included and for the sakeof completness the text. Chapter4 on the representation power of of gives the steadystatemodelsof the synchronous systemcomponents machineand the circuit modelsof compositepower systemsalong with the per unit method.
preface ro rhe Frrst Edition W Chapter5 deals with the performanceof transmissionlines. The load flow problem is introducedright at this stagethroughthe simple twobus systemand basicconceptsof watt and var control are illustrated.A brief treatmentof circle
concept of load flow and line compensation. ABCD constants are generally well covered in the circuit theory course and are, therefore, relegated to an appendix. Chapter 6 gives power network modelling and load flow analysis, while Chapter 7 gives optimal system operation with both approximate and rigorous treatment. Chapter 8 deals with load frequency control wherein both conventional and modern control approaches have been adopted for analysis and design. Voltage control is briefly discussed. Chapters 9l l discuss fault studies (abnormal system operation). The synchronous machine model for transient studies is heuristically introduced to the reader. Chapter l2 emphasisesthe concepts of various types <lf stability in a power system. In particular the concepts of transient stability is well illustrated through the equal area criterion. The classical numerical solution technique of the swing equation as well as the algorithm for large system stability are advanced. Every concept and technique presented is well supported through examples employing mainly a twobus structure while sometimes three and fourbus illustrations wherever necessary have also been used. A large number of unsolved problems with their answers are included at the end of each chapter. These have been so selected that apart from providing a drill they help the reader develop a deeper insight and illustrate some points beyond what is directly covered by the text. The internal organisation of various chapters is flexible and permits the teacher to adapt them to the particular needs of the class and curriculum. If desired, some of the advanced level topics could be bypassed without loss of continuity. The style of writing is specially adapted to selfstudy. Exploiting this fact a teacher will have enough time at his disposal to extend the coverage of this book to suit his particular syllabus and to include tutorial work on the numerous examples suggestedin the text. The authors are indebted to their colleagues at the Birla Institute of Technology and Science, Pilani and the Indian Institute of Technology, Delhi for the encouragement and various useful suggestionsthey received from them while writing this book. They are grateful to the authorities of the Birla lnstitute of Technology and Science, Pilani and the Indian Institute of Technology, Delhi for providing facilities necessary for writing the book. The authors welcome any constructive criticism of the book and will be grateful for any appraisal by the readers. I J NlcRArH D P KorHlnr
Contents
Preface to First Edition vn
1 . Introduction
I
r.4
1 . 1 A Perspective I 1 . 2 Structureof Power Systems I0 1 . 3 Conventional Sourcesof Electric Energy I3
RenewableEnergy Sources 25 29
1 . 5 Energy Storage 28 1 . 6 Growth of Power Systemsin India 1 . 7 Energy Conservbtion 3I
r.8
Deregulation 33
1 . 9 Distributed and DispersedGeneration 34 1.10 Environmental Aspects of Electric Energy Generation 35 1.11 Power SystemEngineersand Power SystemStudies 39 T . I 2Use of Computers and Microprocessors 39 1.13 ProblemsFacing Indian Power Industry and its Choices 40
References 43 2. Inductance and Resistance of Transmission Lines
45
2 . 1 Introduction 45 2 . 2 Definition of Inductance 45 2 . 3 Flux Linkages of an Isolated
CurrentCtrryingConductor 46
2.4 Inductanceof a SinglePhase TwoWire Line 50 2 . 5 ConductorTypes 5I 2 . 6 Flux Linkages of one Conductorin a Group 53 2 . 7 Inductanceof CompositeConductorLines 54 2 . 8 Inductanceof ThreePhase Lines 59 2 . 9 DoubleCircuitThreePhaseLines 66 2 . 1 0Bundled Conductors 68 2 . l I Resistance 70 2 . r 2 Skin Effect and Proximity Effect 7I
Problems 72 References 75
3. Capacitance of Transmission Lines
3.1 3.2 Introduction 76 Electric Field of a Long Straight Conductor 76
76
fW
3.3 3.4 3.5
. contents
Potential Diff'erencebetweentwo Conductors of a Group of Parallel Conductors 77 Capacitance a TwoWire Line 78 of Capacitanceof a ThreephaseLine with Equilateral Spacing B0 UnsymmetricalSpacing BI Effect of Earth on TransmissionLine capacitance g3
rvleln(Jo or \rlvll, (vloollled)
l t  t l   l a r . ^ / r .
.
6.4 6.5 6.6 6.7 Load Flow Problem 196 GaussSeidelMethod 204 (NR) Method 213 NewtonRaphson DecoupledLoad Flow Methods 222
3.7
J.o "
o
6.9 Control of Voltage Profile 230 Problems 236
IIYJEIETLLCJ
yl
D^t^.
LJ7
)20
3.9
BundledConductors 92 Problems 93 References 94 95
7. Optimal System Operation 7.I 1.2 7.3 7.4 1.5 7.6 7.7 Introduction 242 on Optimal Operation of Generators a Bus Bar 243 Optimal Unit Commitment (UC) 250 ReliabilityConsiderations 253 Optimum GenerationScheduling 259 Optimal Load Flow Solution 270 of System 276 OptimalScheduling Hydrothermal Problems 284 References 286
242
4. Representation,of Power System Components 4.1 Introduction g5 4.2 Singlephase Solutionof Balanced Threephase Networks 95 4.3 OneLineDiagram and Impedanceor Reactance Diagram 98 4.4 Per Unit (PU) System 99 4.5 ComplexPower 105 4.6 Synchronous Machine 108 4.7 Representation Loads I2I of Problems 125 References 127 5. Characteristics and Performance of power Transmission Lines 5.1 5.2 5.3 5.4 5.5 5.6 5.1 5.8 5.9 5 .1 0 Introduction 128 Short Transmission Line 129 Medium Transmission Line i37 The Long Transmission LineRigorous Solution I 39 Interpretationof the Long Line Equations 143 FerrantiEffect 150 TunedPowerLines 151 The Equivalent Circuit of a Long Line 152 Power Flow througha Transmission Line I58 Me th o d s l ' V o l ra g e o n trol 173 o C Problems 180 References 183
8. Automatic Generation and Voltage Control 8.1 Introduction 290 8.2 Load FrequencyControl (Single Area Case) 291 8.3 Load FrequencyControl and Economic DespatchControl 305 8.4 TwoArea Load FreqlrencyControl 307 8 . 5 Optimal (TwoArea) Load FrequencyControl 3I0 8 . 6 Automatic Voltage Control 318 8 . 7 Load Frequency Control with Generation Rate Constraints(GRCs) 320 8 . 8 SpeedGovernor DeadBandand Its Effect on AGC 8 . 9 Digital LF Controllers 322 8 . 1 0DecentralizedControl 323 Prohlents 324 References 325 9. Symmetrical Fault Analysis 9.1 Introduction 327 Line 328 9.2 Transienton a Transmission Machine 9.3 ShortCircuit of a Synchronous (On No Load) 330 Machine 339 9.4 Short Circuit of a LoadedSynchronous of Circuit Breakers 344 9.5 Selection
291'l
128
321
327
6. Load Flow Studies 6.1 6.2 lntrotluction 184 NetworkModel Formulation I85
t84
rffi#q
I
confenfs
9.6 Algorithm for ShortCircuit Studies 349 9.7 Zsus Formulation 355 Problems 363 References 368 '
Symmetrical Com 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9 Introduction 369 SymmetricalComponentTransformation370 Phase Shift in StarDeltaTransformers 377 Sequence Impedances TransmissionLines 379 of Sequence Impedances Sequence and Network of Power Systern 381 Sequence Impedances Networks of and Synchronous Machine 381 Sequence Impedances TransmissionLines 385 of Sequence Impedances Networks and of Transformers 386 Constructionof Sequence Networks of a Power System 389 Problems 393 References 396
12.10MultimachineStabilitv 487 Problems 506 References 508 13. Power System Security 13.1 13.2 13.3 13.4 13.5 13.6 Introduction 510 SystemStateClassification 512 SecurityAnalysis 512 Contingency Analysis 516 SensitivityFactors 520 Power System Voltage Stability 524 References 529 531
14. An Introduction to state Estimation of Power systems l4.l Introduction 531 I4.2 Least SquaresEstimation: The Basic Solution 532 14.3 Static StateEstimation of Power Systems 538 I4.4 Tracking State Estimation of Power Systems 544 14.5 SomeComputational Considerations 544 14.6 External System Equivalencing 545 I4.7 Treatmentof Bad Dara 546 14.8 Network observability and PseudoMeasurementss49 14.9 Application of Power SystemStateEstimation 550 Problems 552 References 5.13 15. Compensation in Power Systems 433 15.1 Introduction 556 15.2 Loading Capability 557 15.3 Load Compensation 557 15.4 Line Compensation 558 15.5 SeriesCompensation 559 15.6 ShuntCornpensators 562 I5.7 ComparisonbetweenSTATCOM and SVC 565 15.8 Flexible AC Transmission (FACTS) 566 Systems 15.9 Principle and Operationof Converrers 567 15.10FactsControllers 569 References 574
ll.
Unsymmetrical Fault Analysis 11.1 Introduction 397 11.2 SymmetricalComponentAnalysis of UnsymmetricalFaults 398 11.3 Single LineToGround (LG) Fault 3gg , 11.4 LineToLine (LL) Fault 402 11.5 Double LineToGround (LLG) Fault 404 11.6 Open Conductor Faults 414 11.1 Bus Impedance Matrix Method For Analysis of Unsymmetrical ShuntFaults 416 Problems 427 References 432
397
550
12. Power System Stability 12.1 12.2 12.3 12.4 I2.5 12.6 12.7 I2.8 Introduction 433 Dynamics of a Synchronous Machine 435 Power Angle Equation 440 Node Elimination Technique 444 SimpleSystems 451 Steady State Stability 454 Transient Stability 459 Fq'ralArea Criterion 461
16. Load Forecasting Technique 16.1 Introduction 575 16.2 Forecasting Methodology 577 timation of Averageand Trend Terms 577 1 6 . 4 Estimation of Periodic Components 581 1 6 . 5 Estimationof y., (ft): Time SeriesApproach 582 1 6 . 6 Estimationof Stochastic Component: Kalman Filtering Approach 583 16.7 LongTerm Load Predictions Using EconometricModels 587 r 6 . 8 ReactiveLoad Forecast 587 References 589 17. Voltage Stability 1 1 .1 17.2 17.3 11.4 In tro d u c ti o n 5 9 1 Comparisonof Angle and Voltage Stability 592 ReactivePower Flow and Voltage Collapse 593 MathematicalFormulationof Voltage Stability Problem 593 11.5 Voltage Stability Analysis 597 17.6 Preventionof Voltage Collapse 600 ll.1 StateoftheArt,Future Trends and Challenses 601 References 603 591 I.T A PERSPECTIVE
Appendix A: Introduction to Vector and Matrix Algebra Appendix B: Generalized Circuit Constants Appendix C: Triangular Factorization and Optimal Ordering Appendix D: Elements of Power System Jacobian Matrix Appendix E: KuhnTucker Theorem Appendix F: Realtime Computer Control of power Systems Appendix G: Introduction to MATLAB and SIMULINK Answers to Problems Index
605 617 623 629 632 634 640
Electric energy is an essentialingredient for the industrial and allround development any country.It is a covetedform of energy,because can be of it generated centrallyin bulk and transmittedeconomically over long distances. Further, it can be adaptedeasily and efficiently to domestic and industrial applications, particularly for lighting purposes and rnechanical work*, e.g. drives. The per capita consumptionof electricalenergyis a reliable indicator of a country's stateof developmentfigures for 2006 are615 kwh for India and 5600 kWh for UK and 15000 kwh for USA. Conventionally,electricenergy is obtainedby conversion fiom fossil fuels (coal,oil, naturalgas),and nuclearand hydro sources. Heatenergyreleased by burning fossil fuels or by fission of nuclearmaterialis convertedto electricity by first converting heatenergyto the mechanical form througha thermocycle and then converting mechanicalenergy through generators the electrical to form. Thermocycleis basicallya low efficiency processhighestefficiencies for modern large size plantsrange up to 40o/o, while smallerplants may have considerably lower efficiencies. The earth has fixed nonreplenishableresourcesof fossil fuels and nuclear materials,with certain countries overendowedby natureand othersdeficient. Hydro energy,thoughreplenishable,is also limited in terms of power. The world's increasingpower requirementscan only be partially met by hydro sources.Furthermore,ecologicaland biological factorsplace a stringentlimit on the use of hydro sources power production. for (The USA has already developed around 50Vo of its hydro potential and hardly any further expansionis plannedbecause ecological of considerations.)
x Electricity is a very inefficient agent for heating purposes, becauseit is generatedby the low efficiency thermocycle from heat energy. Electricity is used for heating purposesfor only very special applications, say an electric furnace.
Introduction with the ever increasingper capita energyconsumptionand g x p o n g n l l a _ _ _ __ _ _   ^ D v^,vt6J vurrJLurlp[rult iltlu exponentially rising population, technologists already r* the end of the earth,s n o n s nc fuel reso.urces*. oil crisis of the 1970s has dramatically llfenislable The drawn attentionto this fact. In fact,we can no lon tor generation of electricity. In terms of bulk electric energy generation, a distinct shift is taking place acrossthe world in favour of coalLJin particular intense pollution in their programmes of energy development.Bulk power generating stations are more easily amenable to control of pollution since centralized onepoint measurescan be adopted. to on consumption a worldwidebasis.This figure is expected rise as oil supply can be expected to for industrial usesbecomesmore stringent. Transportation when nonconventionalenergy go electric in a big way in the long run, resourcesare we[ developedor a breakthroughin fusion is achieved. To understand some of the problems that the power industry faces let us briefly review some of the characteristic features of generation and transmission. Electricity, unlike water and gas, cannot be storedeconomically (except in very small quantitiesin batteries),and the electric utility can exerciselittle control over the load (power demand) at any time. The power system must, therefore, be capable of matching the output from generatorsto the demand at in The difficulty encountered this voltageand frequency. any time at a specified task can be imagined from the fact that load variations over a day comprises three componentsa steady component known as base load; a varying componentwhose daily patterndependsupon the time of day; weather, season, a popular festival, etc.; and a purely randomly varying componentof relatively small amplitude. Figure 1.1 shows a typical daily load curve. The characteristics of a daily load curve on a gross basis are indicatedby peak load and the time of its occurrence and load factor defined as averageload = less than unity maximum (peak)load
Cufiailment
of enerry
consumption
The energyconsumptionof most developcdcorrntries has alreaclyreachecl a level, which this planet cannot afford. There is, in fact, a need to find ways and meansof reducingthis level. The developingcountries,on the other hand,have to intensify their efforts to raisetheir level of energyproduction to provide basic amenities to their teeming millions. of course,in doing ,o th"y need to constantly draw upon the experiencesof the developed countries and guard againstobsolete technology. rntensification of effofts to develop alternative sources of enerw including unconventional sources like solan tidal energy, etc. Distant hopesare pitched on fusion energy but the scientific and technological advanceshave a long way to go in this regard. Fusion when harnessed could provide an inexhaustiblesourceof energy. A breakthrough in the conversion from solar to electric energy could pr*io" another answer to the world,s steeply rising energy needs. Recyclingr of nuclear wastes
B
o tr 6 :o
100
E B o c 6 0
E
Fast breederreactortechnologyis expectedto provide the answer for extending nuclear energy resourcesto last much longer. D e velopm ent an d applicati on of an ttpollu tion techn ologries
l + o
x o
In this regard, the developing countries already have the example of the developed countries whereby they can avoid going through the phases of
*varying estimatcs havebccn put forth for rescrvcs ol'oil, gas and coal ancllissionable rnaterials'At the projectedconsumptionrates,oil and gases are not expectedto last much beyond 50 years; severalcountries will face serious shortagesof coal after 2200 A'D' while fissionable materials may carry us well beyond the middle of the next century. These estimates,however, cannot be regarded as highly dependable. Fig. 1.1 Typical daily load curve The average load determines the energy consumption over the day, while the peak load along with considerations of standby capacity determines plant capacity for meeting the load.
mterconnection,rgreajly in jackinguF tn,a factors aids i at er in&viJJp excess Powerof a plant audng tight"toaa periodsis evacuated throughlong distance high voltage transmissionlo"r,*hl" h"""itr;;;pj;;,:;";# rwervcs '"*Prur power. " Diversity Factor
;;;ilnffi generalhave a time dJveniry,which when"r,#u","'.i.,ii, ;61il ;&;';;"irili,jij
Asindividuar centris rJir o*n road have
. A h i g h l o a d f a c tuvrl,; e l p s i n d rmoren g m o r e e n e r g y * ' n u , ' u " n , n , . u * I ' , i o r h ur (xawrrg a w energy lrom a given installation. I I
i 
I

Tt:"9:"lT:f:,_Tj_.:llT:11,:::."1:,1.j:ij:::.,j:j:*T:j wages and on and o:rynd..o_ltheunitsproduced therefore th€tuel charges the
of the.station .staff. the may Tariff structures be suchasto influence loadcurveandto improve s"j:9.fi",.:''
Tariff should consider the pf (power factor) of the load of the consumer. If it is low, it takes more current for the same kWs and hence Z and D
]
i
i;;;;.i."
and distribution) losses are conespondingly increased. The
more. Luckily, rhefactoris much highertrranunity, loads. "il*t,
transmission rf authedemands ;;" prant. i:"::19'c,.r"d ru"to', i.e. divenitv unitv "r"
the maximum on the ::m:r,:_".*'Ji:[1t z%""l,i""irT3H::Tgl""rii:'J],f;fi:_'J""1tr dividedas sum individual demands consumers, I by **iui;#"#UH:ffiTTfi,1"ff":"il:ffi:i ,r," of :#:*'::i,:"ff"Hf,.j,f.',.d1:$:"*Yil'.:fff:fii*,:.;iff; ,."'.::"'?; ;3:":ffii"T,"jT,ffiH: i:s,?" 
rhisisdenned
::::*1":i""1".:1t:i1;ft;,"
tr,"'iotJ ;;';";;;il.;;;;;;ilTffi?
,;"'11'f,;:
i i
i
ili"tH * ;xl#1rJ,ffi?,1?Til'jl[,"ff"tJ'trJ; tr;'3f:l :,':T":."::i,"":"^::::":":J _ ;;ff; ;;"";;'lil mtharee,lfe may be imposedon the consumer. ::"T:l:_^,:Tl it1:Y_T^:* tlil a pf penaltyclause
po*er factor of his installations.
f;;#;
A high diversity factor could be obtained bv; 1' Giving incentivesto farmers and/or some industries to use electricity in the r u E i r r or lighr load periods. u r u night u r U B I  L r o a o p e f l o o s . 2 using daylight saving as in many other counfies. 3' staggering the offrce timings 4' Having different time zones in the country like USA, Australia, etc. 5' Having twopart tariff in which consumer has to pay an amount dependent on the maximum demand he makes, plus u fo. unit of energyconsumed.sometimes consumer "h.g; "u"t ii charged o? tt" u"si, of kVA demand instead of kW to penalize to"O. of to'* lo*", tin"tor. other factors used frequently are: enuy 3re: plant capacity foctor
I ,
I
(iiD the consumermay be asked to use shunt capacitorsfor improving the
Llg4" L
1'1
L
A factory to be set up is to have a fixed load of 760 kw gt 0.8 pt. The electricrty board offeri to supplJ, energy at the following alb;ate rates: (a) Lv supply at Rs 32ftvA max demand/annum + 10 paise/tWh (b) HV supply at Rs 30/kvA max demand/annum + l0 paise/kwh. rne lrv switchgear costs Rs 60/kvA and swirchgear losses at full load i amount to 5qa Intercst depreciation charges ibr the snitchgear arc l29o of the I capital cost. If the factory is to work for 48 hours/week, determine the more e.conomical tariff. Actual energyproduced 7@ m a x i m u m p o s s i b l e e ' m s o f u t i o n M a x i m u m d e m a n d = 0 3 = 9 5 0 k v A on @ased instarelptant capaciiyy Loss in switchgear= 5% .. InPut dematrd= 950 j= 1000 kvA
_ Average demand Installedcapacity
Plant usef(tctor Actual energyproduced _ _ _ (kWh) . . ,: plant capacity (kw) x Time (in hours)th" plunrh^
Tariffs
I ' b;i" il;;ti""
= of switchgear 60 x 1000 = Rs 60,000 "ost = on Annual charges degeciation 0.12 x 60,000= Rs 7,200
Annual fixed chargesdue to maximum demand correspondingto tariff (b)
= = 30 x 1.000 Rs 30,000 Annual running chargesdue to kwh consumed = 1000x 0.8 x 48 x 52 x 0.10
= Rs 1. 99. 680
The cost of electric power is normally given by the expression(a + D x kW + c x kWh) per annum,where 4 is a rixea clarge f_ ,f," oiifif,'ina"p".a"* of the power output; b depends the maximum on demandon tir" syrie ano
i
Introduction
t
WI
I
.'.600 +0.03 or
dP
4Etooo.r3dE o =
dP
Total charges/annum Rs 2,36,gg0 = Max. demand corresponding tariff(a) j to 950 kVA Annual running chargesfor kWh consumed
l"'
dE=3m dP dE=dPxt
From triangles ADF and ABC,
=950x0.8x48x52x0.10 = Rs t,89,696 Total= Rs 2,20,096
Therefore, tariff (a) is economical.
5,00,000P _ 3000 8760 5,00,000 P = 328,say 330 MW Capacityof thermalplant= 170MW
Energy generatedby thermal plant =
170x3000x1000
A region has a maximum demand of 500 MW at a road factor of 50vo. The Ioad duration curve can be assumedto be a triangle. The utility has to meet this load by setting up a generating system, which is partly hydro and partry thermal. The costs are as under: Hydro plant: Rs 600 per kw per annum and operating expenses 3p at per kWh. Thermal plant: Rs 300 per kw per annum and operating expenses r3p at each, and overall :ffily:f generation cost per kWh. Solution
= 255 x106 kwh Energy generatedby hydro plant = 1935 x i06 kwh Total annual cost = Rs 340.20 x 106/year
generation = cost overall
###P
x 100
= 15.53paise/kWh
Determine the
hydroprT!, rheenergy generated annually by
A generatingstation has a maximum demandof 25 MW, a load factor of 6OVo, a plant capacity factor of 5OVo,and a plant use factor of 72Vo.Find (a) the 'the reserve capacity of the plant, and (c) the daily energy produced, (b) maximum energy that could be produceddaily if the plant, while running as were fully loaded. per schedule, Solution average demand Load factor = maximumdemand average demand 25 Average demand= 15 MW 0.60 = Plant capacity factor = demand average ;#;;..0".,,, l5
Total energy generated per year = 500 x 1000 x 0.5 x g760  219 x 10' kwh Figure 1.2 shows the load duration curve. Since operatingcostof hydro plant is low, the base load would be supplied from the hydro plant and peak load from 500,000 P the thermal plant. Ler the hydro capacity be p kW and the energy generaredby hydro plant E kWh/year. Thermal capacity = (5,00,000_ p) kW Thermal energy = (2lg x107 _ E) kwh B Hours afoo 0 Annual cost of hydro plant =600P+0.03E Fig. 1.2 Load duration curve Annual cost of thermarplant = 300(5,00,000 p) + 0.r3(zrg x r07_ n) Total cost C = 600p + 0.038 + 300(5,00,000 p) _
I
0.50=
installed capacity
+ 0.t3(219 x 107 E) _ For minimum cost, 4Q = 0 dP
= capacity += 30 MW Installed 0.5
Introduction =3025=5MW Daily energyproduced flver&g€ = demand 24 = 15 x 24 x = 360 MWh Energycorresponding installed to capacity day per =24x30_720MWh
axlmum energy t be produced _ actualenergyproducedin a day plant use factor Reserve capacity of the plant = instalredcapacity  maximum demand
N
I
1 . 0 7 5 x 1 0 E x 1 0 6135%o = 0 The annual load factor of the total plant = x 20,000 8760
CommentsThe various plant factors, the capacity of base and peak load units can thus be found out from the load duration curve. The load factor of than that of the base load unit, and thus the generation from the peak load unit is much higher than that cosf of power from the baseload unit.
= :9 = 5ooMWh/day 0.72
From a load duration curve, the folrowing data are obtained: Maximum demandon the sysremis 20 Mw. The load suppliedby the two units is 14 MW and 10 MW. Unit No. 1 (baseunit) works for l00Vo of the time, and Unit No. 2 (peak load unit) only for 45vo of the time. The energy g e n e r a t e d b y u n iIt i s 1 x 1 0 8 u n i t s , a n d t h a t b y u n iz i s 7 . 5 x 1 0 6 u n i t s . F i n d t the load factor, plant capacityfactor and plant use factor of each unit, and the load factor of the total plant. Solution Annual load factor for Unit 1 =
14,000 8760 x
1 x 1 0 8 x 1 0 0:81.54V o
The maximumdemand Unit 2 is 6 MW. on 6000 8760 x Load factor of Unit 2 for the time it takesthe load 6 Annual load factor for Unit 2 = 7 . 5 x 1 0x 1 0 0 = 14.27Vo
There are three consumersof electricity having different load requirementsat different times. Consumer t has a maximum demand of 5 kW at 6 p.m. and a demand of 3 kW at 7 p.m. and a daily load factor of 20Vo.Consumer 2 has a maximum demandof 5 kW at 11 a.m.' a load of 2 kW at 7 p'm' and an averageload of 1200 w. consumer 3 has an averageload of I kw and his (b) maximum demandis 3 kW at 7 p.m. Determine:(a) the diversity factor, the load factor and averageload of each consumer, and (c) the average load and load factor of the combined load. Solution LF 3kw MD5KW (a) ConsumerI ZOVo atTpm at6pm Average load 2kw MD5KW Consumer2 1\2 kW atTpm at 11 am Average load MD3KW Consumer3 1kw atTpm Maximum demand of the system is 8 kW at 7 p'm' = = sum of the individual maximum dernands 5 + 5 + 3 13 kw DiversitY factor = 13/8 = 7.625 ConsumerI Average load 0'2 x 5 = I kW, 2 Consumer Averageload 1.2 kW, Consumer3 Average load I kW, (c) LF= 20Vo
i;;;";l* i " ' '.
7.5x106x100 6000x0.45x8760 = 3I.7 I7o Since no reserveis available at Unit No. 1, its capacity factor is the sameas the load factor, i.e. 81.54vo. Also since unit I has been running throughout the year, the plant use factor equals the plant capacity factor i . e .8 1 . 5 4 V o . 100 Annual plant capacityf'actorof Unit z = lPgx = 8'567o loxg76oxloo Plant use factor of Unit 2 =
(b)
l'2*100024Vo 5 I 10033.3Vo LF= 5x LF=
Combinedaverageload = I + l'2 + l = i . 2 k W Combinedload factor
= =+ x10040Vo
7.5x106x100 = 19.027o 10x0.45x8760x100
Load Forecasting As power plant planning and constructionrequire a gestationperiod of four to energy anrl eight yearsor even longer for the presentday superpower stations, studies. load demandfnrecastingplays a crucial role in power system
ffiil,ftffi
I
powerSyslem Modern nnatysis
lntroduction all system(gtid,area) to another.A distributionsystem connects the loads in lines. a particularareato the transmission in (which will be discussed detail reasons and technological For economical areasor regionalgrids (also called power pools). Each electricallyconnected area or regional grid operatestechnicallyand economically independently,but these are eventuallyinterconnected*to form a national grid (which may even form an internationalgrid) so that each areais contractually tied to other areas features.India is now heading in respectto certaingenerationand scheduling grid. for a national The siting of hydro stations is determinedby the natural water power The choiceof site for coal fired thermalstationsis more flexible. The sources. following two alternativesare possible. may be built closeto coal tnines (calledpit headstations) l. power starions energy is evacuatedover transmission lines to the load and electric centres. Z. power stations may be built close to the load ceutres and coal is transportedto them from the mines by rail road' In practice,however,power stationsiting will dependupon many factorstechnical, economical and environmental. As it is considerablycheaper to transport bulk electric energy over extra high voltage (EHV) transmission lines than to transportequivalentquantitiesof coal over rail roqd, the recent trends in India (as well as abroad) is to build super (large) thermal power stations near coal mines. Bulk power can be transmitted to fairly long kV and above. However, the distancesover transmissionlines of 4001765 belt and some coal are locatedmainly in the eastern country's coal resources regions. fired stationswill continueto be sitedin distantwesternand southern fuel transport by As nuclear stationsare not constrained the problemsof and air pollution, a greater flexibility exists in their siting, so that these while avoiding high densitypollution stationsare locatedclose to load centres areasto reducethe risks, however remote,of radioactivityleakage.
*Interconnectionhas the economic advantageof reducing the reserve generation capacity in eacharea.Under conditionsof suddenincreasein load or loss of generation in one area, it is immediately possible to borrow power from adjoining interconnected lines under faulty causeslarger currentsto flow on transmission areas.Interconnection condition with a consequent increase in capacity of circuit breakers. Also, the
This necessitates long range forecasting. while sophisticated probabilistic methodsexist in literature [5, 16, 28], the simple extrapolationtechnique is quite adequatefor long range forecasting.since weatherhas a much more influence on residentialthan the industrial component,it may be better to prepare forecast in constituentparts to obtain total. Both power and energy uru ractors rnvolved re ng an involved processrequiring experienceand high analytical ability. Yearly forecastsare basedon previous year's loading for the period under considerationupdated by factors such as general load increases,major loads and weathertrends. In shortterm load forecasting,hourbyhour predictions are made for the
decadeof the 21st century it would be nparing2,00,000Mwa stupendous task indeed.This, in turn, would require a correspondingdevelopmeni in coal resources. T.2 STRUCTURE OF POWER SYSTEMS
Generating stations, transmission lines and the distributionsystemsare the main components an electric power system.Generatingstationsand a distribution of systemare connected through transmission lines, which alsoconnectone power
* 38Voof the total power required in India is for industrial consumption. Generation of electricity in India was around 530 billion kWh in 20002001 A.D. compared to less than 200 billion kWh in 198687.
exchangeof power between areas centres.It provides capacity savingsby seasonal having opposing winter and summer requirements.It permits capacity savings from time zones and random diversity. It facilitates transmissionof offpeak power. It also gives the flexibility to meet unexpectedemergencyloads'
In India, as of now, abou 75voof electric power used t is generated thermal in plants(including nuclear).23vofrommostly hydro stationsandZvo.comefrom
:^:yft.s
and.others. coal is the fuer for most of the sream plants,the rest
on wherethe reductionis to a rangeof 33 to 132 kV, depending the substation, may require power at these voltage transmissionline voltage. Some industries level. Normally, two The next stepdownin voltage is at the distribution substation. voltage levels are employed: distribution l. The primary or feeder voltage(11 kV) 2. The secondaryor consumer voltage (440 V three phase/230V single phase). The distribution system, fed from the distribution transformer stations, supplies power to ttre domestic or industrial and commercial consumers. Thus, the power system operatesat various voltage levels separatedby transformer.Figure 1.3 depicts schematicallythe structure of a power system. Though the distribution system design, planning and operation are subjects of great importance,we are compelled,for reasonsof space,to exclude them from the scope of this book. 1.3 CONVENTIONAL SOURCES OF ELECTRIC ENERGY
a .qiaji,
Generating stations at 11kV  25kv 'qff9a, O
Transmission level (220kv  765 kV)
Tie linesto othersystems
Thermal (coal, oil, nuclear) and hydro generations are the main conventional sources of electric energy. The necessity to conserve fosqil fuels has forced scientists and technologists across the world to search for unconventional sources of electric energy. Some of the sourcesbeing explored are solar, wind and tidal sources.The conventional and some of the unconventional sourcesand techniquesof energy generation are briefly surveyedhere with a stresson future trends, particularly with reference to the Indian electric energy scenarioTtrermal Power StationsSteam/Gasbased
Large consumers
Small consumers Fig. 1.3 schematic diagram depicting power system structure
The heat releasedduring the combustionof coal, oil or gas is used in a boiler to raise steam. In India heat generationis mostly coal basedexcept in small sizes, because of limited indigenous production of oil. Therefore, we shall discuss only coalfired boilers for raising steam to be used in a turbine for electric generation. The chemical energy stored in coal is transformed into electric energy in thermal power plants. The heat releasedby the combustion of coal produces steam in a boiler at high pressureand temperature,which when passedthrough a steamturbine gives off some of its internal energy as mechanicalenergy. The axialflow type of turbine is normally used with several cylinders on the same shaft. The steam turbine acts as a prime mover and drives the electric generator (alternator). A simple schematic diagram of a coal fired thermal plant is shown in Fig. 1.4. The efficiency of the overall conversionprocessis poor and its maximum of because the high heat lossesin the combustiongasesand value is about 4OVo
E
rt^r^
h.^
^
rntroduction
Effi
and the large quantity of heat rejectedto the condenser which has to be given off in cooling towers or into a streamlake in the case of direct condenser cooling' The steam power station operateson the Rankine cycle, modified to
\vv'yvrDrwrr ur r'.lr. r.u Inecnanlcal energy) can be increased by using steam at the highest possible pressure and temperature. with steam
perhaps increase unit sizes to several GWs which would result in better generatingeconomy. Air and thermal pollution is always present in a coal fired steam plant. The COz, SOX, etc.) are emitted via the exhaust gasesand thermal pollution is due to the rejected heat transferred from the condenserto cooling water. Cooling towers are used in situations where the stream/lake cannot withstand the thermal burden without excessivetemperaturerise. The problem of air pollution can be minimized through scrubbers and elecmostaticprecipitators and by resorting to minimum emission dispatch [32] and Clean Air Act has already been passedin Indian Parliament. Fluidizedbed Boiler
turbines of this size, additionalincreasein efficiency is obtainedby reheating the steam after it has been partially expanded by an ext;;; i"ui"r. rn" reheated steam is then returned to the turbine where it is expandedthrough the final states of bleedins. To take advantageof the principle of economy of scale (which applies to units of all sizes),the presenttrend is to go in foilarger sizesof units. Larger units can be installed at much lower cost per kilowatt. Th"y are also cheaper to opcrate because of higher efficiency. Th"y require io*", labour and maintenance expenditure. According to chaman Kashkari [3] there may be a saving of as high as lvo in capitalcost per kilowatt by going up from a 100 to 250 MW unit size and an additional saving in fuel cost of ubout gvo per kwh. Since larger units consumeless fuer pJr kwh, they produce ress air, thermal and waste pollution, and this is a significant advantage our concern in for environment' The only trouble in the cai of a large unit is the tremendous shock to the system when outage of such a large capacity unit occurs. This shock can be tolerateclso long as this unit sizeloes not exceed r}vo of the online capacity of a large grid.
Stack
max). The main problem with coal in India is its high ash content (up to 4OVo bed combustion technology is being developed and To solve this, Jtuidized perfected.The fluidizedbed boiler is undergoingextensivedevelopmentand is being preferred due to its lower pollutant level and better efficiency. Direct ignition of pulverized coal is being introduced but initial oil firing support is needed. Cogeneration Considering the tremendous amount of waste heat generatedin tlbrmal power generation,it is advisable to save fuel by the simultaneousgenerationof electricity and steam (or hot water) for industrial use or space heating. Now called cogeneration,such systemshave long been common, here and abroad. Currently, there is renewedinterest in thesebecauseof the overall increasein energy efficiencies which are claimed to be as high as 65Vo. Cogeneration of steam and power is highly energy efficient and is particularly suitablefor chemicals,paper,textiles,food, fertilizer and petroleum refining industries.Thus theseindustriescan solve energyshortageproblem in a big way. Further, they will not have to dependon the grid power which is not so reliable. Of course they can sell the extra power to the government for use in deficient areas.They may aiso seil power to the neighbouring industries,a concept called wheeling Power. and As on 3I.12.2000, total cogenerationpotential in India is 19,500MW actual achievementis 273 MW as per MNES (Ministry of NonConventional Energy Sources,Government of India) Annual Report 200H1. There are two possible ways of cogenerationof heat and electricity: (i) Topping cycle, (ii) Bottoming cycle. In the topping cycle, fuel is burnt to produce electrical or mechanical power and the waste heat from the power generationprovides the processheat.In the bottoming cycle, fuel first produces is process heat and the waste heat from the process6s then used to produce power.
Stepup Ah uE transformer 1030 kv /
Coolirrgtower
Condenser Burner
Preheated air Forced draft fan
mill
Flg. 1.4 schematic diagram a coarfiredsteamprant of In India, in 1970s the first 500 Mw superthermal unit had been commissioned at Trombay. Bharat Heavy Electricals Limited (BHEL) has produced severalturbogenerator setsof 500 MW capacity.Today;s maximum generator unit size is (nearly 1200 Mw) limited by the permissible current cjensitiesused in rotor and stator windines. Efforts are on to develoo srDer.
H
I
Coalfired plants share environmental problems with some other types of fossilfuel plants; these include "acid rain" and the ,,greenhouse,, effect. Gas Turbines With increasing availability of natural gas uangladesh) primemovers basedon gas turbines have been developedon the lines similar to those used in aircraft. Gas combustion generates high temperatures and pressures, so that the efficiency of the las turbine is comparable to that of steam turbine. Additional advantageis that exhaust gas from the turbine still has sufficient heat content, which is used to raise steam to run a conventional steam turbine coupled to a generator. This is called combinedcyclegasturbine(CCGT) plant. The schernatic diagram of such a plant is drawn in Fig. 1.5.
The oldest and cheapestmethod of power generationis that of utilizing the potential energy of water. The energy is obtained almost free of nrnning cost and is completely pollution free. Of course, it involves high capital cost requires a long gestation period of about five to eight years as compared to four to six years for steamplants. Hydroelectric stationsare designed,mostly, as multipurpose projects such as river flood control, storage of irrigation and drinking water, and navigation. A simple block diagram of a hydro plant is given in Fig. 1.6. The vertical difference between the upper reservoir and tail race is called the head.
Headworks Spillway
Valve house
Surgechamber
Reservoir
Pen stock Powerhouse
pond Tailrace Fig. 1.6 Generator A typical layout for a storage type hydro plant
Steam
Fig. 1.5 CCGTpowerstation CCGT plant has a fast start of 23 min for the gas turbine and about 20 minutes for the steam turbine. Local storage tanks Jr gur ured in caseof gas supply intemrption. The unit can take up to ITVo "uiu" overload for short periods of time to take care of any emergency. CCGT unit produces 55vo of CO2 produced by a coal/oilfired plant. Units are now available for a fully automatedoperation for 24h or to meet the peak demands. In Delhi (India) a CCGT unit6f 34Mw is installed at Indraprasthapower Station. There are culrently many installationsusing gas turbines in the world with 100 Mw generators.A 6 x 30 MW gas turbine station has already been put up in Delhi. A gas turbine unit can also be used as synchrono.r, .ornp"nsator to help maintain flat voltage profile in the system.
Hydro plants are of different types such as runofriver (use of water as it comes), pondage (medium head) type, and reservoir (high head) type. The reservoir type plants are the ones which are employed for bulk power generation.Often, cascaded plants are also constructed, i.e., on the sa.me water stream where the discharge of one plant becomesthe inflow of a downs6eam plant. The utilization of energy in tidal flows in channets has long been the subject of researeh;Ttrsteehnical and economic difficulties still prevail. Some of the major sites under investigation are: Bhavnagar,Navalakhi (Kutch), Diamond Harbour and Ganga Sagar. The basin in Kandala (Gujrat) has been estimated to have a capacity of 600 MW. There are of course intense siting problems of the basin. Total potential is around 9000 IvftV out of which 900 MW is being planned. A tidal power station has been constructed on the La Rance estuary in northern France where the tidal height range is 9.2 m and the tidal flow is estimated to be 18.000 m3/sec. Different types of turbines such as Pelton. Francis and Kaplan are used for storage,pondageand runofriver plants, respectively. Hydroelectric plants are
W
t


powersystem Modern Anarvsis
daily load demand curve. Some of the existingpumpedstorageplantsare I100 MW Srisailem in Ap and 80 MW at Bhira in Maharashtra. p=gpWHW where W = dischargem3ls through turbine p = densiry 1000 kg/m3 11= head (m) 8 = 9.81 mlsz Problems peculiar to hydro plant which inhibit expansion are: 1. Siltingreportedly Bhakra dead storagehas silted fully in 30 years 2. Seepage 3. Ecological damageto region 4. Displacement human habitation from areasbehind the dam which of will fill up and become a lake. 5. Thesecannotprovide baseload, mustbe usedfor peak.shaving and energy saving in coordination with thermal plants. India alsohas a tremendous potential (5000MW) of having large number of micro (< 1 Mw), mini (< 15 Mw), and,small (< 15 Mw) Mrl plants in Himalayan region, Himachal, up, uttaranchal and JK which must be fully exploited to generate cheapand clean power for villages situated far awayfrom the grid power*. At present 500 MW capacityis und"r construction. In areaswheresufficienthydro generationis not available,peak load may be handled by means of pumped storage. This consists of un ,rpp". and lower reservoirs and reversible turbinegeneratorsets, which cun ulio be used as motorpump sets.The upper reservoir has enough storagefor about six hours of full load generation.Such a plant acts as a conventional hydro plant during the peak load period, when production costs are the highest. The iurbines are driven by water from the upper reservoir in the usual manner. During the light load period, water in the lower reservoir is pumped back into the ipper one so as to be ready for use in the next cycle of the peak ioad p.rioo. rn" generatorsin this period change to synchronousmotor action and drive the turbineswhich now work as pumps. The electric power is supplied to the sets from the general power network or adjoining thermal plant. The overall efficiency of the sets is normarly as high ut 607oEo. The pumped srorage scheme,in fact, is analogousto the charging and discharging or u battery. It has the added advantage that the synchronousmachin", tu1 be used as synchronous condensersfor vAR compensationof the power network, if required. Ina way, from the point of view of the thermal sector of the system,
* Existing capacity (small hydro) is 1341 MW as on June 200I. Total estimated potentialis 15000 MW. Waterintake
Nuclear Power Stations With the end of coal reservesin sight in the not too distant future, the immediate practical alternative sourceof large scale electric energy generation is nuclear energy.In fact, the developedcountries have already switched over in a big way to the use of nuclear energy for power generation.In India, at present, this sourceaccountsfor only 3Voof the total power generation with nuclear stations at Tarapur (Maharashtra),Kota (Rajasthan),Kalpakkam (Tamil Nadu), Narora (UP) and Kakrapar (Gujarat). Several other nuclear power plants will be commissioned 20I2.In future,it is likely that more and more power by will be generatedusing this important resource (it is planned to raise nuclear power generationto 10,000MW by rhe year 2010). When Uranium235 is bombardedwith neutrons,fission reaction takesplace releasingneutrons and heat energy. These neutronsthen participate in the chain reaction of fissioning more atoms of 235U.In order that the freshly released neutronsbe able to fission the uranium atoms, their speeds must be ieduced to a critical value Therefore,for the reaction to be sustained, nuclear fuel rods must be embeddedin neutronspeedreducing agents(like graphite, hqavy water, etc.) called moderators.Forreaction control, rods made of n'eutronabsorbing material (boronsteel) are used which, when inserted into the reactor vessel, control the amount of neutron flux thereby controlling the rate of reaction. However, this rate can be controlled oniy within a narrow range. The schemadc, diagramof a nuclear power plant is shown in Fig. 1.7. The heit releasedby the 'uclear reaction is transported a heat exchangervia primary to coolant (coz, water, etc.). Steam is then generated the heat exchanger, in which is used in a conventionalmanner to generateelectric energy by meansof a steam turbine. Various types of reactorsare being used in practicefor power plant pu{poses, viz., advancedgas reactor (AGR), boiling water reactor (BwR), und h"uuy water moderated reactor.etc.
Control rods
Fuelrods_
W
ModernPo*", system An"tysis
iiiriociucrion

CANDU reactorNatural uranium (in cixideform), pressurized heavy water moderatedis adopted in India. Its schematic diagram is shown in Fig. 1.8.
require that they be normally located away from populated areas. Demerits Nuclear reactors produce radioactive fuel waste, the disposal poses serious environmentalhazards. 2. The rate of nuclear reaction can be lowered only by a small margin, so that the load on a nuclear power plant can only be permitted to be marginally reduced below its full load value. Nuclear power stations must, therefore, be realiably connected to a power network, as tripping of the lines connecting the station can be quite serious and may required shutting down of the reactor with all its consequences. 3 . Because of relatively high capital cost as against running cost, the nuclear plant should operate continuously as the base load station. Wherever possible, it is preferable to support such a station with a pumped storage schemementioned earlier.
4. The greatestdangerin a fission reactor is in the caseof loss of coolant
in an accident. Even with the control rods fully lowered quickly called scrarn operation, the fission does continue and its afterheatmay cause vaporizing and dispersalof radioactive material. The world uranium resourcesare quite limited, and at the presentrate may not last much beyond 50 years.However, there is a redeemingfeqture. During 235U,some of the neutrons are absorbed by the fission of lhe more abundant 23sUwhile 238U uranium isotope lenriched uranium containsonly about 3Voof 238U)converting it to plutonium ("nU), which in itself is a most of its is fissionablematerial and can be extractedfrom the reactor fuel waste by a fuel reprocessing plant. Plutonium would then be used in the next generation reactors (fast breeder reactorsFBRs), thereby considerably extending the life of nuclear fuels. The FBR technologyis being intenselydevelopedas it will extend the availability of nuclear fuels at predicted rates of energy consumption to several centuries. Figure 1.9 shows the schematicdiagram of an FBR. It is essentialthat for breeding operation, conversion ratio (fissile material generated/fissilematerial consumed) has to be more than unity. This is achieved by fast moving neutrons so that no moderatoris needed.The neutrons do slow down a little through collisions with structural and fuel elements.The energy densitylkg of fuel is very high and so the core is small. It is therefore necessarythat the coolant should possessgood thermal properties and hence liquid sodium is used. The fuel for an FBR consists of 20Voplutonium phts 8Vouranium oxide. The coolant, liquid sodium, .ldaves the reactor at 650"C at atmospheric pressure.The heat so transportedis led to a secondarysodium circuit which transfers it to a heat exchanger to generate steam at 540'C.
Containment
Fig. 1.8 CANDUreactorpressurized heavywater rnoderatedadopted in In d i a The associated merits and problems of nuclear power plants as compared to conventional thermal plants are mentioned below. Merits 1. A nuclear power plant is totally free of air pollution. 2. It requireslinle fuel in terms of volume and weight, and thereforeposes . no transportation problems and may be sited, independentlyof nuclear
t _
Modprn
rrtyyvrr.
pnrrrar
r vrrvr
vyglgttl
errolam
nttdtvsts
Anal.,^l^
at Kalpakkam alongside nucrear a powerplant.FBR technology i,
conventional thermal plants.
with a breeder reactor the release of plutonium, an extremely toxic material, would make the environmentalconsiderationsmost stringent. been An experimentalfast breeder test reacror (FBTR) (40 MW) has built
"*f..l"J
engineering experience. The present cost of nuclear wlm coalttred power plant, can be further reduced by standardisingpl4nt design and shifting from heavy wate,r reactorto light water reactortechnology. Typical power densities 1MWm3) in fission reactor cores are: gas cooled 0.53, high temperaturegas cooled 7.75, heavy warer 1g.0, boiling iut., Zg.O, pressurizedwater 54.75, fast breederreactor 760.0. Fusion Energy is produced in this processby the combination of two light nuclei to form a single heavier one under sustained conditions of exiemely high temperatures millions of degree centigrade).Fusion is futuristic. Genera(in tion of electricity via fusion would solve the longtenn energy needs of the world with minimum environmental problems. A .o"i.iul reactor is expectedby 2010 AD. Considering radioactive wastes, the impact of fusion reactors would be much less than the fission reactors. In case of successin fusion technologysometime in the distant future or a breakthroughin the pollutionfree solarenergy,FBRs would becomeobsolete. However, there is an intense need today to develop FBR technology as an insuranceagainst failure to deverop these two technologies. \ In the past few years, serious doubts have been raised.about the safety claims of nuclear power plants. There have been as many as 150 near disaster nuclear accidents from the Threemile accident in USA to the recent Chernobyl accident in the former USSR. There is a fear.that all this may pur the nuclear energy developmentin reversegear. If this happensthere could be serious energy crisis in the third world countries which have pitched their hopes on nuclear energy to meet their burgeoningenergy needs.France (with 78Voof its power requirement from nuclear sources)and Canadaare possibly the two countries with a fairty clean record of nuclear generation.India needs to watch carefully their design, constructionand operating strategiesas it is committed to go in a big way for nuclear generation and hopes to achieve a capacity of 10,000 MW by z0ro AD. As p.er Indian nuclear scientists, our heavy waterbased plants are most safe.But we must adopt more conservative strategies design, constructionand operationof nuclearplants. in World scientistshave to adopt of different reaction safety strategymay be to discover additives to automatically inhibit feaction beyond cr;ii"at rather than by mechanically inserted control rods which have possibilitiesof several primary failure events. Magnetohydrodynamic (MHD) Generation
suitedfor India, with poor quality coal,inadequare hydro potentiaiilentiful reserves uranium(70,000tons) and thorium,and many yearsof nuclear of

Core Coolant Containment
Fig. 1.9 Fastbreeder reactor (FBR) An important advantageof FBR technologyis that it can also use thorium (as fertile material) which gets convertedto t33U which is fissionable. This holds great promise for India as we have one of the world's largest deposits of thoriymabout 450000 tons in form of sand dunes in Keralu una along the Gopalpfur Chatrapurcoastof Orissa. We have merely 1 per cent of the world's
In thermal generation of electric energy, the heat released by the fuel is converted to rotational mechanical energy by means of a thermocvcle. The
ry
Modern Power System Anatysis
Introduction
w I
mechanicalenergy is then used to rotate the electric generator.Thus two stages of energy conversion are involved in which the heat to mechanical energy conversion has inherently low efficiency. Also, the rotating machine has its associated lossesand maintenance problems.In MHD technology, cornbustionof fuel without the need for mechanicalmoving parts. In a MHD generator,electrically conducting gas at a very high temperature is passed in a strong magnetic fleld, thereby generatingelectricity. High temperature is needed to iontze the gas, so that it has good eiectrical conductivity. The conductinggas is obtainedby burning a fuel and injecting a seeding materials such as potassium carbonate in the products of combustion. The principle of MHD power generation is illustrated in Fig. 1.10. Abotrt 50Vo efficiency can be achievedif the MHD generatoris operated in tandem with a conventional steam plant.
are volcanic regionscan be utilized. Since the pressureand temperatures low, fossil fuelled plants,but the the efficiency is even less than the conventional capital costs are less and the fuel is available free of cost. I.4 RENEWABLE ENERGY SOURCES
development, the importance of To protect environmentand for sustainable It renewableenergysourcescannot be overemphasized. is an established and forms of energy will play acceptedtact that renewableand nonconventional an increasingly important role in the future as they are cleanerand easier to benign and are bound to becomeeconomicallymore use and environmentally use. viable with increased internaBecause of the limited availability of coal, there is considerable tional effort into the development of alternative/new/nonconventionaUrenewable/cleansourcesof energy. Most of the new sources (some of them in fact have been known and used for centuries now!) are nothing but the manifestationof solar energy, e.g., wind, sea waves, ocean thermal energy conversion (OTEC) etc. In this section, we shall discuss the possibilities and potentialities of various methods of using solar energy. Wind Power Several Winds are essentiallycreated by the solar heating of the atmosphere. attempts have been made since 1940 to use wind to generateelectric energy and developmentis still going on. However, technoeconomicfeasibility has yet to be satisfactorily established. it Wind as a power source is attractivebecause is plentiful, inexhaustible and nonpolluting. Fnrther, it does not impose extra heat burden on the Control environment.Unlbrtunately, it is nonsteadyand undependable. equipment has been devised to start the wind power plant wheneverthe wind constant speedreaches30 kmftr. Methods have also been found to generate and consequently varying speeds frequencypower with varying wind speeds of wind mill propellers. Wind power may prove practical for small power needs in isolated sites. But for maximum flexibility, it should be used in conjunction with other methods of power generationto ensurecontinuity. For wind power generation, there are three types of operations: 1. Small, 0.510 kW for isolated single premises i 2. Medium, 10100 kW for comrnunities 3. Large, 1.5 MW for connectionto the grid. The theoreticalpower in a wind streamis given by P = 0.5 pAV3W
where
Gas flow at 2,500'C
Strong magnetic field
e generati on F i g .1 .1 0 T h e p ri n c i p lof MH Dpow er Though the technologicalfeasibility of MHD generationhas been established, its economicf'easibilityis yct to be demonstrated. lndia had starteda research and developmentproject in collaboration with the former USSR to install a pilot MHD plant based on coal and generating2 MW power. In Russia, a 25 MW MHD plant which uses natural gas as fuel had been in operation for some years. In fact with the developmentof CCGT (combined cycle gas turbine) plant, MHD developmenthas been put on the shelf.
Geothermal
Power Plants
In a geothermal power plant, heat deep inside the earth act as a source of power. There has been some use of geothermalenergy in the form of steam coming from undergroundin the USA, Italy, New Zealand,Mexico, Japan, Philippines and some other countries. In India, feasibility studies of 1 MW station at Puggy valley in Ladakh is being carried out. Another geothermal field has been located at Chumantang. There are a number of hot springsin India, but the total exploitableenergy potential seemsto be very little. Ttre present installed geothermal plant capacity in the world is about 500 MW and the total estimatedcapacityis immenseprovided heat generated the in
p = densityof air (1201 g/m' at NTP) V _ mean air velocity (m/s) and A = sweptarea (rn").
Introduction pot Tota l solar ener gy ent ial I ndia is 5 x lO ls kwh/ yr . Up r o 31. t 2. 2000. in 462000solar cookers,55 x10am2solar thermai system collector area,47 MW of SPV power, 270 community lights, 278000 solar lanterns(PV domestic lighting units),640 TV (solar), 39000 PV streetlights and 3370 warer pumps MW of grid connected solar power plants were in operation. As per one estimate[36], solar power will overtakewind in 2040 and would become the world's overall largest source of electricity by 2050. Direct Conversion to Electricity (Photovoltaic Generation)
2. Rural grid systems likely to be 'weak, in theseareas. are since retatrvely low voitage supplies(e.g. 33 kV). 3. There are always periods without wind. In India, wind power plants have been installed in Gujarat, orissa, Maharashtra and Tamil Nadu, where wind blows at speedsof 30 kmftr during summer' On the whole, the wind power potential of India has been estimated to be substantial and is around 45000 Mw. The installed capacity as on Dec. 2000 is 1267 Mw, the bulk of which is in Tamil Nadu (60%). The conesponding world figure is 14000 Mw, rhe bulk of which is in Europe (7UVo). Solar Energy The average incident solar energy received on earth's surface is about 600 W/rn2 but the actual value varies considerably. has It the advantageof beingfree of cost,nonexhaustible completelypollutionfree. and On the other hand,it has severalcrrawbacksenergy density pei unit area is very row, it is available for only a part of the day, and cl,oud and, y hazy atmospheric conditions greatly reduce the energy received. Therefore, harnessing solar energyfor electricitygeneration, challengingtechnological problemsexist,the most important being that of the collection and concentration of solar energy and its conversion to the electrical form through efficient and comparatively economical means. At present, two technologiesare being developedfor conversion of solar energyto the electrical form.'In one technology,collectors with concentrators are employedto achievetemperatures high enough(700'C) to operatea heat engrne at reasonable efficiency to generate electricity. However, there are considerableengineeringdifficulties in building a single tracking bowi with a diarneter exceeding30 m to generate perhaps200 kw. The schemeinvolves large and intricate structuresinvoiving lug" capital outlay and as of today is f'ar from being competitive with Jlectricity generation. "otru"titional The solar power tower [15] generates steam for electricity procluction. ]'here is a 10 MW installationof such a tower by the Southern California EdisonCo' in USA using 1818plane rnirrors, each m x i 7 m reflecting direct racliation thc raisecl to boiler. Electricity may be generated from a Solar pond by using a special .low temperature' heat enginecoupledto an electricgenerator. solar pond at A Ein Borek in Israel procluces steady150 kW fiorn 0.74 hectare a at a busbarcost of abo u t$ O.tO /k w h . Solar power potential is unlimited, however, total capacityof about 2000 MW is being planned.
This technologyconvertssolar energyto the electrical form by meansof silicon wafer photoelectriccells known as "Solar Cells". Their theoreticalefficiency is about 25Vobut the practical value is only about I5Vo. But that does not matter as solar energy is basically free of cost. The chief problem is the cost and maintenance solar cells. With the likelihood of a breakthrough the large of in scale production of cheap solar cells with amorphoussilicon, this technology may competewith conventional methodsof electricity generation,particularly as conventional fuels becomescarce. Solar energy could, at the most, supplementup to 5r0vo of the total energy demand.It has been estimated that to produce 1012 kwh per year, the necessary cells would occupy about0.l%oof US land areaas againsthighways which occupy 1.57o(in I975) assumingI07o efficiency and a daily insolation .\ of 4 kWh/m'. In all solar thermalschentes, storage necessary is because the fluctuating of nature of sun's energy. This is equally true with many other unconventional sourcesas well as sourceslike wind. Fluctuatingsources with fluctuating loads complicatestill further the electricity supply. Wave Energy The energyconientof sea wavesis very high. In India, with severalhundreds of kilometersof coast line, a vast sourceof energyis available. The power in the wave is proportionalto the squareof the anrplitudeand to the period of the motion. Therefore,rhe long period ( 10 s), large amplitude( 2m) waves are of considerable interest for power generaticln, with energy fluxes commonly averagingbetween50 and 70 kW/m width of oncoming wave. Though the engineeringproblems associated with wavepowerare formidable, the amountof energythat can be harnessed large and development is work is (alsoseethe sectionon HydroelectricPower Generation, in progress page 17). Sea wave power estimated poterrtial is 20000 MW. Ocean Thermal Energy Conversion (OTEC) The ocean is the world's largest solar coilector. Temperaturedifference of 2O"Cbetween\,varrn, solar absorbingsurfacewater and cooler 'bottorn' water
ffiffi
Modem Pow'er system Anatysis
lntroduction solar. The most widely used storagebatteryis the lead acid battery.invented by Plantein 1860.Sodiutttsulphur battery(200 Wh/kg) and other colrbinations of materialsare also being developed get more output and storageper to unit weisht. Fuel Cells A fuel cell converts chemicalenerry of a fuel into electricity clirectly, with no intermediatecotnbustioncycle. In the fuel cell, hyclrogen supplied to the is negative electrodeand oxygen (or air) to the positive. Hydrogenand oxygen are combined to give water and electricity. The porous electrodesallow hydrogen ions to pass.The main reason';rhy fuel cells are not in wide use is their cost (> $ 2000/kW). Global electricity generatingcapacity from full cells will grow from just 75 Mw in 2001 ro 15000MW bv 2010.US. Germanvand Japan may take lead for this. Hydrogen Energy Systems
can occlrr.This can provide a continuallyreplenished storeof thermal energy which is in principle available fbr conversion to other energy forms. OTEC refers to the conversion of someof this thermal energy into work and thence 50,000 Mw. A proposedplant using seaiemperaturedifferencewould be situated 25 km cast ol'Mianii (USA), wherethe temperature clil'l'eronce 17.5"C. is Biofuels The material of plants and animals is called biomass, which may be transformed by chemical and biological processesto produce intermediate biofuels sttch as methane gas, ethanol liquid or charcoal solid. Biomass is burnt to provide heat for cooking, comfort heat (space heat), crop drying, tactory processes and raising steamfor electricity production and transport. In I ndia p o te n ti aI' ttlb i o E n e rg ys 1 7 0 00MW and thatfbr agri cul tunrl i rstci s l i w about 6000 MW. There are about 2000 community biogas plants and tamily size biogas plants are 3.1 x 106. Total biomass power harnessedso far is 222 MW . Renewableenergy programmes specially designedto meet the growing are energy needs in the rural areas for prornoting decentralized and hybrid dcvelopment as to stem growing migration of rural populationto urban st.l areasin searchof better living conditions. It would be through this integration of energy conservationefforts with renewable energy programmesthat India would be able to achievea smooth transition from fossil fuel economy to sustainablerenewableenergy basedeconomy and bring "Energy for ali" for ec;uitable and environrnental friendly sustainable development. 1.5 ENERGY STORAGE
Hydrogen can be used as a medium for energy transmissionand storage. Electrolysis a wellestablished is commercial process yielding pure hydrogen. Ht can be convertedvery efficiently back to electi'icityby rneans fuel ceils. of Also the use of hydrogena.s fuel for aircraftand automcbilescould encourase its large scaleproduction, storageand distriburion. 1"6 GROWTH OF POWER SYSTEII{S IN INDIA
is a lol ol problenrin storing clectricity in largc quantities.Enclgy wliich can be convertedinto electricity can be storedin a number of ways. S t or ag eo f a n y n a tu rei s l ro w e v e r ery costl y arrcl ts cconomi csmust be v i worked out properly. Various options available are: pLrmped storage, c:onlpressedair, heat, hydrogengas,secondary batteries, flywheels and superconduc t in gc o i l s . As already mentioned, gas turbines are normally used for meeting peak loads but are very expensive. A significant amount of storage capable of instantancous would be betterway of meetingsuch peak loads, and so far use the most importantway is to have a pumped storage plant as discussed earlier. Other methods are discussed below very briefly. Secondary Batteries Large scale battery use is almost ruled out and they will be used for battery powered vehicles and local fluctuating energy sourcessuch as wind mills or
'l'here
India is fairly rich in natural resources like coal and lignite; while sorne oil reserveshave been discoveredso far. intenseexplorationis being undertakeri in vitriousregitlnsof thc country. India has immensewater power l.csources also of which only around25To have so farbeen utiliseci, i.e.,oniy 25000 t\,IW has so far beencommissioned to the end of 9th plan. As per a recentreport up of tlre CEA (Ccntlal Flectricit,v Authority),the total potentialof h1,dro power is 84,040Iv{W at ('L't% load factor. As regardsnuclearpower, India is cleflcient in uranium,but has rich deposits thoririm of rvhichcan be utilisedat a future clatc in l'ast brccclorrci.tctor.s. Since indepcndcncc, thc coulltry has nnde tremendous progress the development electricenergyand todayit has the in of largest systemamong the developingcountries. When lndia attained independence, installeclcapacity was as low as the 1400 MW in the early stagesof the growth of power system,the major portion of generationwas through thermal stations,but due to economicalreasons. hydro development receivedattentionin areaslike Kerala, Tamil Nadu. Uttar Pradesh and Punjab. In the beginningof the First Five Year Plan (195156), the rotal installed capacity was around2300 MV/ (560 MW hydro, 1004MW thermal,149 MW through oil stations and 587 MW through nonutilities). For transportingthis
Introduction
power to the load centres,transmissionlines of up to 110 were constructed.
year 201112 [19]'
HE FI
and peak load in the regions of the country with projectedenergy requirement
io ororrcrt crcncreri nuclear power' At the During the Fourth Five Plan, India startedgenerating units were comrnissionedin AprilMay Tarapur i\uclear Plant 2 x 210 MW American design. By . This station uses two boiling water reactorsof
Northern region 308528 (49674)
MW* 9
.,.
region Western 299075 (46825)
commissionedbY 2012. projection for 2011The growth of generating capacity so far and future 1'1' 2012 A.D. are given in Table capacityin lndia (ln MW) Tabte 1.1 Growthof Installed
Year
Hydrtt
Nuclear
Thermal
DieseI
Total
19707t 197879 198485 200001
6383 l 1378 t4271 25141
420 890 1095 2720
7503 t6372 27074 71060
398
\'./
Fig. 1.11 in projected energyrequirement five Mapof Indiashowing regional and parkloadin MW for year201112' MkWh
=2700 MW renewable
r4704 28640 42240 101630
The emphasisduring the SecondPlan (195661) was on the developmentof and thus there was a need to step up power basic ancl heavy inclustries total installedcapacity which was around 3420 MW at the end generation.The of tn" First Five year Plan became 5700 MW at the end of the SecondFive voltage came up in Tarnil year plan. The introduction of 230 kv transmission
was: Domestic Pattern of utlization of electrical energy in 199798 and othersis industry 35'22Vo 6.917o, inigation 30.54Vo, {O.6g\o,commercial 200405' in 6.657o.It is expectedto remain more or less same out all over the be selfsufficient in power' BHEL has plants spread To viz' turbo sets'power equipment, country ancltheseturn out an entirerangeof tiigft pi".ture boilers, power transformhydro sets,turbinesfor nuclearplants, in range of equipment'BHEL's ers, switch gears,etc. Each plant specializes a was cornmissionedat singrauli' Today BIIEL first 500 MW turbogenerator manufacturersin the is consideredone of the major power plant equipment world. T.7 ENERGY CONSERVATION
new sourceof energy' we should resort is Energy conservation the cheapest earlier), and such as cogeneration(discussed to various conservationmeasures
,r32 I
Modernpower Svstem Analvsis Load Management
lntroduction
use energy efficient motors to avoid wasteful electric uses.We can achieve considerable electricalpower savings by reducing unnecessary high lighting levels,oversizedmotors,etc. A 9 W cornpactfluorescent lamp (CFL) may be used instead of 40 w fluorescenttube or 60 w lamp, all having the same lu a year.Everyoneshouldbe madeaware throughprint or electronicmedia how consumptionlevels can be reducedwithout any essential lowering of comfort. Rate restructuringcan have incentivesin this regard.There is no consciousnesson energy accountability etndno senseof urgencyas in developed yet countries. Transmissionand distributionlossesshoulclnot exceed2OVo. This can be achievedby employing series/shunt compensation,power factor improvement methods, static var compensators,HVDC option and FACTS (flexible ac technology) devices/controllers. Gas turbirre combined with steam turbine is ernployed for peak load shaving. This is more efficient than normal steam turbine and has a quick automated starl and shut doivn. It improves the load factor of the steam staflon. Energy storage can play an important role where there is time or rate mismatchbetween supplyand demandof energy.This has been discussed in Section 1.5. Pumped storage(hyclro)schemehas been consiclered Section in 1.3. Industry
'load management'schemes.It is possibleto As mentionedearlier by various shift demanrlaway frorn peak hours (Section I .1.). A more direct method would be the control of the load either through rnodified tariff structurethat encourage or schedules direct electrical control of appliancein the form of remote timer controlled on/off switches with the least inconvenience io the customer. Various systems for load rnanagementare described in Ref. [27]. Ripple control has been tried in Europe. Remote kWh meter reading by carrier sysrems is being tried. Most of the potential for load control lies in the domestic sector. Power companies are now planning the introduction of systemwideload managementschemes. 1.8 DEREGULATION
In India where most areashave large number of sunny days hot water for bath arrdkitchen by solarwater heatersis becomingcommon for commercial buildings,hotels even hospitals. In India where vastregionsare deficient in electric supply and,aresubjected to long hours of power sheddingmostly random, the use of small diesel/petrol generators and invertersare very conmon in commercialand domestic use. Theseare highly wastefulenergydevices.By properplanned maintelance the downtime of existing large stationscan be cut down. Plant utilization factors of existingplants must be improved.Maintenance must be on schedulerather than an elnerqency. Maintenancemanpower training should be placed on war footing. These actions will also improve the load factor of most power stations,which would indirectly contribute to energy conservation.
For over one hundred years,the electricpower industry worldwide operatedas a regulated industry. In any area there was only one company oI government agency (mostly stateowned)that produced,transmitted,distributed and sold electric power and services.Deregulationas a conceptcame in early 1990s.It competition. to brought in changesdesigneci enc<.rutage of the power industry and reassembly Restructuring involves disassembly into another form or functional organisation. Privatisation started sale by a and operatingeconomy, electric utility assets, governmentof its stateowned private companies.In some cases,deregulationwas driven by privatization to needs. The state wants to sell its electric utility investment and change the rules (deregulation)to make the electric industry more palatable for potential investors,thus raising the price it could expect from the sale. Open accessrs nothing but a common way for a govenlmentto encouragecompetition in the electric industry and tackle monopoly. The consumer is assuredof good quality power supply at competitive price. The structure for deregulation is evolved in terms of Genco (Generation System Company)and ISO (Independent Company),Transco (Transrnission Operator).It is expectedthat the optimal bidding will help Genco to maximize its payoffs. The consumersare given choice to buy energy from different retail energy suppliers who in turn buy the energy from Genco in a power market. (independentpower producer, IPP). The restructuringof the electricity supply industry that norrnally accompanies the introduction of competiiion provides a fertile ground for the growth of embeddedgeneration,i.e. generationthat is connectedto the distributicn systetn. systemrather than to the transmission The earliest reforms in power industrieswere initiated in Chile. They were followed by England, the USA, etc. Now India is also implementing the the to is Lot restructuring. of research needed clearlyunderstand power system operation under deregulation. The focus of, researchis now shifting towards
W
Modernpo*", Syster Anulyri,
I nt r oduct ion If There are already32 million improved chulhas. growing energy needsin the and hybrid enerqy systems(distributed/ rural areasare met by decentralised generation), this can stem growing migrationof rural populationto dispersed urban areasin searchof better living conditions.Thus, India will be able to ableenergy based econolny iind bring "Energy for all" for equitable, development. and sustainabie environmentfriendly, 1.10 ENVIRONMENT/\L GENERATION ASPECTS OF ELECTRIC ENER,GY
finding the optimal bidding methodswhich take into account local optimal dispatch,revenueadequacyand market uncertainties. India has now enactedthe Electricity RegulatoryComrnission'sAct, 1998 and the Electricity (Laws) AmendmentAct, 1998. These laws enablesetting uo of State Electricity RegulatoryComrnissions(SERC) at srate level. 'fhe main purpose of CERC is to promote efficiency, economy and competition in bulk electricity supply. orissa, Haryana, Andhra Pradesh,etc. have started the processof restructuringthe power sector in their respectivestates. 1.9 DISTRIBUTED AND DISPERSED GENERATION
(DG) entailsusing lnany srnallgenerators 250 MW DistributedGeneration of output,installedat variousstrategic points throughout the area,so that each providespower to a small numberof consumers nearby.Thesemay be solar, mini/micro hydel or wind turbine units, highly efficient gas turbines,small combincdcycle plitnts,sincc thcsearo the rnost ccon<lnrical choiccs. Dispersedgenerationreferesto use of still smaller generatingunits, of less than 500 kW output and often sized to serve individual homes or businesses. Micro gas turbines,fuel cells, diesel,and small wind and solar PV senerators make up this category. Dispersedgenerationhas been used for clecades an emergencybackup as power source.Most of theseunits are used only fbr reliability reinfbrcement. Nowadaysinverters are being increasinglyused in domestic sector as an emergency supply during black outs. The distributed/dispersed generatorscan be stand alone/autonomous or grid connecteddepending upon the requirement. At the time of writing this (200i) there still is and will probably alwaysbe some economy of scale favouringlarge generators.But the margin of economydecreased considerably last 10 years [23]. Even if the power itself in c t ls t sa b i t rtttl rcth i tnc c n (r' as ta ti o n o wcr,therei s no nccd < tftransrni ssi on l p lines, and perhapsa reducedneed fbr distribution equipment as well. Another maior advantageof dispersedgene.ration its modularity, porlability and is relocatability. Dispersed generators also include two new types of tbssil fuel unitsfuel cells and microgas turbines. The main challenge today is to upgradethe existing technologies and to proniotedeveloprnent, demonstration, scaling up and cornmercialization of new and emerging technologiesfor widespreadadaptation.In the rural sector main thrust areasare biomassbriquetting,biomassbased cogeneration, etc. In solar PV (Photovoltaic),large size solar cells/modulesbased on crystalline siliconthin films need to be developed. Solarcells efficiencyis to be improved to 15%o be of use at commercial to level.Otherareasare developrnent high of eificiency inverters.Urban and industrial wastesare used for variousenergy applications including power generationwhich was around 17 Mw in 2002.
and health risks involvedin nuclearplants of various As far as environmental discussed Section1.3. The in thesehave already'been kinds are concerned, problerns relatedto largelrydroplantshavealsobeendwelledupon in Section 1.3. Therefore,we shall now focus our attentionon fossil fuel plant including plants. gasbased Conversion of clne lornr ol' energy or anotherto electrical tortn has in unwanted side effects and the pollutants generated the processhave to be disposed off. Pollutants know no geographical boundary, as result the pollution issue has become a nightmarish problem and strong national and international pressuregroups have sprung up and they are having a definite has awareness Governmental of impacton the development energyresources. levels, w[ich power creatednumerouslegislationat national and international have to be fully conversantwith in practiceof their professionand engineers survey and planning of large power projects.Lengthy, time consuming level, PIL (public interestlitigation) and demonstraprocedures governrnent at tive protestshave delayedseveralprojects in severalcountries.This has led of projectsand redevelopment existing sites.But to favouring of smallsize with the increasinggap in electric dernandand production,our country has to move forward fbr severallarge thermal, hydro and nuclear power projects. issucs. t uilt nent t r ansnt issit t n cur of is E ntphasis lr cinglaid on cor ] scr vilt ior t losses, theft, subsidized power supplies and above all on sustainable whercver feasible. It has to be devektpnrenlwittr uppntpriata technolog)' particularly assuredthat no irreversible damageis caused to environment which wouid affect the living conditions of the future generations.Irreversible in by like ozonelayer holesand global warmingcaused increase CO2 damages up. are alreadyshowing in the atmosphere Atmospheric Pollution
We shall treat here only pollutrorras causedby thermalplants using coal as feedstock. Certain issues concerning this have already been highlighted in Section 1.3. The fossil fuel based generatingplants fonn the backbone of power generation in our country and also giobally as other options (like nuclear and even hydro) have even stronger hazardsassociatedwith them.
w_
ffiffi
rr^r^
n tviouern ^ . . . ^  uystem Anaiysts rower^ . ,  r   a   r .   ,
lntroduction Oxides of Carhon (CO, COt)
Also it should be understood that pollution in large cities like Delhi is caused more by vehicrtlar traffic and their emission.In Delhi of course Inderprastha and Badarpur power stationscontributetheir share in certain areas. Problematic pollutants in emission of coalbasedgeneratingplants are.
a a o a
in to CO is a very toxic pollutantbut it getsconverted CO'., the openatmosphere (if available) surroundingthe plant. On the other hand CO2 has been identified developingcountries. Ifydrocarbons During the oxidation process in cornbustioncharnbercertain light weight hydrocarbon may be formed. Tire compounds are a major source of photochemical reaction that adds to depleti,rnof ozone layer. Particulates (fIY ash)
2
NO.r, nitrogen oxides CO
coz
. Certain hydrocarbons o Particulates Though the account that follows will be general, it needs to be mentioned here that Indian coal has comparatively low sulphur content but a very high ash content which in some coals may be as high as 53Vo. A brief account of various pollutants, their likely impact and methods of abatements presentedas follows. are Oxides of Sulphur (SOr)
Most of the sulphur present in the fossil fuel is oxidized to SO2 in the combustion chamberbefore being emittedby the chimney. In atmosphere it gets further oxidized to HrSOo and metallic sulphateswhich are the major sourceof concern as these can causeacid rain, impaired visibility, damageto buildings and vegetation. Sulphate concenffations of 9 10 LElm3 of air aggravate asthma,lung and heart disease. may also be noted that although It sulphur does not accumulatein air, it does so in soil. Sulphur emissioncan be controlledby: o IJse of fuel with less than IVo sulphur; generally not a feasible solution. o LJseof chemical reaction to remove sulphur in the form of sulphuric acid, from combustionproducts by lirnestonescrubbersor fluidized bed combustion. . Removing sulphurfrom the coal by gasificationor floatationprocesses. It has been noticed that the byproduct sulphur could offset the cost of sulphur recovery plant. Oxides of Nitrogen (NO*)
Dust content is particularly high in the Indian coal. Particulatescome out of the stack in the form of fly ash. It comprisesfine particles of carbon, ash and other inert materials.In high concentrations,these cause poor visibility and respiratory diseases. Concentration of pollutants can be reducedby dispersal over a wider area by use of high stacks.Precipitators can be used to remove particles as the flue gasesrise up the stack.If in the stack a vertical wire is strung in the middle and charged to a high negative potential, it emits electrons. These electrons are captured by the gas molecules therebybecomingnegative ions. These ions accelerate towards the walls, get neutralized on hitting the'walls and the particles drop down the walls. Precipitatorshave high efficiency up to 99Vofor large particles, but they have poor performancefor particles of size less than is 0.1 pm in diameter.The efficiency of precipitators high with reasonable in flue gasesbut drops for'low sulphurcontentcoals;99Vofor sulphur content 37o sulphur and 83Vofor 0.5Vosulphur. Fabric filters in form of bag lnuses have also been employed and are located before the flue gases enter the stack. Thermal Pollution
Of theseNOz, nitrogenoxides,is a majorconcernas a pollutant. is soluble It in water and so has adverseaff'ect on human health as it enters the lungs on inhaling and combining with moisture converts to nitrous and nitric acids, which danngethe lungs. At ievels of 25100 parts per million NO, can cause acutebronchitis and pneumonia. Emissionof NO_, can be controlledby fitting advanced technologyburners which can assuremore complete combustion, thereby reducing theseoxides from being emitted. These can also be removedfrom the combustionproducts by absorptionprocessby certain solventsgoing on to the stock.
Steam fronr lowpressureturbine has to be liquefied in a condenser and reduced to lowest possible temperatureto maximize the thermodynamic practicallyachievableis about efficiency. The best efficiency of steamcycle the cycle end must be removed' of 4\Vo.It meansthat60Vo the heat in steamat This is achievedby following two methods' cooling tubes of seaor river 1. Once through circulation through condenser of water where available.This raises the temperature water in these two sources and threatenssea and river life around in sea and downstream in river. ThesE,are serious environmental objections and many times cannot be overruled ard also there may be legislation againstit. 2. Cooling tov,ers Cool water is circulatedrottnd the condensertube to remove heat from the exhaust steam in order to condenseit. The
Gfrfud ffiffii
I
Mociern PowerSysteqAnaiysis
lntrcCuction T.TT POWER SYSTEMENGINEERSAND POWER SYSTEM STUDIES
sffi EEF
circulating water gets hot in the process. is pumped to cooling tower tt and is sprayedthrough nozzlesinto a rising volume of air. Some of the water evaporates providingcooling.The latentheat of water is 2 x 106 J/kg and cooling can occur fast, But this has the disaclvantage raising of
u n o e s t r a o t e Je v e l s l n t h c s u l r f t l u n d l n g r e a s . t a
The power system engineerof the first decadeof the twentyfirst century has abreastof the recent scientific advancesand the latest techniques.On the planning side, he or she has to make decisions on how much electricity to generatewhere, when, and by using what fuel. He has to be involved in and transmission. He constructiontasksof greatmagnitudeboth in generation has to solve the problemsof planning and coordinatedoperationof a vast and complex power network, so as to achieve a high degree of economy and reliability. In a country like India, he has to additionally face the perennial problem of power shortages and to evolve strategies energyconservation for and load management. improvementand expansion a power system, of For planning the operation, a power system engineerneeds load flow studies,short circuit studies, and stability studies.He has to know the principlesof economicload despatchand load frequency control. All these problems are dealt with in the next few chapters after some basic concepts in the theory of transmission lines are discussed.The solutions to these problems and the enormouscontribution problems of to made by digital cornputers solve the planning and operational power systemsis also investigated. I.I2 USE OF COMPUTERS AND MICR.OPROCESSOiTS
course the water evaporated must be macleup in the systemby adcting fresh water from the source. Closed cooling towers where condenr;ate flows through tubcs anclair is blown in thesetubesavoidsthe humidity problembut at a very high cost. In India only v,et towers are being used. Electromagnetic Radiation from Overhead Lines
Biological effects of electromagnetic radiation from power lines and even cables in close proximity of buildings have recently attractedattentionand have also causedsomeconcern. Power frequency(50 or 60 Hz) and even their harmonics are not considered harmful. Investigations carried out in certain advanced countries have so far proved inconclusive. The electrical and electronics engineers, while being aware of this controversy, must know that many other environmentalagentsare moving around that can causefar greater harm to human health than does electromagnetic radiation. As a piece of information it may be quoted that directly under an overhead line of 400 kV, the electricfield strengthis 11000 V/m and magnericflux density (dependingon current) may be as much as 40 ptT. Electric field strengthin the rangeof 1000015000 v/m is consideredsafe. Visual and Audible Impacts These environmentalproblems are causedby the following factors. l. Right of way acquiresland underneath. Not a seriousproblernin India at present.Could be a problem in future. 2. Lines converging a large substation at mar the beauty of the lanclscape around. Underground cablesas alternativeare too expensivea proposition except in congestecl city areas. 3' Radio interference (RI) has to be taken into account and counteredbv varlous means. 4. Phenomenon corona (a sort of electric dischargearound the high of tension line) producesa hissingnoise which is aucliblewhen habitation is in close proximity. At the to'wers great attention must be paid to tightness of joints, avoidance of sharp edges and use of earth screen shielding to lirnit audible noise to acceptable levels. 5' Workers inside a power plant are subjectedto various kinds of noise (particularly near the turbines) and vibration of floor. To reduce this uoise to tolerable level foundations and vibration filters have to be designed properly and simulation studiescarried out. The worker nlust be given regularmedical examinations and sound medical advice.
were AC and DC lirl Jlhef irst rnethos solvingvariouspowcr systemproblenis were used for load in developed early 1930s.AC analysers network analysers florv and stability studieswhereasDC were preferredfor shortcircuitstudies. were developed 1940sand were used in conjuncin Analogue compLrters tion with AC network analyserto solve variousproblemsfor offlinestudies. In 1950s many analogue devices were developed to control the online Ii'equencyand tieline controt. tunctions such as genelationrontrol, The 1950s also saw the advent of digital computerswhich were first used to solve a. load flow problem in 1956. Power system studiesby computers gave greater flexibility, accuracy,speedand economy.Till 1970s,there was a widespreaduse of computersin systemanalysis.With the entry of microprocessors the arena,now, besidesmain frame compLlters, mini, micro and in personalcomputersare all increasinglybeing used to carry out various power systern studies and solve power system problems for offline and online applications. Offline applications include research, routine evaluation of system performanceand data assimilationand retrieval. It is mainly usedfor planning and arralysing some new aspects of the system. Online and real time applications include datalogging and the monitoring of the system state.
rytrfi\ r
powerSvstem tutodern Anaivsis
'.1g.r.,'""
A large central computer is used in central load despatch centres for cc<ln<lmic securc and controlof'largc integrated systems. Microprocessors ancl computersinstalledin generating stations control various local processes such as startingup of a generator from the cold state,etc. Table 1.2 depictsthe time microprocessors. some of these problemsare tackled in this book. T a b l e1.2
Tirne scale Milliseconds 2 s 5 minutes 10 minfew hours  dofew hoursl week I month 6 months I yr 10 years Control Problems Relaying and system voltage control and excitation control AGC (Automatic generation conrrol) ED (Economic despatch) Securityanalysis UC (Unit commitment) Mai ntcrrancc schedLrl i ng Systernplanning (modification/extension)
F
1.13 PROBLEMS FACING INDIAN POWER INDUSTR.Y AND ITS CHOICES The electricity requilements of .[ndia have giown tremendously anC the demand has been running ahead of supplyl Electricity generation and t r ans m i s s i op ro c c s s cis In d i a a rc v c ry i neffi ci cnt n c< l l npari soni tl r those n n i w of somedeveloped countries. per one estimate, India generating As in capacity is utilized on an average 360t) hours out of 8760 hclursin a year, r,vhilein for Japanit is rrsed 5 t00 hours.ll' the utilizationlactor could be increascd, lbr it should be possibleto avoid power cuts.The transmission loss in 199798 on a national basiswas 23.68Vo consisting both technicallossesin transmisof sion lines ancltransfonners, also nontechnical and lossescaused energy by thefts and meters not being read correctiy.It should be possibleto achieve considerable savingby leducingthis lossto 1570by the end of the Tenth Five Year Plan by rrsing well known ways and nreans and by adooting sound commercial practices.Further, evcry attempt should be made to improve system load factors by flattening the load curve by giving proper tariff incentives and taking other administrativem.easures. As per the Central Electricity Authority's (CEA) sixteenthannual power survey of India report, the all India load factor up to 199899was of the order of 78Vo.In future it is likely to be 7I7o.By 200i,5.07 lakh of villages(86Vo) havebeenelectrified and 117 lakh of pumpsetshave been energized. Assuming a very modest averageannualenergy growth of 5Vo,India's electrical energy requirementin the year 2010 will be enormouslyhigh. A difficult and challengingtask of planning,engineeringand constructingnew power stationsis imrninentto rneetthis situation. The governnlent has bLrilt
Farakka severalsuper thermal stationssuch as at Singrauli (Uttar Pradesh), (AndhraPradesh)and (Madhya Pradesh), Rarnagundam (West Bengal),Korba Neyveli (Tamil Nadu), Chandrapur (Maharashtra)all in coal mining areas, 2000 MW*. Manv more super thermal plants would be built in future. Intensive work must be conductedon boiler furnaces to burn coal with high ash content. Nationai Thennal Power Corporation(NTPC) is in chargeof theselarge scale generationprojects. Hydro power will continue to remain cheaper than the other types for the next decade.As mentioned earlier, India has so far developed only around l87o of its estimatedtotal hydro potentialof 89000 MW. The utilization of in this perennialsource of energy would involve massiveinvestments dams, system. The Central Electricity Authorchannelsand generationtransrnission ity, the Planning Commissionand the Ministry of Power are coordinating to work out a perspectiveplan to develop all hydroelectric sourcesby the end of this century to be executed by the National Hydro Power Corporation (NHPC). NTPC has also startedrecently developmentof hydro plants. Nuclear energy assumesspecialsignificancein energy planning in India. and its poor quality, India has no choice but Becauseof limited coal reserves keep going on with its nuclear energy plans. According to the Atomic to to will increase 10000 Energy Commission,India's nuclearpower generation to MW by year 2010. Everything seems be set for a take off in nuclearpowel' production using the country's thorium reservesin breederreactors. \other nonIn India, concerted efforts to develop solar energy and the growing so sourcesof energy needto be emphasized, that conventional may be conserved.To can clemancl be met and depletingfbssil fuel resources that the coal productionwill have meet the energyrequirement,it is expected 2005 lts cotttpltrcdto q)orc than .150nrillion totts itt 200'+ to to be ipclcascd 180 million tonnesin 1988. A number of 400 kV lines are operating successfullysince 1980s as a This was the firsi stepin working towards nationalgrid. alreacly. mentioned to go in for even higher voltages(800 kV). It is There is a need in future expecredrhat by the year 2Ol112,5400 ckt krn of 800 kV lines and 48000 Also lines may be sericsand ckt kni gf 400 kV lines would be in operation. to carry huge blocks of power with greaterstability. There shunt compensated is a needfor constructingHVDC (High Voltage DC) links in the country since more power at the samevoltageand require DC lines can carry considerably fewer conductors. A 400 kV SingrauliVindhyachal of 500 MW capacity by has beencommissioned NPTC (National first HVDC backtobackscheme power Transmission Corporation) followed by first pointtopoint bulk + of EHVDC transmission 1500 MW at 500 kV over a distanceof 915 km on from Rihandto Delhi, Power Grid recentlycommissioned 14'Feb.2003 a
'k NTPC has also built seven gasbased combined cycle power stationssuch as Anta and Auraiya.
t
2000 MW TalcherKolar + 500 kV HVDC bipole transmissionsystem thus enabling excesspower from East to flow to South. 7000 ckt km of + 500 kV HVDC line is expected by Z0llI2. At the time of writing, the whole energy sce is so clouded with future. However, certain trends that will decide the future developments of electric power industry are clear. Generally,unit size will go further up from 500 MW. A higher voltage (7651 1200 kV) will come eventually at the transmissionlevel. There is little chance for sixphasetransmission becomingpopular though there are few suchlines in USA. More of HVDC lines will do. in operation.As populhtion has already touched the 1000 million mark in India, we may see a trend to go toward undergroundtransmissionin urban areas. Public sector investment in power has increasedfrom Rs 2600 million in the First Plan to Rs 242330 million in the SevenrhPlan (1985 90). Shortfall in the Sixth Plan has been around 26Vo. There have been serious power shortagesand generationand availability of power in turn have lagged too much from the industrial, agricultural and domestic requiremeni. Huge amounts of funds (of the order of Rs. 1893200million) will be required if we have to achievepower surplusposition by the time we reach the terminal year to the XI Plan (201I2012). Otherwise achieving a rarget of 975 billion units of electric power will remain an utopian dream. Power grid is planning creation of transmissionhighways to conserve Rightofway. Strong national grid is being developedin phasedmanner. In 20Ol the interregional capacitywas 5000 MW. It is Lxpecredthat by 2OlI12, it will be 30000 Mw. Huge investmentis planned to the tune of us $ 20 billion in the coming decade.presenr figures for HVDC is 3136 ckt km, 800 kV is 950 ckt km, 400 kV is 45500 ckt krn and.220/132 is 215000 kv ckt km. Stateoftheart technologieswhich are, being used in India currently are HVDC bipole, HVDC backtoback, svc (static var compensator), FACTs (Flexible AC Transmissions) devices etc. Improved o and M (Operation and Maintenance) technologieswhich are being used tgday are hotline maintenance,emergencyrestoration system, thermovision scanning, etc. Because of power shortages,many of the industries, particularly powerintensive ones,have installed their own captive power plants.* Curcently 20Vo of electricity generatedin lndia comesfrom the captive power plants and this is bound to go up in the future. Consortiumof industrial .onru.rs should be encouragedto put up coalbasedcaptive plants. Import should be liberalized to support this activity.
x Captive diesel plants (and small diesel sets for commercial and domestic uses) are very uneconomical from a national point of view. Apart from being lower efficiency plants they use diesel which should be conservedfor transportationsector.
real time control of power system.It may also be pointed out that this book will also help in training and preparingthe large number of professionals trained in computeraided power system operationand control that would be required to handle v
CES REFEREN
Books l. Nagrath,I.J. and D.P. Kothari, Electric Machines,Tata McGrawHill. New Delhi. 3rd edn, 1997. 2. Eilgerd, O.1., Basic Electric Power Engineering,Reading,Mass., 1977. 3. Kashkari,C., Energy Resources, Demandand Conservation with SpecialReference to India, Tata McGrawHill, New Delhi, 1975. 4. Parikh,Kirit, .sacondIndia studiesEnergy, Macmillian, New Delhi, 1976. 5. Sullivan, R.L, Power System Planning,McGrawHill, New york, 1977. 6. S. Krotzki, B.G.A. and W.A. Vopat, Power Station Engineeringand Economy, McGrawHill, New York. 1960. 7 . Car,T.H.,Electric Power Stations, vols I and lI, Chapman Hall, London, 1944. and 8. Central Electricity GeneratingBoard, Modern Power Station Practice, 2nd edn, ' Pergamon, L976. 9t Golding, E.W., The Generationo.f' Electricitlt b1t Wind Power, Ctnpman and Hall, London,1976. i 10. McMillan, J.T., et. al., Energy Resoorces and Supp\t, Wiley, London, 1976. I L Bennet, D.J., The Elements NuclearPoveer, oJ' Longman,1972. 12. Berkowitz,D.A:,. Power Generationand Environmentalchange, M.I.T. press, Cambridge, Mass., 1972. 13. Steinberg, M.J. and T.H. Smith, Econonrloadingof Power Plants and Electric Wiley, New York, 1943. Systems, 14. Power System Planning and Operations:Future Problemsand ResearchNeeds. EPRI EL377SR, February1977. 15. Twidell, J.w. and A.D. weir, Renewuble Energy Resources, and F. N, spon, E. London. 1986. 16. Mahalanabis, A.K., D.P. Kothari and S.l. Ahson, Conxputer Aided Power,S),srenr Analysisand Control, Tata McGrawHill, New Delhi, 1988. 17. RobertNoyes (Ed.), Cogeneration Steamand Electric Power, NoyesDali Corp., of usA, 1978. 18. weedy, B.M. and B.J. cory, ElectricPower svstems,4th edn, wiley, New york, 1998. 19. cEA 12 Annual survey of Power Report,Aug. 1985; l4th Report,March l99l; 16th ElectricPower Survey of India, Sept 2000. 20. Kothari, D.P. and D.K. sharma (Eds), Energy En.gineering. Theory and practice, S. Chand,2000. 21. Kothari, D.P. and I.J. Nagrath, Basic Electrical Engineering, 2nd edn, Tata McGrawHill, New Delhi, 2002. (Ch. 15).
ffiffi
powerSystem N4odern Alqtysis
,
20. Kothari, D.P. and D.K. Sharma(Eds), Energy Engineering. Theory and practice, S. Chand, 2000. 21. Kothari, D,P. and I.J. Nagrath, Basic Electrical Engineering, 2nd edn, Tata McGrawHill, New Delhi, 2002. (Ch. l5). 22. Wehenkel,L.A. Automatic Learning Techniques Power Systems, in Norwell MA: Kiuwer, i997. 23. Philipson,L and H. Lee Willis, (Jnderstanding Electric Utilities and Deregulation, Marcel Dekker Inc, NY. 1999. Papers 24. Kusko, A., 'A Predictionof Power System Development,19682030', IEEE Spectrum, Apl. 1968,75. 25. Fink, L. and K. Carlsen,'Operating underS/ress and Strain',IEEE Spectrum,Mar. r978. 26. Talukdar, s.N., et. al., 'Methodsfor Assessing energyManagement options', IEEE Trans.,Jan. 1981,PAS100, no. I, 273. 27. Morgen, M.G. and S.N. Talukdar, 'Electric Power Load Management: some Technological,Economic, Regularity and Social Issues',Proc. IEEE, Feb. L979, vol.67,no.2,241. 28. Sachdev, M.S.,'LoadForecastingBibliography, IEEE Trans., PAS96, 1977,697. 29. Spom, P., 'Our EnvironmentOptions on the Way into the Future', ibid May 1977,49. 30. Kothari D.P., Energy ProblemsFacing the Third World, Seminar to the BioPhysics Workshop,8 Oct., 1986 Trieste Italy 31. Kothari, D.P,'Energy system Planning and Energy conservation',presented at
YYIV Nntinnnl fnnvontinn n s ^F vJ llltr ttrL, NI^r', r \w w ntt y v l hlil ,
^a
j
2.I
INTRODUCTION
32. Kothari, D.P. et. a1., 'Minimization of Air Pollution due to Thermal Plants'. ,I1E (India), Feb. 1977, 57, 65. 33. Kothari, D.P, and J. Nanda, 'Power Supply Scenarioin India' 'Retrospects and Prospects', Proc. NPC Cong., on Captive Power Generation,New Delhi, Mar. 1986. 34. NationalSolarEnergyConvention, Organised SESI, 13 Dec. 1988,Hyderabad. by 35. Kothari, D.P., "Mini and Micro FlydropowerSystemsin India", invited chapterin the book, Energy Resources and Technology,Scientific Publishers,1992, pp 147158. 36. PowerLine, vol.5. no. 9, June2001. 37. United Nations.'Electricity Costs atxd Tariffs: A General Study; 1972. 38. Shikha,T.S. Bhatti and D.P. Kothari,"Wind as an Ecofriendly Energy Sourceto meet the Electricity Needs of sAARC Region", Proc. Int. conf. (icME 2001), BUET, Dhaka, Bangladesh. Dec. 2001, pp 1116. 39. Bansal,R. C., D.P. Kothari & T. S. Bhatti, "On Some of the Design Aspectsof Wind Energy ConversionSystems",Int. J. of Energy Conversion and Managment, Vol. 43, 16, Nov. 2002,pp.21752187. 40. l). P. Kothari and Amit Arora, "Fuel Cells in TransporationBeyond Batteries", Proc. Nut. Conf. on Transportation Systems, Delhi, April 2002, pp. 173176. IIT 41. Saxena,Anshu, D. P. Kothari et al, "Analysisof Multimedia and Hypermediafor ComputerSimulation and Growtft", EJEISA, UK, Vol 3, 1 Sep 2001, 1428.
Ek I wu,
lOQO I 7QL,
line as an the performance of a transmission The four parameterswhich affect e l e m e n t o f a p o w e r S y s t e m a r e i n d u c t a n c e , c a p a c ito leakale s i s t alinee a n d t a n c e , r e over n c , .rL..++ ^^nrrrrnrAnce which is normally due
illffilf?; r
L*",,;i;;.";;r""i"J
#;;t;
lines' transmission rhis ii overhead
1'e' lnductance and resistance'
r 'i+r +ha caricq line narameters,
ffiffir"##r'"*
the atong lineandthevtogether oi.itiuu,.d
th" ,"ri., imPedance of the line' i"; from a power system the most dominant line parameter Inductar,c. i, ;;f; e n g i n e e r , s v i e w p o i n t . A s w e s h a l l s e e i n l a t e r c hline' e r s , i t i s t h e i n d u c t i v e apt capacity of a reactancewhich limits the transmission 2,2 DEFINITION OF INDUCTANCE
given by Voltage induced in a circuit is
, =VY
(2.r)
(WbT)' of the circuit in weberturns ", ,ti" flux linkages (2.2)
in anceof the circ'lit in henrys' which near magneticcircuit, i'e'' a circuit current such that Ies vary linearly with
This can be written in the form drb di , di., ,  : L  : v e = dr dt dr
+f ,,il
po**, Syrtm An"lyri, Modern L = ! H
lnductance Resistance Transmission and of Lines 
ffffi
or
I
(2.3)
,f\]y
i i r ' i+r
1 t \ \.
t
A = LI
where ) and I arc the rms values of flux linkages and current respectively. These are of course in phase. Eq. (2.1) by ir, we get the steady state AC volrage drop i + due to alternating flux linkages as Y= jwLI = jt^r) V e.5) On similar lines,the mutual inductancebetweentwo circuits is defined asthe flux linkages of one circuit due to current in another,i.e.,
)t, M r rL= n Iz ,
e.4)
,'.rt l
t, ?5
\ \ ll! t'l
l
t.rtt
i I
Replacing
ir."
r1/
(2.6)
where
Fig. 2.1 Flux linkages due to internal flux (crosssectional view)
The voltage drop in circuit 1 due to current in circuit 2 is V, = jwMnlz = 7tl\12 V
(2.7)
H, = magnetic field intensity (AT/m) /y = current enclosed(A) By symmetry, H, is.constantand is in direction of ds all alongithe circular path. Therefore,from Eq. (2.8) we have
The conceptof mutual inductance requiredwhile consideringthe coupling is betweenparallel lines and the influence of power lines on telephonelines. 2.3 FLUX LINKAGES OF AN ISOTATED CURRENT. CARRYTNG coNqucroR
2rryH,=1,
Assrrmino rrniform crrrrenf dcnsifv*
(2.e)
(2.10)
Transmissionlines are composedof parallel contluctorswhich, for all practical purposes, can be considered infinitely long. Let us first developexpressions as for flux linkages of a long isolatedcurrentcanying cylindricat conductor with return path lying at infinity. This systemforms a singleturncircuit, flux linking which is in the form of circular lines concentricto the conductor.The total flux can be divided into two parts,that which is internal to the conductor and the flux externalto the conductor.Such a division is helpful as the internal flux progressivelylinks a smalleramountof current as we proceedinwards towards the centreof the conductor,while the external flux alwayslinks the total current inside the conductor. Flux Linkages due to Internal Flux
\rrr') \r") FromEqs.(2.9) and(2.10),we obtain H,.=)!t AT/m
2Tr"
,' " = (  ) t : [ 4 ) ,
(2.rr)
(2.r2)
The flux density By, y metres from the centre of the conductorsis
Bu=pHu=:+ Wb/m2 z ztr
r ntI
Figure 2.1 shows the crosssectional vi,ew of a long cylindrical conductor carrying current 1. The mmf round a concentricclosed circular path of radius y internal to the conductor as shown in the figure is {nr.ds =Iy (Ampere'slaw)
where p is the permeability of the conductor. Consider now an infinitesimal tubular elementof thicknessdy and length one metre. The flux in the tubular element dd = Bu dy webers links the fractional trrrn (Iril  yzly'l resulting in flux linkages of
(2,8)
*For power frequency of 50 Hz, it is quite reasonableto assume uniform current density. The effect of nonuniform current density is consideredlater in this chapter while treating resistance.
ModernPqwer System Analysis
of Lines and lnductance Resistance rransmission
ffi.$ t
(2.r3)
Integrating,we get the total internal flux lin
The flux dd containedin the tubular elementof thickness dy is
dd = +dy (2.14)
Wb/m lengthof conductor
^^, = I #f d y :ffw a t^
^rn =T*10/ and Lint= ]xto7 z Flux Linkage to Conductor rVm Two Points External
f 
For a relative permeability lf,, =  (nonmagnetic conductor), 1t = 4n x l0'[Vm. therefore wbT/m
The flux dQbeing external to the conductor links all the current in the conductor which together with the return conductor at infinity forms a single return, such that its flux linkages are given by d)=1 xd6=FI d, ' 2ny Therefore, the total flux linkages of the conductor due to flux between points P, and Pr is
p D " , , f n ' tn' \," =  dy  t" I ln "2 wbT/m ,1Dt2 n.v 2r Dr
(2.1s) (2.16)
due to Flux Between
where ln standsfor natural logarithm*. Since Fr=I, F = 4t x107
n
Figure 2.2 showstwo pointsP, and Prat distances and Drftoma conductor D, which carries a cunent of 1 amperes. the conductoris far removed from the As return current path, the magnetic field external to the conductor is concentric circles around the conductor and therefore all the flux between P, and Pr lines within the concentriccylindrical surfacespassingthrough P, and P2.
(2.r7) ) r z= 2 x l} t l ln =L wb/ m Dr The inductanceof the conductor contributedbv the flux included between points P, and Pr is then
Lrz = 2 x I01 fn ?
or L,n = 0.461 los.D' mH,/km LL
"Dl
Dl
fV*
(2.18)
(2.1e)
Point
Flux Linkages
due to Flux up to an External
Let the external point be at distanceD from the centre of the conductor.Flux linkages of the conductor due to external flux (from the surface of the conductor up to the externalpoint) is obtainedfrom Eq. (2.17)by substitutin D t = r and E Dz = D, i'e', D (2.20) r Total flux linkages of the conductor due to internal and externalflux are ).*,= 2x701 lln
)= )in,* )"*,
Fig.2.2 Fluxlinkages to flux between due points PI, P2 external Magnetic field intensity at distancey from the conductor is I H.,= ' Z r r y AT/m
= I x 1 o  7 + Z x r o  : I. I n 2 2 r
*Throughout the book ln denotes natural logarithm (base e), while log denotes logarithm to base 10.
W
Modern Power systgmAnalysis
 
Lines of and lnductance FlesistanaeTransmission
li$Bfr f
 2x = . ^  1 0  , / t +4r n 2 ) [r )
x ro_71 *J_r, t"
Iet ,t  ,rr/4 = 0.7788r
\ = 2 x ro4rh + wbT/m rl L= 2x 1o7 Ir w^ n
(2.Zra)
Inductance of the conductor due to flux up to an external point is therefore
(z.zrb)
Here r' can be regarded as the radius of a fictitious conductor with no internal inductancebut the same total inductanceas the actual conductor. 2.4 INDUCTANCE OF A SINGLE.PHASE TWO.WIRE LINE
To start with, let us considerthe flux linkages of the circuit caused by current in conductor 1 only. We make three observationsin regard to these flux linkages: 1. External flux from 11to (D  ,) links all the current It in conductor 1. 2. External flux from (D  r) to (D + rr) links a current whose magnitude progressivelyreducesfrom Irto zero along this distance,becauseof the effect of negative current flowing in conductor 2. 3. Flux beyond (D + 12)links a net cunent of zero. For calculating the total inductance due to current in conductor 1, a simplifying assumption will now be made. If D is much greater than rt and 12 (which is normally the casefor overheadlines), it can be assumedthat the flux from (D  r) to the centre of conductor 2 links all the current ^Ir and the flux from the centre of conductor2 to (D + rr) links zero current*. Based on the above assumption,the flux linkages of the circuit causedby current in conductor 1 as per Eq. (2.2Ia) are
Considera simple twowire line composed solid round conductors of carrying currents1, and 1, as shown in Fig. 2.3.n a singlephase line, 11+ Ir= Q Iz=  It
) r = 2 x 1 0  7 1 ,l n  L r\
(2.22a)
The inductance of the conductor due to current in conductor 1 only is then
l" Lt= 2 x 107 + f'1
(2.22b)
Similarly, the inductanceof the circuit due to current in conductor 2 is ' D . (2'23) Lz=2x107h r,2
r l oriEn c vurr ruh a csrvrr n c r r r n s i f i n n u r r r v f! ahve^n^r^ et r n r v v ^ fhe flrrx linkaoes and likewise fhe indttcfances
of the circuit causedby currentin each conductorccnsideredseparatelymay be addedto obtain the total circuit inductance.Therefore,for the complete circuit L= Lt+ 4=
Drz D D+rz
4 x 10'ln
FVm
(2.24)
If r/r= r'z= /; then L= 4 x 10'ln D// Wm
(2.25a)
Fig. 2.3 Singlephase twowire and the magnetic line fielddueto current in conductor only 1 It is important to note that the effect of earth's presenceon magneticfield geometry* is insignificant. This is so because relative permeabilityof earth the is about the sameas that of air and its electricalconductivitv is relativelv small.
*The electric field geometry will, however, be very much affected as we shall see later while dealing with capacitance.
(2.zsb) L  0.92t log Dlr' mHlkm Transmission lines are infinitely long comparedto D in practical situations and therefore the end effects in the above derivation have been neglected.
2.5 CONDUCTOR TYPES
So far we have considered transmission lines consisting of single solid cylindrical conductors for forward and return paths. To provide the necessary flexibility for stringing, conductorsused in practiceare always strandedexcept *Kimbark [l9] has shownthat the resultsbasedon this assumption fairly are even when D is not much larger than 11and 12. accurate
52

I
Vodern PowerSystemAnalysis
lnelr rn+ana^ ^^f rrrLrL.vtclttutt ct'ttu a t ^ ^ l ^ . ^   nt'stl'latlug
ut
^t
r^
I lallsrnlsslon
^:
!

,
r
.

fo, .r.ry small crosssectionalareas.Stranded conductors are composed of strands of wire, electrically in parallel, with alternate layers spiralled in opposite direction to prevent unwinding. The total number of strands (M) in concentrically stranded cables with total annular space filled with strands of (rD uniform diameter is given by (2.26a) N=3x'3x+l where x is the number of layers wherein the single central strandis counted as the first layer. The overall diameter(D) of a strandedconductoris
Unes
2.6 FIUX LINI{AGES OF ONE CONDUCTOR IN A GROUP
l*
!
5.* "
^
As shown in Fig. 2.5, considera group of n pnallel round conductorscarrying phasor currents Ip 12,, I, vvhose sum equals zero. I)istances of these an expression for the total flux linkages of the ith conductor of the group consideringflux up to the point P only.
lt ult,..t un.
us oDtam
P(2xr)d
(2.26b) (,
3 2
o
n
Aluminium is now the most commonlyemployedconductormaterial.It has the advdntages being cheaperand lighter than copper though with less of conductivity and tensile strength.Low density and low conductivity result in larger overall conductordiameter,which offers anotherincidental advantagein high voltage lines. Increased diameter results in reduced electrical stress at conductor surfacefor a given voltage so that the line is coronafree. The low tensile strengthof aluminium conductors made up by providing central is strandsof high tensilestrengthsteel.Sucha conductoris known as alurninium conductor steelreinforced (ACSR) and is most commonly used in overhead transmission lines. Figure 2.4 shows the crosssectional view of an ACSR conductor wrth 24 strands of aluminium and 7 strands of steel.
Steelstrands
4 I
Fig. 2.5 Arbitrary groupof n parallel roundconductors carrying currents The flux linkagesof ith conductordue to its own current1,(self linkages)are given by [see F,q. (2.21)]
)ii= 2 x 107 h! t,
ri
WbT/m
(2.27)
The flux linkages of conductor i due to current in conductor 7 1rlf"r to Eq. ( 2 . 1 7 ) li s
5,,= 2 x l;il,fn a
Dij
WbT/m
(2.28)
Aluminium strands
where Du is the distanceof ith conductor from 7th conductorcarrying current 1r.From F,q. (2.27) and by repeateduse of Eq. (2.28), rhe rotal flux linkages of conductor i due to flux up to point P are
)i = Xir + )iz + ... + )ii *... * )in
view steelstrands, aluminium 24 strands Fig.2.4 Crosssectional of ACSR7 line, expandedACSR conductors In extra high voltage (EHV) transmission are used. Theseare provided with paper or hessian between various layers of strandsso as to increasethe overall conductordiameter in an attemptto reduce electrical stressat conductor surfaceand prevent corona. The most effective per coronafree EHV linesis to provide several conductors way of constructing phase in suitable geometrical configuration. These are known as bundled conductors and are a common practice now for EHV lines.
= 2 x rca(r, t'*
\^
Dt
+ 4 u !  + . . . + ,,ln D i 1 ' rl. D,z
+... + In
The above equation can be reorganizedas
)i = 2xro'[[r,
h + +h n [ * . +
I,m]+..+
But,
+ (/, ln D r + I r h D z + . . . + I i k D , + . . + I n ^ O , )   (1, + Iz +... + In). I,
lnductanec
anrl
Fls ai r an , r a r q orru g a a
. ^i T^! ur I rallsmlssfon

Substiiuti'g for /n in the secondterm of Eq. (2.29)and simplifying, we have
Llnes
f.. k,...i<Er
I 
)i' = 2 * totl t, ml * L,h L_.+...+1, I rna Dir'""'Diz""''''^'rti
L(' * I,l n = e+ . ..* l i l n at. . . * I ,  r 1 nD r ) 1) '4 t ) In order to accountfor total flux linkages of conductor i, let the point p now recedeto infinity. The terms such as ln D1/Dn,etc. approachln = t o. Also for the sake of symmetry, denoting ,.{^ D,,,wi have
f/
I a /1 lA\t^cr Applying Eq. (230) ro filamenr i of conductor A, weobtain its flux tintages u,
zxfir !
h+a6Ja...*rnl*...*rn I
, (
Dir, ^^' 6  J  1 . . .' *' l'r t ) n " Diz, D,^, )
 2 x t o 4 l t " ]  * 7
^, \
= 2 x lo716 (P,r:D,r,...D,t
The inductanceoffilament f is then
(D,rD,r. . . D,,.. . D,n)r'"
Wb_T/m
l i = z x 1 o rI t r n * *
U Dt
r r ' , ' !  + 1 , l n 1 " '
D,2 Dii
+...+I^t" wbr/m +)
2.7 INDUCTANCE OF COMPOSITE CONDUCTOR LINES
(2.30)
(Dn'"' D,,' D,^')1/^' "' FVm 7/n (D,rD,r...D,, D '* . The average inductance the filaments composite of of conductor is A Lt+L2+4+"'+Ln Luu,=
L i = ] !
: 2nx70t,n
(2.3r)
Sinceconductor is co#posed n filaments a of electrically parallel,its in inductance is
_A
We are now ready to study the inductanceof transmissionlines composedof compositeconductors. Future 2.6 showssuch a singlephase line comprising compositeconductors and B with A having n paraliel filaments A and B having mt parallelfilaments.Though the inductance each filament will of be somewhat different(theirresistances be equalif conductordiameters will are chosento be unttorm), it is sufficiently accuratefo assum.e the currentis that equally divided among the filaments of each compositeconductor.Thus, each filament of A is taken to carry a current I/n, while each filament of conductor B carries the return current of  Ihnt.
, _Lu,,_ \+ 12+...+L,
=
.*pr"f,rion rir"#i"t indu*ance Eq.(2.31)i" for from E;. 3'::r', *ru"rorr:rr"rh"
l ( D n ,. . .D rj , . . . D r . , ) . . . ( D ^.,. .p r j ,. . . D , ^ , ) . . . L e = 2 x 1 0  7l n (Dnr, . Dnj,... Dr^,17r/ntn ..
o
2
o
o
[ ( D n . .D u . . . D r n ) . . . ( D i t . . n , t . . 4.
(Dnt... Dni ... D rn)]r,n'
Htm (2.33)
o
2l
11?x.::ir:l
I
,he,arsumenrthelogarithm Eq. (2.33) of in is rhem,nth :f
t:?,T,:!"^: u.]tl*'of conductor iri"L.il Atim'
"il"#il#:ili:ffi;
" of
lf
r )
m
int p,;d;;;._,i;ff ffiilil:; l:#1:# *: jr:lr* nryn2.thofnz to fii";;;;rrj"",i,sts setof product pertains yflr,""1i) term
T:l
Composite conductor A
Fig. 2.6 singlephase consisting two composite line of conductors
!:{Kr,"li:t:t:y? GMD is also called
rnl denominaror u, I!:::::1li:*Ti:r{. of conduc A,and isdefin"o ;;;w;;:;;:;,1;::; tor ,s "r';#;;;":::"*
geometric mean radius (GMR). "J In terms of the above symbols, we can write Eq (2.33) as
auureviate; ;,
56 ,1
I
todern FowerSystemAnalysis
Inductance and Resistance Transmission of Lines (2.34a)
L t = 2 x 1 0  7 h+ '^
Dr.t
Iilm
^Hn^
Note the similarity of the above relation with Eq. (2.22b), which gives the inductanceof one conductorof a singlephase line for the specialcaseof two solid, round conductors.In Eq. (2.22b) r\ is the self GMD of a single conductor and D is the mutual GMD of two single conductors. The inductanceof the composite conductor B is determined in a similar manner, and the total inductance of the line is L= Le+ Ln
srngrelayer oI alurrunlum conductorshownin Fig. 2.8 is 5.04 cm. The diameterof eachstrand 1.6g cm. is Determinethe 50 Hz reactance I rn spacing;neglectthe effectof the central at strand of steel and advance reasonsfor the same. Solution The conductivity of steelbeing much poorer than that of aluminium and the internal inductance of steel strandsbeing ptimes that of aluminium strands,the current conductedby the central strandsof steel can be assumedto be zero. Diameterof steel strand= 5.04 2 x 1.68= 1.68 cm. Thus, all strands are of the same diameter, say d. For the arrangement of strandsas given in Fig. 2.8a, Dtz= Drc= d Drt= Dn= Du= 2d Dr= Jld
(2.3s)
A conductor is composed of seven identical copper strands, each having a radius r, as shownin Fig. 2.7. Frnd the self GMD of the conductor.
(l(+) dTil,eilf')
 D^^= 2..fir
Fig.2.7 Crosssection a sevenstrand of conductor Solution The self GMD of the sevenstrandconductor is the 49th root of the 49 distances. Thus D, = (V)7 (D1zD'ruD r)u (zr)u )t'on rp Substituting the values of various distances, p.. ((tJ.7788r)7 (2212 3 x 22f x 22rx 2r x 2r)u)t'o' x
(b) Line composed of two ACSR conductors (a) Crosssection ACSR conductor of
, _ us _ 2 r(3 (0 .7 7 8 8))tt7 2.t77r 6U+e
Fig. 2.8
5t
I
I
ModernPower Svstem Analvsis
Lines oi and Resistance Transmission lnductance The self GMD fbr side A is D,A = ((D nD nD n)(DztDzzDn)(D3tDtrDtr))''e Here. D,, = Doc = Dat = 2.5 x 103x 0.7788m and in the Substituting values of various interdistances self distances D16, we get D,A= (2.5 x 103x 0.778U3x 4a x 8\tte = 0. 367 m Similarly, D, B= ( ( 5 x 10 3x 0. 778q2, 4') t 'o = 0 . 1 2 5m Substitutingthe values of D^, D6 andDr, in Eq. (2.25b),we get the various as inductances 8'8 = 0.635 mHlkm L^ = 0.461 los ^ " 0.367 8'8 = 0.85 mH/<m L. = 0.461 toe D 0J25 L = Lt + Ln = 1.485 mH/km If the conductors in this problem are each composed of seven identical as strands in Example2.1, the problemcan be solved by writing t[e conductcr self distancesas Dii= 2'177rt where r, is the stranci raciius. 2.8 INDUCTANCE OF THREE.PHASE LINES
Substitutingd' = 0.7188dand simplifying D , = l . l 5 5 d = 1 . 1 5 5 1 . 6 8= 1 . 9 3c m x D ^ = D s i n c eD > >d 1P L = 0 .4 6 t to e " 1.93 0.789mH /km Loop inductance 2 x 0.789 = 1.578 mHlkm
The arrangement conductorsof a singlephase of transmissionline is shown in Fig.2.9, whereinthe forward circuit is composed three solid wires 2.5 mm of in radius and the return circuit of twowires of radius 5 mm placed symmetrically with respectto the forward circuit. Find the inductanceof each side of the line and that of the complete line. Solution The mutual GMD between sides A and B is D.= ((DMD$) (Dz+Dz) 1D3aD3))tt6
;;;,;,; I
= Loop reactance 1.578 x 314 x 103  0.495 ohms/lcm
()z
4m
4m
l
I
s ( t"l ' I "' t
i
lines. The basic equations So far we have considered only singlephase of to can,however,be easilyadapled the calculation the inductance developed line lines. Figure 2.10 shows the conductorsof a threephase of threephase spacing. with unsymmetrical
Side A
Side B
Dn
Dzs
Fig. 2.9 Arrangement conductors Example of for 2.3 From the figure it is obvious that D t q = D z q = D z s  D . u = J O am Drs=Dy = 10m = D ^ = (6 8 2x 1 0 0 )l /6 8.8 m
spacing view of a threephaseline with unsymmetrical Fig. 2.10 Crosssectional
Ugdern Power SystemA Assume that there is no neutralwire, so that Ir,+Ir+ Ir0 Unsymmetrical spacingcauses flux linkagesand thereforethe inductance the of eachphaseto be differentresultingin unbalanced receivingend voltageseven
en senolnge tages and llne currents are balanced. AIso voltages will be induced in adjacent communication lines even when line currents are balanced. This problem is tackled by exchanging the positions of the conductors at regular intervais aiong the line such that each conductor occupies the original position of every other conductor over an equal distance. Such an exchange of conductor positions is called transposition. A complete transposition cycle is shown in Fig.2.11. This alrangement causes each conductor to have the same average inductance over the transposition cycle. Over the length of one transposition cycle, the total flux linkages and hence net voltage induced in a nearby telephone line is zero.
1
Lines and Resistance Transmission of lnductance But, 1, I I, =  /n, hence A o =2 x 1 0  7I o k (D"D'trDt')'''
r'a
equilateral spacing D"o (DtzDnDrr)t''  equivalent L o = 2 x t 0  7h +
f
= 2 x f O  ?n { : r F V m f
' o D ,
(2.36)
This is the same relation as Eq. (2.34a) where Dn, = D"o, the mutual GMD = lf conductors. ro = 11, r6t we have betweenthe threephase Lo= LO= L, practiceto transpose power lines at regularintervals. the It is not the present is in However,an interchange the positionof the conductors madeat switching stations to balancethe inductanceof the phases.For all practical purposesthe and the inductancc an untransposecl can of line dissynrnrctry bc neglectcd can line. be taken equal to that of a transposed If ttre spacingis equilateral, then D"o= D and
b
Fig. 2.11 A complete transposition cycle To find the averageinductance each conductorof a transposed of line, the flux linkages of the conductor are found for each position it occupies in the iransposeci cycie. appiying Eq. (2.s0) to conciuctor of Fig. z.lI, for section a 1 of the transposition cycle whereina is in position1, b is in position2 and c is in position3, we get
ra = If ro = 11, r,, it follows from F,q. (2.37) that Lo= Ltr= L,
L o = 2 x l 0  7 mI
fmn
(2.37)
) u r= 2 x t o  r,f, , r " *
\" For the second section
''r,
I I", , m ]  * l
Dr,
(
I n 1 j * o  r r , r ,
Dt,)
Show that over the length of one transpositioncycle of a power line, the total line are z,ero, flux linkagesof a nearby telephone ftrr balanced threephase currents.
For tr,ira := the ,".*
'x to?[r" + * 16 k tn . '' t" wuvb: ],;i 16 tn*+ r.,nll " D, uzr *orr
(


\
(l)
c'l
b
c i') \7
\ Average flux linkages of conductor a are
r" h = )o3 z x ro71 !t f',,
/
()
 1 u
r c t l I , , h ! r , , t nD n D r D l t ) t / 3 (
\ 4 ,
+1.ln
(Dt2D2.)3t)t/3
T2
on F19.2.12 Effect of transposition Inducedvoltage of a telephoneline
62

Modern Power Svstem Anatrrqic.
solution Referring to Fig. 2.r2, the flux linkages of the conductor r, of the telephone line are
lnductance and Resistance Transmission of Lines (ii) third and multiple of third harmonic currenrsunder healthy .ondirion, where 1 , , ( 3 ) +I o Q ) + I , ( 3 ) = 3 1 ( 3 ) The harmonic line currentsare troublesome two wavs: in (i) Inducedernf is proportionalto the frequency. (ii) Higher frequencies come within the audiblerange. Thus thereis need to avoid the presence suchharmonic currentson power of Iine from considerations the performance nearbytelephone of of lines. It has been shown above that voltage inducedin a telephoneline running parallel to a power line is reduced to zero if the power line is transposed and provided it carries balanced currents. It was also shown that ptwer line transpositionis ineffective in reducing the inducedtelephoneline uoitug. when power line currents are unbalanced or when they contain third harmonics. Power line transpositionaparrfrom being ineffectiveintroducesmechanical and insulation problems.It is, therefore,easierto eliminate induced voltages by transposing telephone the line instead.In fact, the readercan easily verify that even when the power line currents are unbalanced or when th"y cgntain harmonics,the voltage induced over complete transpositioncycle (called a barrel) of a telephone line is zero.Someinducedvoltagewill alwaysbe present on a telephoneline running parallel to a power line becausein actu4l practice transposition never completelysymmetrical.Therefore,when the lines is run parallel over a considerable length, it is a good practiceto transpose both power and telephone lines. The two transposition cyeles ^re staggered. and the telephoneline is transposed over shorter lengths comparedto the power line.
rc7 , " ^ * + 1 ,
Similarly,
ln *.1
t"*
WbT/m (2.38)
) t 2 = 2 x t o t l, " L n 5
( " ),= ),t)tz
Doz
o * r 6t n + * 1 . 1 n + l w b"_ r i' 'm ( 2 . 3 s ) t v Du, D,, )
The net flux linkages of the telephoneline are
= 2 x tor( h D:, * 16 t" tn? *r" ,"+) wbrim(2.40) " D" ) rhe emfinduced ,n! ,.,"p'#ir"t* ," "3Jti, E,= Zrf),\lm
fjnder balanced load conditions, ), is not very large because there is a cancellationto a great extent of the flux linkages due to Io, 16 and 1r. Such cancellation doesnot takeplacewith harmonr. which are multiplesof "u.r.nts three and are thereforein phase.Consequently,these frequencies, if present. may be very troublesome. lf the power line is f'ully transposedwith respectto the telephone line
),, =
where ),r(I), )/r(II) ancl),,(III) are the flux linkagesof the telephoneline r, in the three transposition sections the power line. of Writing for ),,(l), ,\/2(II)and ),r([l) by repeared of Eq. (2.3g), use we have
A,t= 2 x l0/(tn+ 16+ 1,,)ln (DntDutD,.itt3
)" (I)* 4, (tr)* ),,(III) 3
i,."rp," u fI iA threephase, Hz,15 km long line hasfour No. 4/0 wires (1 cm dia) 50 spaced horizontally 1.5 m apart in a plane. The wires in order are carrying currents 1o, Iu and I, and the fourth wire, which is a neutral, carries ,".o iu.."nt. The currents are:
Similarly,
=2x104(1,+.Ib+1,)ln
(Dn2Db2D,,r)t,,
Io= 30 + 750 A Iu= 25 + j55 A I"= 55  j105 A The line is untransposed. (a) From the fundamentalconsideration, find the flux linkages of the neutral. Also find the voltage induced in the neutral wire. (b) Find the voltage drop in each of the threephase wires.
)r= 2 x 10711,It,+ I,)ln(D"2Db2D,)rt: + ' (2.4r) (D orD arD ,r)r,, If Io + Iu + I, = 0, ), = 0, i.e. voltage induced in the telephone loop is zero over one transpositioncycle of the power line. It may be noted here that the condition Iu+ Iu+ I, = 0 is not satisfiedfor (i) rpwer frequency LG (linetoground fault) currents. where Io+ Iu* Ir=J[o
64
I
I
t a
Modern Power System Analysis
b
lnductance and Resistance Transmission of Lines
n
(:;
(. 1
,r"n.'f(')
c
I n2 ln Dlr l ln2
voltage
(:)
. tr.s n.' t1.5 ml*Solution (a) From Frg.2.I3,
l r.sn1
p o
s e 4 r s c a ulated below:
Fig. 2.13 Arrangement conductors Example of for 2.5
Avo = j2xralx x rs x rd(6 !9e (30+j50)+0.6e3(25+is5)) 3r4 =  (348.6 j204) V +
r \ lwbT/m
D,n)
D o n = 4 .5 m, D b n= 3 m, D rn = 1.5 m Flux linkages of the neutral wire n are  ( I+ 1 ^ l n ' )  = 2 x 1 0  71  l n 1
\" Don "
1* l l n ' '
Drn
Substitutingthe valuesof D,,n,Dg, and D,.n,and simplifying, we get An=  2 x lO' 0.51 I, + 1.1 Iu + 0.405 1") WbTim Since I, =  (Io+ I) (this is easily checkedfrom the given values), 2,, =  2 x l0' (1.1051o 0.695I) WbT/m + The voltage induced in the neutral wire is then Vn = j w ),n x 1 5 x 103 V =or j 3 1 4 x 15 x 103x Z x 107(1 J05Io+ 0.695 I) V y Vn =  j 0 .9 4 2 (1 .105 Io + 0.695I)
A singlephase Hz power line is supported a horizontal crossann. 50 on The spacing between the conductors is 3 m. A telephone line is supported symmetricallybelow the power line as shown in Fig. 2.14. Find the mutual inductancebetweenthe two circuits and the voltageinducedper kilometre in the telephoneline if the current in the power line is 100 A. Assume the telephone Iine current to be zero. Solution Flux linkaees of conductor Z,
( ),r = 2 x Io7,rntl6t) Dt \
Flux linkages of conductor 7, \,2= 2 x 1071lnD' D^
Dr)
= 2 x ro1 h D' I
D1
Substituting the valuesof 1, and 16,and simplifying Vn = 0 .9 4 2 x 1 06 = 100 V (b) From Eq. (2.30), the flux linkages of the conductor a are
(
\ The voltagedrop/metre phase canbe writtenas in a av,,=2xr07iw(,rn
\."
)o=2x ro?[ l* ru rn**r rn*  wurr,rn " r'o D 2D)
I
f'o
r
I
r
\
n + 1' ^ l n l * 1  l c I ) '
D
tr,o
2D)
Since Ir=  (lr,+ Iu), and furthersince ro= rb= rr= r, the expression for AVo can be written in simplified form
\)d/
\ll
Avo= 2 x toty,(r, rn!*
romz) v/m
Similarly, voltage drop/petre of phases b and c can be written as
t " l*O.e
, 'Yt'
I
Fig.2.14 Power and telephone linesfor Example 2.6 Total flux linkage of the telephonecircuit ),=\,r\rz=4x10'lln
D2 Dr
AVo= x r;a iulbo_ 2 # A V , = 2x t o t / , ( 1 ,f n2 + 1 , " + )
Using matrix notation, we can present the result in compact form
66
i I
Modernpower SystemAnalysis
Lines of and Resistance Transmission lnductanee will lead to eachother. (The readercan try other configurations verity that these equivalentequilateralspacingis to low D..) Applying the methodof GMD, the
D rn (DnbDbrD.n)''t
A4pt=4 x tli n ! Dl
Ul^
(2.42) a andb in sectionI of the
D r = Q . 1 2+ 2 \ t / 2 = ( 5 . 2 I ) t / 2 D 2 = (L 9 2 + Z \r/2 = (7.61)t/2 v o t= o .g 2 r b s (Jsf)t" = r' \ 5 2 1) 0.0758mH /km
cYcle transposition = (DpDp)rr+ 7DP)tt2 Dt,=mutual GMD betweenphasesb and c in section 1 of the cYcle transposition = (DP)r;z D,o=mutual GMD betweenphasesc and a in section I of the cYcle r transPosition = (2Dh)v2 Hence D"o 2rt6Drt2pr/3htt6 (2'43)
Voltage inducedin the telephone circuit V,= jttMr,I lV,l = 314 x 0.0758x l03 x 100 _ 2.3,/9 Vlkm 2.9 DOUBI"E.CIRCUIT THREE.PHASE LINES
It is commonpracticeto build doublecircuit threephase lines so as to increase transmission reliability at somewhatenhancedcost. From the point of view of rrowertransferfrom one end of the line to the other (see Sec. 12.3), it is desirable build the two lines with as low an inductance/phase possible.In to as order to achievethis, self GMD (D") should be made high and mutual GMD (D') should be made low. Therefore,the individual conductorsof a phase should be kept as far apart as possible (for high self GMD), while the distance between phasesbe kept as low as permissible (for low mutual GMD). Figure2.15 showsthe threesections the transposition of cycleof two parallel circuit threephase lines with vertical spacing (it is a very commonly used configuration).
c b ' b
the sallle in each section of the It uray be noteclhere that D"u t'eurains cycle, as the conductorsof eachparallelcircuit rotate cyclically, transposition The reader is advisedto verify this for sections2 and so do D,,h, Dbrand D,.,,. cycle in Fig' 2'15' 3 of the transposition a Self GMD in section1 of phase a (i.e., conductors and a/) rs Drr,= (r'qy'q)t'o= (r'q)'/z
GMD *;:;=' Serr "t
respectiverv ,*;;';:^':'ffi'1'are
'i
Dr r = ( y'qr 'q) ''' = ( r 'q) ''t Equivalent self GMD D, = (Dr,,DrbD,,)rt3 (2.44) rotation of conductorsof eachparallel circuit over the Becauseof the cyclic The section. cycle,D. alsoremainsthe samein eachtransposition transposition verify this for sections2 and 3 in Fig. 2.15. reader should The inductancePer Phaseis
= (r')t''qtt3hrt6
o
Section 1 Section2
\0,
Section 3
L = 2 x 1 0  7l o
D',n Ds
o
'l'ltl"tlt)u = z x ,ot ,n't'u
(r')t'' qr/3hrt6
Fig. 2.15 Arrangementof conductorsof a doublecircuit threephaseline
It may be noted here that conductors a and a' in parallel compose phase a and sirnilarly b and b'compose phase b and, and c'compose phase c. in order c to achieve high D" the conductors of two phases are placed diametrically opposite to each other and those of the third phase are horizontally opposite to
[' \r') \q) by of The self inductance each circuit is given (2)'''p
Lr = 2 x 10 7ln
=2x107h
( 2,,u(q')"'[a)"' ] rvpnur.rrn(2.4s)
r
Lines of and Resisiance Transmission lnductance Equation (2.45) can no\,v be written as n of 810 times the conductor'sdiameter,irrespective the numberof conductors in the bundle.
L=!
=
[, ",u
' ^ c[.2 *2xro7rn (t)'^l ,r.ou,
,(t,
+ M) \./
r\
di
tt,d
dl I
where M is the mutual inductancebetweenthe two circuits. i.e.
d
/\
(}
1)
d
\_..
M = z x 1 o ^ ( z \ ' ' ' 7
\ q ) This is a well known result for the two coupled curcuits connected in parallel (at similar polarity ends).
F i g . 2 . 1 6 Configurationof bundled conductors ] s = 0.4m""  s = 0 .4 m ! .
\ q ) becomes zero. Under this condition
/ \ rf h >> L ll unaM '+0, themutuar o,[ i.e. impedance between cucuits the
,() I ou',
l_
f"=o.aml,
I
be) I On'
d =7 m
"\ '
>l
'/^' :' 6 /
d =7 m*]*
:
L=I
IJ2D x 1 0" ' l n " ' ; 
line threephase conductor Fig.2.17 Bundled
(2.47)
The GMD method, though applied above to a particular configuration of a double circuit, is valid for any configuration as long as the circuits are electricallyparallel. While the GMD method is valid.for fully transposed lines, it is commonly applied for untransposed lines and is quite u.rrrui" for practical purposes.
2 s . r v1 n ElrTl.trrr r:rn svlrrrrr.q.Lr AAr? tetJM,rUUlL,t(S
Further, because of increased self GMD line inductance is reduced considerablywith the incidental advantageof increasedtransmissioncapacity of the line.
It is economical transmit large chunksof power 10 over long dista'ces by employing EHV lines. However, the line voltagesthat can be usedare severely limited by the phenomenon corona.corona,ln fact, of is the result of ionization of the atmospherewhen a certain field intensity (about 3,000 kv/m at NTp) is reached'Corona dischargecausescommunication interference and associated power loss which can be severein bad weather conditions. Critical line voltage for formation of corona can be raised considerably by the use of bundled conductorsa group of two or more conductorup". phase. This increasein criticalcoronavoltageis clepenclcnt numberof on concluct<lrs thc group,thc in clearance between them and the distance between the groups forming the separate phases*. Reichman [11] has shownthat the spacingof conductors in a bundle affects vortagegradient and the optimum spacing is of the order of
The bundle usually comprises two, three or four conductors arrangedin configurationsillustratedin Fig' 2.16. The current will not divide equally amongthe conductors of the bundle unlessconductors within the bundle are ruly transposed. The GMD method is still fairly accurate for all practical purposes. 
Find the inductive reactancein ohms per kilometer at 50 Hz of a threephase per phaseas shown in Fig. z.ri. bundled conductorline with two conductors All the conductorsare ACSR with radii of 1.725 cm' (exceptwhen they Even though the power lines are not normallytransposed is sufficiently accurate to assume enter and leave a switching station), it complete transposition(of the bundles as well as of the conductorswithin the bundle) so that the method of GMD can be applied' The mutual GMD between bundles of phasesa and b Dob= @ @ + s) (d  i d)rt4 GMD betweenbundles of phasesb and c Mutual Dh, = D,,,, (bY sYmmetrY) a Mutual GMD betweenbundles of phasesc uruJ  t)Zd)tt+ D,n = Qd (2d + s ) (2d D, o= ( Dr PaPr o) t 't + = @f@ + s)z(d s)2(2a i(Zd  t))tt" = GQ)6Q .q2 6.0)2(14.4)(13.6))','',2 = 8 . 8 1m
The more the number of conductors in a bundle, the more is the self GMD'
70
I
n llarlarn
Dnre,o'
erra+^
^ ^r,^.
D .,= (r' s r' r)' ' o =(rk)r/z = (0.77ggx 1.725x l 02 x 0.4\t/z = 0 .0 7 3m Inductivereactance phase per
lnductance and Resistance Transmission of Lines A = crosssectional area, ln2 The effective resistance given by Eq. (2.48) is equal ro the DC resistanceof the conductor given by Eq. (2.49) only if the current distribution is uniform throughoutthe conductor. r small changesin temperature, resistanceincreases the with temperature in accordancewith the relationship R, = R ( 1 + oor ) where R = resistanceat temperature0"C ao = temperaturecoefficient of the conductorat 0"C
8.8r
0.073
= 0 .3 0 1 o h m/k m In most cases' is sufficiently accurate usethe it to centreto centredista'ces betweenbundlesratherthan mutual GMD between bundles for compu ting D"n. with this approximation,we have for the example in hand Drn=exTxl4yrrtg.g2m Xr= 3 I4 x 0 .4 6 1 x 103l os 8' 82 " 0.073 = 0.301 ohmlkm Thus the approximate method yielclsalmostthe samereactance value as the exact method'It is instructive to comparethe inductive reactanceof a bundled conductor line with an equivalent heuristic (on basis) singleconductor line. For the example in hand, the equivalent line will have d = 7 m and conductor diameter(for sametotal crosssectional area)as JT x 1.i25 cm Xr= 3 1 4 x 0 .4 6 1 x l 03 6n = 0 .5 3 1o h m /k m
lltlA iiiiv' Ac \i olran.l" diiuduJ ^lr^l
(2.s0)
Equation (2.50) can be used to find the resistance Ro at a temperature /2, if resistanceRr1 at temperature tl is known
(2.s ) 1
2.T2 SKIN EFFECT AND PROXIMITY EFFECT
(7 x7 xl 4)t/3 0.7799x J2 xl.725x 103 for a bundledconductor
conductor line
This is 76'4rvo higher than the colrespondingvalue
liuirit€c out'
.r
iower
r
increases transmission its capacity. 2.TT RESISTANCE
reactance of a bundled
Though the contribution of line resistanceto series line impedance can be neglectedin most cases'it is the main sourceof line power loss. Thus while considering transmission line economy,the presence line resistance of must be considered. The effectiveAC resistance given by is O_ average power lossin conductor watts in ohms (2.48)
where 1is the rms current in the conductoi in amperes. Ohmic or DC resistanceis given by the formula
nl R'n  ' ' " o h l n s A
(2.49)
where
p = resistivity of the conductor,ohm_m / = length,m
The distribution of current throughout the crosssectionof a conductor is uniform only when DC is passing through it. on the contrary when AC is flowing through a conductor,the currentis nonuniformlydistributedover the crosssection a mannerthat the current density is higher at the surface of in the conductor compared to the current density at its centre.This effect becornes Inore pronounced frequencyis increased. as This phenonlenon cilled stirr is qffect.It causeslarger power loss for a given rms AC than the loss when the Sairr€ vaiueof DC is flowing ihroughthe conciuctor. Consequently, effective the conductor resistance more fbr AC then fbr DC. A qualitative is explanation of the phenomenon as follows. is Imagine a solid rottnd conductor (a round shape is considered for convenience only) to be composed annularfilamentsof equalcrosssectional of area. The flux linking the filaments progressively decreasesas we move towards the outer filaments fbr the simple reasonthat the flux inside a filament does not link it. The inductive reactance the inraginaryfilaments therefore of decreases outwards with the result that the outer filaments conduct more AC than the inner filaments (filaments being parallel). With the increase of frequency the nonuniformity of inductive reactanceof the filaments becomes more pronounced,so also the nonuniformity of current distribution. For large solid conductors the skin effect is quite significant even at 50 Hz. The analytical study of skin effect requires the use of Bessel's functions and is beyond the scope of this book. Apart fronl the skin effect, nonuniformity of current distribution is also causedby proximity eJJ'ect. consider a twowire line as shown in Fig. 2.1g. Each line conductorcan be divided into sectionsof equalcrosssectionat area (say three sections).Pairs aat, bbt and, can form threeloops in parallel. cct The
72
I t

ModernPowerSystemAnaiysis
Inductance and Resistance Transmission of Lines
b
flux linking loop aat (and thereforeits inductance)is the least and it increases somewhatfor loops bbt and ccl. Thus the density of AC flowing through the (au') of the conductors is the least and is at conductors highest the inneredges at the outer edges (cc').This type of nonuniform AC current distribution
Decomes more pronounceo as me olstance Detween conouctors ls reouceo. LlKe skin effect, the nonuniformity of current distribution caused by proximity effect
also increasesthe effective conductor resistance.For normal spacing of overhead lines, this effect is always of a negligible order. However, for whereconductors locatedcloseto eachother,proximity are underground cables etfect causesan appreciable increasein effective conductorresistance.
Fig. p2.5 2.6 Two threephase linesconnected parallelhaveselt'reactances X, and in of X2. If the mutual reactancebetween them is Xp, what is the effective reactance between the two ends of the line? 2.7 A singlephase Hz power line is supportedon a horizonralcrossarm. 50 The spacing between conductorsis 2.5 m. A telephoneline is also supported a horizontalcrossarm the samehorizontalplane as the on in power line. The condttctors the telephrlncline are of solid copper of spaced 0.6 m between centres. The distance between the nearest conductors the two lines is 20 m. Find the mutual inductance of between the circuitsand the voltageper kilometreinducedin the teiephone line for 150 A current flowing over the power line. 2.8 A telephone line runs parallelto an untrasposed threephase transmission line, as shown in Fig. P2.8.The power line carriesbalanced current of 400 A per phase. Find the mutual inductancebetweenthe circuits and calculatethe 50 Hz voltageinducedin the telephone line ptsrkm.
a b c h b
Fig. 2.18 Both skin and proximity effects depend upon conductor size, fiequency, distance between conductors and permeability of conductor material.
LE S PROB IVI
2 . 1 Derive the formula for the internal inductancein H/m of a hollow
r;onductor havinginsideradiusr, and outsideradiusr, andalsodetermine line in for the expression the inductance H/rn of a singlephase consisting of the hollow conductors described above with conductors spaced a distance D apart. 2 . ? Calculate the 50 Hz inductive reactanceat I m spacing in ohms/km of a core. The cable consistingof 12 equal strandsaround a nonconducting diameterof each strandis 0.25 cm and the outsidediameterof the cable i s 1 .2 5c m . 2 . 3 A concentriccable consistsof two thinwalled tubes of mean radii r and for Derive an expression the inductanceof the cable per It respectively. unit length. leadsheathed 50 2 . 4 A singlephase Hz circuit comprisestwo singlecore cableslaid side by side; if the centresof the cablesare 0.5 m apart and each sheathhas a mean diameterof 7.5 cm, estimatethe longitudinal voltage induced per km of sheathwhen the circuit carries a current of
( r
f^
( r
l
( ,
L
, .
15m *1rn.__
5m+++5mf 
Fig. P2.8 Telephoneline parallelto a power line 2.9 A 500 kV line has a bundling arrangement of two conductors per phase as shown in Fic. P2.9.
Fig. P2.9 500 kV,threephase bundled conductor line Computethe reactance per phaseof this line at 50 Hz Each conductor carries50Voof the phasecurrent.Assume full transposition.
800 A. conductors carrycurrents + 1 and 1. What is the of 2.5 Two long parallel in at magnetic intensity a point P, shown Fig. P2.5? field
lgq er1tgrygf .!yg!U_An4ygp
2.10 An overhead line 50 kms in length is to be constructed conductors of 2.56 cm in diameter,for singlephase transmission. The line reactance must not exceed31.4 ohms.Find the maximum permissiblespacing. 2.11In Fig. P2.1I which depictstwo threephase circuitson a steeltower there
cal cenre tlnes.
Inductance and Resistance Transmission of Lines
NCES REFERE
threephase circuit be transposedby replacing a by b and then by c, so that the reactances the threephases equal and the GMD method of of are reactancecalculationscan be used. Each circuit remains on its own side of the tower. Let the self GMD of a single conductorbe 1 cm. Conductors a and at and other corresponding phase conductors are connected in parallel. Find the reactance per phaseof the system.
Electric'ttl Transrnission and Distributiort Book, Westinghouse Electric and Manufacturing Co., East Pittsburgh, Pennsylvania, 1964. 2. Waddicor, H., Principles of Electric Power Transmission, 5th edn, Chapman and Hall, London, 1964. 3. Nagrath, I.J. and D.P. Kothari, Electric Machines, 2nd edn, Tata McGrawHill, New Delhi, 1997. 4. Stevenson,W.D., Elements of Power System Analvsis,4th edn, MccrawHill, York, 1982. New
l.
C) j I o iz.sm Io c ) l1o l m4m
b b ' 4 m
c'
5. Edison Electric Institute, EHV Transmission Line Reference Book, 1968. 6. Thc Aluminium Association, Aluminium Electrical Conductor Handboo,t. New York, 1971. 1. Woodruff. L.F., Principles of Electric Pov,er Trun.snissiorr,John Wiley & Sons, New York, 1947. 8. Gross, C.A., Power System Analysis, Wiley, New York, 1979. 9. Weedy, B.M. and B.J. Cory Electric Power Systems,4th edn, Wiley, New york, 1998. 10. Kimbark, E.W., Electrical Transmission of Power and Signals, John Wiley, New York, 1949.
l
I
I
')
r.___zsm___l
.)
Paper
I l. Reichman.J., 'Bundled Conductor Voltage Gradient Calculations," AIEE Trans.
Fig. P2.11 ).12 A doublecircuit threephase line is shown in Fig. P2.I2. The conductors a, a/l b, bt and c, c/ belong to the same phase respectively. The radius of each conductor is 1.5 cm. Find the inductance of the doublecircuit line in mH/km/phase.
a b
l I t l
1959, Pr III. 78: 598.
a/
l
(_)
() i 1m i'1m
a) \..,_/
1m l ,r
b/
[r \. _./
l
lI
1m
>l< I
Fig. P2.12 Arrangement conductors a doublecircuit of for threephase line 2.13 A threephase line with equilateralspacingof 3 m is to be rebuilt with horizontal spacing(.Dn = ZDn  ZDrr).The conducrors are to be fully transposed.Find the spacingbetween adjacentconductorssuch that the new line has the sameinductanceas the original line. 2.14 Find the self GMD of three arrangernents bundledconductorsshown in of Fig. 2.16 in termsof the total crosssectional A of concluctors (same area in each case)and the distance d between them.
Lines of Capacitance Transmission
t
: V , "= 6 e a v 6    q  d v V
tL r r  .', L ),,, /t'\V
Fig. 3.1
Electricfield of a lclngstraightconductor
3. 1
IN T R O D U C T IO N
The capacitance togetherwith conductance forms the shunt admittanceof a transmission line. As mentioned earlierthe conductance the result of leakage is over the surface of insulatorsand is of negligible order. When an alternating voltageis appliedto the line, the line capacitance drawsa leading sinusoidal current called the charging current which is drawn even when the line is open circuited at the far end. The line capacitance being proportionalto its length,the chargingcurrentis negligible lines lessthan 100km long. For longerlines for the capacitance becomes increasingly importantand has to be accounted for. 3.2 ELECTRIC FIELD OF A LONG STRAIGHT CONDUCTOR
path of As the potential difference is independent of the path, we choose the the path PP2lies along an integration as PrPP2 shown in thick line. Since equipo',entral,Vrris obtained simply by integrating along PyP, t.e.
vrz= l:,'trav:ht"?u
3.3
(3.1)
POTENTIAL DIFFERENCE BETWEEN TWO CONDUCTORS OF A GROUP OF PARALLEL CONDUCTORS
Imagine an infinitely long straightconductorfar removedfrom other conductors (including earth) carrying a unifbrrn charge of 4 coulomb/metrelength. By symmetry,the equipotentialsurfaces will be concentriccylinders,while the lines of electrostaticstresswill be radial. The electric field intensitv at a distancev from the axis of the conductoris ,= Q v/^ 2nky
that the It Figure 3.2 showsa group of parallelchargedconductors. is assumed ground and are sufficiently removed from conductorsare far removed from the between i.e. eachother, the conductorradii are much smallerthan the distances lines always power transmission usedin overhead them. The spacingcommonly imply that the charge on Further, these assumptions meets theseassumptions. remainsuniformly distributedaroundits peripheryand length. eachconductor
nn
where ft is the permittivity* of the medium. As shown in Fig. 3.1 consider two pclintsP, and P, locatedat clistances D, and Dr respectively from the conductor axis. The potential difference Vn (betweenP, and Pr) is given by
* In SI units the pennittivity of free space is permittivity for air is ft, = klko = 1. ko = 8.85 x 1012 F/m. Relative
i,o  
Fig. 3.2 A group of parallelcharged conductors
7S
'l
Modernpower SystemAnalysis or
Cot
Capacitance Transmission of Lines
The potentialdifferencebetweenany two conductors the group can then be of obtainedby adding the contributionsof the individual chargedconductors:by repeated application of Eq. (3.1). so, the potential difference between conductorsa and b (voltagedrop from a to b) is
 r 79 ( 3. 4b)
l' .^
0.0121
log (D / (ror)ttz)
trtF/km
tn *lcln tn qo *;. n,h ++.. qn +) +. "
v
(3.2)
The associated line charging current is
I,= ju.Co6Vnu Allvn
(3.4c)
Each term in Eq. (3.2) is the potentialdrop from a to b causedby charge on one of the conductors of the group. Expressionson similar lines could be written for voltage drop between any two conductorsof the group. If the chargesvary sinusoidally, do the voltages(this is the casefor AC so transmissionline), the expressionof Eq. (3.2) still applies with charges/metre length and voltagesregarded phasorquantities. as Equation(3.2) is thus valid for instantaneous quantities and for sinusoidal quantities as well, wherein all chargesand voltagesare phasors. 3.4 CAPACITANCE OF A TWO.WIRE LINE
(3.5)
"O___l
(a)
[__{b
i
Linetolinecapacitance
lr )o
. . 
c",
Considera twowire line shownin Fig. 3.3 excitedfrom a singlephase source. The line developsequaland oppositesinusoidalcharges the two conductors on which can be represented phaso$ QoNd qb so that as eo = _ eu.
(b)
Cnn
C.n=Cbr=2C"a Linetoneutralcapacitance Fig. 3.4
l*l

J
l
Fig. 3.3 Crosssectional view of a two_wireline The potential difference Vo6 can be written in terms of the contributions made by qo and q6 by use of Eq. (3.2) with associated assumptions (i.e. Dlr is large and ground is lar away). Thus, V 0t), = ' Since
As shown in Figs. 3.a @) and (b) the linetoline capacitancecan be equivalentlyconsidered two equal capacitances senes.The voltageacross as is the lines dividesequallybetweenthe capacitances suchthat the neutralpoint n is at the groundpotential.The capacitance eachline to neutralis then given of by
C,= Co,= Cb,= 2Cou= ,!#A
pflkm
(3.6)
" , l ; t ( n " h*D ^ + )
27rk rorb
(3.3)
The assumptions inherent the abovederivation in are: (i) The charge on the surfaceof each conductoris assumedto be uniformly distributed, but this is strictly not correct. If nonuniforrnityof chargedistributionis takeninto account,then C,
Qr,=  qu, we have
0.0242
pFkm )
(3.7)
v o b =+ m L !
r c 2L(+ ( 4  , ) " ' ) [ nr
) lf D/2r >>1, the above expressionreducesto that of Eq. (3.6)and the error causedby the assumptionof uniform chargedistribution is negligible. (ii) The crosssection both the conductorsis assumed be circular, while of to in actual practice strandedconductors are used. The use of the radius of the circumscribingcircle for a strandedconductorcauses insignificanterror. \4r'
The line capacitance Cnh is then
"ab 
Qo Vot
In (D / (ror)ttz1
F/m length of line
Q.aa)
a
s
1
^

4
q0 3.5

Modern pb*gl Jysteq 4!gly..!s For air medium (k, = l),
es
CAPACITANCE OF A THREEPHASE LINE WITH EOUILATERAL SPACING
LgT '
(3.14b)
,,=#ffi p,Flkm
1o (line charging) = ju,CnVnn
Figure3.5 showsa threephase composed threeiclentical line of concluctors of
(3.1s)
\D
Vac,/
Vab+ Vac= 2 rt cos 30" V"n =3V"n
A \_/b
a/
Fig. 3.5
crosssectionof a threephaseline with equilateralspacing
,''h
I I
Using Eq. (3.2) we can write the expressions for Vu,,and V,,..as
t " , " I (3.8) (3.e)
n". Vca 1 I
1",
vub= *(0"
6 2 * t1,, tn , r, ,, t +)
\V'o
30
IF
I
vn,= i ! * ( n " 6 P t q 1 , t n # * r ,^ r j )
Adding Eqs. (3.8) and (3.9),we get
Itvun
= v,,t, v,,, *
,'oolrr,,
'l )t,, ,, D + (qu ,1,
; ]
(3.0) r
v",
6 D
vb
Vr
!b
Since there ilre no othcr chargesin the vicinity, thc sunr of'clrar_{es on the three conductors is zero. Thus q6 * Qr=Qu, which when substituted in Eq. (3.10) vields
Fig. 3.6 phasordiagram baranced of threephase vortages 3.6 CAPACITANCE OF A THREE.PHASE LINE WITH UNSYMMETRICAL SPACING
= V,,h Vn, !+ r" 2 I zTTk r
(3.11)
With balanced threephase voltages applied to the line, it follows from the phasor diagrarn ol'I,rig. 3.(r that Vou I Vor= 3Von
(3.r2) (3.13)
Substituting for (Vot,+ V,,,) trom Eq. (3.12) i n E q . ( 3 . 1 1 ) , e g e r w
v^_ 4o tn2
27Tk
r
For the first,sectionof the transposition cycle
The capacitance of line to neutral immediately follows as /" Qo Von 2 Tlc ln (D/r)
= v o b + ( e o t r n ! * q u r  'n f t
(3.l 4a) zTvc \ r
Dr z
+e,rrn})
Dr , )
(3.r6a) \
.
l
E2' I
t
ModernPower SystemAnalysis Capacitance Transmission of Lines I
t
m
*lr"ho3L+aatnL
D rn = (D nDnDT)tl3
(3.18)
r'o,= l / Fig. 3.7 of Crosssection a threephaseline with asymmetricalspacing (fully transposed)
lfiln,
( ^ r ^ D " n' tn;*n,
j , r" t ,* )
')
(3.1e)
Adding Eqs. (3.18) and (3.19),we get vot # vo, = *( / r ^,\, 6 r, 2r t r ( qt  r q") r n t f %)
For the second section of the transposition cycle
zTrK\
f
 t e.zt" +lz ) (3.16b) uzt ut
(3.16c)
(3.20)
As per Eq. (3.12) for balancedthreephase voltages V n ti and also Vor='3Vnn (qu + qr) =  qo
cycle For the third sectionof the transposition
vob=.(r"rln &r* q6rtn!i e,tt"+l Dr,) zi'!\ Dy
Use of these relationships Eq. (3.20) leadsro in Qo ln D"n l/ o n^.. (3.2r) Znk" r The capacitance line to neutral of the transposed of line is then giveh by C. = 3s: 2nk F/m to neutral " Von ln (D"o/ r)
If the voltage drop along the line is neglected, Vno is the same in each transposition cycle. On similar lines three such equations can be written for Vbr= Vut, I 120. Thrce more equations can be written equating to zero the summation of all line charges in each section of the transposition cycle. From these nine (independent) equations, it is possible to determine the nine unknoWn charges. The rigorous solution though possible is too involved. With the usual spacing of conductors sufficient accuracy is obtained by assuming
Qat= Qa2= Qa3= Q"i Qut= Qaz= Quz= Qo, 4ct= Q,'2= 4,3= 4r'
(3.22a)
Forair medium 'o;=
= \n 'uut" log (D,o /r) p,F/km to neutral
(3.22b)
(3.t7)
This assumptionof equaicharge/unitlength of a line in the three sectionsof the transposition cycle requires,on the other hand, three diff'erent values of Vnu The solution can be designated Vo61, as Vo62andVoo,in the three sections. as of considerablysimplified by taking Vou the average thesethreevoltages,i.e. v , q , @ v g ) =! { v , , u r + v o u r + v o o z )
J
It is obvious that for equilateralspacing D,, = D, the above (approximate) formula gives the exact resdlt presentedearlier. The line charging current for a threephase line in phasor form is
Io Qine charging) = jut,,Vnn Alkm
(3.23)
3.7
EFFECT OF EARTH ON TRANSMISSION LINE CAPACITANCE
or
'/ nb=
6 q1aln'
.1*o" I l 'n ( DtzD?tDy rnI '
[ ,, ) "'
r'
\ Dt2D23D3t )
)
( DrrDr^D^,\1 +q^lnl:ll \ DnDnD3t ))
In calculatingthe capacitance transrnission of lines,the presence earthwas of ignored, so far. The effect of earth on capacitance can be conveniently taken into account by the method of images. Method of Images The electric field of transmission line conductorsmust conforrn to the presence of the earth below. The earthfor this purposemay be assumedto be a perfectty
E4

ModernPowerSYstemAnalYsis
we get ubstituting the valuesof different chargesand simplifying'
Vob=
conducting horizontal shdet of infinite extent which therefore acts like an surface. equipotential unit and potential plane midway betweenthe conductors as is such that it has a zero shown in Fig. 3.8. If a conducting sheetof infinite dimensionsis placed at the zero potential plane, the electric field remains undisturbed.Further, if the conductor carrying charge q is now removed, the electric field above the conducting sheet stays intact, while that below it vanishes.Using these well of we known resultsin reverse, may equivalentlyreplacethe presence ground below a chargedconductorby a fictitious conductor having equal and opposite chargeand locatedasfar below the surfaceof ground as the overheadconductor above itsuch a fictitious conductor is the mirror image of the overhead of conductor.This method of creatingthe same electric field as in the presence by images originally suggested Lord Kelvin. earth is known as the method of
2hD 'l rn r(4hz + D2)'tz rk
r Radius
(3.2s)
l
h
t
,/'7
Zero potential p l a n e( g r o u n d )
I
I I
lmage charge
Fig'3.9SingephasetransmissionIinewithimages It immediatelY follows that irk
wab 
F/m linetoline
(3.26a)
h
\
,
I t__
F i g . 3 . 8 Electricfield of two long, parallel,oppositelycharged conductors
and
Cn= 
ln
2nk D
F/m to neutral
(3.26b)
41+tO't4h21stt2
Capacitance
of a Single'Phase
Line
its to line Considera singlephase shownin Fig. 3.9.lt is required calculate by the methodof images of taking the presence earth into acoount capacitance describedabove. The equationfor the voltage drop Vo6as determined by the b, two chargedconductorsa and, and their images a'and b' canbe written as follows:
t"gt#Y = m2 +nrrn e,,, vub i* *1, "
*q,,tnGFfUl
(3.24)
Capacitanee Transmission of Lines 86 ,*  Modernpower SvstemAnalvsis images.with conductor a in position1, b in position2, and c in position 3, 1 2 7rl( ln r  ln h, The equation for the averagevalue of the phasor %. ir found in a similar manner. Proceedingon the lines of Sec. 3.6 and using Vou* Vo, = 3Von and Qo* Qt * Qc = 0, we ultimately obtained the following expression for the capacitanceto neutral.
(3.27)
Similar equationsfor Vo6can be written for the second and third sectionsof the transposition cycle.If the fairly occurate assumption constant of chargeper unit length of the conductor throughoutthe transmissioncycle is made,the average value of Voufor the three sectionsof the cycle is given by
tn
D'u r
rr( rn"n"\'),)!t \
( (hrlhhs)''' )
F/m to neutral
(3.29a)
C,
0.0242
D'n lon o ron@"httu')'!t o
r (4h24)tt3
pPttcm neutral (3.29b) to
(3.28)
where D, = (DnDnDrr)"t
Comparing Eqs. (3.22a) and (3.29a), it is evident that the effect of earth is to increase the capacitanceof a line. If the conductors are high above earttr comparedto the distancesamong them, the effect of earth on the capacitance of threephase lines can be neglected.
Calculate the capacitanceto neutrallkm of a singlephaseline composed of No. 2 single strand conductors(radius = 0.328 cm) spaced3 m apart and,7.5 m above the ground. compare the results obtained by Eqs. (3.6), (3.7) and (3.26b). Solution (1) Neglecting the presenceof earth tEq. (3.6)l 0.0242 n, ^ C,=31fitkm log 
o.oi+z =ffi=o'00817
0.328 By the rigorousrelationship [(Eq. (3.7)]
cn
P'Flkm

o.0242
"'(+.(#')"')
C" = 0.00817 pFkm
+ negligible.
Since
=915, the effect of nonuniformity of chargedistribution is almost
Flg. 3.10 Threephase line with images
_uu I 
r
r
l
powel€ysteryr_Anslygls Modern
b c a
Capacitance Transmission l of Lines

89
(2) Considering eff'ect earthandneglecting the of nonuniformity charge of distribution [Eq. (3.26b)] 0.0242
r(l*( 300
/4h"))  897
d=8m
,Jto4
o32BJLo4
c' o:o?? = o'0082 P'Fkm " 2.9s3
Note: The presenceof earth increases the capacitanceby approximately 3 partsin 800.
Examp 3.2 le
A threephase Hz transmissionline has flat horizontal spacing with 3.5 m 50 betweenadjacentconductors. The conductorsare No. 2/0 harddrawn seven strandcopper (outsideconductor diameter= 1.05 cm). The voltage of the line is 110 kV. Find the capacitance to neutral and the charging current per kilometre of line. Solution D"o= (3.5 x 3.5 x '7)t't= 4.4 m
 h=6m
Fig. 3.11 Crosssection a doublecircuit of threephase line Solution As in Sec. 3.6, assumethat the charge per conductoron each phase is equal in all the three sectionsof the transpositioncycle. For section / of the transpositioncycle V,,n(l)=
0.0242 tog(440/0.525)
.'"f z*l*("i*'';)+c,(rn; )
*n.(,n j+r";)]
(3.30)
For section II of the transpositioncycle
v n"  l _ 1 0 6 u,Cn 314.o.oos%
= 0.384 x 106O/km to neutral
v"b = ,lAln.("i* r"#)+c, .'" Gr) (rn; *)
+a,(rnj.,":)]
For section III of the transpositioncycle
charging cunent +
(l 19l ]e) x l000
(33r)
0.384x 106 = 0.1I Aftm
Xn
(rrr) vtu =*lr. ('" .^ a, ^I) j i)+ (rn I. +a.(rn;.r"f)f,"r,
Average value of Vo6over the transpositioncycle is given by
The six conductorsof a doublecircuitthreephase line having an overall radius of 0.865 x 102m are arrangedas shown in Fig. 3.11. Find the capacitive reactanceto neutral and charging current per kilometre per conductor at 110k V, 5 0 H z .
= (av v,* s) iL*[n.'n r,^(###)] (ttrj+*
90

Modernpower SystemAnalysis
=fi,^,,)h(ffi)"'
i'B' jh
Capacitance Transmission of Lines
(3.33)
l'.
h=6m
I 
gf
,'f 'd
vou* 3von= vo,= q,)tn (ffi)''' fien" Qt3von=*r"(:#)
(3.3s)
Fig. 3. 12
Capacitanceto neutral per conductor =
2dc
(3.36)
3.8
METHOD OF GMD (MODTFTED)
Total capacitance neutral for two conductorsin parallel to
Cn=
4rk
(3.37)
Now h= 6 m; d = 8 m; / = 8 m. Referring to Fig. 3.I2, we can write
f , ..? .
\ 2 )) L\2) = (j ' + h 2 )rt 2 l 0 m = f
: r=lf/)+(oh\"1Jim
,
,,121'1/2
g = (72 + 42)rt2=J65 Conductorradius (overall) = 0.865 x 102m Substitutingthe valuesfor various distances. have we cn 4 r x 7 x 8 . 8 5 1 0  1 2 1 0 6x 1 0 0 0 x x
A comparison of various expressions for inductance and capacitance of transmission lines [e.g. Eqs. (2.22b)and (3.6)] brings our rhe facr that the rwo are sirnilarexceptthat in inductance expressions haveto use the tictitious we conductorradius rt = 0.7788r, while in the expressionsfor capacitanceactual conductorradius r is used.This fact suggeststhat the method of GDM would be applicable the calculations capacitance well providedit is modified in for as by using the outer conductor radius for finding D,, the self geometric mean distance. Exarnple3.3 can be conveniently solved as under by using the rnodifled GMD method. For the first section of the transpositioncycle mutual GMD is Dub= ((ts) (ts))tta_ liglttz Db' = Qil''' Drr, = ( jh)'''
1 n r z * o s * s t o fr o o) 3  l ' / 3 f
[ 100x8 \0.86s/ 
= 0.018 1F/km I Q C , = 3 1 4 x 0 ' 0 1 8 1x 1 0  6 = 5.68 x 106 Ulkm Chargingcurrent/phase = "t#* Charging current/conductor 0'361 =
L
pF /km
(D"pop,o)r,t = [(i, grjh)r,t)t,, "n ln the first section of the transposition cycle self GMD is D,o = ?f rf )rt4 = ?f)''' D D'l' = QAt'' Dr, = (At'' D, = (D,oDroD,,)''3= l?3f Arl3fitz
Now
Cn=
2d(
x 5.6g x 10{ = 0.361
2nk
= 0.1805A/km
4 ltk
F/m
FThis result
Example 3.3. 3.9
ai
,'lt I
obviously checks with the fundamentally derived expressionin
LsL
PROB MS IE
BUNDTED CONDUCTORS
A bundled conductorline is shown in Fig. 3.13. The conductorsof any one bundle are in parallel, and it is assumedthat the charge per bundle divides equally among the conductors the bundle as Drr> of r/. Also Drr d x D* + d x D12for the samereason.The results obtainedwith theseurru'rnptions are fairly accurate for usual spacings. Thus if the charge on phase a is qo, the conductorsa and a'have a chargeof qolz each; similarly the charge is equally divided for phasesb and c.
applied voltage is baranced threephase, Hz. Take the voltage 50 of phase a as referencephasor. All conductors have the sameradii. Also find the charging current of phase a. Neglect the effect of grouno.
j;;:i;g;:;fiIH#,",iHffi T,.$l;:i,h*j :::f ::i::t"^tll:;
[? aQ 1O,, I
DP l' Fig' 3'13
oQ eu, I >fDy 
Fqi
Dzs
_ "6 iec/____ '..l
Fd+
Fig. p3.1 3'2 A threephase doublecircuitline is shown in Fig. p3.2.The diameterof each conductoris 2.0 cm. The line is transffio and carries balanced load' Find the capacitanceper phaseto neutral of the line. aQ
2m
I
Crosssection a bundled conductorthreephase of transmissionline
Now, writing an equ'ation for the voltage from conductor a to condu ctor b, we get
T
I
Qc'
,f o Q
2m
vob= *lotn,(rn4.*ro , ou* *o.Sq"lr" rn4. ) \
D, D t, ) or
Qu,
_t
= r t v,,h ,*lr,
/
,n'.r q n t n g . n " t n D r r \ +ro ,., Dtz . Dzt)
(3.3e)
Consideringthe line to be transposed proceeding and in the usual manner,the f inal r e s u l t w i l l b e
',=
^ffi
p,Fkmto neutral
(3.40)
where Do, = (DnDnD3)In It is obvious from Eq. e.aD that the method of modified GMD is equally valid in this case (as it shouldbe).
3'3 A threephase, Hz overheadline 50 has regularly transposed conductors equilaterally spaced 4 m apart The of such a line is 0.01 "upu.Itun.e tFkm'Recalculate the capacitanc.p"i kilometre to neutral when the conductorsare in the same horizontai prane with successive spacing of 4 m and are regularly transposed. 3'4 consider the,500 kv, threephase bundled conductorline as shown in Fig' P2'9' Find the capacitivi reactance to neutral in ohms/km at 50 Hz. 3.5 A threephase trernsnri.ssion has r,rat, rinc horizontarspacingwith 2 rn betweenadjacentconductors.The radius of eachconductoris 0.25 cm. At a certain instant the chargeson the centre conductor and on one of the outside conductors are identical and voltage drop betweenthese identically chargedconductorsis 775 v.xegtecithe effect of ground, and find the value of the identical chargein coulomblkm at the instant specified. 3'6 Find the 50 Hz susceptance neutral to per kilometre of a doublecircuit threephaseline with transposition shown as in Fig. p_3.6.Given D = Jm and radius of each of the six conductors is 1.3g cm.
94
i
t

Modern Po*", Syrt"r An"lysi.
6
l'
o  +
6
o*_
6
D >)< D 'l oJ
5
6
5
. P3.6 Doublecircuit threephaseline with flat spacing
3.7 .{ single conductorpower cable has a conductorof No. 2 solid copper (radius = 0.328 cm). Paper insulation separatingthe conductor from the concentric lead sheath has a thickness of 2.5 mm and a relative permittivity of 3.8. The thickness of the lead sheath is 2 mm. Find the capacitive reactance per kilometre between the inner conductor and the lead sheath. 3 . 8 Find the capacitanceof phase to neutral per kilometre of a threephase line having conductors 2 cm diameterplacedat the cornersof a triangle of with sides5 m, 6 m and 7 m respectively. Assume that the line is fully transposedand carriesbalanced load. 3 . 9 Derive an expressionfor the capacitanceper metre length between two long parallel conductors,each of radius,r, with axes separated a by distanceD, where D ,, r, the insulating medium being air. Calculatethe maximum potentialdifferencepermissible betweenthe conductors, the if electric lield strengthbetweenthem is not to exceed25 kY lcm, r being 0.3 cm and D = 35 cm.
4.I
INTRODUCTION
NCES REFERE
Books
l . Stevenson,w.D., Elements of Power SystemAnalysis,4th edn, McGrawHill, New York, 1982.
2 . Cotton, H., and H. Barber,
The Transmission and Distribution of Electrical Energy,3rd cdn, Hodder and Stoughton, 1970. 3 . Starr, A.T., Generation, Transmission and Utilization of Electric Power, Pitman, 1962.
A complete diagram of a power system representing all the three phases becomes complicatedfor a systemof practicalsize, so much so that it may too no longer convey the information it is intended to convey. It is much more practicalto represent power systemby meansof simple symbbls for each a componentresultingin what is called a onelinediagram. Per unit system leads to great simplification of threephasenetworks involvingtransformers. impedance An diagram drawn on a per unit basisdoes not requireideal transfbrrners be includedin it. to An importantelementof a power systemis the synchronous machine,which greatly influencesthe systembehaviourduring both steadystateand transient conditions. The synchronous machinemodel in steadystateis presented this in chapter.The transient model of the machine will be presentedin Chapter 9. 4.2 SinglePhase Solution of Balanced ThreePhase Networks
Papers
A T.
Parton, J.E. and A. Wright, "Electric Stresses Associated with Bundle Conductors", International Journal of ELectrical Engineering Education 1965,3 :357. 5 . Stevens, R.A. and D.M. German, "The Capacitance and Inductance of Overhead Transmission Lines", International Journal of Electrical Engineering Education, 1 9 6 5 , 2: 7 1 .
The solution of a threephase network under balancedconditions is easily carriedout by solving the singlephase networkcorresponding the ref'erence to phase. Figure4.1 showsa simple,balanced threephase network.The generator and load neutrals are therefore at the samepotential, so that In = 0. Thus the neutral impedance Zn does not affect network behaviour. For the reference phase a En= ( Zc+ ZL) I ' (4.1) The currentsand voltages in the other phases have the samemagnitudebut are progressively shifted in phaseby 120".Equation(4.1) conespondsto the singlephase network of Fig. 4.2 whose solution completely determinesthe network. solution of the threephase
Modernpower SystemAlelygis
Ia
Representation PgryerSystemComponents of
l
.*In=o
lrJ'ZL
\; \e,
* c ) 'trb t L_l zn 16
,",
If the transformeris YIA connectedas inFig. 4.4a, the delta side has to be replacedby an equivalentstar connectionas shown dotted so as to obtain the singlephaseequivalent of Fig. 4.4b. An important fact has, however, to be
AN m0 lrne culTent /A
Z1
!:
Fig. 4.1
Z6
Ea
Ic
Balanced threephase network
have a certain phaseangle shift from the star side values Vorand Io(90" for the phase labelling shown). In the singlephaseequivalent (Vew, I) are respectively in phasewith (Von,/o). Since both phasevoltage and line current shift through the samephaseangle from star to delta side, the transformerper phaseimpedanceand power flow are preserved the singlephase in equivalent. In most analytical studies,we are merely interestedin the magnitude of voltages and currents so that the singlephaseequivalentof Fig. 4.4b is an acceptable proposition. Wherever proper phaseangles of currents and voltages are needed, correction can be easily applier after obtaining the solution through a singlephasetransformerequivalent.
Fig' 4.2
Singlephase equivalentof a balancedthreephasenetwork of Fig. 4.1
(a) Y/A transformer withequivalent connection star
€ l > Ia
>, Ia
A
n_
(b) Singlephase equivalent of Y/A transformer
(a) Threephase transformer Y/y
Fig.4.4
It may be noted here that irrespective of the type of connection, the transformation ratio of the singlephaseequivalent of a threephasetransformer is the same as the linetoline transformationratio.
(b) Singlephaseequivalentof 3phase y/y transformer
Fig. 4.3 ' See Section 10.3.
tro
Yli 4.3
I
i I ONE.IINE
Modern Power System Analysis DIAGRAM AND IMPEDANCE OR REACTANCE
DIAGRAM A onelinediagramof a power system shows the main connectionsand
showrrdependingon the information required in a system study, e.g. circuit breakersneednot be shown in a load flow study but are a must for a piotection study. Power system networks are represented oneline diagrams using by suitable symbols fbr generators,motors, transformersand loads. It is a convenient practical way of network representationrather than drawing the actual threephasediagram which may indeed be quite cumbersome and confusing for a practical size power network. Generator and transformer connectionsstar, delta,and neutralgrounding are indicatedby symbols drawn by the side of the representation these elements.Circuit breakers are of represented rectangularblocks. Figure 4.5 shows the oneline diagram of a as simple power system.The reactancedata of the elementsare given below the diagram.
T1
99 t basisfor use under balancedoperating The impedancediagramon singlephase drawn from the onelinediagram.For the systemof conditionscan be easily transtbrmer diagramis drarvnin Frg. 4.6. Singlephase Fie. 4.5 the impedance as ideal transformers with transformer impedances equivalents are shown of Representation Power SystemComponents This is a fairly good approximationfor most power system have been neglected. as are The generators represented voltagesourceswith seriesresistance studies. in modelwill be discussed Sec. (synchronous machine and inductivereactance in by line is represented a zimodel(to be discussed 4.6). The transmission (not involving rotatingmachines) to Chapter5). Loads are assumed be passive and are representedby resistanceand inductive reactancein series.Neutral do grounding impedances not appearin the diagram as balancedconditions are assumed. Three voltagelevels(6.6. I 1 ancl33 kV) are presentin this system.The
('Y (
I
l
o
A
ft
__'l
r
fl
T2 .tF
Y r
l F' t l at
rYA l
v
i
r
throughout this book. 4.4 PER UNIT (PU) SYSTEM \ It is usual to express voltage, current, voltamperes and impedance of an valuesof these of electricalcircuit in per unit (or percentage) baseor ret'erence value of any quantity is defined as: quantities. The per unit* the actualvaluein anYunits the baseor referencevaluein the sameunits The per unit method is particularly convenientin power systemsas the various sections of a power system are connectedthrough translormersand have different voltage levels. system.Let Consider first a singlephase = (VA)s VA Base voltamPerss Basevoltage= Vu V Then
L
q
of , Fig. 4.5 Onelinerepresentation a simple power system G e n e r a t oN o . 1 : 3 0 M V A ,1 0 . S V ,X ' = 1 . 6o h m s r k r 1 G e n e r a t oN o .2 : 1 5 M V A ,6 . 6k V ,X ' = 1 . 2o h m s G e n e r a t oN o .3 : 2 5 M V A ,6 . 6k V ,X ' = 0 . 5 6o h m s r Transformer (3 phase) 15MVA,33/1 kV,X = 15.2ohmsper phase hightension 11 : 1 on side Transformer Tr(3 phase): MVA,3316.2 X= 16 ohmsper phaseon frign 15 kV, tensionside Transmission 20.5ohms/phase line: LoadA : 15 MW, 11 kV, 0.9 lagging powerfactor L o a dB : 4 0 M W ,6 . 6 k V ,0 . 8 5 l a g g i np o w e r a c t o r g f Note: Generators specified threephase are in MVA,linetoline voltage and per phase reactance (equivalent star). Transformers specified threephase are in MVA, linetoline transformation ratio,and per phase(equivalent star)impedance one side.Loads are on specifiedin threephase MW,linetoline voltageand powerfactor. it 'litrd'r,n'rq J'rn,A,n?fid\ I r_1 Intrn^'"1'"rr' f I 1 r i , I
Basecurrent/u = [4)e \v/B
A
(4.2a)
Genl'
 > F r  F .T r a n s f o r m e r l+' l
I Load A
J;ffi ' 'Tr. ns : :n e  , Fr _ . a  : f o r m e r T 2  t C ; / C " ' S Li
Load g
cent value = per unit value x 100. Per cent value is not convenient for use as the factor of 100 has to be carried in computations.
*Per
Fig. 4.6
lmpectancediagram of the power system of Fig. 4.5
I
tir.n' l .,rvq..il
I
MooernHowersystem Analysis
Baseimpedance = Y4: z,
" IB
Vi
(VA )"
ohms
(4.2b)
'er unit imPedanceZ (Pu) =
x Z (oltrns) (MVA)' @
(4.e)
If the actual impedanceis Z (ohms), its per unit value is given by z(pu) = JZB Z(ohms)x (vA)' V;
Z(ohms)x (kVA),
,r trt2 , i n^n
(4.3)
For a power system,practical choice of base values are: Base megavoltamperes (MVA)B = or Base kilovoltamperes= (kVA)B Base kilovolrs = (kv)a Base current 1, 1,000x (MVA)u (kV)a
to trom (MVA)r, oto (MVA)a' n'*' andtV base when MVA baseis changecl from Eq' the (kv)r, orotJ (kV)n, new' nt* p"t unit impedance s changedfrom .4.9) is given bY
=z(pu)o.,. Z(pu)n"* ffi"ffi
Per Unit Representation of a Transformer
(4.10)
(4.4)
Baseimpedance,u 1'ooq] GV)r z _ __Ia _ x GV)3 _ 1,000 (kv)1 ohms
(kvA)8 Z (ohms)x (MVA),
(MVA)B Per unit impedancez (pu) 
(4.s)
(4.6)
transformer forming part of It has been said in Section 4.2 that a threephase ty a singlephasetransformer in a threephase system can be represented The deltaconnectedwinding of the obtaining per phasesolution of the system' so that the transformationratio of transformer is replacedby an "quiuaient star transformer is always the linetoline voltage ratio the equivalent singlephase transformer' of the threePhase Fi gure4. Tar cpr esent sasingle phaset r ansf gr m lr int er m sof pr im ara' yand andan ideal transformer of ratio 1 : secondary leakagi reactancesZp artdZ, of Let us chooseavoltamperebase The magnetizing impedanceis neglected. two sides of the transformer in the ratio of (vA)a and voltage bases on the transformation,i.e. V r s ! Vt, a
{_J>
(kv)"
(4.11a)
Z(ohms)x (kVA)u
(kV)? 1,000 x
In a threephasesystemrather than obtaining the per unit valuesusing per phase base quantities,the per unit valuescan be obtained directly Uy u.ing threephase base quantities.Let Threephasebase megavoltamperes (MVA,)B = Linetoline base kilovolts = (kV)B Assuming star connection (equivalentstar can always be found),
zs
12
transformer of (a) Representation singlephase '' neglected) impedance (mignetizing
Base current Ir
l'ooox(MVA)u
Jr 1rv;u
l'ooox(kv)u J3rB
o
(4.7)
Baseimpedance 7o "
_
x GD3 _ 1,ooo(kv)2, ^L_: onms (MVA)' G"Ab
transformer of circuit singlephase (b) Perunitequivalent
(4.8)
Fig. 4.7
W
I
I
todern
Power Srrstem Analrrsie
Yvr v.vr.; , rr rqt ystr
Therefore
'I 2 8
= a (as (VA)a is common) Vrn z 1,,
.
of Representation Power System Uomponents (4.1i b) (4.11c)
[,.t'ff.
l
LaD


Vza
zz(pu)= zB ' +++o'1' zzu
zru
1,,
From Fig. 4.7a we can write Vz=(VtIflp)alrZ, We shall convertEq. (4.12) into per unit form Vz(pu) = [Vr(pu) Vw  I r(pu)I Vzn 6Zo(pu)Zru]a Ir(pu)IruZ,(pu)Z* Dividing by vzn throughout and using base relations (4. rra, b, c), we get Vz(pu)= Vr(pu)  I,(pu)Zr(pu)_ Ir(pu)Z,(pu) (4.13) (4.12)
is of Thus the per unit impedance a transformer the samewhethercomputed from primary or secondaryside so long as the voltage baseson the two sides are in the ratio of transformation(equivalentper phaseratio of a threephase transformer which is the same as the ratio of linetoline voltage rating). transformeris conveniently The pu transformerimpedanceof a threephase obtained by direct use of threephaseMVA base and linetoline kV base in relation (4.9). Any other impedanceon either side of a transformeris converted to pu value just like Zo or Zr. Per Unit Impedance Diagram of a Power System
Now
or
+=+=,
Ir I'o 12 I'u I r ( p u ) = 1 2 ( p u ) =1 ( p u )
From a oneline diagramof a power systemwe can directly draw the impedance diagramby following the stepsgiven below: common MVA (or kVA) basefor the system. 1. Choosean appropriate 2. Consider the system to be divided into a number of sections by the Choose an appropriatekV base in one of the sections. transfbrmers. Calculate kV bases of other sectionsin the ratio of transformation. in 3. Calculateper unit valuesof voltagesand impedances eqch sectionand connectthem up as per the topologyof the onelinediagram.The result per unit impedancediagram. is the singlephase The above stepsare illustrated by the fallowing examples.
Equation (4.13) can thereforebe written as = Vz@u) Vr(pu) (pu)Z(pu) where
(4.r4)
Z(pu)=Z,,(pu)+ Z,(pu) Equation (4.I4) can be represented the simple equivalent circuit by of Fig. 4'7b which doesnot requirean idealtransfurmet. Coniia.rablesimplification has thereforebeenachievedby the per unit method with a common voltampere base and voltagebaseson the two sidesin the ratio of transformation. Z(pu) can be determined directlyfrom the equivalentimpedance primary on or secondaryside of a transformerby using the appropriaie impedancl base. On primary side: Zr=Zp+ Z,/a2
j Example4 1
diagram of the power system of Fig. Obtain the per unit impedance(reactance) 4.5. diagramof the power system of Fig. 4.5 Solution The per phaseimpedance has been drawn in Fig. 4.6. We shall make some further simplifying assumptions. so are as and resistance neglected that it is represented 1. Line capacitance a series reactanceonly. 2. We shall assumethat the impedancediagram is meant for short circuit studies.Current drawn by static loadsunder short circuit conditions can be neglected.Loads A and B are thereforeignored. to Let us convert all reactances per unit form. Choosea common threephase MVA base of 30 and a voltagebaseof 33 kV linetolineon the transmission line. Then the voltagebasein the circuit of generator1 is 11 kV linetoline and 2 that in the circuits of generators and 3 is 6.2 kV. of The per unit reactances various componentsare calculatedbelow:
Z ( p=+ : !  + L * I ' u)
zru Zro
But a2Ztn = Zzr
z,o"a2
Zr(pu)= Zo(pu)+ Z,(pu) Z(pu) On secondary side: zz= Z, + o2zo
(4.1s)
Transmission line: Transform T,: er
TransformerTt: Generator1: Generator2: Generator3:
2o.5x3o GT2
= 0.564
I iliffij IExample 4.I, we now calculatethe pu values of the reactances transfonners of and generatorsas per relation (4.10): Transformer Z : Transformet Tr:
Flepresentation power System eomponents of
l!,? l_0_ 0 . 4 1 8 I = 16x30
0.209 x
aiT
1.6 30 x 1.2 30 x
= 0.44
0.22 ff =0.++ x
0.43s x (10'5i1 =0.3e6
(ll)'
u t7 = o'396
Generator1:
tazr
= o'936
Generator2: 0 . 4 1 3r * I) *(6'612 =0.936 6.2)'
0.56 30 x = 0'437 AZI{)frL_/XfX)<1 _"
The reactance diagramof the systemis shown in Fig. 4.g.
JU U000._J
64
0.44
I
i
(6'6)1 = 0.431 0.3214 * i9 x /.r 6.21' obviously these values are the same as obtained already in Exampre 4.r. Generator 3: 4.5 Complex Power
Consider a singlephaseload fed from a source as in Fig. 4.9. Let
v tvt16 r _tn t (6_0)
Fig.4.9 Reactance criagram the systemof Fig. 4.5 (roads of neglected) Et' Ez and E, are per unit values of voltagesto which the generators are in ashort
iJ"',ltl3;3llllrlt""
Example4.2
circuit study, wil beraken t /.,"pu (no these ui
Source
(a)
The reactancc data of gencrators and transtbrmers usually specified are in pu (or per cent) values, based on equiprnentratings rather than in actual ohmic valuesasgiven in Exampl 4.7; *iit" e rhetransmirr;;; hne irnpedances nray be us
Fig. 4.9 Complex power flow in a singlephase load When d is positive, the current lags behind voltage.This is a convenient choice of sign of 0 in power systems where loads have mostly lagging power factors. Complex power flow in the direction of current indicatedis siven bv S=VI* =lVlllll0 = l V l l 1 lc o s d + j l v l l 1 l s i n 0 = P + i e or @.17)
;"";:,[J:Ti]#?let
Transformer 0.209 T,: TransformerT): 0.220 Generator Gr: 0.435 Generator Gr: 0.413 Generator G3: 0.3214
rcsotve rJxarnpre b; assuming rbuowing 4.1 rhe
With a baseMVA of 30. base volta_ee of I I kV in the circuit of generator I md b:isevoltageof 6.1 k\; in the circuit of generators2 and 3 ,r; i; ";
t S l= ( p 2 + e \ t , ,
I Here
Modern Power system Anatysis S  complex power (VA, kVA, MVA) lSl = apparent power (VA, kVA, MVA); it signifies rating of
I
nepresentation powersystemeompqlents ot
As per Eq.(4.19), Kirchhoff's current law appliesto complex power (also applies separatelyto real and reactive powers). In a series RL load carrying current {
T
l.i; ffii
V=l(R+jxr)
P = I"R = active power absorbedby load power absorbedby load Q  IzXr = reactive In case of a series RC load carrying current I P _I2R power absorbedis negative) O  IzX, qreactive Consider now a balanced threephaseload representedin the form of an equivalentstar as shown in Fig. 4.L2. The threephase complex power fed into load is given by
P = lVl l1l cos 0  real (active) power (watts, kW, MW) Q = lVl l1l sin 0 = reactivepower  voltamperesreactive (VAR) = kilovoltamperesreactive (kVAR) = megavoltamperesreactive (MVAR) It irnmediatelyfollows from Eq. (4.17) that Q, thereactivepower, is positive for lagging current (lagging power factor load) and negativefor leading cunent (leading power factor load). With the direction of current indicated in Fig. 4.9, .9= P + iQ is supplied by the source and is absorbedby the load. (4.17)can be represented the phasordiagramof Fig.4.10 where Eqrration by , i1 n = positive for lagging current 0 = Lan'' (4.18) P :gativefbr leadingcurrent
S = 3vplt= 3 lvptl6pl; : JT lvrl zOrti
If Ir =llil I (6p A .S = ,'5 lvLl lILl I 0
(4.20)
Then
 Ji tvLtvLtcosd + iJT t
0=P+iQ@.21)
Fig. 4.10 Phasor representation complex of power(lagging load) pf If two (or more) loads are in parallel as in Fig. 4.ll Fig. 4.12 Complex powerfed to threephase load Here
s=vF_v(i+i) f Y r i . . r ' r(' pr + p r+ j ( e t +e z ) ) !
(4.1e)
tsl = Ji tvrt ttrl P  Ji tvLltILtcosd e = Ji lvLltILtsin d
d  power factor angle lf vL, the line voltage,is expressed kv; andIy,the line currentin amperes, in s is in kvA; and if the line current is in kiloamperes, s is in MvA.
In terms of load impedanceZ,
N = rotor speed (synchronousspeed)in rpm P = numberof poles
,  v, _lvLll6P r L_ Jiz z Substituting I, in Eq. (4.20) for
f
r  V'l' " r
winding
(4.22a)
Field winding ..
tr V,.is in kV, ,Sis now given in MVA. Load impedance if required be Z can calculatedfrom
1)xr
\Xp>{
o
F
,_lvrl'_ lvl "  Jt T1O
4.6 SYNCHRONOUSMACHINE
(4.22b)
vl
I T:t F
I
The synchronousmachineis the most important elementof a power system.It converts mechanical power into electrical form and feeds it into the power network or, in the caseof a motor, it draws electrical power from the network and converts it into the mechanicalform. The machine excitation which is controllable determinesthe flow of VARs into or out of the machine.Books on electrical machines 1151 may be consulted for a detailed account of the synchronousmachine. We shall presenthere a simplified circuit model of the (undertransient which with suitable machine modifications wherever necessary conditions) will be adoptedthroughout this book. Figure 4. 13 shows the schematic crosssectional diagramof a threephase synchronousgenerator(alternator)having a two pole structure.The stator has a balanced threephasewindingaat, bbt and cct. The winding shown is a concentratedone, while the winding in an actual machineis distributed across the stator periphery. The rotor shown is a cylindrical" one (round rotor or nonsalient pole rotor) with rotor winding excited by the DC source. The rotor winding is so arrangedon rotor periphery that the field excitation produces nearly sinusoidally distributedflux/pole (d) in the air gap.As the rotor rotates, threephase emfs are produced in stator winding. Since the machine is a balancedone and balanced loading will be considered, can be modelled on it per phase basis for the referencephase a.
Tn o mqnhinc rrrifh mnra fhqn f r r r nv nnlec l/vrvo, fhp Lllv qlrnrre quv vv ApfinpA svrlllvu cfnrnfrrro JLr uvLur v ronpafc rvyvolo
lot t'. \
NN
'\r \lQ I \,
generator diagram a roundrotorsynchronous of Fig. 4.13 Schematic On no load the voltage EJ inducedin the referencephasea lags 90" behind dywhich producesit and is proportionalto dyif the magneticcircuit is assumed is This phasorrelationship indicatedin Fig. 4.14. Obviously to be unsaturated. the terminal vclltage V, = Er
Qr
I
electricallyfor every pair of poles.The frequencyof inducedemf is given by
l

Ef=Vt
f =ffi nz
where . Highspeedturbogenerators have cylindrical rotors and Iow sppedhydrogenerators have salient pole rotors.
relationship Fig.4.14 Phasor between fuand E, stator winding, the As balancedsteady load is drawn from the threephase stator currents produce synchronously rotating flux Q/poIe (in the direction of rotation of the rotor). This flux, called armature reaction flux, is therefore stationarywith respectto field flux Qy.It intuitively fbllows that Qois in phase with phase c current 1o which causesit. Since the magnetic circuit has been
,liO* 
Modern powsr Syglem_Anatygis
urrurnJdto be unsaturated, superposition the principleis applicableso that the resultant air gap flux is given by the phasor sum
d' = d1+ Q,, @,3) Further assumingthat the armatureleakagereactance and resistance are
re eml whtch equals the termin tage V,. Phasor diagram under loaded (balanced) conditions showing fluxes, currents and voltagesas phasorsis drawn in Fig. 4.15.
NEffi t The circuit of Fig. 4.L6 can be easilymodifiedto include the effect of armature leakage reactance resistance and (these series are effects)to give the complete circuitmodelof the synchronous generator in Fig. 4.I7. Thetotal as
+xl=Xslsc synchronous reactance of the machine. Equation(4.24) now becomes V, = Et  jlo X,  IoRa (4.2s)
Rapr"."n,",ion po*"r' Syrr"r Cornp.on"n,, of
jlX"=  E"
Fi g.4. 16
Fig. 4.1S phasor diagram synchronous of generator Here d  power factor angle 6  angle by which Et leads v, called load angre or torque angle We shall see in Sec.5.10 that dmainly determines power deliveredby the the generatorand the magnitude E, (i.e. excitation)determinesthe VARs of deliveredby it. Becauseof the assumed linearity of the magneticcircuit, voltage phasorE, Eo and v, are proportionalto flux phasors dr, doand d, respectively;furthei, voltage phasors lag 90' behind flux phasori. It therefore easily follows from Fig.4.15 that phasorAB = Eois proportionalto (o (and therefore Io)and is 90' leading d" @r 1,). With the direction of phasorAB indicatedon the diagram AB = jlo Xo where X" ir the constantof propotionality.
In terrnc nF thp ql n r q urv vrv o rlofiiri^uvruulrurr ^s vL v ,lra, wE ualt li^^rrurr€utly ^rrwll[c) 4tule r rt l0ilowlng
This model of the synchronousmachinecan be further modified to account for the effect of magnetic saturation where the principle of superposition does not hold.
&
Fig 4.17 circuitmodelof roundrotorsynchronous generator Armature resistance Rn is invariably neglectedin power system studies. Therefore,in the place of the circuit model of Fig. 4.I7, the simplified circuit model of Fig. 4.18 will be usedthroughout thisbook.The corresponding phasor diagram is given in Fig. 4.i9. The fieici induceciemi Ey ieacisthe terminal condition voltage by the torque (load) angle d This, in fact, is the for acrive power to flow out of the generator.The magnitudeof power delivered depends upon sin d In the motoring operationof a synchronous machine,the current 1,,reverses as shown in Fig. 4.20, so that Eq. @.25) modifiesro Ef = V,  jIoX, (4.26) which is represented the phasordiagramof Fig. 4.2I.It may be noted that by V, now leads lby d, This in fact is the condition for power to flow into motor terminals.
(4.24) where Ef = uolrage induced by field flux Q, alone = uo load emf The circuit model of Eq. (4.24) is drawn in Fig. 4.16 wherein X, is interpreted as inductive reactancewhich accountsfor the effect of armature reactionthereby avoiding the needof resortingto addition of fluxes l&q.@.23)1.
expressionfor voltageswithout the need of invoking flux phasors. V, = Ef  jloXo
the'i I
Modern Power System Analysis

machine The flow of reactivepower anclterminal voltage of a synchronous in excitation.This is discussed detail in by is mainly controllecl meansof its regulated Section5.10. Voltage and reactivepower flow are often automatically
uy YUIL4S\/ IvSurqlvro \uvv vvve^v^r vr
Components System of RepresentationPower = poweroutput lvtl ll,,lcos d = constant active
FIIS':
and by automatic tap changing devices on transformers.

 '/ 6666
l
E1
( ;
I
x
"
Fig.4.22Synchronousmachineconnectedtoinfinitebus l/ol cos dof the phasor Io on V' It rneansthat since lV,l is fixed, the projection is varied' Phasordiagramscorresponding whiie the excitation remains constant, in presented Fig' 4'23' T\e phasor to high, medium and low excitationsare to the unity power factor case'It is obvious diagramof Fig. 4.23(b)colresponds frot the phasor diagram that for this excitation c lEJl os 5=lV)
E1
I
t
of diagram synchroFig. 4.19 Phasor nousgenerator
., I" 0
Ia
l "' /
it"x'
c Fig. 4 . 1 8 S i m p l i f i e d ircuit model of round rotoI synchronous generator
(a) Overexcited
Ef'
1I J
Vt',
jIJ'
( b ) N o r m a le x c i t a t i o n
Fig. 4.20 MotoringoPerationof machine sYnchronous
of Fig. 4.21 Phasordiagram motoring oPeration in generatoroperates parallel with other generators Normally, a synchronous simplicity of operationwe shall considera to connected the power system.For bus generatorconnectedto an inJinite bus as shown in Fig' 4'22' As infinite independent remainconstant voltageand frequency whose a means largesystem machine and the bus' and of the power exchangebetweenthe synchronous machine. the excitationof the synchronous of independent activepower into an feedingconstant generator consider now a synchronous In and its infinite b's bar. As the machineexcitation is varied, armaturecurrent keep angle g, t.e. power factor, changein such a manner as to
(c) Underexcited
Fi1.4.23 P h a s o r d i a g r a m s c f s y n c h r o n o u s g e n e r a t o r f e e d i n g c o n s t a n t
power as excitation is varied
case(Fig' a'23a)'i'e' Forthe ovetexciterl This is definedas normal excittttittrt feedspositive reactive thatthe generator cos ,8,.1 6>v),1, lags behind V, so powerintothebus(ordrawsnegativereactivepowerfromthebus)'Forthe
^
# . . . . . . . ' @ , ^
^
l L ^
"
j
a
^
{

A
l 4
^
l i ^ / \ ^ ^
A a A u
t
_
^
^

;iiii
Components of Representation PowerSystem I
t
Modern Power systemAnalysis
T.._
I iilis'.$
or
underexcited case (Fig. 4.23c), i.e. lErl cos 6 < lV), 1o leads V, so that the generator feedsnegativereactivepower into the bus (or drawspositivereactive Figure 4.24 shows the overexcitedand underexcitedcasesof synchronous motor (connectedto infinite bus) with constantpower drawn from the infinite bus. In the overexcited case,Io leads Vu i.e. the motor draws negativereactive power (or suppliespositive reactive power); while in the underexcitedcase .Io lags V, r.e. the motor draws positive reactive power (or suppliesnegative reactivepower).
ln,l cos l1,l e: E: sin6
(4.27)
(4.28)
The plot of P versus { shown in Fig. 4.25, is called the power angle curve. The maximum power that can be delivered occurs at 6 = 90" and is given by
(4.2e)
For P ) P** or for 6> 90' the generatorfalls out of step. This problem (the at stability) will be discussed length in Chapter 12.
E1 (a) Overexcited

V1
generator anglecurveof a synchrcnous Fig. 4.25 Power Power Factor and Power Control
(b) Underexcited
Fig.4.24
Phasor diagramsof synchronousmotor drawing constantpower as excitationis varied
From the above discussionwe can draw the general conclusion that a synchronousmachine (generating or motoring) while operating at constant power suppliespositive reactive power into the bus bar (or draws negative
r c q e t t v cI
v
While Figs 4.23 and 4.24 illustrate how a synchronousmachine power factor thesedo not give us a clue changeswith excitation for fixed power exchange, regarding the quantitativevalues of llnl and d This can easily be accomplished by recognizing from Eq. (4.27) that lEll sin 6 llolX, cos d PX" = constant (for constantexchangeof power to = # lyrl (4.30) infinite bus bar) , Figure 4.26 shows the phasor diagram for a generatordelivering constant power to infinite bus but with varying excitation.As lEtl sin dremainsconstant, the tip of phasor Ermoves along a line parallel to y, as excitation is varied. The direction of phasor1ois always 90o lagging jI"X, andits magnitudeis obtained Figurc 4.27 shows the caseof limiting excitation with d= 90". from (l1olX5)/X5. For excitation lower than this value the generatff becomesunstable.
nrtrx/cr
vv
frnrn
fhc
hrrc
hsq r \ v U
rYrvr rhvarn r r
n r r cvrvc\v lnvi v .a r l f Y l l v L v
/
Al n ll
r rl nvrt v avrl to u n i f p r l ln vw ul u
lll4vllltlv
mqnl"io
on the oih.. hand, feeds negative reactive power into the bus bar (or draws positive reactivepower from the bus bar). Considernow the power deliveredby a synchronousgeneratorto an infinite bus. From Fig. 4.19 this power is P = lVtl llol cos 0 The above expression.can written in a more useful form from the phasor be geometry.From Fig. 4.19
sin (90" + 0)
lnA _ rl,lx, sin 6
ModernPower System Analysis
of Ftepreseniation Fqwer SystemComponents Salient Pole Synctrronous Generator
1!1 
machine,as shown in Fig. 4.29, is distinguished A salient pole synchronous from a round rotor machine by constructional features of field poles which
/.{1
\3
62
Iaz
Fig. 4.26
Effect of varying excitationof generator deliveringconstant power to infinitebus bar
employed in machinescoupled to hydroelectric turbines which are inherently slowspeedones so that ttre synchronousmachine has rnultiple pole pairs as different from machines coupled to highspeed steam turbines (3,000/1,500 Salientpole machineanalysisis rpm) which have a two or fourpolestructure. tworeaction theory outlined below. made through the
Direct axis
I
/,"
V1 Fig. 4.27 Case of limitingexcitationof generator deliveringconstant power to infinitebus bar
Similar phasor diagrams can be drawn for synchronous motor as well for constant input power (or constant load if copper and iron losses are neglected and mechanical ioss is combined with load). Another important operating condition is variable power and fixed excitation. In this case lV,l and lE1trare fixed, while d and active power vary in accordance with Eq. (a.28). The corresponding phasor diagram for two values of d is shown in Fig. 4.28. It is seen from this diagranr that as d increases, current magnitude increases and power t'actor improves. lt will be shtlwn in Section 5.10 that as dchanges, there is no significant change in the flow of reactive Power' of Locus Er Er', 4, ,' l\jluzX"
l/
'
l
Fig. 4.29 Sallent pole synchronousmachine (4polestructure) In a round rotor machine, armatLlrecurrent in phase with field induced emf Ey or in quadrature (at 90") to S, produces the same flux linkages per arnpere ai the air gap is uniform so that the armature reaction reactance offered to inphase or quadrature current is the same (X,, + X1 = Xr), In a salient pole
machrne at gap ls nonunllorTn arong IULOI'ljcrlPilury.
l  ,   1   ^ ^  :  L ^   . f t
, n'
rL ls Lllc rtrilsL .lruug trrtr
i ^
+ L ^
l ^ ^ ^ +
^ l ^  ^
+ L ^
Fig. 4.28
generator with variablepower and fixed Operationof synchronous excitation
axis of main poles (called direct axis) and is the largestalong the axis of the interpolar region (called quadrature oxis).Armature current in quadraturewith El producesflux along the direct axis and the reluctanceof flux path being low (becauseof small air gap), it produces larger flux linkages per ampere and X, hencethe machinepresentslarger armaturereactionreactance (called direct componentIl of armaturecurrent 1o. axis reactance)to the flow of quadrature On the other hand, armaturecurrent in phasewith { producesflux along the of quadratureaxis and the reluctanceof the flux path being high (because large
4
r
i5 I
I
rl,lr
ffil
Po*r. syrt"t Rn"tyri, Mod"rn
Representation power of
Resultant
interpolar air gap), it producessmaller flux linkages per ampereand hencethe machine presents smaller armature reaction reactance Xu (guadrature axis reactancea X) to the flow of inphase component Io of armature current /o. Since a salientpole machineoffers different reactances the flow of Il and to 1ocornponentsof armature current Io, a circuit model cannot be drawn. The phasordiagramof a salientpole generator shown in Fig. 4.30.It can be easily is drawn by following the stepsgiven below:
Fig. 4.31 power angrecurvefor sarient poregenerator In this book we shall neglect the effect of sariency and take X'= X't in all types of power system studiesconsidered. During a machine transient, the direct axis reactancechanges with time acquiring the following distinct values during the completetransieht. X/ = subtransientdirect axis reactance Xh = transient direct axis reactance X,r = steady state direct axis reactance The significance and use of thesethree values of direct axis reactancewill be elaboratedin Chapter 9. Operating Chart of a Synchronous Generator
generator polesynchronous Fig. 4.30 Phasor diagram salient of 1. Draw % *d Io at angle 0 2. Draw IoRo.Draw CQ = .il,X,t(L to 1,,) 3. Make lCPl  llol Xq and draw the line OP which gives the direction of Ey phasor 4. Draw a I from Q to the extended line OP such that OA = Ef It can be shown by the abovetheory that the power output of a salient pole generatoris given by
lv,l lE,l ,=1;sind+
lv,l'(xo xn)
2XdXq
si n 26
(4.31)
The first term is the same as for a round rotor machine with X, = Xa and constitutesthe major part in power transfer. The second term is quite small (about I020Vo) comparedto the first term and is known as reluctancepower. P versus d is plotted in Fig. 4.31. It is noticed that the maximum power 34 (change power per unit in output occurs at 6 < 90' (about 70'). Furt1t"r ' d5' change in power angle for small changes in power angle), called the power.cofficient,in the operatingregion (r< 70') is larger in synchronizing a salient pole niachin.: than in a round rotor machine.
while selecting a large generator,besidesrated MVA and power factor, the greatestallowable stator and rotor currents must also be consideredas they influence mechanical stresses and temperaturerise. Such timiting parameters in the operationare brought out by meansof an operating chart orperformance chart. For simplicity of analysis,the saturationeffects,saliency,and resistanceiue ignored and an unsaturated value of synchronous reactanceis considered. Consider Fig. 4.32, the phasor diagram of a cylindrical rotor machine. The locus of constantllolx,V) and henceMVA is a circle centered M. The locus at of constantlEtl (excitation) is also a circle centeredat O. As Mp is proportional to MVA,QP is proportional to MVAR and Me to MW, all to the same scale which is obtained as follows.
I \Y/rl n r l or r r r , v u ! rn
D n rYrY v r v )r 'rs r a mr  , r Arnq rl t r cri u Q ci v r I v r ar r a ru c

2.Opu excitation
. 0.85 pf lagging
E
N
l
a
o
(E
i\
Locus11, X" I (circlecentreM)
I
C]'
(U o c) F
generator of Fig.4.32 Phasordiagram synchronous For zero excitation. i.e. lE.l = 0  i Io X r' = Y, or Io = jV,lX, to i.e. llol =lV)lXr leading at 90" to OM which corresponds VARs/phase. now the chart shown in Fig. 4.33 which is drawn for a synchronous Consider machinehaving Xt = 1.43pu. For zero excitation,the currentis 1.01I.43 0.J pu, so that the length MO conespondsto reactivepower of 0.7 pu, fixing both active and reactivepower scales. With centre at 0 a number of semicircles are drawn with radii equal to different pu MVA loadings.Circles of per unit excitation are drawn from centre to M with 1.0pu excitationcorresponding the fixed terminalvoltageOM. Lines may also be drawn from 0 conesponding to various power factors but for clarity only 0.85 pf lagging line is shown. The operationallimits are fixed as fbllows. Taking 1.0 per unit active power as the rnaximum allowable powel', a that the horizontal limirline ubc is drawn throughb at 1.0 pu. It is assumed machineis ratedto gire 1.0 per unit active power at power factor 0.85 lagging value and this tixes point c. Limitation of the statorcurrentto the corresponding requires limitline to becomea circular arccd aboutcentre0. At point d the the rotor heatingbecomesmore important and the arc de is fixed by the maximum to excitationcurrent allowable,in this caseassumed be lEtl = 2.40 pu (i.e.2.4 at times ly,l). The remaininglimit is decidedby loss of synchronism leading power factors. The theoreticallirnit is the line perpendicularto MO at M (i.e. d= 90o), but in practice a safety margin is brought in to permit a further small ln increase load belore instability. Fig. 4.33, a 0.1 pu margin is employed in and is shown by the curve afg which is drawn in the following way.
\
0.7
I , ,
M
9o.s
Leading
,
tl
0.5
1.0
> Reactive power(pu)lagging
Fig. 4.33 operating chartfor rargesynchronous generator , Considera point h onthe theoreticallimit on the lETl 1.0 pu excitations = arc, the pcrwer Mh is reducedby 0.1 pu to Mk; the op"ruiing point must, however, still be on rhe on the desiredlimiting curve. This is repeatedfor other excitationsgiving the curve afg. The completeworking area.shown shaded. gfabcde.,{ is working point placed within this area at once defines the MVA, Mw, MVAR, current, power factor and excitation.The load angle 6 canbe measured as shown in the figure. 4.7 REPRESENTATION OF LOADS
Load drawnby consutners the toughest is parameter assess to scientifically. The magniiudeof the ioad, in iact, changes continuouslyso that the load forecasting problern is truly a statistical one. A typical daily load curve is shown in Fig' 1.1. The loads are generally composed of industrial and domestic components. industrial load consists An mainly of largethreephase induction nlotors with sulficient load constancy and predictable duty .y.lr, whereasthe domestic lclad mainly consists of lighting, heating and many single_phase devicesusedin a randomway by houscholders. designun6 upJrotion The of power systemsboth economically and electrically are greatly influenced by ttrp nature and magnitude of loads.
.sa ._IaZ II
Modern Power SvstemAnatrrsic
Representationp of Sotution Base A = 645,3_phase MV BasekV = 24, lineto_line
Load volt
In representation loads for various systemstudies such as load flow and of stability studies, is essential know the variationof real and reactivepower it to with variation of voltage. Normatly in such studies the load is of composite nature with both industrialand domesticcomponents. typical composition A of Induction motors 55757o Synchronous motors 5:75Vo Lighting and heating 2030Vo Though it is always better to consider the PV and of QV characteristics each of these loads for simulation, the analytic treatment would be very cumbersome and complicated. In most of the analytical work one of the following three ways of load representation used. is (i) Constant Power Representation
JlfZJ.t
ASa= i
)L
= 1 pu
Synchronous reactance = +# X, = 1.344pu 'r" (24)z
Full load (MVA) = I pu, 0.9 pf lagging Load current = generatorcurrent Io= 7 pu, 0.9 pf lagging = 0.9  7 0.436 pu (a) Excitation emf (see Fig. 4.Ig) Ef = V,+ j XJ" = 1 1 0 " + j 1 . 3 4 4( 0 . 9  j 0 . 4 3 6 ) = 1.586 j l.2l = 199 137.1" E, (actual) = 1.99 x 24 = 47.76 kV (line) 6= 3j. 1" ( leading) (b) Reactive power drawn by load Q = VJ,, sin r/ = 1 x I x 0.436 = 0.436 pu or 0.436 x M5 = 281 MVAR
This is usedin load flow studies.Both the specifiedMW and MVAR are taken to be constant. (ii) Constant Current Representation
Here the load current is given by Eq. (4.17), i.e. " I' = P : i Q  t n ' l ( 6  0 ) V{< where V = lVl 16and 0= tanl QlP is the power factor angle. It is known as constantcurrent representation because magnitudeof current is regarded the as constantin the .study. (iii) Constant Impedance Representation
This is quite olten used in stability studies.The load specified in MW and MVAR at nominal voltageis used ro compurethe load impedance(Eq. (4.?2b)). Thus I "
z = !I:  P = J O w*
f ll Fvattrt v F^sr..lsr
I
l v l 2 t PJQ:T
The generator of Example 4.3 is carrying full load at rated voltage but its excitationemf is (i) increased 20vo and (ii) reducedby by 20vo. Calculatein each case (a) load pf (b) reactive power drawn by load (c) load angle 6 Solution Full load, P _ 1 x 0 . 9 = 0 . 9 p u Ef=
which then is regarded as constant throughout the study.
li .l ar v e z I
I
t


A synchronous generator rated 645 MVA , 24 kv,0.9 pf lagging. It has a is syrrchronous reactancel.z o. The generatoris feeding full loadat 0.9 pf lagging at rated voltage.Calculate: (a) Excitation emf (E1) andpower angle 6 (b) Reactivepower drawn by the load Carry out calculationsin pu form and convert the result to actual values.
T
r.99
lM
I
ftrodernPower System Analysis
tation of Power (b) or
V,=7
(i) Et is
x 0.9 x sin 1.5= 0.024 Q = 0.024x 645= 15.2 MVAR
at d by 20Vo samereal load.Now
As per Eq. (a.28) P= l E', l l v', l
x,
sind
(i)
PROB S IEIvI
4'r Figure P4'l shows the schematic diagram of a radial transmission system' The ratings and reactances the various of componentsare shown rherein.A load of 60 MW at 0.9power factor ragging is tappedfrom the 66 kv substation which is to be mainraineo alt oo kv. calcurate rhe terminal voltage of the synchronous machine.Representthe transmission line and the transforrnersby series reactances ontu. . 11t220 kv 220t66kV
0.9= (2'388xr) ,i' d \ 1.344)
or or
sin d= 0.5065 6 = 30.4" 2.388130.4"110" jr.344
= 0 . 8 9 j 0 . 7 9 = 1 . 1 8 3l 4L2"
(a) pf = cos 4I.2" = 0.75 lagging (b) Reactivepower drawn bY load Q = lV,lllolsin / =1x1.183x0.659 = 0.78 pu or 502.8 MVAR (ii) or by E, decreased 20o/o E f = 1 . 9 9x 0 . 8 = 1 . 5 9 Substitutingin Eq. (i)
V1
100MVA X = 10o/o
dh ___EF_l___> 1 6 0 5F  100MVA V2 X = Bo/o
Fig. p_4.1
kv
60 MW 0.9pf tagging
4.2 Draw the pu inrpedancediagram fbr the power system shown in nig. e4.2. Neglectresistance, usea baseof ioo and vrre , 220kv in 50 () rine. The ratings of the generator,motor and transformers are Generator40 MVA, 25 kV, Xu = 20Vo Motor 50 MVA, I I kV, X,t = 30Vo YIltransformer, 40 MVA, 33 y_220 y kV, X = I5Vo Yl transformer, MVA, ll L_220 y 30 kV, X = l5*o
o e =( L 2 \ L ) , i n 6
\ 1.344 ) which gives 6  49.5"
In=
.l
,50o UvGir+1,__iFrF
l   Y Y I
Fig. p4.2
.*fF.,
[y^
l_1 f 1 \M ) 'Y 2
I
) /
t.59149,5"rlo" .i1.344 =0.9j0.024 = 0.9 l1.5"
4 . 3 A synchronousgenerator is rated 60 MVA, 1r kv. It has a resistance Ro =
0l pu and xo 7 r.65 pu. It is feedinginto an infinite bus bar ar 11 kV delivering a cirrrent 3.15 kA at 0.g ff hgging. (a) Determine E, and angte d (b) Draw a phasor diagram for this operation. (c) Bus bar voltage fails to 10 kv while the mechanicalpower input to generator and its excitation remains unchanged. wtrat is the value and pf of the current delivered to the bus. In this case assumethe
(a) pf = cos 1.5" = 1; unity pf
I
126 
ModernPower SystemAnalysis Representationoflgygr_gyg!"_!L_qg[p!g$q
T
generator resistance be negligible. to 4.4 A 250 MVA, 16 kV ratedgeneratoris feedinginto an infinite bus bar at 15 kV. The generator a synchronous has reactance 1.62pu.lt is found of that the machine excitation and mechanicalpower input are adiustedto give E, = 24 kY and power angle 6 = 30o. (a) Determine the line current and active and reactivepowers fed to the bus bars. (b) The mechanicalpower input to the generatoris increasedby 20Vo from that in part (a) but its excitation is not changed.Find the new line current and power factor. (c) With referenceto part (a) current is to be reduced by 20Voat the same power factor by adjusting mechanical power input to the generatorand its excitation. Determine Ey, 6 and mechanicalpower rnput. (d) With the reducedcurrent as in part (c), the power is to be delivered to bus bars at unity pf, what are ttre correspondingvalues of El and d and also the rnechanical power input to the generator. 4.5 The generatorof Problern4.4 is feeding 150 MVA at 0.85 pf lagging to infinite bus bar at 15 kV. (a) Determine Ey and d for the above operation.What are P and Q fed to the bus bars? (b) Now E, is reducedby l0o/okeeping mechanicalinput to generator same, find new dand Q delivered. (c) Et is now maintainedas in part (a) but mechanicalpower input to generatoris adjustedtill Q = 0. Find new d and P. (d) For the value of Eyin part (a) what is the maximum Qthat can be deliveredto bus bar. What is the corresponding 6and {,? Sketchthe phasor diagram for each part.
REFERE I.ICES
Books
l' Nagrath,I'J. and D.P Kothari, Electric Machines,2nd edn Tata McGrawHill, New Delhi, 7997. 2' van E' Mablekos, Electric MachineTheory PowerEngineers,Harperlno for Raw, New York, 1980. 3. Delroro, v., Electric Machines and power systems,prentice_Hall, Inc., New Jersey,1985. 4' Kothari, D.P. and I.J. Nagrath, Theory and.Problems of Electric Machines, 2nd Edn, Tata McGrawHill, New De\hi, 2002. 5. Kothari, D.p. and I.J. Nagrath, Basic Electicar Engineering, 2nd Edn., Tata McGrawHill, New Delhi. 2002. Paper 6' IEEE CornitteeReport,"The Effect of Frequency and Voltage on power System Load", Presented IEEE winter po,ver Meeting,New york, at 1966.
Answers
4.1 12kV 4.3 (a) 26.8 kV (line),42.3" leading (c ) i .i 3 l 2 8 .8 " k A ; 0 .8 7 6l a g 4.4 (a) 0.5ll"l 25.6"kA; 108 MW, 51.15MVAR (b) 6.14 kA, 0.908lagging ( c ) 1 . 5 7 8 ,1 3 . 5 o5 3 . 3 M W , ( d ) 1 8 . 3 7k v , ' 3 5 . 5 " ,9 6 M W 4. 5 (a ) 2 5 .2 8 k V , 2 0 .2 ' ,1 2 7 .5M W, 79.05MV A R (b) 33.9", 54 14 MVAR (c ) 4 1 .1 " , 1 5 0 .4MW (d) 184.45,MVAR,53.6", 7 0.787 pu
eharacteristics Perforr"nance povrerTransmission and of Lines The following nomenclature has been adoptedin this chapter: z = seriesimpedance/unitlength/phase y = shunt admittance/unitlength/phase neutral to
;=;l_*#:,T.":Jff::
C = cepacitance/unit length/phaseto neutral / = transmissionline length Z = zl = total series impedance/phase Y = ll = total shunt admittance/phase neutral to Note: Subscript,Sstands a sendingend for quantityand subscriptR standsfor a receivingend quantity 5.1 INTRODUCTION 5.2 This chapter deals primarily with the characteristics and performance of transmission lines.A problemof major importance power systems the flow in is of load over transmissionlines such that the voltage at various nodes is maintained within specified limits. While this general interconnectedsystem problem will be dealt with in Chapter 6, attention is presently focussed on performance of a single transmissionline so as to give the reader a clear understanding the principle involved. of Transmission lines are normally operated with a balancedthreephase load; the analysiscan thereforeproceedon a per phasebasis.A transmissionline on a per phasebasis can be regardedas a twoport network, wherein the sendingend voltage Vr and current 15are related to the receivingendvoltage Vo and current 1o through ABCD constants"as SHORT TRANSMISSION LINE
For short lines of length 100 km or less, the total 50 Hz shunt admittance* QwCl) is small enoughto be negligible resultingin the simple equivalentcircuit of Fig. 5.1.
Fig. 5.1 Equivalent circuit a shortline of This being a simple series circuit, the relationship between sendingend receivingend voltagesand currents can be immediately written as:
lyrl lA BlIy*l L1,J Lc plLroJ
Also the following identity holds for ABCD constants: ADBC=l
(sr)
/< t\
=I l z f l v o l Lr,J Lo rJLr.J
lyrl
(5'3)
These constants can be determined easily for short and mediumlength lines by suitable approximations lumping the line impedance and shunt admittance. For long lines exact analysis has to be carried out by considering the distribution of resistance, inductance and capacitance parameters and the ABCD constants of the line are determined therefrom. Equations for power flow on a line and receiving and sendingend circle diagrams will also be developed in this chapter so that various types of end conditions can be handled. *R"f'.. to Appenclix B.
The phasor diagram for the short line is shown in Fig. 5.2 for the lagging current case.From this figure we can write lV5l = l(ly^l cos /o + lllR)z + (lV^l sin dn + lllxyzlr/2
l v 5l = [ t v R P + l t 2 ( R z+ f ) l
* z t v R  i l ( R c o sQ ^ + X s i nf{ ) r l 2 ( 5 . 4 ) u w
It2
. rl/ = tvnt * ?JlJr cos A + 4#rin" Qp 1il21n2+ x2 r l t / f r l V R l /o , I / i "L tvRP J
; For overhead transmissionlines, shunt admittanceis mainly capacitive susceptance (iwcl) as the line conductance(also called leakance)is always negligible.
Lines J"#, of Characteristics Performance PowerTransmission and 1 The last term is usually of negligible order. l/l R cos/^+l1l X sind^ x 100 I yRl
(s.7)
positive for a lagging load. In the above derivation, Q*has been considered Expanding binomially and retaining first order terms, we get
per cenrregularion !!4 9"t!t:l n x tin 4r x 100 lvRl
(for leading load)
(s.8)
l=l rfr Ivs v^ .
ff
cos + /^
ff
sin f^)'''
(s.s)
Voltage regulationbecomesnegative(i.e. load voltage is more than no load voltage), when in Eq. (5.8) X sin Qo> R cos /p, or tan </o(leading) >+ X It also follows from Eq. (5.8) that for zero voltage regulation
l V 5= l V ^ l + l X ( R c o s/ o + X s i n / o ) l The above equation is quite accurate for the normal load range.
tan/n= i.e.,
X
=cot d e
/^(teading= [
(5.e)
where d is the angle of the transmission line impedance. This is, however, an approximate condition. The exact condition for zero regulation is determinedas follows:
current Fig. 5.2 Phasordiagramof a shortlinefor lagging Voltage Regulation
Voltage regulationof a transmissionline is defined as the rise in voltage at the of as receivingend,expressed percentage full load voltage, when full load at a specifiedpower factor is thrown off, i.e. = Per cent regutation
where
condition underzeroregulation diagram Fig. 5.3 Phasor Figure 5.3 shows the phasor diagram under conditions of zero voltage regulation, i.e. lV5I = lVpl or O C= O A AD  AClz _ llllzl sn /AOD _ zlvRl lyR oA
''^?):'Yu'x lvRLl
100
(5.6)
= lVool magnitude of no load receivingendvoltage lVprl = magnitude of full load receivingendvoltage (at a specified power factor) 'u1,'%' lvRl
= For short line, lV^61 lVsl, lVsl = lVpl = Per cent regutation
', U!4 IAOD= sinr zlvRl lt follows from the geometry angles A, that for zero of at voltaee regulation,
/p (leading) =
Lines of and Characteristics Performance PowerTransmission or new value of lVr  = 11.09 kV
133
or
Figure 5.4 showsthe equivalentcircuit of the line with a capacitivereactance placed in parallel with the load.
R+jx
(5.10)
l
From the above discussion it is seen that the voltage regulation 6f a line is heavily dependent upon load power factor. voltage regulation improves (decreases) as the power factor of a lagging load is increased and it becomes zero at a leading power factor given by Eq. (5.10).
l
!
l
,L

l
Fig. 5.4 Assuming cos y''^now to be the power factor of load and capacitive reactance taken together, we can write
kv.
A singlephase Hz generator 50 suppliesan inductive load of 5,000 kw at a power factor of 0'707 lagging by means of an overhead transmissionline 20 km long. The line resistance and inductanceare 0.0195ohm and 0.63 mH per km' The voltageat the receivingend required is to be kept constantat 10
(11.09 10) x 103= l1n ( R cos d^+ X sin dn) does not draw any real power, we have Since the capacitance 10 x cos/^ Solving Eqs. (i) and (ii), we get cos dn= 0'911 lagging and l l a l = 5 4 9A Now Ic= In I = 549( 0. 911 j0'412) 707( 0. 107  j0. 70'7) = 0.29 + j273'7
l/ol=
(i)
5000
(ii)
Find (a) the sendingend voltage and voltage regulation of the line; (b) the value of the capacitors be placed in parailet viittr to the load such that the regulation is reducedto 50vo of that obtained in part (a); and (c) compare the transmissionefficiency in parrs (a) and (b). Solution The line constantsare R = 0 .0 1 9 5x 2 0 = 0.39 f) X = 3I4 x 0.63 x 103x Z0 = 3.96 e (a) This is the case of a short line with I = Ia= 1, given by l 1 l=   5 0 0 0 = 7 0 7 A 1 0 x 0 .7 0 i From Eq. (5.5), l V 5l = l V o l + l 1 l ( R c o sQ * + X s i n / ^ ) = 1 0 ,0 0 0+ 7 0 7 (0.39 0.701 + 3.96 x x y 0.707,\ = 72.175kV
due the approximationin (i) Ignoring it, Note that the real part of 0.29 appears we have I, = j273.7 A
'
or (c)
'YL ^
:ltl loxlooo
3wxc ll. I 273'7
Voltage regulation= pfTL l9 x roo :Zt.j7vo 10 (b) Voltage regulariondesired= ?+t = l0.9Vo
C81 P'F Efficiency of transmission;
Case (a)
lyst 10 = 0.109 t0
characteristics and Pedormance Power Transmission of Lines LI{S Case (b) Per unit transformerimpedance,
f*
r/= '
5000  g7.7%o ' 5000 (549)2 0.39 103 x + x
, uapacltor ln parallel wlth the load, the
Zr=
j ( 0 . 0 2 +0 . 1 2 ) x 5 ( 0 . 5 + 1 3 . 7 5 ) x 5 _ 6.q2 Q'2
/u Lrr6f.L prdvrtg uJ
receivingend power factor improves (from 0.707 iug to 0.911 lag), the line current reduces(from 707 A to 549 A), the line voitage regulationdecreases (one half the previous value) and the transmission iproves (from "ffi"i"nJy 96'2 to 97'7vo)' Adding capacitors parallel with load is in a powerful method of improving the performanceof a transmissionsystem and will be discussed further towards the end of this chapter.
= (0.0046+ j0.030) pu = Total seriesimpedance (0.037 + j0.0115)+ 2(0.0046+ j0.030) = (0.046 + j0.072) pu Given: Load MVA = 1 pu = = 0. 91 pu Loadvolt age + 6. 6 Load current= .1 = 1.1 pu 0.91 Using Eq. (5.5),we get lVs = 0. 91 + 1. 1( 0. 046 0. 85 + 0. 072x 0. 527) x = 0.995 pu = 0.995 x 6.6  6.57 kV (linetoline)
A substation shown in Fig. 5.5 receives MVA at 6 kv, as 5 0.g5lagging power factor on the low voltage side of a transforner from a power stationthrough a cablehaving per phaseresistance reactance 8 and 2.5 and of ohms,respectively. Identical 6.6/33 kV transfoffnersare installed at each end of the line. The 6'6 kV side of the transfonnersis delta connectedwhile the 33 kV side is star connected. The resistanceand reactance the star connectedwindings of are 0.5 and 3'75 ohms, respectivelyand for the delta connectedwindings arJ0.06 and 0.36 ohms. what is the voltage at the bus at the power station end?
6.6/33 kV
33/6.6kV
Fig. 5.5 Solution lt is convenient here to employ the per unit method. Let us choose, BaseMVA = 5 Base kV = 6.6 on low voltageside = 33 on high voltage side Cabie impeciance (8 + jZ.S) e/phase =
Input to a singlephase short line shown in Fig. 5.6 is 2,000kw at 0.8 lagging power factor. The line has a seriesimpedance (0.4 + j0.a) ohms. If the load of voltage is 3 kV, find the load and receivingend power f'actor.Also tind the supply voltage.
I
2,000 I kw
at 0.8 pf *Vs
lassins
3kv
I
L
Fig. 5.6
__1
I
_ ( 8 + r 2 . s ): x s (0'037 = + io'0r15) Pu (33)'
Equivalent star impedance of 6.6 kv winding of the transformer
Solution It is a problem with mixedend conditionsload voltage and input power are specified.The exact solution is outlined below: Sendingendactive/reactive power = receivingendactive/reactivepower + activekeactiveline losses For active power l ys I l1l cos ds= lVRllll cos / a + l1l2p For reactivepower l y s I l / l s i n g 5 5 =l V p ll 1 l s i n Q o + l t l 2 X ( i) (ii)
=
(O.OU 7O.36) (0.02+ /0.12) O/phase = + f
'F6, I
Modern PowerSystem Anatysis
Squaring(i) and (ii), adding and simplifying, we get lvrl2 lll2 = lVnlzlll2 + zlvRl lll2 (l1lR cos + tItX sh /n) + tlta @2 + f) the numerical values given l Z l 2 =( R 2 +f ) = 0 . 3 2 lysl l1l2,ooox1o3  2 , 5 0 0x 1 0 3 0.8
5.3
/o (iii)
Lines l. l3?.,{ Transmission of and Characteristics Performance Power IMEDIUM TRANSMISSION I,INE
due For lines morethan 100 km long, chargingcurrents to shuntadmittance
100 km to 250 km leneth. it is accurate to lump all the line admittance at the receivingend sufficiently resulting in the equivalent diagram shown in Fig. 5.7. Starting frorn fundamental circuit equations,it is fairly straightforward to write the transmissionline equationsin the ABCD constantform given below:
lVs I l1l cos /,  2,000 x 103 l y s l l l l s i n / 5 = 2 , 5 0 0x 1 0 3x 0 . 6 = 1 , 5 0 0x 1 0 3 From Eqs. (i) and (ii), we get 2000x1030.4ltP ,n l 1c o s P o = F l (iv)
=l'*;'llVl [:]
(s.1 ) 1
capacitance loadend line,localized Fig. 5.7 Medium l1l sin /o
 0 1500x103 .4tIf 3000
(v)
Nominalf
Representation
Substituting the known valuesin Eq. (iii), we have all
p'{lt (2,500 103;2 (3,000)'til2 2 x 3,000 = x + t210.4*29W"19! 3000 L 1s00xlq1r0.4112 +0.4x l+ o.zztt ta 3000 J
Simplifying, we get 0 . 3 2 V f  1 1 . 8x 1 0 6l l l 2 + 6 . 2 5 x 1 0 1 2 0 = which upon solution yields t_725 A Substituting l1l in Eq. (iv), we ger for cos Qp= 0.82 Load Pn = lVRl lll cos /a = 3,000 x 725 x 0.82 = 1,790kW Now l V5  = l 1 l c o s d s = 2,000
l V 5l =
If all the shunt capacitanceis lumped at the middle of the line, it leads to the nominalZ circuit shown in Fig. 5.8.
representation line,nominalT Fig. 5.8 Medium For the nominalZ circuit, the following circuit equationscan be written: Vc= Vn+ I o( Zl2) Is = In + VrY = In r Wo+ Vs = Vc + it (ZiZ) Substitutingfor Vg and 1, in the last equation, we get IR(ZL)Y
YvR) + . vs= vn+ I^ (zt2) (zD) [r^(t +)+ = v n ( , . t { ) + rn z ( t . t f )
7 2 5 x 0 .8
2000 :3 . 44 kY
of and Performance Power Tralq4lqslon Lines [l$k Characteristics Rearrangingthe results, we get the following equations
t
l.
r
MVA at 0.8 lagging power factor to a balancedload at 132 kV. The line conductorsare spacedequilaterally3 m apart. The conductorresistanceis 0.11 ohmlkm and its effective diameteris 1.6 cm. Neglect leakance.
+
Nominal zr Representation
(s.12
2 = 0.0094 pFlkrrr
R = 0. 11 x 250 = 27. 5 C) X  Zr f L = 2r x Y = jutl 50 x I . 24 x 10 3x 250 = 97. 4 Q Z= R + jX = 27.5 + j97^4 = I0I.2 174.T Q = 314 x 0.0094 x 10{ x 250 lg0" U
In this methoc the total line capacitance divided into two is equal parts which are lumped at the sending and receivingendsresulting in the nominal zr representation shown in Fie. 5.9. as
= 7.38 x 104 lW
Fig. S.9 Mediumline,nominal_7r representation From Fig. 5.9, we have _ t Is= In + VoY + Vs= l 2Vsy
li6.g"o = 109.3 369" A I r^ = #9 "13xl32 = vo (per phase) (I32/d, ) 10" = 76.2 l0 kv
z v s =[ t + + Y Z l v R + I *
/
1
\
vo+ eo*,)vov>zvn(r.* = tvoy + t;trr(t+vz)
\ 2 ) " r 1 x I i +: x 7.38 ioa 190"x10r.2134.2"\to.z \ 2 ) + 101.2174.2" x 109.3x 103l36.9"
1s=r**
7 6 . 2+ 2 . 8 5l 1 & . 2 ' + 1 1 . 0 6 3 7 . 3 " 1 j7.48  825 15.2" + 82.26 lV5 (line)= 82.6xJl  143 kv + ll&.2" = 0.964 70.01 I + L Y Z = 1 + 0.0374 2
v o v t + t ^ v z ) .+ ( , + (
Finally, we have
)vz)
'r i4 I ('*'t ', 
f / 1 \
/
z l,u.r ., ll "  
= (rine rv^or noload)
(5.13)
= 148.3 kv ffid:;# l t z
x 100  l2.3vo
LrU+;rt) [t*r")]L1nr
It sl uld b e noted tha nominalzand nominalrrwith the above hou J t l d at constants are not equ vale to eachother. The readershould verify this fact lurva 3nt e t by applying star_ delt trans ormati< to either one. :a tr nsfc ion t 5.4
Voltage regulation = W#2
THE LONG TRANSMISSION LINERIGOROUS
SOLUTION
Using the nominaiz method, find the sendingendvoltage and voltage regulation of a 250 km, threephase, Hz, transmission 50 line delivering 25
of For lines over 250 km, the fact that the parameters a line are not lumped but distributeduniformally throughoutits length, must be considered.
149 1
I
mooetn Po*", Sv
sis
characteristics and Performance Power Transmission of Lines
C17e)r'Czle)'21., 9t ,r, C, ,',', Z, Z,
(5.1e)
Fig. 5.10 Schematic diagram a longline of Figure 5.10 shows one phase and the neutral return (of zero impedance) of a transmission line. Let dx be an elemental section of the line at u dirtunr. " from the receivingend having a series impedance zdx and a shunt admittance ydr. The rise in voltage* to neutral over the elemental section in the direction of increasing "r is dV". We can write the following differential relationships acrossthe elemental section:
(5.20)
The constantsC, and C2 may be evaluatedby using the end conditions, i.e. when .r = 0, Vr= Vn and 1r= In. Substitutingthese values in Eqs. (5.17) and (5.19) gives Vn= Cr + Cz
dVx = Irzdx o, Yd,lx= v*ldx o, !1'
= ZI, = yvx
r^=!
(ct cz)
(s.14)
Lc
which upon solving yield
(s.1s)
,r=
Cr=
It may be noticed that the kind of connection (e.g. T or r) assumedfor the elemental section, does not affect these first order differential relations. Differentiating Eq. (5.14) with respecrro tr,we obtain drv, dT = dI.
*
1
(vn+zJn)
ZJ*)
2Un
i''
with 9r ant c, as determined above,Eqs. (5.r7) and (5.19) yield the solutionfor V.and 1. as
Substituting valueof + from Eq. (5.15),we ger the dx d2v
Ea
= rZv'
(5.16)
,,=(Yn+/o),,. *(h.? ),,. ,._ (bITk),,._(w*),,.
(s.2r)
This is a linear differentiai equation whose general solution can be written as follows:
V*=CpI**Cre1x
(s.r7)
(s.18)
where
7= ,lW
and C, and C, are arbitraryconstants be evaluated. to Differentiating (5.17)with respect x: Eq. to *_
Here Z, is called the charqcteristic impedanceof the line and 7is called the propagatton constant. Knowing vp, In and the parametersof the line, using Eq. (5.21) complex number rms values of I/, and I, at any distance x along the line can be easily found out. A more convenientform of expression voltageand currentis obtainedby fbr introducing hyperbolic functions. Rearranging Eq. (5.21), we get
v*=vnt+l+
\ 2 I. = VoL( " " 2 " \ "* 2
( o7* +
o7'\
/
r"^' 2 , ( e : J : : ) (
2 ) )* ," ( "\ ) e1' + et' \ 2 )
/
+
.t"
Here V' is the complex expressionof the rms voltage, whose magnitude and phase vary with distance along the line.
These can be rewritten after intloducing hyperbolic functions, as Vr= Vn cosh 1r + I^2, sinh 1r I,= Io cosh rr + V l.i when x = l, Vr= V, Ir= Is
Characteristics Performance PowerTransmission and of Lines ffi (5,22)
lines and
sinh  .+*4* 7/ 3!
he convenienflw
)!
..=Jyz(H+)
\^'
^l^i^.^
6.)

(s.28a) \
This seriesconvergesrapidly for values of 7t usually encounteredfor power r,Lqrrrrrnwio.l.r
can
expressionsfor ABCD constanis art
':::i;:l;;t =l;::,, Hl
A=D cosh 7/ B = Z, sinh 7/
A=D=l*
YZ 2
(s.23)
Bx,z (t.+)
c x Y(r*V\
(s.28b)
Here
(s.24)
\ 6 ) The above approximation is computationally convenient and quite accuratefor lines up to 400/500 km. Method 3 tat"iot +'?t'iot _= 1
c = Jsinh :r/ Z,
*In case [vs 1s] is known, fvn Inl can be easily found by inverting Eq. (5.23).
:l =li, [:] Hl
of ABCD Constants ^ 7 a+ jp
cosh (o/ + i7t) 
Z
;(e" ){r",
tpt + e"1 lBt1
(s.2s)
5.5
sinh (o/ + jpl) 
,at"i0t _e:deipt
(s.2e)
: tpt  td lBt1
Evaluation
INTERPRETATION
OF THE LONG LINE EOUATIONS
The ABCD constantsof a long line can be evaluatedfiorn the results given in F4. 6.24). It must be noted that 7 Jw is in generala complex number and can be expressedas
As alreadysaid in Eq. (5.26), 7is a comprexnumber which can be expressed as 7= a+ jp The real part a is called the attenuation constantand the imaginary part ^ Bis called the phase constant.Now v, of Eq. (5.2r) can be writtin as Z,lRlr.,,ritgx+n V^ = I V R + , *lVnZ,lnl
(s.26)
The hyperbolic function of.complex numbers involved in evaluating ABCD constantscan be computedby any one of the three methods given uJtow. Method I

2


z
I'
oc,x,tt/,xe2)
where cosh (cr./+ j1l) = cosh ul cos gl+ j sinh a/ sn pt sinh (a/ + jQl) = sinh al cos gl + j cosh a/ sn pt Note that sinh, cosh, sin and cos of real numbers as in Eq. (5.27) can be looked tp in standard tables. Method 2 (5.27)
Qr= I (Izn+ I&,) dz= I (Vn InZ,) The instantaneous voltagev*(t) canbe writtenfrom Eq. (5.30)as v* (t) = xd Ol! + 3/tl L t 2 ,o' ,i@t+r,,.+o,) l
(5.30)
= cosh r ++* 7/ .+) #*. .=(t
144 
nrodern Power System Analysis *l ^* j(at/tx+t , nlV^ +Jlltrle*ett''tLtrh)  ZrI '1
characteristics and Performance PowerTransmission of Lines
I
(5.31)
 A t ( f+ 4 0
Sendingendx=/
of voltageconsists two termseachof which is a function The instantaneous two travelling waves, distance.Thus they represent of two variablestime and
Vx= Vrl * VxZ
(s.32)
Now
 JIlh?\,* "',,
+ gx+ h) "or{'l
(5.33)
Fig. 5.12 Reflected wave If the load impedance Zr = +
.
At any instant of time t, v.rl is sinusoidally distributed along the distance from the receivingendwith amplitude increasingexponentially with distance, as shown in Fig. 5.11 (a > 0 for a line having resistance).
Envelop eox
IR
= 2,, i.e. the line is terminated in its
urAf
B
Sendingendx=/ x=0Receiving end ' ' Direction of travellingwave
w F i g . 5 .1 1 l nci dent ave After timeAt, the distribution advancesin distancephaseby (u'Atlfl. Thus and is the incidentv'ave'Line the this wave is travellingtowarcls receivingend exponentiallyin gclingl}onr tht: sendirlg to causeits arnplitucle decrease losses to the receivingend. Now
characteristic impedance,the reflected voltage wave is zero (vn zJn= 0). A line terminatedin its characteristicimpedanceis called the infinite line. The incident wave under this condition cannot distinguish betweena termination and an infinite continuation of the line. Power systemengineers normally caII Zrthe surge impedance.It hasa value of about 400 ohms for an overhead line and its phase angle normally varies from 0" to  15o.For undergroundcables Z. is roughly onetenthof the value for overheadlines. The term surge impedanceis, however, used ih connection (dueto lightningor switching)or transmission with surges lines.wherethe lines loss can be neglected such that
z, = z,=
(
;,.,1
rl/2
i/1
l:i:r)
rcsist''cc (;)"'. ,, purc
Elt+tlu^ u,r=
(az dz) cos  0x+
(s.34)
SLtrge Impetlance Loading (SIL) of a transmission line is.defined as the power delivered by a line to purely resistive load equal in value to the surge impedance of the line. Thus for a line having 400 ohms surge impedance,
retardsin distancephaseby (uAtl4. After time At the voltage distribrrtion trom the receivingendto the sendingend This is the reflecterlwave travelling to in exponentially going from the receivingend the with amplitudedecreasing in Fig' 5.12. as sendingend, shown At any point along the line, the voltage is the sum of incident and reflected voltage waves presentat the point tEq. (5.32)1.The same is true of current waves.Expressionsfor incident and reflected current waves can be similarly current from Eq. (5.21).If Z" is pure resistance, written down by proceeding wavescan be simply obtainedfrom voltage wavesby dividingby Zr.
SIL= JT y!lvo loook tv, I x t00okw J3 x aoo
= 2.5 lyRl2kw
(s.3s)
where lVol is the linetoline receivingend voltagein kV. Sometirnes, is it found convenientto expressline loading in per unit of SIL, i.e. as the ratio of the power transmittedto surge impedanceloading. At any time the voltage and current vary harmonically along the linc with respectto x, the space coordinate.A completevoltageor currentcycle along the line correspondsto a change of 2r rad in the angular argument Bx. The correspondingline length is defined as the wavelength. It 0 i.s expressedin radlm,
iAd I
Po*g!Sy$e!l Anelysis Modern (s.36)
147
)Zn/gm line for a typicalpowertransmission Now length)= 0 g (shuntconductancelunit
= 7= (yz)1/2 Qu,C(r+ juL))rt2
' =
of wave along the line would be The actual velocity of the propagation less than the velocityof light' somewhat
*fu=
of velocitvlight
(s.42)
 iu (LQ'''(r t
i)'''
7= a+ jg = ju\Lc),,,(tt#)
r ( C\ltz a   l  l
pointed out heyethat the vyaves hundredkilometres). It needsto be Figs,5.11and5.]'2areforitlustrationonlyanddonotpertainnareal power transmissionline'
., = 1" lot = 6,000 km "50(usuallya few lines are much shorterthan this Practicaltransmission tlrau'ttin
(s.37) (s.38)
at the line is 400 km long' The voltage 50 A threephase Hz transmission x = 0.4 ohm/ are r = o.!25 ohnr/km, d sending.en ts 220kV. The line parameters tm ani y = 2.8 x 106rnho/km' Find the following: (i )Thesending enclcur r ent andr ebeiving endvolt agewhent her eisno load
2\L)
0 = a (Lq'''
Now time for a phasechangeof 2n is 1f s, where/= cul2r is the frequency in cycles/s.During this time the wave travels a distanceequal to ). i.e. one wavelensth. \ ' (5.3e) of Velocity of propagation wave, , = 4=: f^ m/s "' ri f which is a well known result. line (R = 0, G = 0), transmission For a lossless  iu(LC)tlz ,= 7yz)''' such that e. = A, 0 = . (Lq'''
)2110 ,?n=,=:
and v = fA = ll(LC)rlz ms transmissionline For a singlephase L= lto ,n D r' 2r
2ffi0
1,, m
(s.40) (s.41)
are: Solution The total line parameters R  0.125 x 400 = 50.0 f) X = O.4 x 400 = 160'0 fl U  l'12 x IA3 lxf Y = 2.8 x 106 x 400 lg)" = 172'6" Q Z = R + i X = ( 5 0 ' 0+ j 1 6 0 ' 0 ) 1 6 8 ' 0 /90" x 168 172'6" YZ = l.l2 x 1O3 = 0.1881162'6" (i) At noload Vs= AVn' ls = CVa
C_
l n D /r
v=4 '
(Pol^D
2*o
It;,.t
'lt"nG
)t/2
'
A and C are comPutedas follows:
/A = L 
)
l*
l t
Y Z= l +
l
Since r and rt arequite close to each other, when log is taken, it is sufficiently q, = h D/r. accurateto assumethat ln
)
*x
2
0 . 1 8 81 1 6 2 ' , 6 "
= 0.91+ j0.028
f .r14E l A l = 0 .9 1
Modern Power Svstem Analvsis Characteristics. Performance Power Transmission and of Lines I 14g^ tsimplifying, we obtain the maximum permissiblefrequency as f = 57.9 Hz
C = Y(l + YZ/6) = = 1.09x 103 I
'
= lvnhn' #:#
= 242 kY
1 1 l5 l c l l V R l = 1.09 x 103xry = * 103= 152A "J3 It is to be noted that under noload conditions, the receivingend voltage (242 kV) is more than the sendingend voltage. This phenomenon known as is the Ferranti elfect and is discussed length in Sec. 5.6. at (ii) Maximum permissible noload receivingend voltage = 235 kv.
If in Example 5.5 the line is open circuited with a receivingendvoltag e of 220 kV, find the rms value and phase angle of the following: (a) The incident and reflected voltages to neutral at the receivingend. (b) The incident and reflected voltages to neutral at 200 km from the receivingend. (c) The resultant voltage at 200 km from the receivingend. Note: Use the receivingendline to neutral voltage as reference. solution From Example 5.5, we have following line parameters: r = 0.725 Qlkm; x = 0.4 Olkm; y = j2.g x 10{ Uncm z = (0.125 + j0.4) Olkm = 0.42 172.6 CI/km = y 1yz)t'2 (2.8 x 106x 0.42 /.(g0 + 72.6))t/2 = 1 . 0 8x 1 0  3l g I . 3 = ( 0. 163+ i1. 068) x t O    0+ j{J r a = 0 . 1 6 3x t O  3 ; = f (a) At the receivingend; For open circuit 1n= 0 Incident vcrltage Vn I ; ?cl n 2 1 . 0 6 gx l 0  3
% ^ ,= llV " l l :2f2f0i = 0 . e 3 6
,r, Now
1 A = l + L Y Z
2
1 ^ = 1 + ; t', .i2.8 x 106 x (0.125+ 70.4) z = (1  0._56 t06P1+ j0.t75 x r06P x S in c eth c i rn a g i n a ry a rt w i l l b c l essthan, p approxirnated as th of the real part. lzll can be
,l
V,,
l A l =   0 . 5 6x l r J 6 P= 0 . 9 3 6 p_ r0.936 x 0.56 l0 6
 220/J3 = 63.51 kV (to 10" neurral) 2
Reflected voltage vRzclR
/ = 3 3 8k m ;;,, ,A,= 220= 0.88 250
I Al*;xi1.l2xl03x Neglecting the imaginary part, we can write
)

VR
= 63.51 10" kV (to neurral) (b) At 200 km from the receivingend:
L( to+160,.I) s0) J0\
Incident voltage = :!u+urltlxl 2 lr:2oorm
l
/
l
= tAl 1 +" r.r2x1o3 160 x x &=
0.88
= 63. 51exp ( 0. 163x l0 3 x 200) x exp 01.068 x 103x 200)
ffi0'.I
f
Modern Po*e, System Anatysis
= 65.62 112.2" kV (to neutral)
Reflectedvoltage
Ys"'uritt'l 2 l' ,,,n,u
characteristics performance and of power Transmission I rsr Lines / v ^ , \ u"'J andturns rhrough positive a pr angle (represenred by phasor oB); I i.
while the reflected voltage wave decreasesin magnitude exponentiaily
YEu,
It is apparentfrom the geometry of this figure that the resultantphasor voltage Vs QF) is such that lVol > lysl. A simple explanationof the Ferranti effect on an approximate basis can be advancedby lumping the inductanceand capacitance parameters the line. As of shown in Fig.5.14 the capacitance lumpld at the ieceivingend is of the line.
= 6I.47 l12.2" kV (to neutral) (c) Resultantvoltage at 200 km from the receivingend  65.62 112.2" + 61.47 I  12.2" = 124.2+ j0.877 = 124.2 10.4" Resultantlinetoline voltage at 200 km = 124.2 x J3  215.1 kV 5.6 FERRANTI EFFECT
As has beenillustratedin Exarnple5.5, the eff'ectof the line capacitance to is cause the noload receivingend voltage to be more than the sendingend voltage.The effect becomesmore pronouncedas the line length increases. This phenomenonis known as the Ferranti. effect. A general explanation of this effect is advanced below: = / and In = 0 (noload)in Eq. (5.21), we have Substitutingx
t7 /5 
Fig.5. 14 Here Ir= ,
Vn
,at4gt *

Vn
\ juCt ) Since c is small comparedto L, uLl can be neglectedin comparisohto yc,tl. Thus 15  jVruCl Now Vn= Vs  Is QwLl) = V, + V,tj CLlz
(+.
V,
2
,at
ifl
(< A2\
\J.tJ I
lncreasino/ \.Locus of V5 wlth /
= vs0+ Jctt2)
Magnitude of voltage rise _ lvrltJ CLf
(s.M)
=olv.rt+
D Vpfor I = 0
(s.4s)
where v = 7/J LC. i1 velocity of propagationof the electromagneticwave the along the line, which is nearly equal to trr" velocity of light. 5.7 TUNED POWER LINES
\
En= Ero= Vpl2 Increasing/ Fig. 5.13
Equation (523) characterizes performanceof a long the line. For an overhead line shunt conductanceG is always negligible and it is sufficiently accurateto neglect line resistanceR as well. with this approximation 7 Jyz = jalLC cosh 7/ = cosh jwlJTC = cos Lt,lrc
The above equationshows that at I = 0, the incident (E,o)and reflected (E o) voltage waves are both equal to V^/2. With reference to Fig. 5.13, as I increases, the incident voltage wave increases exponentiallyin magnitude
;fiTi I
I
power uooern system Rnarysis
sinh ?/ = sinh jalJLC = j sin wtJLC
eharacteristics and performance power Tr@ of t
Z = Z, sinhl/ =

Hence Eq. (5.23) simplifies to
cos LC ulJ
. rT sina'tJ LC Z,
jZ, sinutJE
cosutE
=!
.o,nh..,tt2
_1
,,1  (t. It,t,) _ 
=
_ _ ___{l
t(tU Y ganh1il2) 2)
N o wi f a l J t C = h r , n = I , 2 , 3 , . . . lV5l= lVpl llsl = llal i.e. the receivingend voltage and current are numerically equal to the colresponding sendingend values,so that thereis no voltagedrop on load. Such a line is called a tuned line. For 50 Hz, the length of line for tuning is
Fig. 5.15 Equivarentz network a transmission of rine For a zrnetwork shown in Fig. 5.15 [refer ro Eq. (5.13)].
Lr,J +!v,2,) (r*ir,r,)lL,_) Lr,(,
According to exact solurion of a long line [refer to Eq. (5.23)].
z,
.Jru. ,
( 5. 48)
1 !r_.
Since ll^frc
2nrfJ LC
= y, the velocity of light
,i^,^,?rx... r=+@))=
= 3,000km, 6,000 km,...
(s.47)
For exact equivalence, must have we Z/= Z, sin h 7/ f * *YtZ=cosh 2 From Eq. (5.50) 7/
. 7/ L1,_, I z, ___.. cosh J JLl^
= fi:IfilH;':"! l'lt,^t
(s.4e)
(s.s0) (s.51)
It is too long a distance of transmissionfrorn the point of view of cost and efficiency (note that line resistancewas neglectedin the above analysis).For a given line, length and freouencytuning can be aehievedby increasing or L C, i.e. by adding series inductances shunt capacitances severalplaces or at along the line length. The methodis impracticaland uneconomical power for frequencylines ancl is adoptedfor tclephonywhere higher frecluencies are employed. A methodof tuning power lines which is being presentlyexperimentedwith, uses seriescapacitorsto cancel the effect of the line inductanceand shunt inductors to neutralize line capacitance. long line is divided into several A sectionswhich are individuatly tuned. However, so far the practical method of improving line regulationand power transfercapacity is to add seriescapacitors to reduceline inductance;shunt capacitorsunder heavy load conditions;and shunt inductors under light or noload conditions. 5.8 THE EOUIVALENT CIRCUIT OF A LONG LINE
Z' = ,l:. sinn1/ = Z, tlnh'' : z[*init rr 1 ( vY tJyz 7t )
(5.52)
11"'' sln!24 is thc fitct.r by which thc scrics irupcdunce ol'the no'ri'alz must be multiplied to obtain the z parameter of the equivalenta Substituting 7 from Eq. (5.50) in Eq. (5.51),we ger 1*
I
; YtZ, sitrh 7/ = cosh fl
So far as the end conditions are concerned, the exact equivalentcircuit of a transmissionline can be established the form of a T or zrnetwork. in The parameters the equivalentnetwork are easily obtainedby comparingthe of perfbrmance equationsof a znetwork and a transmissionline in terms of end quantities.
Lv,2
(s.53)
Wl
I
Anqlysis Power System Modern
(tanhfll2\. d anr ol' the rnus I is thc lactor by which thc shunt aclmittancc
\ il/z )
/ 1
VR= 20 +
VJ
 l z 7 1 0k v
nominalnmustbe multiplied to obtain the shunt parameter (Ytl2) of the z. equivalentNote that Ytl  + +Y' Z' I : j\ 4 ) 2 , of abovevalues Y/and Z/.
\ l
(a) Short line approximation:
equationin termsof the sinh 7/ is a consistent
Vs= t27 + 0.164/._36.9 x 131.2 172.3 = 14514.9 = lYsltin" 25L2 kV Is= In= 0.764/_36.9kA Sendingend powerfactor= cos (4.9"+ 36.9"_ 41.g.) = 0.745laggrng Sendingend power_ JT x 251.2 0.764x x 0.745 = 53.2MW x@)Nominaltrmethod: I ' A = D = l + ! y z = . + _x l0r I lg0" x l3l .2 /.72.3" 2 = 1 + 0.0656 L162.3 = 0.93g/_L2" B=Z=I3L.2 172.3"
sinh 7/ = tuthl!/2 1 so that the 1 and For a line of medium length
fll2
1t
equivalent n network reduces to that of nominaln.
Fig. 5.16' EquivalentTnetwork of a transmissionline EquivalentT network parameters of a transmission line are obtained on similar lines. The equivalentT network is shown in Fig. 5.16. As we shall see in Chapter 6 equivalentr (or nominalr) network is easily
^.1^^+^l <fllIWPLVLT +^ L\J l^^.1 l\rat.l {'l^,r, llLrW otrr.lioo JLUUM o 1 q ll.u io lot f}roref'nro lltvlvlvlvt r 4nri lr r p r c q l l . , (r l v vrLtqrrJ pmnlnrrerl vr^lHrvJ vs'
c = v ( r + ! v z ) =y + l f z
\ 4 / 4
5.7 Exirmple
line 300 km long has a total seriesimpedanceof 40 + A 50 Hz transmission j125 ohms and a total shunt admittanceof 103mho. The receivingendload is voltage, 50 MW at220 kV with 0.8 lagging power factor.Find the sendingend current,power and power factor using (a) short line approximation, (b) nominalzrmethod, (c) exacttransmission line equation lBq. (5.27)1, (d) approximation[Eq. (5.28b)]. Comparethe resultsand comment. Solution Z = 40 + iI25  131.2 172.3" Q
= 0.00 190"+  I . t061J g0. x 131.2 L72.3" 4 = 0.001/.90 7s= 0'938 1r.2 x r27 + r3r.2 172.3" x 0.164L36.9" = 719.7/7.2 + 21.5135.4"= 73j.4 /.6.2 = lVsllin" 238 kV 1s= 0.001Z90 x 127+ 0.93gZl.Z x 0.164/._ 36,9" = 0.127l9A + 0.154l_35.7"= 0.13 /.16.5" Sendingend= cos (16.5. 6.2) = pf _ 0.9g4 leacling Sendingendpower= * 23gx 0.13x JT 0.gg4 = 52.7 MW (c) Exacttransmission equarions line (Eq. (5.29)). fl = al + jpt =JyZ
Z= 10' 190" U load The Leceivingend is 50 MW at ?20 kV,0.8 pf lagging. 1o= l36.9" = 0.164 l36.9" kA ,=+ J3 x220x0.8
 155 

PowerSystem Modern 4nalysis = Jrotlgo"*l3t.zl7z.3" = 0.0554 i0.3577 + = 0.362 8L2" 1
characteristics and Performance Power Transmission of Lines = 0.938 ll.2 (alreadycalculatedin part (b))
cosh(al + i 0l) = !k"'
lgt + e*t lBt1
B= z(t*!Z\= \ 6 )
Z+  yf 6
'
gl = a.5'fi1 (radians)r'6.4,g"
20.49" = 0.886  j0.331
= lzo.49"= 0.99+ j0.37 1e0.49") 1.057 eo.0s54
e4'oss4l20.49" = 0.946 l= gosh 7/ = 0.938+ iO.O2 0.938 11.2" sinh 7/ = 0.052 + 70.35 = 0.354 181.5"
L..= 1
= 731.2172.3 + { x 1o3 lgo" x (13l.D2 1144.6. 6 = 131.2172.3"+ 2.87 l125.4" = 128.5172.7'
c= y(w!Z') = o.oo1 + ]x to{ lrgo xr3t.2172.3" tetr \ 6 )
6
= 0.001l9O"
8.85'
E
Y 10'190" 1 A = D = c o s hf l = O : 9 3 8 L . 2 "
"
lY
: t
l3nnn"
= soz.2I l
.,^
Vs= 0.93811.2"x 127 10" + 128.5 172.7'x 0.164I 36.9" = ll9.I3 11.2" + 21.07135.8" 136.2 it4.82 + = 137 16.2" kV = I YsIrin" 237 kY .3 /s = 0.13 116.5"(same calculated part (b)) as in Sendingend = cos (16.5'  6.2 = 10.3") 0.984leading pf = Sendingend power= Jl 'fhe * ?313 x 0.t3 x 0.984
t
Now
x tsr t 88s" 0354 s' =7::: ,lr ul"u"'
Vs= 0.93811.2"x 127 10" + 128.2172.65"x 0.164 l36.9"' = 1 1 9 . 1 3 1 . 2 "+ 2 L 0 3 1 3 5 . 7 5 " 1 = 136.97 16.2" kV lVs hin.= 81,23W C = Lsinh .y/ z, f x 0.354 181.5" 362.2118.85"
= 52.58MW results tabulated are below:
Short line uppntximatktn Nominalr Exact Approximation
(s.28h)
238 kV o.I3 116.5"kA 0.984leading 52. 7M W 237.23 kV 0.1286 1t5.3" kA 0.987leading 52. 15 W M ?37.3kV 0. 131t 6. 5" kA 0.984leading 52.58MW
.\ Is = 9 .J 7 x l O a 1 9 0 .4 " x 127+ 0.938 11.2" x0.164 l 36.9 = 0.124 190.4" + 0.154"  35.7" = 0.1286 115.3"kA pf Sendingend = cos (15.3'  6.2"  9.1") = 0.987leading power = Jt Sendingend x 237.23 x 0.1286 x 0.987
= 9.77 x 104 190.4"
l y s l h n "2 5 I . 2 k v I, 0J& l36.9" kA p.f, 0.745 lagging P" 53.2 MW
Comments We find from the above example that the results obtained by the nominal zr method and the approximation (5.28b) are practically the same and are very close to those obtained by exact calculations(part (c)). On the other hand the results obtained by'the short line approximation are in considerableerror. Therefore, a line of this length(about300 knr), it is sufficientlyaccurate ftlr to use the nominalr (or approximation(5.28b)) which results in considerable savingin computational eflort.
= 52.15 MW (d) Approximation(5.28b): A'=D =  +  Yz )
15ElI 5.9
ModernPowerSystemAnalysis LINE
characteristics and performance power Transmission of Lines
POWER FLOW THROUGH A TRANSMISSION
sn= t0 lvRt al!lrv^tz{1*t] ti+l vstt13= 'ul'll^' d) l4v^t2l1p t(/r Similerly,
in equationwas presented the form of line performance So far the transmission voltase and current relationshipsbetween sendingandreceivingends.Since in loadsare more often expressed terms of real (wattslkW) and reactive (VARs/ line equationsin the to kVAR) power, it is convenient deal with transmission form of sending and receivingendcomplex power and voltages. While the problem of flow of power in a general network will be treated in the next chapter,the principles involved are illustrated here through a single transmissystem)as shown in Fig. 5.17. sion line (2nodel2bus
ay
t, = I?l,r,f tur  u) lll*lf t@+6)
(s.se)
In the above equationsso and s, are per phasecomplex voltamperes,while v* and vr are expressed per phasevolts. If yR and 7, are in expressed kv line, in then the threephase receivingend complex por., is given by
x so(3phase = t {l#%+191 t( s r,  l4l tvot2to6,, ,, a)F vA) '/ .'"'lEl L\tl ,i Jt *JT lBl L 3
Ss= iQs
Sp = Pp +lQP
s^(3phase = MVA)
T#
4p  , l*lvopt1p a1
(s.60)
system F i g .5 .1 7 A tw obus voltage as areferencephasor(Vn=lVRl 10") Let us take the receivingend and let the sendingendvoltagelead it by an angle 6 (Vs = lVsl 16). Tlp angle d is known as the torque angle whose significance has been explained in Chapter 4 and will further be taken up in Chapter 12 while dealing with the problem of stability. The complex power leaving the receivingendand entering the sendingend
This indeedis the sameas Eq. (5.59).The sameresult holds for sr. Thus we see that Eqs. (5.58) and (5.59) give rhe threephase MVA if vs ina vo *" expressed kV line. in If Eq' (5'58) is expressed real and imaginary parts, we in can write the real and reactive powers at the receiving_endas
basis) as line of the transmission canbe expressed (on per phase = Pn + .iQn Vnfn Sn Ss= Ps + iQs = YsI;
r* = JI]lv^]cos  a/  l{l rv^r' (B a) (/ v cos " tBt lrl "
/1 Qn= ff lyslly*l . ,n ^ sin1/ 6t
(5.61) (s.62)
(J.J+/
/ E
E
A \
(5.s5)
sin(/ a) lliivni'?
lAl .__."
similarly, the real and reactive powers at sendingend are
t" can, however, .*pr"rsed in terms of currents Receivingand sendingend voltages[see Eq. (5'1)] as receiving and sendingend
* vR t,r= ivr,=
,r=*r, *ro
p, = I ?l v,P cos o) lvsllv'l (o' lBl J
pr Qs =
(s.s6) (s.s7)
at
tBicos(0*A
(5'63)
6.e) lsrtl It is easyto seefrom Eq. (5.61) that the receivedpower po will be maximum
tvsttvRt ' .'., (/1+ b) l*l  'ur,t sin (p  a) If
Let A, B, D, the transmissionJine constants,be written as A = lAl la, B = lBl lP, D = lDl lo (since A = D) Therefore, we can write
6=0
such that
t^=
r1r "r t v 5tt( , _ 0 ) _ l A l r v4 a _ 0 ) rBl 'l ; l ' (
tp r,= v,t 6,l+lvRt i+l t(a+
for 1o in Eq. (5.54)'we get Substituting
lBThe corresponding en @t max po ) is
'u:'ll^'t Attv'P Po(max) c,os _ a) (/1 v) .= I Bt
o ^ = l4LYry . i n ( 0  u ) A n =   f f f : !t s
:
(5.65)
f60 
I
Modern Po*e, Sy.t"r Analysir
_
Thus the load must draw this much leading MVAR in order to receive the maximum real power. Consider now the specialcaseof a short line with a seriesimpedanceZ. Now AD=I l0:B=Z=lzlle Substitutingthesein Eqs. (5.61) to (5.64), we get the simplified resultsfor the short line as
t Equation (5.72) be furthersimplified assuming can by cos 6 = r,since dis normallysmall*. Thus
Let lvtl  lvRl= lAv1, the magnitudeof voltage drop acrossthe rransmission line.
p o  Y r i l l o l c o s ( g  6 \  l v R Pc o s o lzl tzl i g o = l v ' ) l Y * l  s(n  b )  l v * l ' , i n o d lzl tzl for the receivingend for the sendingend and
(5.66)
n^= #tavl
(s.74)
(s.67)
ps=lI:1.o,etl1'lunl @+6) .o,
" lzl tzl
(s.68)
d_'u:'lI.' (d* e,_ ry_srn l z l sin o vo \ lzl
(s.6e)
The above short line equation will also apply for a long line when the line is replacedby its equivalentr (or nominalr) and the shunt admittances are lumped with the receivingendload and sendingendgeneration.In fact, this technique is always used in the load flow problem to be treated in the next chapter. From Eq.(5.66),the maximum receivingend power is received,when 6 = 0
= sothatP^ (rnax)
Now cos d= RllZl,
lv]\Yol Ytc's tzt tzl
//
l'rt o = (max)'tr',,o' Pp
(5.70)
Normally the resistanceof a transmission line is small compared to its reactance(since it is necessary maintain a high efficiency of transmission), to so that 0 = tant xlR = 90"; where z = R + jx.The receivingend Eqs. (5.66) and (5.67) can then be approximated as
(s.1r)
(s.12)
several importanrconclusionsthat easily follow from Eqs. (5.71 to (5.74) ) are enumeratedbelow: I' For R = 0 (which is a valid approximation for a transmissionline) the real power transferred the receivingend proportional to is to sin 6 (= 6for small values of d ), while the reactiu. po*", is proportional to the magnitude of the voltage drop acrossthe line. 2. The real power received is maximum for 6 = 90o and has a varue lvsllvRvx. of course, d is restricted to varues weil below 90o from considerations stability to be discussed of in Chapte 12. t 3. Maximum real power transferred for a given line (fixed X) can be increasedby raising its voltage level. lt is from this consideration that voltage levels are being progressivelypushedup to transmit larger chunks bf power over ronger distances wananted ty l*g; ;irtlln"rutirrg stations. For very long lines voltage level cannotbe raised beyond the limits placed ' 4'1'presentday high voltagetechnoiogy. increase To power transmitted in suclt citscs,tltc only choicc is to reducethe line reactance. This is accomplished adding series capacitors the by in line. This idea will be pursuedfurtherin chapter 12. Seriescapacitors woulclof c<lurse incrcos!. the severityof line over voltagesunder switching conditions. 4. As said in 1 above,the vARs (lagging reactive power) deriveredby a line is proportionalto the line voltage drop and is independent of d Therefore, in a transmissionsystem if the vARs demand of the load is large, the voltage profile at that point tends sag rather to sharply. To maintain a desiredvoltageprofile, the vARs demanJof the load must be met locally by employing positive vAR generators(condensers). This will be discussed lengthin Sec. 5.10. at A somewhat more accurateyet approximateresult expressingiine voltage drop in terms of active and rlactive powers can be written directly from E q. (5.5), i. e. lAVl= llnl R cos Q + llol X sn Q lvRl dis necessary considerations system from of stability whichwill be discussed at length in Chapter12. *small
r i
I
i
I I
162  "odern
power SvstemAnalvsis
characteristics and Performance Power Transmission of Lines
R!ry* XQn lvRl Thisresult reduces thatof Eq. (5.74) R = 0. to if = j Example5.8
(s.75)
= Case (a): Cable impedance 70.05 pu. Since cableresistance zero, thereis no real power loss in the cable.Hence is Pcr t Pcz= Por * Poz = 40 pu Pot= Pcz= 20 pu The voltage of bus 2 is taken as ref'erence, i.e. V, 10" and voltageof bus 1 is V1 16r. Further, for flat voltage profile lVll = lV2l  L Real power flow from bus I to bus 2 is obtained from Eq. (s.68) by recognizingthat since R = 0, 0= 90". Hence l D _ D = fv t l l v lsin 6, Ps= Pn
) =
An interconnector cablelinks generating stations1 and 2 as shownin Fig. 5.18. The desired voltageprofile is flat, i.e.lVrl=lVzl = 1 pu. The total demandsat the two busesare Spr=15+75Pu So z = 2 5 + 7 1 5 P u The station loads are equalizedby the flow of power in the cable. Estimate the torqueangleand the stationpower factors:(a) for cable z = 0 + 70.05 pu, and (b) for cable Z  0.005 + 70.05 pu. It is given rhat generator G, can generatea maximum of 20.0 pu real power. Solution The powers at the various points in the fundamental(twobus) systemare defined in Fig. 5.18(a).

SlIl
lxl
0.05
d,
or From Eq. (5.69)
4 = I4.5" vr= I lL4'5"
@
I Sn.' = P61+jQ61
lvlt51
Qt=
^'
I
lv,l'
X
lv,llvl ,:
cosdt

Q v u1.
 ,D  1 r i A  D . 'l\1D1
Jp2 = Ip2+ J\Jp2
x 0.e68= 0.638pu
U.UJ
From Eq. (5.67)
":
U,\,J
c)
SDI= 15 +/5
(a)
f2o=
lvtllv, Tcosl
d'
tv12 Qs =  0.638 Pu
X
Reactivepower loss* in the cable is
__>
Sn=5l
(b)
G)
{ + zo.ro i16.12
V 2 =1 . 0l o "
pu QL= Qs en _ 2es _ 1.276 Total load on station = (15 + i5) + (5 + j0.638) 1 = 20 + j5.638 Powerfactorat starion = cos [run' t'f:t') = 0.963lagging I \ 2 0 )  (25 + jls)  (5  j0.638) Total load on station 2 =20+i15.638
15+j5
(c)
25 + j1S
Fig. 5.18 Twobus system
Reactivc powcr loss can also be cornputedas l/l2X=
:t +(9.!1q)1 o.os= pu. t.z7 " I
164 I
toctern power Svstem ,Analvsis
characteristics and Performance power Transmission of Lines
165
') Powerfactor ar station2 = coS (ron*r 15'638 = O.zgslagging \ 2 0 ) The stationloads,load demands, and line flows are shownin Fig.5.1g(b).
6r = 14.4" Substituting in Eqs.(ii), (iii) and (iv), we ger dr , = Q c t = 5 . 1 3 , c 2= 1 6 . 1 2P G z 2 0 . 1 0 Q
^'^D'
Case(b): the cable resistance causesreal power loss which is not known a priori. The real load flow is thus not obvious as was in the case of R = 0. we specify the generation at station I as Pcr= 20 Pu The considerationfor fixing this generation is economic as we shall see in Chapter 7. The generationat station 2 will be 20 pu plus the cable loss. The unknown variables in the problem are P62, 6p Qcp Qcz Let us now examine as to how many system equationscan be formed. From Eqs. (5.68)and (5.69)
ir; { iz i, r"rp rt v. t "" case = Cableimpedance 0.005 +70.05 = 0.0502 lB4.3 pu. In rhis""r
It may be noted that the real power loss of 0. i pu is supplied by Gz(Pcz  20.10). The above presentedproblem is a twobus load flow problem. Explicit solution is always possiblein a twobus case.The readershould try the case when Q c z = 7 1 0 a n d l V 2 l= ' The general load flow problem will be taken up in Chapter 6. It will be seen that explicit solution is not possiblein the generalcaseand iterative techniques have to be resortedto.
"'^'
= Pct Por ', =
#cos
d V#*s@+
61)
(i)
A 275 kV transmission line has the following line constants: A = 0.85 15": B  200 175" (a) Determine the power at unity power factor that can be received if the \ voltage profile at each end is to be maintaine at 275 ky . d (b) What type andrating of compensation equipment would be requiredif the load is 150 MW at unity power factor with the samevoltage profile as in part (a). (c) With the load as in part (b), what r,vould thc receivingend bc voltageif the compensation equipmentis not installed? Solution (a) Given lV5l= lVal = 215 kY; e, = 5o,C = J5". Since the power is receivedat unity power factor, Qn= o Substitutingthesevalues in Eq. (5.62), we can write
5 = = cos 84.3"U.U)UZ
cos (84.3"+ 4) ^= U.U)UZ
e c r  e o t =e s = f f r t ^ t  t r r , \ i , s i n ( d +4 )
sin 84.3" =+sin Qcr  S = ;j0.0502 0.0502 FromEqs.(5.66) and (5.67) (84.3" Ur) * (ii)
lv c P = l P o z  P z =P n  l vftf l l v z c o s( d  6 r )   1 :'P o s e G 2s  Pc2= (84.3" A,l 84.3o (iii)
o  !5x]75200 whichgives
6 ))" From Eq. (5.61)
sin (75" d) 
*;rcos
ij;
x Q75)2 (75"5") sin
#rcos
0=378sin(75"A302
eorecz= Uffiri, (o 6t>,Yl!rr", en= lzl
15 Qcz= #, sin(84.3" 4) #, sin84.3" (v)
p52, 51, Thus we havefour equations, Eqs. (i) to (iv), in four unknowns e61, thoughtheseare nonlinearalgebraicequations, Q62.Even solutionis possible in this case.Solving Eq. (i) for d,, we have
Pn= "
275^x275 cos(75o 22") 9 85 * (2712cos70o 200 200  227.6 109.9 117.7 = D{W
166 
I
todern power SvstemAnalvsis
(b) Now lV5l= lVpl = 275 kV Power dernanded load = 150 MW at UPF by P n = P R = 1 5 0 M W ;Q o = 0
characteristics and Performance Power Transmission of Lines
Ifil (ii)
0 = !:y*t sin (75' .5) *tvot2 200 200 From Eq. (ii), we get sin(75" A=0.00291VR1 0 . 8 5 . ^  , ) x (275)'cos70o ,*
sin 70'
150=
215zq7cos (75o 5)200
1 5 0 = 3 7 8 c o s ( 7 5 " 4  1 1 0 or From Eq. (5.62) 5 = 28.46"
_ 150 = 1.375 lynt (1 _ (0.002912 lv^121u2 0.00l45tyRl2 Solving the quadratic and retaining the higher value of lv^, we obtain lVal = 244.9 kV Note: The second and lower value solution of lVol though feasible, is impractical as it correspondsto abnormally low voltage and efficiency. It is to be observedfrom the results of this problem that larger power can be transmitted over a line with a fixed voltage profile by installing compensation equipmentat the receivingendcapableof feedingpositive VARs into the line. Circle Diagrams It has been shown above thatthe flow of active and reactive power over a transmission line can be handledcomputationally.It will now be shown that the Iocusof complex sendingand receivingend power is a circle.Sincecircles are convenientto draw, the circle diagramsare a useful aid to visualize the load flow problem over a single transmission. The expressions complex numberreceivingand sendingend for powers are reproduced below from Eqs. (5.58) and (5.59).
en=
sin (75.  28.46") 0'85x e75)2 sin 70o 200 "#  274.46 302=  2l.56 MVAR
Thusin orderto maintain2T5kV at a receivingend, en= 27.56 MVAR must be drawn alongwith the real power of Po = 150 MW. The load being 150 MW at unity power factor, i.e. Qo = 0, compensation equipmentmust be installed at the receivingend. With reference Fig. 5.19, we have to  2 7 . 5 6 +Q c = 0 or Qc = + 27.56 MVAR i.e. the compensationequipment nnustfeed positive VARs into the line. See subsection 5.10 for a more detailedexplanation.
1 5 0 j 2 7 . 5 6 1 5 0+ / 0
sR= rv^r' r. +P 4r b) lfl 4/rt(rn=l+lvs( o)+3 4/r+ o
F i g .5 . 1 9 (c)
Now,
(s.s8) (5.5e)
Since no compensation equipment provided is P n = 1 5 0 M W ,Q n = 0 l V 5 = 2 7 5 k V , l V a l =? l
The units for Sp and S, are MVA (threepha,se) with voltages in KV line. As per the above equations, So and ,9, are each composedof two phasor componenfs6ne a constantphasorand the other a phasorof fixed magnitude but variable angle. The loci lor S^ and S, would, therefore,be circles drawn from the tip of constantphasorsas centres. It follows from Eq. (5.58) that the centre of receivingendcircle is located at the tip of the.phasor. IB l in polar coordinates in terms of rectangular or coordinates, Horizontal coordinate of the centre
l+l vRP  a) ' \ t(r
(s.16)
Substituting this data in Eqs. (5.61)and (5.62),we have
= t':\!'"os (75" A  ggtv^t2 cos 150 70" 200 200
(i)
= llrvf cos a)MW (//
(s.77)
e.'1,,
.d6Eif
,
.
I
_
Modernpower SystemAnalysis
.,.,,
characteristics and performance power Transmission of Lines Vertical coordinate of the centre
t69
Vertical coordinateof the centre lBl The radius of the receivingendcircle is tysllyRlMVA tBl
= i{ltv*t2sin(tJ a) MVAR
tB t The radiusof the sendingend circleis (5.78)
= l+ilvrt2 (F a) MVAR sin
(s.81)
The sendingend circle diagramis shown in Fig. 5.2I.Thecenrre is located by drawing OC, at angle rt? a) from the positive MWaxis. From the centre the sendingend circle is drawn with a .uoi.rr${e(same as in the case of \ lBt receivingend). The operatingpoint N is located by measuringthe torque angle d(as read from the recefvingendcircle diagram) in ttredirection indicated from thc re'fi'rcncc' Iinc.
MVAR
MVAR
C5
Roforonco lino  for angle d
lAll t2 u"l lall
Radius lYsllVnl lBl Phasor55 = P5 +ie5
Referenceline f o r a n g l e6
Fig. 5.20
Receivingend circle diagram
1Qs
F'or constant lVol, the centre Co rernains fixed and concentric circles result for varying l7rl. However, for the case of constant ll{l and varying lvol the centrcso1'circlesmovc along the line OCoaru),have raclii in accordance to ltzrl lvRtABt. Similarly, it follows from Eq. (5.59) that the centre of the sendingend circle is located at the tip of the phasor
+
Fig. 5.21 Sendingendcircle diagram
IB l
t(0. a) l4l,u,,'
(s.7e)
in the polar coordinates or in terms of rectangular coordinales. Horizontal coordinate of the centre
=
(f cos  a)Mw Bltv;2
=u':;=o i,i,=',!"i
(s.80)
in Figs5.22 and5.23.
The corresponding receiving and sendingend circle diagrams have been clrawn
characteristics and Performance power Transmission of Lines [giM t Resistance= 0.035 Olkm per phase Inductance= 1.1 mHlkm per phase Capacitance= 0.012 pFlkm per phase If the line is supplied at 275 kV, determine the MVA rating of a shunt
lvnl2
tzl
Pa=oK Qn= KM
thereceivingend whentheline is delivering load.Usenominal method. no zr Solution R=0.035x400=14Q X = 314x 1.1x l03 x 400 = 138.2O Z = 14 + 7138 138.1184.2" Q Y= 314 x 0.012x 106 400 lg0"  1.507 103/_W U x x A = ( t + L v z \ = 1 +  l  x t . 5 0 7x 1 0  3 r 3 g . i l r i 4 . z " x \ 2 ) 2 = (0.896 70.0106) 0.896 lj.l' = + B=Z138.7 184.2" lV5 = 275 kV, lVpl= 275kV lysllyRl 275x275  545.2MVA Radius receivingend of circletBl 138.7 Location the cenre of receivingend of circle, ,, = 275x275x0.896= 488.5r l l = ' r d d ' ) MVA l4:l1 i l l
IB I 138J \
Fig.522 Receivingend circre diagram a short rine for
MVAR
Ps=oL Qs=LN lvsl2
l@  a) = 84.2" 0.7" = 83.5"
i,\/A P
lzl
55 MVAR
_
> M W
L
Fig. 5.23 Sendingend circlediagram a stror.t for line
488.5 MVA
oJl:
useof circlediagrams i$ustrated means therwo examples is by of given
Cp
.tuou.rrro
Fi1.5.24 Circlediagram Example for 5.10 A 50 Hz, threephasg,275 kY,400 km transmissionline has the following parameters: From the circle diagramof Fig. 5.24, + 55 MVAR must be drawn from the receivingendof the line in order to maintain a voltageof 275 kV. Thus rating of shunt reactor needed= 55 MVA.
X12,:jl Example5.11
power System Modern Analysis
Lines  ,i?3;of and Characteristics Performance PowerTransmission
t
load of 250 MW at 0.85 (a) Locate OP conespondingto the receivingend laggingpf (+ 31.8). Then , ^ or^ lysllyRl 275lvsl
A  0.93 11.5", B = Il5 ll7" If the receivingend voltage is 2'r5kV, determine: (a) The sendingend voltagerequired if a load of 250 MW at 0.g5 lagging pf is being delivered at the receivingend. (b) The maximum power that can be delivered if the sendingendvoltage is held at 295 kV. (c) The additional MVA that has to be provided at the receivihgend when delivering 400 MVA at 0.8 lagging pt the suppty voltage being maintained at 295 kV. Solution In Fig. 5.25 the centre of the receivingendcircle is located at
k lVsl = 355.5 V (b) Given lV5  = 295 kV. t".l?"  705.4 MVA 115 Drawing the receivingend circle (see Fig. 5.25) and the line C^Q parallel to the MWaxis, we read PR o = RQ = 556 M W Radius of circle diagram(c) Locate OPt conespondingto 400 MVA at 0.8 lagging pf (+ 36.8"). Draw P/S parallel to MVARaxis to cut the circle drawn in part (b) at S. For the Therefore, additional specified voltage profile, the line load should be O^S. MVA to be drawn from the line is P/S = 295 MVAR or 295 MVA leading 5.10 METHODS OF VOLTAGE CONTROL
 2t5x275x0'93 611.6 MVA l4li Rt'' lBl tt5 c o s l 0 .8 5= 3 1 .8 " l@a ) = 7 7 o  1 . 5 "= 7 5 . 5 "
MVAR
I
Practically each equipmentusedin power systemare ratedfor a cprtain voltage with a permissible band of voltage variations.Voltage at various buses must, therefore,be controlled within a specified regulation figure. This article will discussthe two methodsby meansof whieh voltage at a bus can be controlled. lvslt6
lvRltj
a)
I
p".io, P^iio* 
system Fig. 5.26 A twobus Considerthe twobus systemshown in Fig. 5.26 (akeady exemplifiedin Sec. 5.9). For the sake of simplicity let the line be characterizedby a series Further, since the torque angle d is reactance(i.e. it has negligible resistance). under practicalconditions,real and reactivepowersdeliveredby the line small voltagel{  can voltage lVrl and a specifiedreceivingencl for fixed sendingend bewritten as below from Eqs. (5.71) and (5.73).
(5.82)
Fig. 5.25 Circle diagram for ExampleS.11
r rl ef, 'li, ,,u, rv,i '= X
Equation (5.83) upon quadraticsolution*can also be written as
nsmtssion l"ffie Lines
(5.83)
Reactive Power Injection It follows from the above discussion that in order to keep the receivingend voltageat a specifiedvalue l{1, a fixed amountof VARs drawn tai I *;;; from the line. To accomplish this under conditions Qn, a local VAR generator (controlled reactive power source/compensating equipment) must be usedas shown in Fig. 5.27.fle vAR balanceequation at the receivingendis now
rr{ r=
}vrt
. +tys  (1  4xesn Avrtzlt/z
(5.84)
Since the real power demandedby the loacl must be delivered by the line, Pn= Po varying real power demandp, is met by consequent changesin the rorque angle d. It is, however' tobe noted that the receivedreactive power of the line must remainfixed at esnas given by Eq. (5.g3) for fixed rv, I and specifiedr4r. il" line would, therefore,operatewith specified receivingendvoltage for only one value of Qo given by Qo = Qsn Practical loads are generally lagging in nature and are such that the vAR demandQn may exceed et*.rt easily follows from Eq. (5.g3)that for or; otthe receivingendvoltagemust changefrom the specified value 'n'i some value lTol to meet the demandedVARs. Thus lv*l 'Jo= '(lYsl  lVol) for (QD> Q^= o^= Qn = QsR) ;i The modified lVol is then given by
Oi * Qc= Qo
Fluctuationsin Qo ue absorbedby the local vAR generator o6 such that the vARs drawn from the line remain fixed at esn.The receivingendvoltage would thus remain at l4l (this of a fixed sending_end {1ed voltage lVrl). L,ocal VAR compensation"ourr"lrrumes made can, in fact, be automatic by using the signal from the VAR meter installedat the receivingend of the line.
= rvlql
lvrt

+ty3
(1  4xeRltvrt )r,,
(s.85)
Fig. szr use of rocarvAR generator the roadbus at Trryotypes of vAR generators are employed in practicestatic type and rotating type. These are discussedbelow. Static VAR grenerator
crrmparisonof Eqs. (5.84)ancl(5.85) rcvcalsthut r^r. n.  n  .,),s tttt: YD vR vp' ,!,., receivingend voltageis r{r, butior bo= Oo; A:,"'
tvo < t4l
Thus a VAR demandlarger than Qf is met by a consequent fall in receivingfrom the specifiedvalue. similarly, if the vAR demandis less than 11d ":t!q," Q " *, it fo l l o w s th a t
tyRt tr4l >
Indeed, under light load conditions, the charging of the line may "upu.lance cause the VAR demand to become negative resuliing in the receivingend voltage exceeding the sendingendvoltage (this is the Ferranti effect already illustraredin Section5.6). In order to regulate the line voltage under varying demandsof VARs, the two methodsdiscussed below are employed.
sign in the quadratic solution is rejected because otherwise the solution would not match the specified receivingend voltage which is only slightly less than the sendingendvortage(the differenceis ress thai nqo). 'Negative
It is nothing but a bank of threephase static capacitors and/or inductors. With reference Fig. 5.28,if lV^l is in line kV, to and Xg is the per phase capacitive reac_ tance of the capacitor bank on an equiva_ Ient star basis, the expression for the VARs fed into the line can be derived as under.
lIc
Fig. 5.28
Static capacitor bank
r,=iH
'of
kA
course, sincct{tis spccificcl withina buntl, rury vrry withil a corresponding Ql band.
: iii,',1,1
powersystemAnarysis Modern J3
t Figure 5'29 shows a synchronous motor connectedto the receivingend bus bars and running at no load. since the motor o.u*, n.grigible real power from
3ry iQcGPhase) ( IF)
i3x #.HMVA
t v P
:j 3:^t1o1
(s.86)
in pr,use. ; th" ,yn.r,ronous reachnce $",.t":::H:,.,:o,^flusium"a have wtr,icrr i, !:T? T*ly ;rr,r*,n to ",
;r;:r*::':il;
QsQPhase)+
XC
MVAR
.C_ I _
(lvRl  IEGD/0. ,. ^ KA
If inductors are employed instead,vARs fed into the line are
J5;E
Q{3phase)='F''tuo*
XL
(s.87)
t l i e c = 3J v R3 4 oG i l _
= 3 W (  l Y R rr r c l ) J3 ( _jxsJl )
= jlVpt(tE6t _ IVRt)lXs MVA ec= tVRt(EGt _ tVRt)lXs MVAR (5.8g) It immediately follows from the above relationship that the machine feeds positive vARs into the line when rEGt> tv^r (werexcited case) and injects negarive VARs if lEGl continuously adjustableby adjusting machine which controls tE6l. rn contrastto statrcvAR generators, "Jtution the following observationsare made in respect of rotating VAR generators. ., (i) These can provide both positive and negativevARs which are continu_ ously adjustable. (ii) vAR inje*ion ar a given .. excirarionis resst s tlt""r""
Under heavy load conditions,when positive VARs are needed,capacitor banks are employed; while under light load conditions, when negative vARs are needed, inductor banks are switchedon. The following observationscan be rnade for.static vAR generators. (i) Capacitor and inductor banks can be switched on in steps.However, stepless(smooth) VAR control can now be achieved using SCR (Silicon Controlled Rectifier) circuitrv. (ii) Since Qg is proportionalto the squareof terminal voltage, for a given capacitor bank, their effectivenesstends to decreaseas the voltage sags under full load conditions. (iii) If the system voltage containsappreciableharmonics, rhe fifth being the most troublesome,the capacitorsmay be overloadedconsiderably. (iv) capacitors act as short circuit when switched on. (v) There is a possibility of series resonancewith the line incluctance particuia.riyat harmonic frequencies. Rotating VAR grenerator
It is nothing but a synchronousmotor running at noload and having excitation adjustableover a wide range. It feeds positive VARs into the line uncler overexcitedconditionsand f'eeds negativeVARs when underexcited. machine A thus running is called a synchronouscondenser. lvnl
il]ffi'ir'$ t:L::"::ch
Control
6r J":ff i::: {.::: smallerreductionin Qc comparedto the .0.r. or static capacitors. From rhe observarions ua" in ,.rp..t of ,tuti. vAR ibr_"" generators, seems it that rotating vAR g.n"ruro* would ""d;;;;;ing be preferred.However, economic considerations, install.tion and 'rai'r.rrun.. problerns limit their buses in the svstem *r,"i"' a taige u.oun,.or vAR
by Transformers
vol rage. IrzodeJeass and E  rv^rr As r e (r i,Lt
Fig. 5.29 Rotating VAR generation
The vAR injectionmethoddiscussed abovelacksthe flexibility and economy of voltage control by transformer tap changing. The transformer tap changing is obviously rimited to a narrow range of voltage control. If the vortage correctionneeded exceeds this range,tap changing used Is in conjunctionwith the VAR injection method. Receivingendvoltage which tends to sagowing to vARs demanded by the load, can be raised by simultaneousry .t,uogir; th. taps of sending_and receivingend transformers. Such tap changes niust"bemade ,on_road, and can be done either manua'y or automaiically,the oo*io.,o"r being ca'ed a Tap Changing Under Load ifCUf_l transformer.
,tig;'J
I
Power Analysis Modern System
and Performanceof Power Transmission Characteristics Lines fi.l7!J,
I
transformer line the Consider operationof a transmission with a tap changing at eachend as shown in Fig. 5.30. Let /5 and r^ be the fractions of the nominal transformationratios, i.e. the tap ratio/nominal ratio. For example, a transkV input has rr  I2lll = z=R+jx
Thus merely tap setting as a method of l.{Vl which is to be compensated. rvould give rise to excessivelylarge tap setting if voltage drop compensation compensationexceedscertain limits. Thus, if the tap setting dictated by Eq. setting range (usually not more than + 20Vo), it would be necessary to simultaneously inject VARs at the receivingend in order to maintain the desiredvoltage level. Compensation of Transmission Lines
1 : fsnl
at transformer eachend line Fig.5.30 Transmission withtap changing of With reference Fig. 5.30 let the impedances the transformerbe lumped to for tn Z alongwith the line impedance.To compepsate voltage in the line and transformers,let the transformer taps be set at off nominal values, rr and ro. With referenceto the circuit shown. we have trn rVs = t^ n rVo + IZ
(s.8e) (s.e0)
to From Eq. (5.75) the voltage drop ref'erred the high voltage side is given by
The perfonnance of long EHV AC transmissionsystemscan be improved by reactive compensationof series or shunt (parallel) type. Seriescapacitorsand shunt reactors are used to reduce artificially the series reactanceand shunt of susceptance lines and thus they act as the line compensators. Compensation of lines results in improving the system stability (Ch. 12) and voltage conffol, facilitating line energization in increasingthe efficiency of power transmission, and reducing temporaryand transient overvoltages. the Series compensationreduces. seriesimpedanceof the line which causes voltage drop and is the most important factor in finding the maximum power transmission capability of a line (Eq. (5.70)). A, C and D constants are functions of Z and therefore the also affected by change in the value of.Z, but these changes are small in comparison to the change in B as B = Z for the ., for the equivalent zr. nominal rr and equalsZ (sinh 4ll) The voltage drop AV due to series compensationis given by
tAvl = !I,!jIQsNow
lAVl  tsn, lTsl 
AV = 1Rcos S, + I(X,. X.) sin ,!,
(s.e4)
t on,rlVol ton2lVol
trnrlvrl tonrlvol +
RPR+xQR t* n r l V o l
(s.e1)
In order that the voltage on the HV side of the two transformersbe of the sameorder and the tap setting of eachtransformerbe the minimum, we choose (5.92) tstn= 1 we tn= llttin Eq. (5.91) and reorganising, obtain SubstitutinE .r( , RPR +xgo ) _ n2 lvRl
(s.93)
"['
"r"rWW ) ", W
the right hand side of Eq. (5.93) For complete voltage drop compensation, shouldbe unity. It is obvious from Fig. 5.30 that rr > 1 and tn 1 I for voltage drop Equation (5.90) indicatesthat /^ tends to increase*the voltage compensation.
This is so becausefn < 1 increasesthe line current / and hence voltage drop.
Here X, = capacitivereactanceof the seriescapacitor bank per phaseancl rcactance the line/phasc. practice,X. may be so of In X, is thc total incluctive selected that the factor (XL  X.) sin Q, becomes negative and equals (in magnitude) R cos /, so that AV becomes zero. The ratio X=IXL is called "compensation factor" and when expressedas a percentageis known as the "percentagecompensation". The extent of effect of compensationdependson the number, location and of circuit arrangements series capacitor and shunt reactor stations.While planning longdistance lines, besides the average degree of compensation required, it is requiredto find out the most appropriatelocation of the reactors and capacitor banks, the optimum connection scheme and the number of intermediate stations.For finding the operating conditions along the line, the ABCD constantsof the portions of line on eachside of the capacitorbank, and ABCD constants of the bank may be first found out and then equivalent can then be arrived at constantsof the seriescombination of linecapacitorline by using the formulae given in Appendix B. like UP, seriescompensation quite importantsincesuper is In India, in states thermal plants are located (east) several hundred kilometers from load centres (west) and large chunks of power must be transmitted over long distances. Seriescapacitors also help in balancing the voltage drop of two parallel lines.
rhOt" l
uodern Power SvstemAnalysis
Characteristics Performance PowerTransmission and of Lines
r81
is When seriescompensation used, there are chancesof sustainedovervoltage the ground at the series capacitor terminals. This overvoltage can be the to power limiting criterion at high degree of compensation.A spark gap with a high speed contactor is used to protect the capacitors under overvoltage trons. Under light load or noload conditions, charging current should be kept less than the rated fullload current of the line. The charging current is approxiof mately given by BrltA where B. is the total capacitive susceptance the line is and lVl is the rated voltageto neutral.If the total inductive susceptance Br due from line to neutral at to several inductors connected(shunt compensation) appropriateplaces along the line, then the charging current would be
kW at a leading power factor. At what value of P is the voltage regulatior zero when the power factor of the load is (a) 0.707, (b) 0.85? 5 . 2 A l o n g l i n e w i t h A = D = 0 . 9 l 1 . 5 " a n d B = 1 5 0 1 6 5 "C I h a s a t t h e o a r l end a transformerhaving a seriesimpedanceZr = 100 167" Q. The loar form of
and evaluatethese constants.
(BcI,he,= Br) lvl= BclVf[r
+)
(s.es)
Reduction of the charging current is by the factor of (1  Br lBc) and 81lBg is the shunt compensationfactor. Shunt compensationat noload also keeps the receiving end voltage within limits which would otherwise be quite high becauseof the Ferranti Effect. Thus reactors should be introduced as load is proper voltagecontrol' removed,f<lr As mentioned earlier, the shunt capacitorsare used acrossan inductive load so as to provide part'of the reactive VARs required by the load to keep the voltage within desirablelimits. Similarly, the shunt reactors are kept across above, to absorbsome capacitive loads or in light load conditions, as discussed of the leading VARs for achieving voltage control. Capacitors are connected and of eithcrclirectlyto a bus or throughtcrtiarywincling the main transformer and the voltage drop. are placed along the line to minimise losses Ii may be noted that for the same voltage boost, the reactive power capacity of a shunt capacitrtr is greater than that of a series capacitor. The shunt capacitor improves the pf of the load while the seriescapacitor has hardly any impact on the pf. Series capacitors are more effective for long lines for of irnprovement systemstability. of Thus, we seethat in both seriesand shunt compensation long transmission it is possible to transmit large amounts of power efficiently with a flat lines voltage profile. Proper type of compensation should be provided in proper quantity at appropriateplacesto achieve the desiredvoltage control. The reader is enceuragedto read the details about the Static Var Systems (SVS) in 'compensation',the reader References7, 8 and 16. For complete treatmenton may refer to ChaPter 15.
= 5.3 A threephase overheadline 200 km long has resistance 0.16 Qlkrn an conductordiameter of 2 cm with spacing4 m, 5 m and 6 m transpose( Find: (a) the ABCD constantsusing Eq. (5.28b), (b) the V,, 1,, pf,,'I when the line is delivering full load of 50 MW at 132 kV and 0.8 laggin pf, (c) efficiency of transmission, and (d) the receivingend voltag regulation. 5.4 A short 230 kV transmissionline with a reactance 18 O/phasesupplir of a load at 0.85 lagging power factor. For a line current of 1,000 A tt receiving and sendingendvoltages are to be maintainedat 230 k\ Calculate (a) rating of synchronous capacitor required, (b) the loa current, (c) the load MVA. Power drawn by the synchronouscapacitt .\ may be neglected. 'threephase 5.5 A 40 MVA generating station is connectedto a line havin Z = 300 175" Q Y = 0.0025 19tr U. The power at the generatingstation is 40 MVA at unity power factor a a voltage of L20 kV. There is a load of 10 MW at unity power factor a the mid point of the line. Calculatethe voltageand load at the distantenc of the line. Use nominalT circuit for the line. 5.6 The generalized circuit constantsof a transmissionline are A0.93+70.016 B=20+ jI40 The load at the receivingend is 60 MVA, 50 H4 0.8 power factor lagging. The voltage at the supply end is 22O kV. Calculate the load voltage.
ff]= [1 ",]l';l
5 . 7 Find the incident and reflectedcurrentsfor the line of Problem 5.3 at the
receivingendand 200 km from the receivingend. 5 . 8 If the line of Problem 5.6 is 200 km long and delivers50 MW at22OkY and0.8 power tactor lagging,determine sendingend the voltage,current, power factor and power. Compute the efficiency of transmission, impedance, wavelength, characteristic and velocity of propagation. 5 . 9 For Example 5.7 find the parametersof the equivalentn circuit for the line.
!.EMS PROB
voltage of 11 kV is applied to a line having R = 10 f) and 5 . 1 A threephase X = 12 ft p"t conductor.At the end of the line is a balanced load of P
,,1€? T il
uooern power system hnalysis
characteristics and Performance power Transmission of Lines
183
5.10 An interconnector cable having a reactanceof 6 O links generating stations1 and 2 as shown in Fig. 5.18a.The desiredvoltage profile is lVtl = lVzl = 22 kY. The loadsat the twobus bars are 40 MW at 0.8 lagging power factor and 20 MW at 0.6 lagging power factor, respectively. The torqueangle and the stationpower factors. line has following kV, 400 km transmission 5.11 A 50 Hz, threephase,275 parameters(per phase). = Resistance 0.035 Qlkm , In d u c ta n c e = 1 mF l /k m = Capacitance 0.01 p,Flkm If the line is supplied at 275 kV, determine the MVA rating of a shunt reactor having negligible lossesthat would be required to maintain 275 when the line is delivering noload.Use nominalkV at the receivingend, zr method. of of a feederhaving resistance 3 Q and areactance 10 f) 5.12 A,threephase suppliesa load of 2.0 MW at 0.85 lagging power factor. The receivingend voltage is maintained at 11 kV by means of a static condenser voltage and drawing2.1 MVAR from the line. Calculatethe sendingend power factor. What is the voltage regulation and efficiency of the feeder? of and reactance 5 and 20 Q, line has resistance overhead 5.13 A threephase is The load at the receivingend 30 MW, 0.85 power factor respectively. What will be the lagging at 33 kV. Find the voltage at the sendingend. equipmentinsertedat the receivingend kVAR rating of the compensating so as to maintain a voltageof 33 kV at each end?Find also the maximum load that can be transmitted. 5.I4 Constructa receivingendpower circle diagram for the line of Example 5.7. Locate the point coresponding to the load of 50 MW at 220 kV with 0.8 lagging power factor. Draw the circle passingthrough the load point. therefromlVrl. Also draw the sendingMeasurethe radiusand determine end circle and determine therefrorn the sendingend power and power factor. overheadline has resistanceand reactanceper phase of 5 5.15 A threephase and25 f), respectively.The load at the receivingendis 15 MW, 33 kV, 0.8 power factor lagging. Find the capacity of the compensation voltage of 33 equipmentneededto deliver this load with a sendingend
NCES REFERE
Books
l. Tron'smission Line ReferenceBook345 kV and Above, Electric Power Research Institute, Palo Alto calif, 1975. 2. Mccombe, J. and F.J. Haigh, overheadIinepractice, Macdonalel,London, 1966. 3. Stevenson, w.D., Elementsof Power Sy.stem Analysis,4thedn, McGrawHill. New York, 1982. 4. Arrillaga, J., High Vohage Direct Curuent Transmission,IF,E Power Engineering Series 6, Peter Peregrinus Ltd., London, 1983. 5. Kirnbark, E.w., Direct current Transmission,vol. 1, wiley, New york, 1971. 6. IJhlmann,E., Power Transmissionby Direct current, Springerverlag, BerlinHeidelberg,1975. 7. Miller, T.J.E., Reactive Power control in Electric systems,wiley, New york t982. 8. Mathur, R.M. (Ed.), Static Compensators for ReactivePower Control, Context Pub., Winnipeg, 1984. 9. Desphande, M.V., Electrical Power System Design, Tata McGrawHill. New Delhi, 1984. Papers 10. Dunlop, R.D., R. Gutman and D.p. Marchenko, "Analytical Development of Loadability Characteristics EHV and UHV Transmission for Lines", IEEE Trans. PAS,1979,98: 606. 11. "EHV Transmission", (special Issue), IEEE Trans,June 1966,No.6, pASg5. 12. Goodrich, R.D., "A Universal Power circle Diagram", AIEE Trans., 1951, 7o: 2042. 1 3 . Indulkar, c.s. Parmod Kumar and D.p. Kothari, "sensitivity Analysis of a MulticonductorTransmission Line", Proc. IEEE, March 19g2, 70: 299. 1 4 . Indulkar, c.s., Parmod Kumar and D.P.Kothari, "some studies on carrier Propagation overheadrransmission Lines", IEEE Trans.on pAS,No. 4, 19g3, in 102: 942. 1 5 . Bijwe, P.R., D.P. Kothari, J. Nanda and K.s. Lingamurthy, "optimal voltage Control Using ConstantSensitivityMatrix", Electric Power System Research,Oct. 1 9 8 6 ,3 : 1 9 5 . 1 6 . Kothari, D.P., et al. "Microprocessors controlled static var s5istems",proc. Int. conf. Modelling & Simulation,Gorakhpur,Dec. 1985, 2: 139.
kv.
Calculate the extra load of 0.8 lagging power factor which can be delivered with the compensatingequipment (of capacity as calculated above) installed, if the receivingend voltage is permitted to drop to 28 kV.
6.2 NETWORK MODEL FORMULATION
The load flow problem has,in fact, been alreadyintroduced in Chapter 5 with systent,i.e. a twobus problem (see Example 5.8)' the help of a f'undamental life power systemcomprising a large number Eor4 lq4d flqW llq4ypfufg4l of buses, it is necessaryto proceed systematicallyby first formulating the network model of the sYstern. by A power systemcomprisesseveralbuseswhich are interconnected rneans while the into a bus from generators' lines. Power is injected of transmission may be buseswith only generators loads are tappedfrom it. Of course,there and noloads, and there may be others with only loads and no generators. to may also be connected some buses.The surplus Further, VAR generators the buses is transportedvia transmissionlines to buses power at some of deficientin power. Figure6.1ashowsthe onelinediagramof a fourbussystem with generatorsand loads at each bus. To arrive at the network model of a a po*"i system, it is sufficiently accurateto represent short line by a series may be used impedanceand a long line by a nominalzr model. (equivalent7T may be neglectedwith a small loss foi very long lines). Often, line resistance in accuracybut a great deal of saving in computationtime. For systematic analysis, it is convenient to regard loads as negative Thus the and lump together generatorand load powersat the buses. generators into the bus is given by at the ith bus, the net complex power injected S;= Pi + jQi= (Pci Po)+ j(Qci Qo) is where the corrrplexpower suppliedby the generators Pot+ iQai Sci= the ccltnplexpower drawn by the loads is ancl Spi= Por+ iQoi The real anclreactivepowersinjected inttl thc itlt bus arc thcn Pi= Poi P^ i = 1,2, "'' fl (6'1)
6.1
INTRODUCTION
With the backgroundof the previous chapters,we are now ready to study the of operationalfeatures a compositepower system.Symmetricalsteadystateis' in fact, the most importantmode of operationof a power system.Three major problems encounteredin this mode of operation are listed beiow in their hierarchical order. 1. Load flow problem problem 2. Optimalload scheduling control Problem 3. Systems This chapter is devotedto the load flow problem, while the other two llow study in power systetn problems will be treateilin later chapters.Ltrad parlance is the steady state solution of the porver system network. The main information obtained from this study comprisesthe magnitudesand phase angles of load bus voltages,reactive powers at generatorbuses, real and reactine power flow on transmissionlines, other variablesbeing specified.This for information is essential the continuousmonitoring of the current state of the of the systernand for analyzing effectiveness alternativeplans for future system load demand. increased expansionto meet Before the advent of digital computers,the AC calculatingboard was the only meansof carrying out load flow studies.These studieswere, therefore, tedious and time consuming.With the availability of fast and large size digital inciuding load flow, can now be all computers, kinds of power systemstudies, carried out conveniently.In fact, some of the advancedlevel sophisticated studieswhich were almostimpossibleto carry out on the AC calculating board have now become possible.The AC calculating board has been rendered obsoletefor all practicalpurposes.
Qi= Qci Qoi Figure 6.lb showsthe network model of the samplepower systemprepared by on the above lines. The equivalentpower sourceat eachbus is represented sourceat the ith bus injects currentJr into a shadedcircle. The equivalentpower that the structureof a power systemis such that the bus. It may be observed are the Sources alwaysconnectedto a commonground node. all The networkmodel of Fig. 6.lb hasbeenredrawnin Fig. 6.lc afier lurnping at the shunt admittances the buses.Besidesthe groundnode, it has four other is (buses)at which the currentfrom the sources injected into the network. nocies nodes i and k is depictedby !ip= Jri'Further, the between The line admittance mutual admittancebetweenlines is assumedto be zero'
.Line by transformers are represented a seriesimpedance(or for accuraterepresentation by series and shunt impedances,i.e. invertedLnetwork).
apitying Kirchhoff s currentlaw (KCL) at nodesr,2,3 and 4,respectively, we getthe following fbur equations: Jt = Vrlro + (Vr  V) ln * (Vr  Vr) ),: Jz= Vz)zo+ (Vz Vr) ln + (Vz_ V) yzt + (Vz_ Vq)yzq s= vztgo+ (y:  v) ln + (Vt _ V) lzz + (Vt_ Vq) yzq Jq= Vayqo (Vq  Vr) lzc * (Vq _ V) yzq + Rearrangingand writing in matrix form, we get (yrc * tn *fn, .,  Yrz _ !* 0 ln (lzo*ln * hz * lzq, t23 J z + \z  lzt (y:o * yrg *rzt*yy,
( Y q a* y z q lzq * Yz+)
0  !z+ lu
(b) Equivalent circuit
Equation (6.3) can be recognized, be of the standardform to
(6.4)
Comparing Eqs. (6.3) and (6.4), we can write
Yrr=)ro* ln+ Yzz= lzo * ltz t Ytt= ):o* ln* Yu= lqo * lzq * Ytz= Yzt =  lni ln lzt + lzq lzz* lzq ly YZt = YZZ=  lZt Soz
(c) Power network of Fig. 6.1 (b) lumped and redrawn
Fig. 6.1 Samplefourbus system Y l t = Y r c =  l n i Y v = Y q t= )r+ = 0 Yzq=Y+z=lzqiYy= yqz=ly Each admittance y,, (i = r,2,3, 4) is calred the serf admittance (or driving point admittance) of node i and equals the algebraicsum of all the admittances terminaring on the node. Each ofidiagonal ierm v* (i, k = r, 2, 3, 4) is the mutual admittance (transfer admittance) between noim i and ft and equals the negativeof the sum of all admittances connected directly between thesenodes. Further, Yr* = Yri. using index notation, Eq. (6.a) can be written in compact form as
n
Sor v1
Ji= D
k=l (a) Onelinediagram Fig. 6.1 Sample four bus system
y p v p ii = r , 2 , . . . ,f l
(6.s)
(6.6
or, in matrix form = "Inus Isus Vnus
1.,.^L{"
urhanp
'vrrvrv
V J^^+^^ BUS \re,'ulsD
aL^
urc llralnx

^^
:
or
admittancematrix. The dimensionof the y"u5 matrix is (n x n) where n is the numberof buses'[The total numberof nocles e m = ii n +  includingthe ground (reference)node.l As seen above, yru, is a symmetric except when phase shifting transformersare involved, so that only n v + t ) terms are to be stored for an 2 ,?blrs system. Furthermore, yi* = 0 if buses i d k are not connected (e.9. YA = 0). since in a power network each bus is connectedonly to a few other buses(usually to two or three buses),the ys.y5 of a rargenetwork is very sparse'i'e' it has a large number of zeroelementi.itrougtr tf,i, prop"rty is not evident in a small system like the sample system under consideration,in a system containing hundreds of buses,the sparsity may be as high as 90vo. Tinney and associates[22] at Bonnevile Power Authority were the first to exploit the sparsityfeature of zsu5 in greatly reducing numericalcomputations in load flow studies and in minimizing the memory required as only nonzero terms needbe stored. Equation (6.6) can also be written in the form Vnus = Zausleus (6.7) where Zsu5 (bus impedancematrix) = fsLs fbr a network of fbur buses(fbur inclependent nodes)
DUS admlttance
and
is known
as bus
involved algorithms. Furthermore,the impedance matrix is a fuil matrix.** The bus impedancematrix, however, is most useful for short circuit studies as will be seenin Chapters10 and 11. Note: In the sample system of Fig. 6.1. the arDrrrary manner, although in more sophisticated studies of large power systems, it has been shown that certain ordering of nodes produfes faster convergence and solutions.Appendix c dealswith the topics of sparsityand optimal ordering.
6.3 FORMATION OF fsus BY SINGULAR TRANSFORMATION
Graph
(6.8)
rnus yields symmetric Zsus.The diagonal erements flmmetric of Zuu, are called driving point imped,ances the nodes,and th! of offdiagonalelementsare called transfer impedances of the nodes. Zsus need not be obtained by inverting rnus.while y"u, is a sparse matrix,Auris a full matrix. i.e., zero elementsof I'ru, become nonzeroin the .o,,.rp"o"niing Zsu, erements. It is t' be stressed here that yBus/zBus constitute models of the passive portions of the power network. Bus admittancematrix is often used in sorving road flow problem. It hal gained widespreadapplication owing to its simpticity of data preparationand the easewith which the bus admittancematrix can be formed ano moOfied for networkchangesaddition of lines,regulatingtransformers, etc. (seeExamples 6'2 and6'7)' of course, sparsityis one of its greatest advantages it heavily as reducescomputer'memory and time requirernents. In contrast to this, the
zss" can be referred to ground or slack bus. In the former case, it is usually necessary l'o creal'e at least one strong artificial tie to ground to avoid numerical difficulties when obtainingZsg5, because absence in of this, rr* is ilrconditionedor even singular' A large shunt admittanceinserted at the slack bus most simply achieves the desired result [20]. **The disadvantagesof the conventional impedancematrix may be overcome by making use of LU factors of the admittancematrix and bf employing compacr srorage scheme' Piecewise methods or tearing techniques (dialoptics) have recently been applied to overcome the disadvantages excessiv, of ,rorug. requirements ***For Ilg]. convenience, direction is so assigned as to coincide with the assu'red positive direction of the element current.
i90 l O
power Modern System Anatysis r @
Primitive Network
o
Fig. 6.2 Lineargraphof the circuit Fig.6.1c in here that each source and the shunt admittance connected across it are represented a single element. In fact, this combination represents by the most general network element and is describedunder the subheading "primitive Network". A connecteds'bgraph containing all the nodes of a graph but having no closedpathsis called a tree. The elementsof a tree are called branchesor ffee branches. The number of branches b that form a tree are given by b=m * 1 n (numberobuses) f (6.9) Thoseelementsof the graph that are not includedin the tree are called links (or link branches) and they form a subgraph,not necessarily connected,called
vr, = Er
E,
where E, and E" are the voltages of the element nodes r and s, respectively. It may be remembered here that for steady state AC performan.., ull element variables (vr* E, 8", irr,7r.,) are phasors and element parameters (zrr, .rr") ar€ complex numbers. The voltage relation for Fig. 6.4a can be written as
vrr* €rr= Zrri^
Similarly, the current relation for Fig. 6.4b is
ir, * jr, = lrrv^
(6.11)
(6.r2)
I
yr"=E E"
c
@
rn
v
+ o
7
\.e  Branch   Link e =9 m=5 b=m1=5'l=4=n l=eb=5 @
(a) Tree
(a) lmpedanceform
(b) Admittance form
Fig. 6.4 Representationa network of element The forms of Figs. 6.4a and, are equivalentwherein b the parallel sourcecurrenr in admittanceform is related to the seriesvoltage in impedanceform by
(b) Co tree
Jr,= Yrs€rs
Fig. 6.3 Treeand cotreeof the oriented connected graphof Fig.6.2 cotree.The number of links / of a connectedgraph with e elementsis I = e  b = e  m + l (6.10) Note that a tree (and therefbre, cotree) of a graph is not unique.
Also
!r.,= /7r,
A set of unconnectedelements is defined as a primitive network. The performance equations a primitive networkerre of given below: In impedance form
V+E=ZI
(6.13)
'f{W';l
ModernPowersvstemAnatvsis
v = AV,'J'
Load Fiow Stucjies
In admittance form I+J=W
<
i I
flH: ( 6. 16)
(6.14)
where the bus incidence matrix A is
Here V and.Iare the elementvoltage and currentvectorsrespectively,and rl and E are the sourcevectors. z and Y arereferred to as the primitive i admittance matrices,respectively.'These relatedas Z = Yr.If there is no are mutual coupling betweenelements, Z and Y are diagonal where the diagonal entries are the impedances/admittances the network elements and are of reciprocal. NeturorkVariables in Bus Frame of Reference
e
Linear network graph helps in the systematicassemblyof a network model. The main problem in deriving mathematical models for large and complex power networks is to selecta minimum or zero redundancy(linearly independent)set of current or voltage variableswhich is sufficient to give the information about all element voltages and currents. One set of such variables is the b tree voltages.*It may easily be seen by using topological reasoning that these variables constitute a nonredundantset. The knowledge of b tree voltages allows us to compute all element voltages and therefore, all bus currents assumingall element admittancesbeing known. Consider a tree graph shown in Fig. 6.3a where the ground node is chosen as the referencenode. This is the most appropriatetree choice for a power network. With this choice, the b tree branch voltagesbecomeidentical with the bus voltagesas the tree branchesare incident to the ground node. Bus Incidence Matrix For the specific systemof Fig. 6.3a, we obtain the following relations tetween the nine elementvoltagesand the four bus (i.e. tree branch) voltages V1, V2,V3 and Va.
Vut = Vt Vtz= Vz Vut = Vt Vuq= Vq Vts= VtVta= Vt Vn=VtVts= VqVtg= Vtor, in matrix form
Vn
I 2 3 4 ) 6 7
8 9 
1 0 0 0 0 0 l 0 l 0 
0 1 0 0 1 l l 1
0 0 1 0 _ 0 0 1 0
0 0 0 1 l links branches
(6.r7)
0 1 0 0 1
This matrix is rectangular and therefore singular. Its elementsa,oare found as per the following rules: aik = 1 if tth element is incident to and oriented away from the ftth node (bus) =   if tth element is incident to but oriented towards the ftth node = 0 if the ith element is not incident to the kth node SubstitutingEq. (6.16) into Eq. (6.14), we get I+J=yAVsu5 Premultiplying by Ar,
( 6 .l 8 )
(6.15)
v2 v2 v2 vr
ArI+ArJ=AryAV"u, (6.1e) F,achcomponentof the ndimensional vector ATI is the algebraicsum of the elementcurrentsleaving the nodes 7,2, ..., n. Therefore, the application* of the KCL must result in ArI =o (6.20) Similarly, eachcomponentof the vector ATJ recognizedas the algebraic sum of all source "unbe currentsinjected into nodesr,2, ...,n. Thesecomponents are thereforethe bus currents**. Hence we can write *For node l, Arl gives iro+ ir,  l:r= 0 verify this for another Thereadershoulci node. **For node l, AT"/gives j61 =currentinjectedinto bus 1 because otherelements connected bus t haveno sources. to
useful set of network variables are the / link (loop) currents which constitute a zero redundancyset of network variables [6].
*Another

i,il9:4iitl
.
Modern Power Svstem Analvsis
J     "   ' f   
I
Load Flow Studies
I ftt^
I
ArJ = Jsus (6.19)thenis simplified Equation to = ATYAV",,,, ,/eus
J . ^ ^ _ . r r ^ ' v r r r r g v l ' v a
(6.2r)
v rBUS  l r = A T vI,A
(6.22)
* yr:)  Jtz  )r: 0 ( yzo* yn * lzt * tzq,
' ,. .t'23
the samenodal currentequationas (6.6). The bus admittance matrix can then be obtainedfrom the singular transformationof the primitive Y, i.e. Yeus = ATYA (6.23) A computerprogramme can be developedto write the bus incidencematrix A from the interconnectiondata of the directed elementsof the power systetn. Standardmatrix transpose and multiplication subroutines can then be usedto compute Yu* fiorn Eq. (6.23).
0
 lzt
()'lo f .yr * lzt * Jy,
!'tn

}'l
_!zc
*"*'^"r Example 6.1  ,
I
The elements this matrix, of course,agreewith thosepreviouslycalculated of i n Eq. ( 6. 3)
Find the Y6u" using singular transformation for the system of Fig. 6.2. Solution )ro
!zo Vro
!qo
Figure 6.5 showsthe oneline diagramof a simple fourbussystem.Table 6.1 gives the line impedancesidentified by the buseson which theseterminate.The shunt admittance all the busesis assumed at negligible. (a) Find Yuu. assumingthat the line shown dotted is not connected. (b) What modifications needto be carriedout in Yuu, if the line shown dotted is connected.
ltn
Y
J Z )
v,
!n
Y.n
ln Using A fiom Eq. (6.17),we get
3
IJ1'
)ro 0 0 0 0 0
!n 0 .Yrr
0
0
0 lzo 0 J n 0 0  jzz lrz
!24
0 0 0
jqo ft+
Fig. 6.5 Sample system Example for 6.2
Table 6.1 Line, bus to bus
0 l y lzz 0
0
&pu 0.05 0.10 0.15 0.10 0.05
xpu
0.15 0.30 0.45 0.E0 0.15
YA=
0 0
lzq
t2 13 23 24 34
0
l
n
0
trrrr.F,b"iFl
'f/0.!:l t
Modern Power Svstem Anatvcis Table 6.2 Line
f.siffi
where V, is the voltage at the ith bus with respectto ground and ,/, is the source current injected into the bus. The load flow problem is handled more convenientlyby use of "/, ratherthan ,I,t. Therefore,taking the complex conjugateof Eq. (6.24), we have
r2
23 24 34
Gpu 2.0
B,pu  6.0
J.\,
1.0 2.0
 2.0  3.0  6.0
(6.25a)
Substitutingfor J, =
solution (a) From Table 6.1, Table 6.2 is obtained from which yuu, for the systemcan be written as
k:l
t
Y*Vr from Eq. (6.5), we can write A L
k:l
Pi jQi=Vf
Y i l , V p ii = l 2
n
(6.2sb)
Equating real and imaginary parts P; (real power) = Re ]Vi
I
t
(6.26a)
Qi (reactive power) =  Irn
In polar form V; = ll/,1si6t Y r* = lY'oleio'r Real and reactive powers can now be expressed as
n
(6.26b)
addedbetweenbuses 7 and 2. Ytz,n"* = Yrz,on  (2  j6) = Yzr.o"* Xrr,n"* = ytt,ora+ (2  j6) Y2z, = Yzz,ot,t (2  j6) + n"* Modified Y"u, is written below
P, (real power) = lvil D k:l (iv)
lvkl lYiklcos (0,0 6t  6'); + i = r , 2,
..., fl
(6.27)
Qi Qeactivepower)  lvil D
k:l
lvkl lYiklsin (e,1+ 6t  6,);
i=1,2,...,fl
(6.28)
6.4
LOAD FLOW PROBLEM
The complex power injected by the sourceinto the ith bus of a power sysrem
IS
(6.24)
Equations (6.27) and (6.28) represent2n power flow equationsat n buses of a power system (n real power flow equationsand n reactive power flow equations).Each bus is characterized four variables; P;, Qi, l7,l and 6i by resulting in a total of 4n variables. Equations(6.27) and (6.28) can be solved for 2n variables if the remaining 2n variables are specified. Practical allow a power systemanalystto fix a priori two variablesat considerations each bus. The solution for the remaining 2n bus variablesis rendered difficult by the fact that Eqs. (6.27) and (6.28) are nonlinear algebraicequations (bus voltagesare involved in product form and sine and cosine terms are present) explicit solutionis not possible. Solutioncan only be obtainedby and therefore, iterative numerical techniques.
:flg.,0iil
toaern power system Anatvsis state variables. These adjustable independent variables are called control parameters. Vector J can then be partitioned into a vector u of confrol parametersand a vector p of fixed parameters.
Dependingupon which two variablesare specified a priori, the buses are classifiedinto three categories. (I) PQ Bus
Qpi p, g ancl andegile specified). *tnn*i, ars fl/,t The and (po, and arc known
r rr rrrro rJyw vr LruD, Lrrc ucr puwers ri ano ai arc known
l."l pure load bus (no generating 6,.A 1:1t"t:i:.t"
Pp bus. (2) PV Bus/Generator
(6.30)
Control parameters may be voltage magnitudesat pV buses,real powers p,, etc. The vector p includes all the remaining parameters which are uncontrollable. For SLFE solution to have practical significance,all the state and control variables must lie within specified practical limits. These limits, which are dictated by specificationsof power systemhardwareand operating constraints, are describedbelow: (i) Voltage magnirudelV,l must satisfy the inequality lv,l^tn< lvil ( lv,l_* (6.31) The power system equipment is designed to operateat fixed voltages with allowable variations of t (5  l})Vo of the rated values. (ii) Certain of the 6,s (state variables) must satisfy the inequality constraint
f*iftyat the bus,i.e., pcr= eci=0) controlled Bus
is a
Bus/voltagre
At this type of bus P' and eo, are known a priori andrv,r p,(hence p6;) ' and. ' \ are specified.The unknowns are e, (hence eo,) and 6,. (3) Slack Bus/Swing Bus/Reference Bus
In a load flow study real and reactivepowers (i.e. complex power) cannot be fixed a priori at alr the buses as the net complex power flow into the network is not known in advance, systempower loss the being unknown till the load flow study is complete. is, therefore, It n.."rrury to have one bus (i.e. the slack bus) at which complex power is unspecified so that it supplies the difference in the total system load plus losses and the sum of the complex powersspecifiedat the rcrnainingbuses. the same By reasoningthe slack bus must be a generatorbus. The complex power allocatedto this bus is determined as part of the solution. In order that the variations in real and reactive powers of the slack bus during the iterative process be a snrall percentage of its generating capacity, the bus connecteclto the largest generating station is normally selectedas the slack bus. Further, for corivenience the slack bus is numbered bus 1. as Equations(6.27) and (6.28) are referred to as stutic load ftow equations (SLFE)' By transposingall the variables on one side. these equationscan be written in the vector form
16, 6ft1 l6i 6rln,o S (6.32) This constraint limits the maximum permissible power angle of transmission line connecting buses i and ft and is imposed by considerationsof system \ stability (see Chapter 12). (iii) owing to physical limitations of p and/ore generation sources,po, and Ne constrainedas follows: Qci
Pc,, ^in 1 Pc, S Pc,. ,n* Qci, ^rn 1 Qci S Qc,, ^u**
(6.33) (6.34)
It is, of course,obvious that the total generationof real and reactive power must equal the total load demandplus losses,i.e.
P D Po,=t Po,+ ,
l t
(6.35) (6.36)
D
i
Qci=l
i
Qot+Qr
f(x,y)o
where ,f = vector function of dimension 2n x = d_ependent state vector of dimension 2n or (2n unspecifiedvariables)
(6.2e)
where Ptand Qpare systemreal and reactivepower loss,respectively.Optimal sharing of active and reactive power generation between sources will be discussedin Chapter 7.
) = vector of independentvariablesof dimension 2n (2n independent variabreswhich are specified.i prrori)
ata PV bus can be maintained constant only if conrollable esource is available at the bus and the reactive generation required is within prescribed limits.
Voltage
ffi#ilfl,.l.i
Po*"t sv"ttt An"lvtit Mod".n
n n
I "Ihe load flow problem can now be fully defined as follows: Assume a certainnominal bus load configuration. Specify P6i+ iQci at all the pQbuses (this specifiesP, + iQi at thesebuses);specify Pcr (this specifies P,) and lV,l at all the PV buses.Also specify lVll and 6, (= 0) at the slack bus'
T.L,, r.,iolrlac nf thp rrerfnr u ere snecifie.d The 2n SLFE can now be solved
Wffi
Pr = .I Po, D
i:2
i:2
Po,;(Pr= 0). Equations (6.37)can be solvedexplicirly
(noniteratively) for 62,61, ..., d, which, when substitutedin Eq. (6.3g), yields simultaneously but can be solved sequentially[solution of Eq. (6.3g) follows immediately upon simurtaneous sorutionof Eq. (6.37)). Since the sorutionis noniterative and the dimension is retlucecr to (rrl) from Zrt, it is computationallyhighly economical.
(iteratively) to determine the values of the 2n vanables of the vector x comprising voltagesand anglesat the PQ buses,reactive powers and angles at fhepV busesand active and reactive powers at the slack bus. The next logical step is to comPuteline flows.' the So far we have presented, methods of assemblinga Yeusmatrix and load and have defined the load flow problern in its genpralform with flow equations that load flow definitions of various types of buses.It has been demonstrated being essentially nonlinear algebraic equations,have to be solved equations, through iterative numerical techniques. Section 6.5 presents some of the accuracyfor algorithmswhich are used for load flow solutionsof acceptable systemsof practicalsize. At the cost of solution accuracy, it is possible to linearize load flow equationsby making suitable assumptionsand approximationsso that f'astand eiplicit solutionsbecomepossible.Such techniqueshave value particularly for planning studies,where load flow solutions have to be carried out repeatedly but a high degreeof accuracyis not needed. An Approximate Load Flow Solution
madehaveo..ouiLii*i;:;ffi ,"(;:;'iLT,",",T:,:T:J"#il,1
consider the fourbus samplesystem of Fig. 6.6 wherein line reactances are indicatedin pu. Line resistances considerld negligible. are The magnitudeof all the four bus vortages specifiedto be r.0 pu. itJuu, are powersLe specified in the table below:
53= 2 +7O, r J
j0.15 jo.2
in and approximations the load flow Let us make the following assumptions ' (6.27) and (6.28). Eqs. of (shiintconductance overhead being smaii are rreglecie,C (i) Line resistances i.e. P7, the activepower loss of the systemis lines is alwaysnegligible),   90o ' zero. Thus in Eqs. (6'21) and (6.28) 1it = 90' and 1ii  6o) = (6r 6r). This is justified (ii) (6,  6r) is small (< r/6) so that sin (6, from considerationsof stability (see Chapter 72)' i.e. as otherthanthe slackbus (numbered bus 1) are PV buses, (iii) AII buses at voltage magnitucles all the busesincluding the slack bus are specified. Equations(6.27) and (6.28) then reduceto
Pi =lVil
iP,ts
'l.o lVzl=
2
.S.= I + i^ I Uz ,,. 
Fig. 6.6 Fourbuslossless samplesystem
Real demand
Reactive demand
Real generation
Reactive generatrcn
l \ l v k l l Y i k (6 i 6 r); i = 2,3, ..., n ,/r
k:1 n
n
(6.37)
1 2 3 4
Por = Poz = Poz = Poq =
1.0 7.0 2.0 2.0
Qot = Qoz = Qoz = Qoq =
0.5 0.4 1.0 7.0
Pcl ='Pcz = 4'0 Pct=o Pcq=o
061 (unspecified) Q62 (unspecified) O63 (unspecified) 06a (unspecified)
et= 'u,' E
r ) r v k r l y ic o s( 6 ,  6 u + r v , r 2y , , ir ;= r , 2 , . . . ,n ( 6 . 3 9 ) kr
Since lv,ls are specified, Eq. (6.37) representsa set of linear algebraic equationi in 6,sr,vhichare(n  l) in numberas 6, is specifiedat the slack bus (6, = 0). The nth equation correspondingto slack bus (n = l) is redundant as the reat power injected at this bus is now fully specified as
Figure 6.6 indicates bus injections fbr the data specified in the table. As bus voltages are specified,all the busesmust have controllable e sources. from the datarhar buses3 and 4 have onry sources. Further, e Il_r::t:: "_bviyus slnce ffie system is assumeci lossless, the real power generation at bus I is known a priori to be Pct = Por * Poz * pot * poo_ pcz = 2.0 pu Therefore,we have 7 unknownsinsteadof 2 x 4 = 8 unknowns.In the present problem the unknown stateand control variablesare {, e, 60, ect, ecz, ecz and Qc+.
'yili.\
I
Modern PowerSystem Analysis
Though the reali63sesare zero,the presenceof the reactive lossesrequires that the total reactive generationmust be more than the total reactive demand
(2.9 pu). From the datagiven, Yru5can be written as follows:
4
4
e r = D e o , D e p i
_ r.+r+ _ L,> = u.JJ4 pu (viii)
( l z l =X , 0 = 9 0 " )
Now, let us find the line flows. Equation (5.6g) can be written in the form lvi]lvkl Pik =  Pki sin ({  6o)
x,o
Using the above Y"u, and bus powers as shown in Fig. 6.6, approximate load flow Eqs. (6.37) are expressedas (all voltage magnitudesare equal to 1.0 pu)
where P* is the real power flow from bus j to bus k.
(iv) = 0, and solving (ii), (iii) and (iv), we Taking bus 1 as a referencebus, i.". 4 get O.0ll rad = 4.4I" 4. 4=0.074rad=4.23' (v)
( P z =3 = 5 ( 6  6 ) + 1 0( 6  4 ) + 6 . 6 6 76  6 q ) P t =  2  6 . 6 6( 6  4 ) + l 0 ( 4  6 ) 7 6 P + =  2  1 0 ( 4  4 ) + 6 . 6 6(7 q : 6 )
(ii) (iii )
pn =  pz, =+ sin(d, q) sin1.23" 0.492 r/ = pu \r 0.15 0.1:
Pt z =  Pzr=  1 sin ( 4  6) =  $n 4'41o =  0'385 L'L Pu ( ix) 0. 2 02 Pqt= + s i n ( {  6 o )= 1 0 s i n 5 . 1 1 o = 0 . g 9 1p u 0.1 Real power flows on other lines can be similarly calculated. For reactive power flow, Eq. (5.69) can be written in the general form ( l Z l =X , 0 = 9 0 o ) Pru= lvillvkl cos(,{  6o) ei* =W Xik Xik where Q* ir the reactive power flow from bus i to bus ft. "
6q=0.089rad=5.11' Substituting 6s in Eqs. (6.38), we have Qr =  5 cos 4.4I"  6.667 cos 4.23"  10 cos 5.11' + 21.667 Qz =  5 cos 4.41"  10 cos 8.64"  6.667 cos 9.52o+ 21.667 Qz =  6.667 cos 4.23o  10 cos 8.64' + 16.667 Qq =  l 0 c o s 5 .1 1 "  6 .6 6 7cos 9.52o+ 16.667
OI'
I Q p = _1, z r = + Q, cos(d,_ hl = 0.015 pu 0.2 i., @
2 + j0.ET 1 0.891+/O.04
Qr = 0'07Pu Qz = 0'22Pu Qz = 0'732Pu Q+ = 0.132Pu power generation the four busesare at Reactive = Qt + 0.5 = 0.57pu Qa Qcz = Qz + 0.4 = 0.62 ptr pu Qa = Qs + 1.0= 1.L32 = Q++ 1.0= 1.132pu Qc+
* l i't ',t,
, 3 8 5 7 0 .0 1 5 70. 018 0.492 +,p. 01S 0.385
(vi)
3 1 . s 0 2i 0 . 1 1 3
0.8917O.04
1.103 j0.0s2 + 1.103 p. 092
+ 1.502 10. 113 2 t, jD.4
n:n l
(vii),
4
Fig. 6.7 Loadflow solution the four'bus for system
= Q r c Q u=
#
#
( c o s d 1 ,O = 0 . 0 1p u 8
I
.2Oftril I
Analysis PowerSystem Modern
I en*.
carried out at the end of a completeiteration, the processis known as the Gauss iterative method. It is much slower to convergeand may sometimesfail to do so. Algorithm for Load Flow Solution
= *+' eA = e+t = +  1 .o, (6r 64) 0.04 pu 0.1 0.1 powerflows on otherlines can be similarlycalculated. Reactive at and Generations load demands all the busesand all the line flows are 6.5 GAUSS.SEIDEL METHOD
The GaussSeidel(GS) method is an iterative algorithm for solving a set of nonlinear algebraic equations.To start with, a solution vector is assumed, basedon guidancefrom practical experiencein a physical situation. One of the equationsis then used to obtain the revised value of a particular variable by substituting in it the present values of the remaining variables. The solution in vectoris immecliatelyupdatecl respectof this variable.The processis then repeatedfor all the variablesthereby completing one iteration. The iterative processis then repeatedtill the solution vector convergeswithin prescribed accuracy.The convergenceis quite sensitive to the starting values assumed' Fortunately,in a load flow study a starting vector closeto the final solution can be easily identified with previousexperience. To explain how the GS method is applied to obtain the load flow solution, let it be assumedthat all busesother than the slack bus are PB buses.We shall seelater that the methodcanbe easily adoptedto include PV buses as well. The slackbus voltage being specified,there are(n  1) bus voltages starting values These values are then updated of whose magnitudesand angles are assumed. process.During the courseof any iteration, the revised through an iterative voltage at the ith bus is obtained as follows: (6.3e) J, = (P,  jQ)lvi [from Eq. (6.25a)] From Eq. (6.5)
v , = * l L , _v , o v o l ,,' f
,[
I
(6.40)
L T:i I
fbr Substituting J, from Eq. (6.39) into (6.40)
v,= *l '':jq
 t Yit' l = 2'3""' n vr' Y" I vr* ilt I" k*t I I
t^ .
,
l
(6.4r)
Presentlywe shall continueto considerthe casewhere all busesother than the slack are PQ buses.The stepsof a computational algorithm'aregiven below: 1. With the load profile known at each bus (i.e. P^ and 0p; known), allocate* Po, and.Q5; to all generatingstations. While active and reactive generations allocatedto the slack bus, these are are permitted to vary during iterative computation.This is necessaryas voltagemagnitudeand angle are specifiedat this bus (only two variables can be specified at any bus). With this step,bus iniections(P, + jQ) are known at all busesother than the slack bus. 2. Assembly of bus' admittance matrix rsus: with the line and shunt admittance data storedin the computer, Yru, is assembled using the by rule for self and mutual admittances(Sec. 6.2). Alternatively yru, is assembled using Eq. (6.23) where input data are in the form of primitive matrix Y and singularconnectionmatrix A. 3. Iterative computationof bus voltages(V;; L = 2, 3,..., n): To start the iterationsa set of initial voltage values is assumed. Since, in a power systemthe voltagespreadis not too wide. it is normalpracticero use a flat voltage start,** i... initialiy ali voltagesare set equal to (r + 70) exceptthe voltageof the slack bus which is fixed. It shouldbe noted that (n  l) equations (6.41)in cornplexnumbers to be solvediteratively are for findin1 @  1) complex voltages V2,V3, ..., V,. If complex number operations Dot availablein a computer. zlre Eqs (6.41)can be converted rnto 2(n  1) equationsin real unknowns(ei,fr or lV,l, 5) by writing Vr = €i + ifi = lV,l ei6' (6.42) A significant reduction in the computer time can be achieved by performing in advance all the arithmetic operationsthat do not change with the iterations. Define
i = 2,3, ...,ft 'Active
The voltages substitutedin the right hand side of Eq. (6.41) are the most recently calcuiated (updated)values for the correspondingbuses. During each updated throughuse iterationvoltagesat buses i = 2,3, ..., n aresequentially of Eq. (6.41). Vr, the slack bus voltage being fixed is not required to be till updated.Iterations are repeated no bus voltage magnitudechangesby more than a prescribed value during an iteration. The computation processis then said to converge to a solution.
(6.43)
and reactive generation allocations are made on econorfc considerations discussedin Chapter 7. .*A flat voltage start means that to start the iteration set the voltage magnitudesand anglesat all busesother than the PV busesequal to (i + l0). The slack bus angle is conveniently taken as zero. The voltage magnitudesat the PV buses and slack bus are set equal to the specified values.
206 
Modernpower SvstemAnatvsis
I*rl.d
(6.44)
flows on the lines terminatingat the slack bus. Acceleration of convergence
17(r+l) 
vi'
Ai
r  t B , o v o t,( ,, * t , 
'S
n
(Vl")
ilt
k=irl
f
r , o v l , ,i = 2 , 3 , . . . , n (6.4s)
The iterative process is continued till the change in magnitude of bus voltage, lav.('*r), between two consecutiveiterations is less than a certain tolerancefor all bus voltages,i.e. _ _ IAV.G*r)l 1y.t+r) V,e)l < 6, ; i = 2, 3, ..., n (6.46) 4. computation of slack bus power: substitution of all bus voltages computedin step 3 along with V, in Eq. (6.25b) yields Sf = pr jey 5 . computtttion ofline .flows:This is thc last step in thc loaclflow analysis whereinthe power flows on the variouslines of the network are computed. Considerthe line connectingbuses i and k. The line and transformersat each end can be representedby a circuit with series admittance y* and two shunt admittances1l;roand )no as shown in Fig. 6.8.
B u si
l*io
up by the use of the acceleration factor. For the tth bus, the accelerated value of voltage at the (r + l)th iteration is given by y(r+r)(accelerated) V,Q'* = a(v.G+r) V,Ql, where a is a real number called the acceleration factor. A suitablevalue of a for any system can be obtained by trial load flow sfudies. A generally recommendedvalue is a = 1.6. A wrong choice of o. may indeed slow down convergenceor even cause the method to diverge. This concludesthe load flow analysisfor thJ case pe of busesonry. Algorithm Modification when pv Buses are arso present
At the PVbuses, p andrv]rarespecified and ancr dare the unknownsto be e detcnnincd.'l'hererirre, valuesof the e and d are to be updatedin every GS iteration through appropriate bus equations.This is accomplishedin the following stepsfor the ith pV bus. l. From Eq. (6.26b)
Busk
sm
ei =  Im j yr* D, y,ovof
L ft:l )
f
"
)
The revised value of ei is obtained from the above equation by substitutingmost updatedvaluesof voltageson rhe right hand side. In fact, for the (.r + 1)th iteration one can write from the above equation
I
g.(r+t) Ln =
Fig. 6.8 7i'representation line and transformers of a connected between buses two The current fed by bus i into the line can be expressedas Iit = Iitt + Iirc = (Vi  V) !it,+ V,y,oo The power fed into the line from bus i is. S* = Pir* jQiF Vi lfr= Vi(Vf  Vr\ yft+ V!,*y,f, Similarly, the power fed into the line from bus k is
a i y,rv,,(,*t)1y.r,1 y,kvk,,,I ],r,',) L r , D " ,u.ro, I f!, )
(6.47) (6.48)
2' The revisedvalueof {.is obtainedfrom Eq. (6.45) immediately following step 1. Thus 6Q+r)_ ay!+r)
= Ansle fei]l  i Bovo(,+r) B,ovo,,,l  D (6.s1) "t L ( t 1 ' ' ' ) * , r :r k:,+r J
where . Pi?t'rtt _ r^ (. ; +. t )__ 1 r = ,u The algorithm for pe buses remains unchanged.
Sri = Vk (V*k Vf) yfo+ VoVfyi,l,
(6.4e)
The power loss in the (t  t)th line is the sum of the power flows determined frorn Eqs. (6.48) and (6.49). Total transmission loss can be computed by sununingall the line llows (i.e. 5';a Sri fbr all i, /<). +
6.52)
l'fiffi
Y 1 Primitive matrix A matrix 2 Bus incidence ( , 3 S l a c kb u s v o l t a g e l Y 1 l 6 1 ) 4 R e a lb u s p o w e r sP i f o ri = 2 , 3 ' 4 , " " ' n bus powersQi,for I = m + 1" {fO buses) 5 Reactive Vf I for i = 2,....,m(PV buses) o 7 VoltagemagnitudelimitsIVrI min andIVi I max'lor '(J b n"u"iiu" pow"t limitsQi min and Q; max for PV buses
demand at any bus must be in the range e^rn  e^u*.If ai any stage during the computatibn,Q at anybusgoes outsidettreJe timits, it is fixed
At Q^in or Q^o, as the case may be. and the
dropped,i.e. the bus is now treatedlike a pe bus. Thus step 1 above branches out to step 3 below. 3. ff 9.('+r) a 0;,6;n, set e,Q*r)  er,r^n and treat bus f as a pe bus. compute 4(r+1) and y(r+t) from Eqs. (6.52) and (6.45),respectively.If = Otu..:,')Qt,^ , set O(r+l) ei,^*and treatbus i as a pebus. Compure 4.Q+r)6 y(r+l) from Eqs. (6.52) and (6.45), respecrively. Now all the computationalstepsare summarizedin the detailedflow chart of Fig. 6.9 which serves as a basis for the reader to write his own computerprogramme.It is assumed that out of n buses, first is slack the as usual,t hen2, 3, . . . , m ar ePVbusesand t he r em aining + l, . . . , f t m are PQ buses.
tg.
Yr,. r.i"g Eq. (6.2
= for i = 2!:'m Vio Make initialassumptions for I m + 1,"',n and O;0 = i = 1'2" A; Computethe parameters for i ^ t 1,"n .and B;p'for (exceptk = i ) from Eqs (6'43)and (6'44) .;; i: 1,2,'...,n count r = 0 Set iteration
Exampl 6.4 e
For the samplesystemof Fig. 6.5 the generators connected all the four are at buses,while loads are at buses2 and3. Values of real and reactivepowers are listed in Table 6.3. All busesother than the slack are pe typd. Assuming a flat voltage start,f)nd the voltagesand bus anglesat the three busesat the end of the first GS iteration. Solution
Table 6.3 Bus
Pp PU
I
lnput data Vt' Pu Retnarks Slack bus PQ bus Pp bus Pp bus
0u pu  0.2 0.5  0.1
I 2
J
r.u 10"
4
0. 5  1.0 0.3
The l"ur for the samplesystemhasbeencalculatedearlierin Example 6.2b (i.e. the dotted line is assumedto be connected). order to approachthe In accuracy of a digital computer, the computations given below have been performed on an electronic calculator.
T)r, ,i Dusvoltages at rne eno or tne llrst rteratron are calculated using Eq. (6.a5).
vtz=+{WYztvtY"v!
Fig.6.9FowchartforloadflowsoutionbytheGauss.Seide iterative method using YBUS
 'r^'iI
')

l 1 +      1 0 . 5 r i '0 . 2 1' . 0\ 4  z +' rj6 )  (  0 . v v6 6' j 2 ) _ ( _ t+ ' iO  ' ( / \ v'6 v+ r! fll Yzz t
I
fei.fii,',1
I
Analysis Power System Modern

Load Flow Studies
4.246 jrr.04 = pu + 1.019 70.046  jLr
3.666
1  
"
Ytt
* (vl)
 r.04 ( 1 + 13)  ( 0.666 j2) (1.019+ ;0.046) ( 2 * i6) + 
I
T#frL+re
 e2+ i6) + io) (to4
r) j3)(t. io)jl
_
1 Yzt
 ( 0.666 j2)(t + + 70)_ (_ 1+
= ''!',1"'.9?' = 1.028 70.087 pu 3.666 jrr
= t( +'z^?e;t  itt'lzsJ = , (r.rstz+ 70.033e) ( 3.666_jtt )= or 612 1.84658o 0.032 rad = .' . v) = 1. 04( cos 6) * j sin dj) = 1.04 (0.99948+ j0.0322) = L03946 + 70.03351
 r^rr\l
19 + 70.046) I  ( 2 + j6) (1.028 jo.o87)  )
( o , t vt =  l  j ' 1  = i Q  yY"v' y _ v t  v , , 0 I 'l=  . v .  Yszv; _ Yrovl
1 114tr
,l

2.99r j9.2s3pu 1.025 70.0093 3je
3.666in L ,f;;
[  r _ 7 0 . s ( t + i3) 1'04
io)f
 ( 0.666 j2) (1.03s46 + + 70.03351) z + _ (_
r.0317
In Example6.4,let bus 2 be a PV bus now with lV2l= 1.04pu. Once again assuming a flat voltage start, find Qz, 6., V3, V4 at the end of the first GS iteration. Given:0.2 < Q, < l. From Eq. (6.5), ii e", (Note fz= 0, i.e. Vl  I.04 + i0)
 Yo, vJ
= I il [0.3+io.r( I + i3) (1.03e4 + 70.0335) tf:;  ( 2 + j6)(r.03t7 yo.08e3z) _ I
_ _ 2.967r j8.9962 = 0 ' 9 9 8 5  7 O ' 0 0 3 1 3ig permissibte limits on ez (reactive powerinjection) are ",,il1;,;llii,""rrirhe
J
+ ( 0.666+ j2) + ( I + i3)l ) =  Im { 0.0693 j0.20791 0.2079 pu ' O" = 0'2079 Pu FromEq. (6.51)
n,: .,:,,,:.J, : !:: =_ A,,i". :;: ^ ^.,,?,: ^r!:, [J
ModernPowerSystem Analysis 0 .2 5 < Q z < i .0 p u It is clear, that other data remainingthe same,the calculatede2 (= 0.2079)is now less than the Qz, ^in.Hence e, it set equal to ez, _in, i.e. = 0'25 Pu Qz
a p uvrvrvrv, tf
I aar{ Lwqv
Ela.., I tvyy
Or.,r:^^ \)t,U(JlUli
6.6 NEWTONRAPHSON (NR) METHOD
The NewtonRaphson method is a powerful method of solving nonlinear algebraic equations.It works faster and is sure to convergein most cases as
longer remain fiied at I.04 pu. The value of V, ar the end of the first iteration is calculatedas follows. (Note VL = t + 70 bf virrue of a flat start.)
2t
vqLt
llv
vL = l( 'z
'r,
[
Yr,
(Y3)
zt   zr J  '" ^!?,  yztvt yztv? yrovl) ")
_[o.sjo.zs 3 . 6 66j r l L t  r ,
(* 2 + i6)1.04 ( 0.666 .i2) ( r + i 3 ) ] +  4'246 jlr,? = t.05.59 + io.o341 3 . 6 6 6 j t l
(
Considera set of n nonlinearalgebraicequations f i ( \ , x 2 , . . . , x n )= 0 ; i = I , 2 ,  . . , n
(6.s3)
vj = : l'::* .Y,,v,Y,,vJ " r,, (Y,')'
[ 3 . 6 6 6  j r[_t;u.t r Lr  ; o _ (_ r + j3)toa *
Assur.cinitialvulucs u'know's as *l: *'), ..., r"r. Let J.r(1, of Jg, ..., J_rl be the corrections, whichon beingadded the initial guess, to givelhe actual solution. Therefore = J t @ \ + f u t l , * u r + A x l , . . , , x 0 , + A x f ; 1 g ; i = 1 , 2 , . . . ,n ( 6 . 5 4 ) Expanding theseequations Taylor seriesaround initial guess, have in the we
f;(x01,xor,
,,r).[[*)' a*0, .(#)' Axt +
*[ l!t] oi l* n,rn, re,,ns (6.ss) order  o
\ux, ) l
/ ^ \0 I
 ( 0.666 j2X1.0s5e + + 70.0 341) ( 2 + j6) _ 2 . 8 tt 2  j n . 7 0 9 3.666 jrr 1.0347 70.0893 pu
]
= vL ;;W,#
= +,t"=#
 Yo,v,  v*v))  Yo,u
 ( I +,r3) (r'0s0e + i0'0341)
1
( Dr, ) \ / [ ,", ,/ x1, x2, ..., x, evaluat ed ( x! , *1, . . . , *0, ) . at Neglecting higher order terins we can write Eq. (6.55) in matrix fbrm
whcrefg)',( )'. '.f9 )' are the derivativesof It with respectto !f, ' d"r
 ( 2 + j6) (r.034i j0.08e3) I
/,.0630 j9 .42M _
I'rl
or in vector matrix form
f''' l
a'l'
Llxi
(6.56a)
3 je
1.0775 j0.0923 + pu
Ax:
2ll
I
Modern Power Svstem Analvsis

l(, + .f,A*, * 0
(6.56b)
the by ,f is k,o*n as the Jacobianmatrix (obtained differentiating function (6.56b)can be it to vector withrespect x and evaluating at r01. Equation f writtenas
:l_J
m
Approximatevaluesof corrections/r0 can be obtainedfrom Eq (6.57).These being a set of linear algebraic equations can be solved efficiently by triangularizationand back substitution(see Appendix C). Updatedvaluesof x are then
tl ="0 + AxI or, in general, for the (r + 1)th iteration
( 6. 63a)
(6.s8)
Iterationsare continuedtill Eq. (6.53) is satisfiedto any desiredaccuracy,i.e. (6.5e) v lJiG")l < e (a specifiedalue); i = 1,2, ..., n NR Algorithm for Load flow Solution
"(r+l)_r(r)+AxQ)
At that all busesarePQ buses. any PQbus the load flow solution First, assume rnust satisfythe following nonlinearalgebraicequations (6.60a) J'ip (lV, 6) = I'i (sPccificcl) Pi = 0 (6.60b) = Qi (specified)  Qi = o fiq (1v1, 6) for whereexpressions P, and Q, are given in Eqs. (6.21) and (6.28).For a trial to lV;1, the vector of residuals of Eq. (6.57) comesponds 6;, set of variables /0  Pi (calculated) = APi (6.61a) fip= Pi (specified)  Qi @alculated) AQi (6.61b) .fie= Qi(specified)
while the vector of corrections y'xo corresponds to
It is to be immediatelyobserved that the Jacobian elements corresponding the to ith bus residuals andmth bus correctionsare a 2 x 2 matrix enclosedin the box in Eq. (6.62a) where i and m Ne both PQ buses. Since at the slackbus (bus number l), Prand Qr are unspecifiedand lV,l, Q are fixed, there are no equationscorrespondingto Eq. (6.60) at this bus. Hence the slack bus does not enter the Jacobianin Eq. (6.62a). consider now the presenceof PV buses.If the ith bus is a pv bus, e, is unspecifiedso that there is no equation correspondingto Eq. (6.60b) for this bus. Therclbrc, the Jacobianeleurentsof the lth bus become a sinele row pertainingto AP,, i.e.
rthbusF l=
(6.62b)
alvil, a,i vector be written can corrections (6.51)for obtaining approximate the Equation
for the load flow case as mth bus
If the mth brrs also a PV bus, lVrl becomesfixed so that Alvml = 0. We is can now wnte
fthbusl
I I
,lPi aQi
(6.62a)
46^ AlV^
i t h b u s lp , l = l n T:tT' LJ t  rH''l 
I l
(6.62c)
mth bus
I ous jmtn
2L6 
rtrodernPower SystemAnalysis
l ztz
t Ninr= N^i = 0 l,^=J^;=O (6.66)
Also if the ith bus is a PQ bus while the mth bus is a PV bus, we can then write
(6.62d)
It is convenientfor numerical solution to normalize the voltage corrections
corresponding to a parricular vecror of variables tqlv2lq64lval6lr, the vector of residuars[aP2 aez ah ap4 ae4 Apr], and the Jacobian(6 x 6 in this example) are computed. Equation rc.an is then solved by triangularization and back substitution procedure to obtain the vector of
alv^l
lv^l elements become Jacobian asa consequence which,the corresponding of
correctionsoo,# oo1464 /,6rl . corrections thenaddedto are J a+ I lv2l lVor
L I
f
nvt
Atrl
1r
updatethe vector of variables. 2(Pa)
3(PVl
(6.63b)
Expressionsfor elementsof the Jacobian (in normalized form) of the load flow Eqs. (6.60a and b) are derived in Appendix D and are given below: Case I
m t l H,^= Li^= a^fi  b*et Nr*=Jirr= a.er+ bJt 6.64)
Fig. 6.10 Sample fivebus network '* Bus No. 2 3
I l 2
Y
I
Yi^= G* + jB,* Vi= e, + jf,
4
Hzn Nza
5
Hzz Jzz Hsz
Hqz Jqz
Nzz
Hzs
(a* i ib)
= (Gi + jBi*) @* + jk\
case2 I,,==t n,  Biirvirz
Mii= Pi + G,,lV,lz J ti = Pi  Giilvil' Li;= QiBiilvilz
IA A<\
\\J ' \'J,'
Lzz
Jzs
Jz+
Lz+ Hss
o
f,z
c0
J
Nsz Hss Nn
Lqz
Haa
Naz.
H+s J+s Hss
(6.67)
Jcc
Lu
Nsq
l]qll
lalv4ll
[:_]
Hss H5a
An important observation can be made in respect of the Jacobian by examination of the Y"u5 matrix. If buses i and m are not connected, Yi^= 0 (Gi^ = Bi^  0). Hence from Eqs. (6.63) and (6.64), we can write
t Jacobian
(Evaluatedat trial values of variables)
Corrections in variables
i_it_l t
TLLV
,re
I I
rrrrodcrn Power Svstem Analvsis
Iterative Algorithm
Omitting programmingdetails, the iterative algorithm fbr the solution of the load flow problem by the NR method is as follows: L W'ith voittge anclangle (usually f'= €I) at s at atl PQ busesand d at all PV buses' In the absence of anY tlther information flat voltage start is recommended' 2. Compute AP,(for PV and PB buses)and AQ,(for aII PQ buses)from Eqs. (6.60a and b). If all the valuesare lessthan the prescribedtoletance, stop the iterations,calculate P, and Q, and print the entire solution including line flows. 3. If the convergencecriterion is not satisfied, evaluate elements of the usingEqs. (6.64) and (6.65). Jacobian Eq. (6.67)for correctionsof voltageangles and magnitudes' 4. Solve 5. Update voltage angles and magnitudes by adding the corresponding to changes the previousvaluesand returnto step 2' at Note: 1. In step 2, if there are limits on the controllable B sources PV each time and if it violates the limits, it is made equal to buses,Q is computed the limiting value and the correspondingPV bus is made a PQ bus in that computation, Q does come within the prescribed iteration. If in the subsequent the bus is switchedback to a PV bus. limits, 2. If there are limits on the voltage of a PQ bus and if any of these limits is violated, the correspondingPp bus is made a PV bus in that iteration with voltage fixed at the limiting value.
I i:ig,: Find the load flow solution using the NR method.Use a toleranceof 0.01 for powermismatch. Solution Using the nominalrcmodel for transmission lines, I"u, for the given systemis obtained as follows:
eacn ltne
Lcad Ftor; Studies
I 1,sen."= nes ^  z . g 4 r  j 1 1 . 7 & r 2 . r 3 l  7 5 . 9 6 " o. oz+jo. og Each offdiagonalterm =  2.94I + jll.764 Each self term = 2l(2.941  j11.764)+ j0.011 = 5.882 j23.528 = 24.23 l75.95"
0 +,O 2
3
1 . 0 4O " l 1 . 5+ 7 O . 6
Fig. 6.11 Threebus system Example for 6.6
systemof Fig. 6.11. Eachof the three lines has a series Considerthe threebus pu impedanceof 0.02 + 1O.OS and a total shunt admittance of 70'02 pu' The quantitiesat the _b"!91j9!qulated below: specifiecl
RealloadReactiveloadRealpowerReactiveVoltage
To startiterationchoose 4= we get
t +70 and ,4 = 0. From Eqs. (6.27) and (6.2g),
Bus
demand PD
demand Qo
generation PG Unspecified
specification power generation O6 UnspecifiedV'=1.94 * ;g (slack bus)
Pz = lVzllvl lY2l cos(0r1+ 6r $) + lVrlzlyrrl cos0zz lvzl l\l + l Y r r lc o s( Q z t +q  6 ) Pt = lVl lvl l\rl cos(dr, + 6r 6r)+ lVrl lv2ll\zl x cos(ez + h  6r)+ lV,rlz lyrrl cos 0r, Q z =  i v z i i v l t y z l s i n( g r + d t  h )  l v z l z y 2 2xl s n 4 z _ l v 2 l l  6) lv3l lY.,.lsin (0r, + bz Substituting givenandassumed values different of quantities, getthevalues we trf Powers rts = PB. 0.23pu
2 .0 0 .0
l 5
1.0 0 .0 0 .6
0.5 0.0
1.0
Qct=?
Unspecified (PO hus) lvll= 1.04 (PV bus)
nt p rc C o n tro l l a b l c a c ti v o o w o r s o urccl s avai l abl c btts3 w i th the ctl nstrai nt 1.5Pu 0<Qct S
O}nr'l ttv.l
iladan rvruuErrr
Darrrar r VYYEr
Crralam \)yolgltr
Anahraia nrrqryoro
t
Load Ftow Studies
ti^i.
pu P3= 0.12
Q o r = 0 ' 9 6 P u
Sz=0.5+j1.00 Sr=1.5j0.15
Transmission loss = 0.031 Line flows

i zzr 
Power residualsas per Eq. (6.61) are  rt (calcu
( 0.23)= 0.73 = Aror 1.5  (0.12)  !.62
The following matrix shows the real part of line flows
tQT= 1  ( o'e6) t'e6
The changesin variablesat the end of the first iteration are obtainedas follows:
0P, 06, 0P, 06, aQ,
'av 06, 06:'
0P, 061 aP, 061, }Qz
0P, alv2l 7Pu 0lv2l aQ,
The following matrix shows the imaginary partof line flows
I
0.0 0.6s4697  E00 I0.r8422eE00 0.673847 L0.826213^800 E00 0.0 J
I
oo
0.1913r2E00 0.839861E'00l
I
o.o
0.5994&E00 _0.r9178zE00]
0.0 0.39604s800 I l0.60s274800 0.37sr6s800 L0.224742E00 0.0 I
Rectangular PowerMismatch Version
Jacobian elements can be evaluated by differentiating the expressions given above fr>r Pr, Py Qz with respectto 6r, d1 and lVrl and substituting the given and assumed values at the start of iteraticln. The chanses in variables are obtained as
I t d j^l rl I 2 4 . 4^ i^ ^   .
I Aai l:l
t2.23
t/..25
^ ^  l /.4.e) J.u)l
s . 6 4  1  r [ 0 . i 3f 0 . 0 2 3  1 [


l  .  .  t .o^ z l : l  u .^u^o ) 4 1 
L [ n r v , r '  ] u u
) 3 . 0 s z z . s 4L r . q o lI o o a r l
l a ) l I a i  l l  ^ 4 I t  0 .[1 0 0 2I. 3 0 0 2 3 . 1 [ a ]l : l I l * l ^ 4 l : l o l * l  0 0 6 s 4 1 : l  0 0 6 s 4 1 I L'i I Itv,t'j L'y,roJ lnrv.,r'.1 I oosoJ r.08eJ
We can now calculatefusing Eq. (6.28)] Qtt= 0.a671 Q o \= Q\ + Q rt = 0 .4677+ 0.6 = 1.0677 which is within limits. If' the sanre problem is solved using a digital computer, the solution in converges three iterations.The final resultsare given below: Vz= 1.081 lVt = I.M l0.024 rad 0.0655 rad
This versionusese, ancl ,the rear ancr imaginary parts .f of the v.ltages resllectivcly, variables''fhe number of equations as and variables greater is than thart Eq. (6.6r), by the nurnber<ttpi buses. tirr Sinceat pv buses e,and .f; canvary but ,,' +.f,'  lviP, a voltagemagnitude squarecl misnlatch equation is required tbr each PV bus. with sparsityprogramming,this increasein order is of hardly any significance.Indeed eachiterationis rnarginally faster than for Eq' (6'67) sincethereare no timeconsunring sine anclcosineterms.It nray, however,be noted that even the polar versionavoids these as far as possible using rectangular arirhemetic in constructing Eqs (6.64) 9f and (6.65). However'thc rcct:tttgtrlltr vcrsionsccnlsto bc slighiiy t.r, rcliablebut faster in convergence than the polar version. The total numberof nonlinearpower flow equations considered this case in arc fixed iurd cqual 2 (trl). Theselbllow fr.o'i Eqs. (6.26a) and (6.26b)and are P, (specifi"d)I{
k:l
e,(epG11,f,B,t) + fihrG*
t epB,o} 0 : ( 6. 68)
i = l , 2 , . . . ,n i = slack(s) ) B; (specifi"d)
n rt. k=l
= * lf,koG,o  ftB,t1 e;(.f*Gi* eoB,oy]}g (lbr each PQ bus) (6.69)
Q u =  0 . I 5 + 0 . 6 = 0 . 4 5( w i t h i nl i n r i t s ) * S r = 1 .0 3 1 j ( 0 .7 91)
222 
I
Modern Porgl Jyllem Anatysis l(l Load Flow Studies necessary motivation in developing the decoupledload flow (DLF) method, in which P6 and QV problems are solveclseparately. Decoupled Newton Methods
(specified)2 @i2+ f,') = 0 (for each PV bus) (6.70) Using the NR method,the linearisedequationsin the rft iteration of iterative processcan be written as Jr J2
J4 J6
AQ AIVP
F o ri * j
Juj=
J3 Js
 Jqij = Gij ei fi Bijei+ Gilfi 6.72)
Jzij= Jt,j=Jsij= Jo,j = O For i=,r
In any conventionalNewton method, half of the elements of the Jacobean matrix representthe weak coupling referred to above, and therefore may be ignored' Any such approximationreducesthe true quadratic convergence to geometric one, butthere are compensating computationalbenents.i large number of decoupled algorithms have been dlveloped in the literature. However, only the most popular decoupled Newton veision is presented here
t1el.
t;^"'::: .::::!tlr,,, !'u,u' J:, l,)i
Jsii= 2"i J6ii = 2fi (6.73)
In Eq. (6.67), the elementsto be neglectedare submatrices [1v] and [,/]. The resulting decoupledlinear Newton equationsbecome
tApl = lHl lA6l
Ltvt J whereit can be shownthat gU = Lij = lvil lvjl (CUsin 6u B,i cos {r) i * j
H i i =  8 , , l V i P  Qi
( 6. 76) ( 6. 77)
I^Qt tLtl4!1
(6.78)
a, and b, are the componentsof the current flouring into node i, i.e.,
(Eq. (6.6s)) (Eq. (6.6s))
(6.7e)
(6.80)
a,+ jb,=
jBi)@r+jf*) frCo *
k:l
(6.74)
Lii=  8,, lV,lz + Qi
Steps in solution procedureare similar to the polar coordinatescase,except that the initial estimatesof real and imaginary parts of the voltages at the PQ busesare madeand the correctionsrequiredare obtainedin each iteration using
(6.7s)
The correctionsare then applied to e andf and the calculationsare repeatedtill convergenceis achieved.A detailed flow chart describing the procedure for load flow analysis using NR methodis given in Fig. 6.12.
O. I
F E F F A A ' ? h I E h
Equations(6.76) and (6.77) can be constructed and solved simultaneously with each other at each iteration, updating the [H] and [r] matrices in each iteration using Eqs (6.78) to (6.80). A better opproo.h is to conduct each iteration by frst solving Eq. (6.76) for 44 *o use the updared 6 in constructingand then solving Eq. (6.77) for Alvl. This will result in faster convcl'geuce than in the sirnultaneous mode. The main advantage the DecoupredLoad Flow (DLF) as of comparedto the NR method is its reducedmemory requirements storing in the Jacotean.There is not much of an advantage from the point of view of speedsincethe time per iteration of the DLF is almost the same as that of NR method and it always takes more number of iterations to convergebecauseof the approximation. Fast Decoupled Load Flow (FDLF)
L'aUL,UTL1L'
' A A
L('AL'
.F LL'VV
IVI.EI.tsI(,L'Ii
An important characteristic any practicalelectric power transmissionsystem of operatingin steadystateis the strong interdependence betweenreal powers and bus voltagesanglesand betweenreactivepowersand voltagemagnitudes. This interestingproperty of weak coupling betweenP6 and QV variable.s gave the
Furtherphysically justifiablesimplifications may be carriedout to achieve some speed advantagewithout much loss in accuracy of solution using the DLF model described in the previous subsection.This effort culminated in the developmenr the Fast DecoupledLoad Frow (FDLF) merhod of by B. stott in 1974l2ll. The assumptions which are valid in normal power ryr,"n1operation are made as follows:
'*i
::,:,t
I I
vodern Power SvstemAnalvsis
I zasWith theseassumptions, entriesof the [1{ and [L] submatrices the will become considerablysimplified and are given by
count iteration I Aduance I r= r+1
and
cmpute anol printlineflows,l powerloss, I voltages, I elci .]
=''i,,,11,,u" ? =i: :: ,::
= LAPI ltyjt lvjl Bfi AA tAQl= [%t tvjtB,,ij, [#]
(6.82) (6.83)
Matrices [/fl and [L] are squiue mafrics with dimension (npe + npy) and nrn respectively. Equations (6.76) and (6.77) can now be written as
DetermineJ1 .'. comPuteAefr) 2n6 Lfiv)
(6.84) (6.8s)
Are all
maxA within
tolerance
2
max Determine in change power max APr,AO
where 8t,,, B(are elementsof [ B] matrix. Further decouplingand logical simplificationof the FDLF algorithmis achieved by: 1. Omitting from [B/] the representationof those network elements that predominantly affect reactive power flows, i.e., shunt reactancesand transformer offnominal inphase taps; 2. Neglecting from [B//] the angle shifting effects of phaseshifters; 3. Dividing eachof the Eqs. (6.84)and (6.85) by lv,l and settinglVrl = I pu in the equations; 4. Ignoring seriesresistance calculatingthe elementsof tdll which then in becomesthe dc approximation power flow matrix. With the above modifications, the resultant simplified FDLF equations become
I last node
')
Compute
el,),1v112
ls \ Compute = scnd12l%tl2 LlVi,12 1V,
'
gt'r ioi)
IAP lyl I = [B'] [L6] tAQl tr4 I  LB" I t^lytl
(6.86) (6.87)
F i g .6 . 1 2 cos 6, : s i n{ r : 0 G,, sin 6,, < B4;
and
1;
In Eqs. (6.86) and (6.87), both [B/] and lBttf are real, sparse and have the structuresof [I{ and [L], respectively. Since they contain only admittances, they are constantand need to be invertedonly once at the beginning of the study. If phase shifters are not present, both [B'l and lB"] are always symmetrical, and their constant sparseupper triangular factors are calculated and stored only once at the beginning of the solution. Equations(6.86) and (6.87) are solvedalternatively.always employing the most recent voltage values. One iteration implies one solution for [4fl, to update[d] and thenone solutionfor IA lVl] to update[Vl] to be called ldand
lV iterafion Senarafe. "r*^*'c  l n v' e ^ro e^n c e  r^ ^  . s tcetc are annlied fhr the rcql qnA rcqotlvc
power mismatches follows: as (6.81) max [APf I €pi max lAQl S eo where eo and €e are the tolerances. A flow chart giving FDLF algorithmis presented Fig.6.13. in ( 6. 88)
Qi < Biilvilz
,bZCI
powerSvstem Modern Analvsis
LoadFtowStudies
(Start )
I ##
Considerthe threebussystemof Example 6.6. Use (a) DecoupledNR method
Read LF data and form Ysr.
(a) Decoupled NR method: Equationsto be solved are (see Eqs. (6.76) and (6.77)). Substitutingrelevant valuesin Eqs. (6.78)  (6.80) we get H z z = 0 ' 9 6 + 2 3 .5 0 8 24.47 Hzt= Hzz= 1.04 ( Brz) =  1.04 x 11.764=  12.23 Htt= Qt Bn0.0q2 = [ Bsr t\P  42 tv3t Bzzt\P]  Bn (r.0a1.2 =  f 1 .7 6 4x (1 .0 4)2 1t.764 x 1.04+ (l .oq2 x2x 2 3 . 5 0 82 5 . 8 9
Lzz= Qz Bzz= t + 23.508 24.508
_r2.2311\6t\1 _ [ 0.731 [ 24.47 lt.oz) L n.23 zs.ssJ[64trl
L J SolvingEq. (i) we get L6;') =  0.0082 0.0401  0.002
(i)
Solvefor A6f usingEq. ., ( 6 . 8 6i)= 1 , 2 , . .n , i * s
Calculate 6i.1 = 6[+ 66[
tLQzl lz4.srl 't;(i)l ' f+uJ"'l
(ii)
fori=1,2,...,n,is (6.60a) 1. = PV bus Calculatesli l slack bus power i er and all line flov flows and print
Aa{tr =  0.018 0.08  0.062 Qz=  lvzl lvtl lYzl sin(0r, + 6r  6r)  lvzlzV2zlsin 82 lvzl lhl llrrl sin (9zt+ 6,  q) =  1.04x 12.13 (104.04 0  0.115) 24.23 sin + sin(75.95) 1.04x 12.13sin (104.04 .115 3.55") + 12.24+23.50512.39 Qz=  l'125
ttv2Arl 1
X
 '  
r 
/
1
\
r.rLJ)
1a<\
=
^
L.ILJ
1?rE
rcEq.l .,n, I
:us l
<'
I

re
\
fr=
1 
Substituting in Eq. (ii)
; Afr/,R
[alY"(')ll Lz.rz5r[24.5r]
Ltu,a, l
e+
Flg.6. 13
;'..ng I
Analysis System Power Modern AN 9 l= 0.086 = pu lvftt  lvlo)t+atv'tt 1.086

lui:.
completean iteration. This is becauseof the sparsityof the network matrix and the simplicity of the solution techniques. Consequently, this method requires less time per iteration, With the NR method, the elementsof the Jacobianare to be computedin eachiteration,so the time is considerably longer. For typical large systems,the time per iteration in the NR rnethodis roughly equivalent to 7 times that of the GS method [20]. The time per iteration in both these methods increasesalmost directly as the number of buses of the network. The rate of convergenceof the GS method is slow (linear convergence requiring a considerablygreaternumber of iterations to obtain characteristic), a solution than the NR methodwhich hasquadratic convergence characteristics and is the best among all methods from the standpointof convergence.In addition, the number of iterations for the GS method increasesdirectly as the number of busesof the network, whereasthe number of iterations for the NR method remains practically constant, independentof system size. The NR method needs 3 to 5 iterations to reach an acceptablesolution for a large system. In the GS method and other methods,convergence affected by the is choice of slack bus and the presenceof seriescapacitor,but the sensitiviry of the NR method is minimal to these factors which causepoor convergence. Therefore, for large systemsthe NR method is faster, more accurate and more reliable than the GS method or any other known method. In fact, it works for any size and kind of pro6lem and is able to solve a wider variety of illconditioned problems t23). Its programming logic is considerably more complex and it has the disadvantage requiring a large computei memory even of when a compact storage scheme is used for the Jacobian and admittance matrices. In fact, it can be made even faster by adopting the scheme of optimally renumberedbuses.The method is probably best suited for optimal load flow studies(Chapter7) because its high accuracy of which is restricted only by roundoff errors. The chief advantage the GS method is the easeof programming and most of efficient utilization of core menrory.It is, however,restrictedin use of small size system becauseof its doubtful convergence and longer time needed for solution of large power networks. Thus the NR method is decidecllymore suitablethan the GS method for all but very small systems. For FDLF, the convergence geometric,two to five iterations are normally is required for practical accuracies,and it is more reliable than the formal NR method. This is due to the fact that the elementsof [81 and [Btt] are fixed of approximationto the tangents the defining functions LP/lVl and L,QAVl, and are not sensitiveto any 'humps' in the ciefiningiunctions. fi LP/lVl and A^QIIV are calculatedefficiently, then the speedfor iterations I of the FDLF is nearly five times that of the formal NR or about twothirds that of the GS method. Storagerequirementsare around 60 percent of the formal NR, but slightly more than the decoupledNR method.
Q(rl canbe similarly calculatedusing Eq. (6.28).
The matrix equationsfor the solution of load flow by FDLF method are [see Eqs. (6.86) and (6.87)l
8,, )Lz4" l
and
Brr1l af',f
(iii)
lffil=
o.tz I t.oz I  I
 
rBzzt tatrt')tl
fil6411oqrl 23.508.1la4" l
(iv)
 1.04 I L: 1'5571
t l 
(v)
Solving Eq. (v) we get 46;r)   0.003
A6t'' =  0.068
lz.rzsl [23.508]tatvitl alvtl= 0.09
p l v t t =1 . 0 9 u
Now Q3 can be calculated. These values are used to compute bus power mismatchesfor the next 'LAPAVll and iteration. Using the values of lAQAl\l the above equationsare solved alternatively, using the most recent values, till the solution converges within the specifiedlimits. 6.8 COMPARISON OF IOAD FLOW METHODS
6') 
rad; {tl 0.003
0.068 rad
In this section,GS and NR methodsare compared when both use liu5 as the network model. It is experienced that the GS method works well when whereasNR requiresmore memory programmedusing rectangularcoordinates, used. Hence, polar coordinatesare preferred when rectangularcoordinatesare for the NR method.
ilai:i;'l
I Changes in system configurations can be easily taken into account and though adjustedsolutionstake many more iteration.s, each one of them takes less time and hencethe overall solution time is still low. The FDLF can be employedin optimization studiesand is specially used for studies, as in contingencyevaluation for system security assessment and enhancementanalysis. Note: When a series of load flow calculations are performed, the final values of bus voltagesin each case are normally used as the initial voltagesof the next case.This reduces number of iterations,particularly when the there are minor changesin systemconditions. 6.9 CONTROL OF VOLTAGE PROFILE
e
,odern powersystemAnatvsis
system,buseswith generators usually made PV (i.e. voltagecontrol) are buses. Load flow solution then gives the voltagelevels at the load buses. If some of the load bus voltages work out to be lessthan the specified lower voltage limit,
it is indicative of the fact that the reerfivc n^rr/ar ff^.,, ^*^^ie, ^r4^^::
lines for specifiedvoltage limits cannotmeet the reactiveload demand(reactive line flow from bus i1o bus ft is proportional lAvl = lvil  lvkD.This to situation
Control by Generators Control of voltage at the receiving bus in the fundamentaltwobus system was discussedin Section 5.10. Though the same general conclusionshold for an interconnectedsystem,it is important to discussthis problem in greater detail. At a bus with generation, voltage can be conveniently controlled by adjusting generatorexcitation.This is illustrated by meansof Fig. 6.14 where the equivalent generatorat the ith bus is modelled ty a synch.onou, reactance (resistance assumed is negligible)and voltagebehind synchronous reactance. It immediately follows upon application of Eqs. (5.71) and (5.73) that
Pai+ jQet
tAvt tEthttv!= ffi O,
tv';t= lErhl+ n, #l
=tvit* hn,
lVilt6i ti
Xei
Flg.6. 1S
P i = v,l Epl.;;,' 'i, c xo,
eci=
With YYrtrr / P
(6.8e) (6.e0)
Since we are considering a voltage rise of a few percent, lV(l can be further approximated as
lVl= lV,l+ ffiO,
(6.e1)
#r,vit+tEcit)
lr/l ,/ f ^:,^L rrr , ft
\rGi
t 
iVcil
iA
\ ,{
ano i'tlri Loi glven bYihe ioaci ijow
soiution these velrres
Thus the VAR injection of +jQc causesthe voltage at the jth bus to rise approximately by (XhllViDQ.. The voltagesat other load buseswili also rise owing to this injection to a varying but smaller extent. Control by Transformers Apart from being VAR generators, transformers provide a convenient meansof controlling real power' and reactive power flow along a transmission line. As
','d&r1'l
 l
Modern Power Svstem Analvsis Since the transformerls assumedto be ideal, complex power output from it equals complex power input, i.e.
has already been clarified, real power is controlled by means of shifting the phase of voltage, and reactive power by changing its magnitude. Voltage magnitudecan be changedby transforrnersprovided with tap changing under Ioad (TCUL) gear.Transformers specially designedto adjustvoltagemagnitude smail values are calleo re rnxers. Figure 6.16 showsa regulatingtransformerfor control of voltagernagnitude, which is achievedby adding inphaseboosting voltage in the line. Figure 6.17a shows a regulating transformer which shifts voltage phase angle with no appreciablechange in its magnitude.This is achieved by adding a voltage in series with the line at 90" phase angle to the corresponding line to neutral voltage as illustrated by means of the phasor diagram of Fig. 6.17b. Here VLn= (Von* tV6) = Q  jJT) Von= dVon 6.92)
5 1= V , / i = c r v l f
or For the transmissionline I\= ! @V1  V2) or Ir = dl'a Also lo:lzyV, cfyVz
(6.e4)
wher ea = (1 j J l t)= 7 1 ta n t J 1 t since r is small. The presence regulatingtransformers lines modifies the l/uur matrix of in tw a t hc r c byrn < i l i fy i n gth e l o a c lfl o w s o l u ti on.C onsi cl er l i ne, connecti ng o buses, having a regulating transformer with off'nominal turns (tap) ratio a includedat one end as shown in Fig. 6.18a.It is quite accurate neglectthe to translonucr, small impedance the rr:gulating of i.c. it is rcgardcdas an iclcal with line device. Figure 6.18b gives the corresponding circuit 4epresentation by represented a seriesadmittance.
(6.es) I z= l( Vz oV, ) =  WVr + lVz by Equations(6.94) and (6.95) cannot be represented a bilateral network. can The I matrix representation be written down as follows liom Eqs. (6.94) and (6.95).
n.,%n
_o
A/
C...,%n
(a)
.___ Jl"r<_tc,
I
J
(V"n+tV"r)=aV",
/^ ic raal nr rmhor\
(b)
magnitude Flg. 6.16 Regulating transformer control voltage for of
for of transformer control voltagephaseangle Fig. 6.17 Regulating
I
Modernpower SystemAnalysis
Bus,1 Bqs2
ao.O a,r.,.r t,o*
(i ) V/ V\ = l. M or a  l/ 1. 04 (ii) V/V\ = d3 or d = si3o
I
__
f#
neffiting
transformer (a)
aV'l Of__1*<<L
Solution (i) With regulating transformer in line 34, the elements of the corresponding submatrixin (v) of Exa 6.2 are modified as under.
sz
I I
v2
3 j2)
. .
l''r
Y
12
3 4 r.92 516 j10.s47 + js.777
r.92 js.777 3  j e
(b)
Fig. 6.18 Line with regulating transformer and its circuitrepresentation
Modified submatrixi n (v) of Example 6.2 is (
a
J
(6.e6)
The entriesof f matrix of Eq. (6.96)would then be usedin writing, the rru, matrix of the complete power network. For a voltage regulating transformer a is real, i.e. d = e, therefore, Eqs. (6.94) and (6.95) can be represenred the zrnetworkof Fig. 6.19. by
4
e i3" e2
4
r j3)
3 (0.666  j2) 4
(2 j6) 1T G z+ j6)
+ j6)
3.
.3r13 5.887 3 je
3 je
t
c
Fis.6.1e :jiiil:':3;:i;ffiTH#*:
with orrnominar or tap settins
Fig. 6.20 Offnominal transformers both line ends (a., a, real) at ,
If the line shown in Fig. 6.18a is representedby a zrnetworkwith shunt admittance at eachend, additionalshunt admittancelol2ysappears bus 1 yu at and yo at bus 2. The above derivations also apply for a transformer with offnomin al tap setting' where a = (k[nuJ(kD,r:up, a real value.
Ylfvil[Iil I t L  , y ) l v t)  U t ) oI[ ' . lt/q l yv i l =l[o , q ) Lv4) 'lIl r,l ; o: [ o I f 4 l 1 , L l v",)],1 L .;  o
Substituting (ii) in (i) and solving we get
(i) .
(ii )
 4t
The fourbus system of Fig. 6.5 is now modified to include a regulating transformerin line 34 near bus 3. Find the modified rru5 gf the system for With offnominaf **ing transformers eachend of the line as shownin ,up at Fig. 6.20,we can write. Thus
Lotory
"!',vf [q 1: ['"1
J 4y ) LV2 ltz )
/iii\
l '=
o?v
aqztf
l^*,
d; )
(iv)
Note: Solve it for the case when ao &2 are complex.
.
l
236 
I
Modernpower SystemAnqtysis CO N C L U SION
Load Flow Studies Bus I is slack bus with Vr = 1.0 10" Pz+ iQz = 5.96 + j1.46 lVTl= 1102 Assume: \ = 1.02/0" and
1
6. 10
In this chapter, perhapsthe most importantpower systemstudy,viz. load flow has been introduced and discussed detail. Important methodsavailable have in methodsis the best, because the behaviourof different load flow methods is dictated by the types and sizes of the problems to be solved as well as the precise details to implementation.Choice of a particular method in any given situation is normally a compromisebetweenthe various criteria of goodnessof the load flow methods.It would not be incorrect to say that among the existing methods no single load flow method meetsall the desirablerequirementsof an ideal load flow method; high speed,low storage,reliability for illconditioned systems,versatility in handling various adjustmentsand simplicity in programming. Fortunately, not all the desirable features of a load flow method are needed a l l s i tu a ti o n s . in Inspite of a large number of load flow methods available,it is easy to see that only the NR and the FDLF load flow methodsare the most important ones fbr generalpurposeload flow analysis.The FDLF method is clearly superior to the NR methodfrom the point of view of speedas well as storage.Yet, the NR method is stili in use because its high versatility, accuracyand reliability and of as such is widely being used for a variety of system optimization calculations; it gives sensitivity analysesand can be used in modern dynamicresponse and outageassessment calculations. course Of newer methodswould continueto be developed which would either reduce the computation requirementsfor large systems or which are more amenableto online implementation.
vi 710"
tt.o.oz o ottooo .7orr:"
Fig. P6.2
2 3
t'
6.3 For the systemof Fig. P6.3 find the voltage at the receivingbus at the end of the first iteration. Load is 2 + 70.8 pu. voltage at the sending end (slack) is 1 + /0 pu. Line admittance is 1.0  74.0 pu. Transformer reactance 70.4 pu. offnominal turns ratio is lll.04. Use the GS is technique. Assume Vn = ll0.
c>+J=F"d
Fig. P6.3 6.4 (a) Find the bus incidencematrix A for the fourbus sysremin Fig. p6.4. Take ground as a reference. (b) Find the primitive admittancematrix for the system.It is givbn rhat all the lines are charactenzed a seriesimpedanceof 0.1 + j0.7 akrn by and a shunt admittanceof 70.35 x 105 O/km. Lines are rated at220
kv.
PROB IVI TE S
6 . r For the power systemshownin Fig. P6.1,obtain the bus incidencematrix
A. Take ground as reference.Is this matrix unique? Explain.
(c) Find the bus admittancematrix for the system. Use the base values 22OkV and 100 MVA. Expressall impedances and admittances per in unit.
1
()
110 m k
3
Fig. P6.4 Fig. p6='!
6 . 5 Consider the threebussystem of Fig. P6.5. The pu line reactancesare 6 . 2 F<rrthe network shown in Fig. P6.2,obtain the complex bus bar voltage
at bus 2 at the end of the first iteration..use the GS method. Line impedances shown in Fig. P6.2 are in pu. Given: indicatedon the figure; the line resistances negligible.The magnitude are of all the threebusvoltages are specified to be 1.0 pu. The bus powers are specified in the following table.
.?38i I I
Modern Power Svstem Anatvsis
Zrrvz  .ffi () ir o  y=_j5.0
y=j5 0 3 Fig. P6.5
Real demand Reactive demand Real Seneratrcn Reactive generation
,l
\
vel
',\ , =is'o ,lJ /
13
Fig. P6.7 Threebus sample system containing a regulating transformer 6.8 Calculate V3for the systemof Fig. 6.5 for the first iteration, using the data of Example6.4. Startthe algorithmwith calculations bus 3 rather than at at bus 2. 6.9 For the sample system of Example 6.4 with bus I as slack, use the fbllowing nrcthuds obtairr load flow solution. to a (a) GaussSeidel using luur, with accelerationfactor of 1.6 and tolerancesof 0.0001 for the real and imaginary components of voltuge. (b) NewtonRaphson using IBUS,with tolerances 0.01 pu for changes of in the real and reactivebus powers. \ Note: This problem requiresthe use of the digital computer. 6.10 Perform a load flow studyfor the systemof Problem6.4. The bus power and voltage specificationsare given in Table P6.10.
Table P6.10
= PGr ? Qot = 0.6 ect (unspecified) Pcz = 1.4 pn, lunspecified) Qoz= 0 go. lunsp"cified) Pa = 0 Qot = l.O Carry out the completeapproximateload flow solution.Mark generations, load demands and linc flows thc onelinediagram. on 6.6 (a) Repeat Problem6.5 with bus voltagespecifications changedas below: l V t l = 1 . 0 0p u lV2l= 1.04pu lV3l= 0.96 pu Your results should show that no significant change occurs in real power flows, but the reactive flows change appreciably as e is sensitive voltage. to (b) ResolveProblem6.5 assumingthat the real generationis scheduledas follows: Pc t = 1 .0 p u , P c z = 1 .0pu, P ct = 0 The real demandremainsunchangedand the desiredvoltage profile is flat, i.e. lvrl = lv2l = lv3l = 1.0 pu. In this casethe resultswill show that the reactiveflows arc essentially unchangccl, thc rcal flgws but are changed. 6.7 Consider the threebussystemof problem 6.5. As shown in Fig. p6.7 where a regulating transforrner(RT) is now introclucedin the ljne l2 near bus 1. Other systemdata remain as that of Problem 6.5. Consider two cases: (i) RT is a magnituderegulatorwith a rario = VrlVl = 0.99, (ii) RT is a phaseangle regrrlaror having a ratio = V/Vi = ",3" (a) Find out the modified )/ur5 matrix. (b) Solve the load flow equationsin cases(i) and (ii). compare the load flow picture with the one in Problem 6.5. The reader should verify that in case(i) only the reactiveflow will change;whereas in case (ii) the changeswill occur in the real power flow.
I ' 3
Por = 1.0 ',r2 = 0 Po: = 1.0
Bus power,pu B us Voltagemagnitude, pu Bus type
I 2
) 4
a
Unspecified 0.95  2. O  1.0
Unspecified Unspecified  1.0  0.2
1.02 1.01 Unspecified Unspecified
Slack PV PQ PQ
Computethe unspecified voltages, bus powersand all line flows. bus all Assume unlimited Q sources. Use the NR method.
NCES REFERE
Books
l. Mahalanabis, A.K., D.P. Kothari and S.I. Ahson, Computer Aided Power System Analysis and Control, Tata McGrawHill, New Delhi. 1988.
' 2 ___ l _  rA 0 " ' _
M.o dp , ,n .  ' r
pnrrrar , v'rvr
er a r e yro rrra i*l l
A^^r.,A1 tatysls
2' weedy, B.M. and B.J. cory, Erectricarpowersystents,4th Ed., John wiley, New York, 1998. Gross,C.A., power System Analysis,2nd 8d..,John Wiley, New york, 19g6. Sterling, M.J.H., power SystemControl, IEE. England,1978. Hill, New York. 19g2. 6. stagg, G.w. and A.H. EIAbiad, computer Methods in power system Anarysis, McGrawHill, New york, 196g. 7' Rose' D'J' and R'A willough (Eds), sparse Matrices and their Applications, Plenum,New york, 1972. 8' Anderson,P'M., Analysisof Faulted Power systems, The Iowa state university Press,Ames, Iowa, 1923. 9' Brown, H'8., sorution of Large Networks by Matix Methods, John w1rey, New , York, 1975. l0' Knight' rJ'G', Power systemEngineering and Mathematics,pergamonpress,New York 1972. 11. Shipley, R.B., Introduction to Matrices and power systems,John w1ey, New York, 1976. 12' Hupp, H.H., Diakoptics and Networks, Academic press,New york, 1971. l3' Arrillaga, J. and N.R. watson, computer Modeiling of Erectricarpower sltstems, 2nd edn, Wiley, New york, 2001. 14' Nagrath, I.J. and D.p. Kothari, power system Engineering, Tata McGraw_Hilr, New Delhi, 1994. l5' Bergen,A.R., power systemAnarysis, prenticeHail, Engrewood criffs, N.J. r9g6. l6' Anillaga' J and N'R' watson, computer Modelting of Electrical power systems, 2nd edn. Wiley, N.y. 2001. Papers 17. Happ, H.H., 'DiakopticsThe solution ,, of system problems by Tearing proc. of the IEEE, Juty t974, 930. l8' Laughton'M'A', 'Decontposition Techniques power systemNetworkLoad in Flow Analysi.s usi'g the Nodar Impcdance Matrix,, proc. IEE, 196g, r15, s3g 19' stott, B., 'Decouprecr NewronLoad FIow,, IEEE Trans., 1972,pAS_91, 1955. 20. stott, B., 'Review of LoadFrow carcuration Method, proc. IEEE, Jury 1974,916. , 21' stott, B., and o. Alsac, 'FastDecoupled Load Flow,, IEEE Trans., 1974,pAS_93: 859. 22' Tinney' w'F" and J'w. walker, 'Direct solutions of sparse Network Equationsby Optimally Ordered Triangular Factorizaiions,, proc. IEEE. Nnvernhe" ' 1oA1 v 55:1801 " 23' Tinney, w.F. and c.E. Hart, 'power Flow sorution by Newton,s Method,, ,EEE Trans.,November1967,No. ll, pAS_g6:1449. 24' ward, J.B. and H.w. Hare, 'Digitar computerSorutionof power problems,,AIEE Trans.,June 1956,pt III, 75: 39g.
, u.r., L,teclrrc Lnergy System Theom: An Introduction, 2nd Ed, McGraw_
Power Frow Argorithmswith Reference Indian power to systems,, proc. of II symp.on Power plant Dynamicsancrcontror, Hyderabad, 14_i6 Feb. rg:/g, zrg. 26. Tinney, W.F. and W.S. Mayer, .Solution TriangularFactorization', IEEE Trans.on Auto contor, August 1973,yor. AC_lg,
JJJ.
27' Saro, N. and w.F Tinney, 'Techniquesfor Exproiting Sparsity of Network Admittance Matrix', IEEE Trans. pAS, I)ecember L963, g2, 44. 28' sachdev, M.s. and r.K.p. Medicherla, 'A second order Load Flow Technique', IEEE Trans., pAS, Jan/Feb 1977, 96, Lgg. 29. Iwamoto, S.and y. Tamura, ,A Fast Load Flow Method Retaining Non linearitv'. ' IEEE Trans., pAS. Sept/Oct. 197g, 97, 15g6. 30. Roy, L., 'Exact Second Order Load Flow,, proc. of the Sixth pSCC Conf. Dermstadt,Yol. 2, August lg7g, 7Il. 31' Iwamoto, s' and Y. Tamura,'A Load Flow calculation Method for Illconditioned Power Systems',IEEE Trans pAS, April lggl 100, 1736. , 32. Happ, H.H. and c.c. young, 'Tearing Argorithms for Large scare Network Problems', IEEE Trans pA,S,Nov/Dec Ig71, gO, 2639. 33. Dopazo,J.F.O.A.Kiltin and A.M. Sarson,,stochastic Load Flows,, .IEEE Trans. P A S ,1 9 7 5 , 9 4 , 2 g g . 34' Nanda' J'D'P' Kothari and s.c. srivastava,"Some Important observationson FDLF'Algorithm", proc. of the IEEE, May 19g7, pp.732_33. \ 35. Nanda,J., p.R., Bijwe, D.p. Kothari and D.L. stenoy, .,second order Decoupred Load Flow', ErectricMachines and power systems, r2, No. 5, 19g7,pp. vor 301_ 312. 36' NandaJ'' D'P' Kothari and S.c. srivastva,"A Novel secondorder Fast Decoupled Load Flow Method in Polar coordinatcs", Electic Machinesand power,s.ysrems, Vol. 14, No. 5, 19g9,pp 339_351 37. Das, D., H.s. Nagi and D.p. Kothari, .,A Novel Method for solving Radiar DistributionNetworks", proc. IEE, ptc, v.r. r4r, no. 4. Jury r994. pp. zgr_zgg. 38' Das, D', D'P' Kothari and A. Kalam. "A Simple and Efficient Method for Load Flow sorurion of RadiarDistributionNerworks',, Inr. J. of EpES, r995, pp. 335_ 346.
OptimalSystem Operation
A3
pariance caiied is the 'load scheduling' (LS) problem.One must first solvethe UC problem before proceeding with the LS problem. Throughout this chapter we shall concern ourselves with an existing installation,so that the economicconsiderationsare that of operating(running) cost and not the capitai outiay. 7.2 OPTIMAL OPERATION OF GENERATORS ON A BUS BAR
the 'unii commitment' (UC) probiemand the secondis calleci
Before we tackle the unit commitment problem, we shall consider the optimal operation of generatorson a bus bar. Generator Operating Cost
7.1
INTRODUCTION
The major component of generatoroperating cost is the fuel input/hour, while only to a small extent.The fuel cost is meaningtulin maintenance coutributes case of thermal and nuclear stations,but for hydro stations where the energy storageis 'apparentlyfree', the operatingcost as such is not meaningful. A to suitablemeaningwill be attached the cost of hydro storedenergyin Section 7.7 of this chapter. Presentlywe shall concentrateon fuel fired stations.
l I I I
L
The optimal system operation,in general,involved the consideration of economy operation, of systemsecurity,emissions certain fossilfuel plants, at optimalreleases water at hydro generation, of etc. All theseconsiderations may make conflicting for requirements usuallya compromise fo he madefor and has optimalsystem operation. this chapterwe cor..;ider economyof operation ln the only, also called the ec'omonic problem. di.spcttch The main aim in theeconomicdispatch problernis to rninimize thc total cost of generating real power (production cost) at various stations while satisfying theloads andthe losses thetransmission in links.For sirnplicitywe consicler the presence thermalplantsonly in the beginning.In the later part of this chapter of we will considerthe presence hydro plants which operatein conjunctionwith of thermalplants.While thereis negligible operatingcost at a hydro plant, there is a limitationol availability ol'watcr'over a pcriodof tinrewhich nrustbc used to savemaximum fuel at the thermal plants. In the load flow problemas detailedin Chapter6, two variablesare specified at eachbus and the solutionis then obtainedfor the rernainingvariables. The specifiedvariablesare real and reactivepowersat PQ buses,real powersand voltagemagnitudes PV buses,and voltagemagnitude at and angle at the slack bus.The additionalvariables be specifiedfor load flow solution are the tap to of settings regulatingtransformers. the specifiedvariablesare allowed to vary If in a regionconstraineci practicaiconsicierations (upperanciiower iimits on by activeand reactivegenerations, voltagelimits, and rangeof transformer bus tap settings), there resultsan infinite numberof load flow solutions,eachpertaining to one set of valueso1'specified variables. The 'best'choice in sornesense of thevaluesof specifiedvariables leadsto the 'best' load flow solution.Economy of operation naturallypredominantin determining is allocationof generation to eachstationfor varioussystemload levels.The first problem in power system
t
\
al)

; P t r o . c 8 6 o g E
6
f LL
8b= e 5
o O
o
Fig.7.1
(vw)min
(MW)max Power output, MW ' 
Inputoutput curve of a generatrng unit
The inputoutput curve of a unit* can be expressedin a million kilocalories per hour or directly in terms of rupees per hour versus output in megawatts. The cost curve can be determined experimentaily. A typical curve is shown in Fig. 7.1 where (MW)o'in is the minimum loading limit below which it is uneconomical (or may be technically infeasible) to operate the unit and (MW)n,u, is the maximunr output limit. The inpLltoutput curve has discontinuities at steam valve openings which have not been indicated in the figure. By fitting a suitable degree polynomial, an analytical expression for operating cost can be written as A unit consists of a boiler, turbine and generator.
.L4+  I

. .
MOOern rower


A    !  
DySIeIrl Arlaly
A  ^ r   ^ : 
Ci(Pc)
Rs/hourat outPutPc,
of Considerations spinning reserve,to be explained later in this section,require that
where the suffix i standsfor the unit number.It generatly suffices to fit a second polynomial,i.e. degree
_Q;
^*) Po D Po,,
(7.6)
2
b,Pc, + d, Rs/hour
(7.r)
is called the incremental The slope of the cost curve, i.".43 Juel cost(lQ, dPo, in and is expressed units of rupeesper megawatthour (Rs/lvIWh).A typical plot fuel cost versus power output is sho.wnin Fig. 7.2.If the cost of incremental as curve is approximated a quadraticas in Eq. (7.1), we have
(lc)i= aiP"t + bt
marsin. i.e. Eq. (7.6) must be a strict inequality. Since the operatingcost is insensitiveto reactive loading of a generator,the rnannerin which the reactive load of the station is sharedamong various ondoes not afl'ectthe operatingeconomy'. line generators 'What is the optimal manner The question that has now to be answeredis: by the generatorson the bus?' in which the load demand Po must be shared This is answeredby minimizing the operating cost
k
(7.2)
c = D ci(pci)
f:l
( 7. 7)
l I
i.
i.e. the of underthe equalityconstraint meeting load demand, (
f 
(J
I I
L i:t
\' G D
i l P o = O
(7.8)
o o o c
E >B ,d
Etr o E E ()
o o
(MW)min MW Power output,
(MW) max
where k = the number of generatorson the bus. Further, the loading of each generator is constrained by the inequality constraintof Eq. (7.5). Since Ci(Pc) is nonlinear and C, is independentof P6t (i+ i), this is a nonlinear programming problem. separable \ If it is assumedat present,that the inequality constraint of Eq. Q.q is not effective,the problem can be solved by the method of Lagrangemultipliers. Define the Lagrangian as
f 
poweroutputfor the unitwhose fuel Fig.7.2 Incremental costversus curve is shownin Fig.7.1 inputoutput i.e. a linear relationship.For better accuracy incremental fuel cost may be expressedby a number of short line segments (piecewise lineanzation). IC Altcrnativcly,wc can fit a polynomial of suitabledegreeto represent curve in the inverseform
Pc;i= a, + {),(lC)i + 1,QC)', + ... Optimal Operation
it (Pci)^[f ""] "",
(7.e)
where X is the Lagrange multiplier. Minimization is achieved by the condition
of, =o
dPo, or where lci dCi ' dPc,  ) i i = 1 , 2 , . . . ,k (7.1o)
(7.3)
itre that it is known a priltri which generutors t<lrtln to ntccta [,et us assulne on Ubvtously load clenrand the statton. p:.rrticular (7.4) DPr,,, ) P,
d4,, a Iurctio' ol'.gencra,il;,":':t
cost of the ith generator (units: Rs/TvIWh), is the incremental
E q u a t i o n ( 7 . 1 0 ) c a n be written as dC^ dPoo \
,'',,*
is where Pci, ,r.,,"* the rated real power capacity of the ith generatorand Po is the total power demandon the station.Further,the load on each generatoris to irc constrainedwithin lower and upper limits, i.e.
Pcr, .in 1 Po, 1 Po,, rn.*, I = L, 2, "', k
:":
d4(il dr*The
(7. l) r
(7.s)
effect of reactive loading on generatorlosses is of negligible order.
n.:^r
A^^r
^
..
L
^
cost of the plant colrespondsto that of unit 2 alone. When rhe plant load is 40 Mw, each unit operatesat its minimum bound, i.e.2o Mw wiitr plant \ = Rs 35/I4Wh. Computer solution fbr optimal loa<ling of generators can be obtained iteratively as follows: 1. Choose a trial value of ), i.e. IC = (IC)o. 2. Solve for P", (i = 1, 2, ..., k) from Eq. (7.3). 3. If ItPc,Otherwise, 4. Increment(lC) bv A (1"), t I: Po, ,rl I < 0 or decrement (/c) by A(tr) fl  Pr] r 0 and repeat from step 2. This step is possible if [D Pc, because P., is monotonically increasingfunction of (1g). consider now the effecr of the inequality constraint (7.5). As (1c) is increasedor decreased the iterative process,if a particular generatorloading in P", reachesthe limit PGi,^o or P6;, min, its loading from now on is held fixed at this value and the balance load is then shared between the remaining generators on equal incremental cost basis. The fact that this operation is optimal can be shown by rhe KuhnTucker theory (seeAppendix n;. Pol < e(a specified value), the optimal solution is reached. When dczldPcz= Rs 44/MWh, 0.25PG2+30  44 pnt = JI= 56 MW 0. 25 The total plant output is then (56 + 20) = 76 MW. From this point onwards, the values of plant load sharedby the two units are found by assumingvarious valuesof \. The results are displayedin Table 7.1. or Table 71 Outputof each unit and plant output for variousvaluesof ) for Example7.1
Plant ),
Unit I
Pcl, MW
RszMWh 35 44 50 55 60 61.25 65
Unit 2 Pcz, NfW
Plant Output
40.0 76.0 130. 0 175.0 220.0\ 231.25 250.0
Incremental fuel costs in rupeesper MWh for a plant consisting of two units are:
dt
Figure 7.3 shows the plot of the plant .trversusplant output. It is seenfrom Table 7.7 that at .\ = 61.25, unit 2 is operatingat its upper limit and therefore, the additional load must now be taken by unit 1, which then determines the plant ).
60
i*o.2opct+40.0 G1 dcz= o.2,pcz+3o.o
dPo, Assume that both units are operating at all times, and total load varies from 40 MW to 250 MW, and the maximum and minimum loads on each unit are to be I25 and 20 MW, respectively.How will the load be sharedbetween the two units as the systemload vanes over the full range?What are the corresponding values of the plant incrementalcosts? Solution At light loads, unit t has the higher incrementalfuel cost and will, therefore,operareat its lower limit of zo Mw, for which dcrldpcr is Rs 44 per MWh. When the ourput unit 2is20 MW, dczldpcz= Rs 35 p"i UWt. Thus, of with an increasein the plant output, the additionalload should be borne bv unit
Fig.7.3
o @
+t
I
l c c
8 5 0
^
45 i= a>
EP 40 t O c
E
J
c
Plantoutput, MW Incremental fuel cost versus plant output,as found in Example7.1
t:Hzut"l .Z.48: i
Mocjern Power Systemnnaiysis
To find the load sharingbetweenthe units for a plant output of say 150 MW, we find from the curve of Fig. 7.3, that the correspondingplant X is Rs 52,22 per MWh. Optimum schedules each unit for 150 MW plant load can now for be found as P c t = 6 1 ' 1 1M W 0.2Pa*4052.22;
Net savingcaused optimumscheduling by is = 772.5 721.875 50.625 Rs/lr assuming Totalyearlysaving continuous operation
This savingjustifies the need for optimal load sharing and the ddvices to be installed for controlling the unit loadings automaticallv.
0.25PG2 30= 52.22; + P c r + P c z = 1 5 0M W
Pcz = 88'89 MW
Proceeding on the above lines, unit outputs for various plant outputs are computedand have beenplottedin Fig. 7.4. Optimum load sharingfor any plant load can be directlv read from this fieure.
I = 100
I tzs
Let the two units of the svstem studiedin Example7.1 have the following cost curves. Cr = 0.lPto, + 40Pc + 120 Rs/hr Cz= 0.l25Pzcz+ 30Po, + 100 Rsftrr 250 tI 200 I 3 150
220 MW
Eru
o
E 5 0
f
0
5
0
150 100 Plantoutput,MW _>
200
250
E 100 o
J
50
Sunday
Monday
Flg.7.4
Output of each unit versus plant output for Example 7.1
12
(noon) Fig. 7.5 PM Time Daily load cycle
l
1 2 (night)
l
6 AM
For the plant describedin ExampleT.l find the saving in fuel cost in rupeesper hour for the optimal schedulingof a total load of 130 MW as comparedto equal distribution of the same load between the two units. Solution Example 7.I revealsthat unit I should take up a load of 50 MW and unit 2 should supply 80 MW. If each unit supplies65 MW, the increasein cost for unit 1 is
165rnnn  \'U . L I ' r : r t . ,^\rn *U)JIl nr = rr.'tnZ (U.Ifnt
; lf\r1 t'tUI'6yll rl
165
=
Jso
"'
Fn^ a llL.J
D^tLl\S/I[
lso
Let us assume daily load cycle as given in Fig. 7.5. Also assume a that a cost of Rs 400 is incurredin taking either unit off the line and returning it to service after 12 hours. Considerthe 24 hour period from 6 a.m. one morning to 6 a.m. the next morning. Now, we want to find out whether it would be more economical to keep both the units in servicefor this 24hour period or to remove one of the units from service for the 12 hours of light load. For the twelvehourperiod when the load is 220 MW, referring to Table 7.1 of Example 7.1, we get the optimum schedule as = 120 M W Pcr = 100 M W' Pcz Total fuel cost for this period is + [ 0 . 1x 1 0 0 2 + 4 0 x 1 0 0 + 1 2 0 + 0 . 1 2 5 x 1 2 0 2 3 0 x 7 2 0 + 1 0 0 ]x 1 2 = Rs. 1, 27, 440
Similarly, for unit 2,
(0. dpor= r25PGz+ :oro;1" J*co.rs" + 30) ",
  721.875 Rs/hr
t
ltgtiFl
power Modern svstem nnAVsis
Dynamic Programming Method t 
If both units operatein the light load period (76 MW from 6 p.m. to 6 a.m.) also, then from the same table, we get the optimal scheduleas Pcr = 20 MW, Pcz = 56 MW Total fuel cost for this period is then
(0.1 x 20' + 40 x 20 + 120+ 0.125x 5 6 + 3 0 x 5 6 + 1 0 0 ) x 1 2
= Rs 37,584 Thus the total fuel cost when the units are operating throughout the 24 hour period is Rs I,65,024. If only one of the units is run during the light load period, it is easily verified that it is economicalto run unit 2 andto put off unit 1. Then the total fuel cost
during periot:tXf ': this 762+ 3o x 76+ 100)x rz = Rs 37,224
Total fuel cost for this case= L,27,440+ 37,224 = Rs 1,64,664 Total operatingcost for this casewill be the total fuel cost plus the startup cost of unit l , i .e . 1,64,6& + 400 = Rs 1,65,064 Comparingthis with the earlier case,it is clear that it is economicalto run both the units. It is easy to see that if the startupcost is Rs 200, then it is economical to run only 2 in the light load period and to put off unit 1. 7.3 OPTTMAL UNrT COMMTTMENT (UC)
As is evident,it is not economicalto run all the units availableall the time. To determine the units of a plant that should operate for a particular load is the problem of unit commitment (UC). This problem is of importancefbr thermal plants as for other types of generationsuch as hydro; their operating cost and startup times are negligible so that their onoff statusis not important. A simple but suboptimal approach to the problem is to impose priority ordering, wherein the most efficient unit is loaded first to be'followed by the less efficient units in order as the Ioad increases. A straightforward but highly timeconsuming way of finding the most economicalcombination of units to meet a particular load demand,is to try all possiblecombinationsof units that can supply this load; to divide the load optimally among the units of each combination by use of the coordination equaiions,so as to finci the most economicaioperatingcost of the combination; then,to determinethe combinationwhich has the least operatingcost among all these.Considerablecomputationalsavingcan be achievedby using branch and bound or a dynamic programming method for comparing the economics of combinations certaincombinations as neetnot be tried at all.
In a practical problem, the UC table is to be arrived at for the complete load cycle. If the load is assumedto increasein small but finite size stensldvnamin prograrrurung can be used to advantage for computing the uc table, wherein it is not necessaryto solve the coordination equations; while at the same time the unit combinations to be tried are much reduced in number. For these reasons,only the Dp approachwill be advancedhere. The total number of units available, their individual cost characteristics and the load cycle on the station are assumed be known a priori.Further, to it shall be assumedthat the load on each unit of combination of units changei'in suitably small but uniform steps of size /MW (e.g. I MW). Starting arbitrarily with any two units, the most iconomical combination is determined for all the discrete load levels of the combined output of the two units. At eachload level the most economic answermay be to run either unit or both units with a certain load sharing betweenthe two. The most economical cost curve in discreteform for the two units thus obtained, can be viewed as the cost curve of a single equivalent unit. The third unit is now added and the procedurerepeated find the cost curve of the threecombined to units. It may be noted that in this procedurethe operatingcombinationsof third and first, also third and secondare not required to be worked out resulting in considerable saving in computationaleffort. The processis repeated,till all available units are exhausted.The advantage of this approach is that having oitiined the gPliu. way,of loading ft units, it is quite easy ro determine the Jptimal manner of loading (ft + 1) units. Let a cost function F" (x) be defined as follows: F,y (x) = the minimum cost in Rs/hr of generating .r MW by N units, fN 0) = cost of generating y MW by the Nth unit F*{x  y)  the minimum cosr of generating (.r  y) Mw by the remain_ ing (1/  t) units Now the application of DP results in the following recursive relation
FN@)= TnVn9) * Fur @  y)
(7.r2)
canaeifies
Using the aboverecursiverelation, we can easily determinethe combination of units, yielding minimum operating costsfor loads ranging in convenient steps from the minimum.permissible load of the smallest unit to the sum of the
qrroiiollo . n f q l l $vs^rqurv r r  i + o rrr + L ; ^   ^ ^ ^  ^ uurrD. i  LrrlD PruuttJs . 1 ^  1  1 r nunlmum ure total oDerating eost
and the load sharedby each unit of the optimal combination are ;il.u,i" determined for each load level. The use of DP for solving the UC problem is best illustratedby meansof an example. Considera sample system having four thermal generating units with parameterslisted in Table 7.2.It is required to determinr th. mosteconomical units to be committed for a load of 9 MW. Let the load changes be in steps of I MW.
.
ModernPow for unit parameters the samplesystem Table 7.2 Generating
Capacity (MW) Cost curve pararneters (d = 0)
Unit No.
The effect of step size could be altogether eliminated, if the branch and bound technique [30] is employed. The answer to the above problem using branch and bound is the samein terms of units to be committed, i,e. units 1 and 2, but with a load sharing of 7 .34 MW and 1.66 MW, respectively and a total
graung cost oI I<s z,y.zL /J/nour. In fact the best scheme is to restrict the use of the DP method to obtain the
1 2 3 4 Now
1.0 1.0 1.0 1.0 Ft@) = ft@) fr(9) = f{9)=
1 2 .0 1 2 .0 1 2 .0 12.0
0.77 1.60 2.00 2.50
23.5 26.5 30.0 32.0
UC table for various discrete load levels; while the load sharing among committed units is then decidedby use of the coordination Eq. (7.10). For the example under consideration,the UC table is prepared in steps I of MW. By combining the load range over which the unit commitment does not change, the overall result can be telescopedin the form of Table 7.3. Table 7.3 Status*of units for minimum operatingcost (Unit commitment table for the samplesystem)
Unit number Load range
LorP'ot+ btPcr
= ollss x 92 + 23.5 x9 = Rs 242.685lhour From the recursive relation (7.12), computation is made for F2(0), Fz(l), ..., Fz(2), Fz(9).Of these FzQ) = min tt6(0) + Ft(9)1, VzG) + Ft(8)l' VzQ) + Ft(7)1, VzQ) + Fr(6)l' Vz@)+ F1(5)l'
r
l5 6r3 t 4 18 1948
2
3
0 0 1 1
4
0 0 0 1 \
+ VzT + Fr(2)1, VzG) Ft(3)1, t6(5)+ Fr(4)1, vzg) + Fr(O)l) tfr(s)+ F,(1)1,
On computing termbyterm and compdng, we get FzQ) = Vz(2) + Ft(1)) = Rs 239.5651how Similarly, we can calculate Fz(8),Fz(1), ..., Fz(l), Fz(O). Using the recursiverelation(7.12),we now computeFl(O), F:(1), ...' F3(9). Of these
*l = unit running;0 = unit not running.
Fse)= min {t6(0)+ Fr(9)f, t6(1)+ Fl8)1, ...'[6(9)+ rr(0)]] = [6(0)+ FzQ)l= Rs 239.565ftour
Proceeding similarly, we get FoQ) = [f4(0) + Fr(9)] = Rs 239.565lhour Examinationof Fr(9), Fz(9),Fl(9) and Fa(9) leads to the conclusion that optimum units to be lommitted for a 9 MW load are 1 and 2 sharing the load ur Z l,tW and 2 MW, respectively with a minimum operating cost of Rs 239.565/hour. 'u\e It must be pointed out here, that the optimai iiC tabie is inclependentof numberingof units, which could be completely arbitrary. To verify, the reader rnay solvethe above problem onceagain by choosing a different unit numbering scheme. If a higher accuracyis desired,the step size could be reduced(e.g. * t*r, with a considerableincreasein computation time and required storage capacity.
' The UC table is preparedonce and for all for a given set of units. As the load cycle on the station changes,it would only mean changes in starting and stopping of units with the basic UC table remaining unchanged. Using the UC table and increasingload in steps,the most economical station operatingcost is calculatedfor the completerange of stationcapacity by using the coordination equations. The result is the overall station cost characteristic in the form of a set of data points. A quadratic equation (or higher order can then be fitted to this data for later use in economic equation,if necessary) load sharing among generating stations. 7.4 RELIABILITY CONSIDERATIONS
With the increasing dependence of industry, agriculture and daytoday household comfort upon the continuity of electric supply, the reliability of power systemshas assumedgreat importance. Every eleciric utilify is normaily under obligation to provide to its consumersa certain degreeof continuigl and quality of service (e.g. voltage and frequency in a specifiedrange). Therefore, economy and reliability (security) must be properly coordinated in arriving at the operationalunit commitment decision.In this section,we will seehow the purely economic UC decision must be modified through considerations of reliability.
In order to meet the load demand under contingency of failure (forced outage) of a generator or its derating caused by a minor defect, static reserve capacity is always provided at a generating station so that the total installed capacity exceedsthe yearly peak load by a certain margin. This is planning a In arriving at the economicUC decisionat any particular time, the constraint taken into account was merely the fact that the total capacity on line was at least equal to the load. The margin, if any, between the capacity of units committed and load was incidental. If under actual operation one or more of the units were to fail perchance(random outage), it may not be possible to meet the load requirements.To start a spare (standby) thermal unit* and to bring it on steam to take up the load will take severalhours (28 hours), so that the load cannot be met for intolerably long periods of time. Therefore, to meet contingencies, the capacity of units on line (running) must have a definite margin over the load requirementsat all times. This margin which is known as the spinning reserve ensurescontinuity by meeting the load demand up to a certain extent of probable loss of generationcapacity.While rules of thumb have been used, based on past experienceto determinethe system's spinning reserve at any time, Patton's analytical approach to this problem is the most promising. Since the probability of unit outageincreaseswith operatingtime and since a unit which is to provide the spinning reserve at a particular time has to be startedseveralhours ahead,the problem of security of supply has to be treated in totality over a period of one day. Furthefinore, the loads are never known with complete certainty. Also, the spinning reserve has to be provided at suitable generating stations of the system and not necessarily at every generatingstation.This indeedis a complex problem. A simplified treatmentof the problem is presented below:
f1(down) f2 (down) f3 (down)
periods are a random phenomenon with operating periods being much longer than repair periods. When a unit has been operating for a long time, the random phenomenoncan be describedby the following parameters. 'up' time), Mean time to failure (mean
No. of cycles 'down' time), Mean time to repair (mean Et, (down) Z (down) No. of cycles
(7.13)
(7.r4)
Mean cycle time = f (up) + Z(down) Inverseof these times can be defined as rates [1], i.e. Failure rate, A = IIT (up) (failures/year) Repair rate, trt= llT (down) (repairs/year) Failure and repair rates are to be estimated from the past data of units (or other similar units elsewhere) by use of Eqs. (7.13) and (7.I4). Sound engineeringjudgement must be exercised in arriving at these estimates.The failure rates are affected by preventive maintenance and the repair rates are .r scnsitiveto size, composition and skill of repair teams. probability, we can write the probabiliiy of a unit being By ratio definition of in 'up' or 'down' statesat any time as p (up) =
r(up)
z(up)* z(down) Z(down) Z(up)* Z(down)
_
l.t
p+A )
tr+ A
(7.rs)
(7.16)
P (down) = Obviously,
p (up) + p(down) = 1 p (up) andp (down) in Eqs. (7.15) and (7.16) are also termed as availability and unavailability, respectively. When ft units are operating, the system state changesbecauseof random of outages.Failure of a unit can be regardedas an event independent the state 'down' state of other units. If a particular systemstate i is defined asX, units in
an0
1
Time>
Fig. 7.6 Randomunit performance recordneglecting scheduled outages A unit during its useful life span undergoesalternateperiods of operation and repair as shown in Fig. 7.6. The lengths of individual operating and repair
If hy^dro.generation available in the system, it could be brought is on line in a matter of minutes to take up load. .
I j ln
r,
'
(
t tlr Up. Stale \K =
\/
i + I )t
.
rz\
!l
urs PIUDaUITILy Ur urtr systelll
1l:1:
^t
L1^^
l^:^
Utrltr$ ttl ulIS Slaltr
2
tLl
is
pt = {r,r;(ul),{* ot(down)
(7.r7)
gW
I
Modern PowersystemAnalysis
the sample system for the load curve of Fig. 7.7
Unit number
WWt
Patton's Security Function A breach of system security is defined as some intolerable or undesirable condition.The only breachof securityconsidered here is insufficient generation probability that the available generation capacity (sum of capacitiesof units cbmmitted) at a particular hour is less than the system load at that time, is defined as [25] (7.18) S = Ep,r, where p, = probability of system being in state i [see Eq. (7.17)] r, = probability that system state i causesbreach of syStem securlty. When system load is deterministic(i.e. known with complete certainty), r, = 1 if available capacity is less than load and 0 otherwise. S indeed, is a quantitative estimateof system insecurity. Though theoretically Eq. (7.18) must be summed over all possible system states (this in fact can be very large), from a practical point of view the sum needsto be caried out over statesreflecting a relatively small number of uniqs on forced outage, e.g. stateswith more than two units out may be neglected as the probability of their occurrencewill be too low; Security Constrained Optimal Unit Commitment
0 (noon)
0L
t
A B C D E F
2
3
I I 0 I 0 0
4
1 0 0 0 0 0
Once the units to be committed at a particular load level are known from purely economic considerations, security function S is computed as per Eq. (7.18). the This figure should not exceed a certain maximum'tolerable insecurity level (MTIL). MTIL for a givcn systemis a management decisionwhich is guided by past experience.If the value of S exceeds MTIL, the economic unit commitment schedule modified by bringing in the next most economicalunit is as per the UC table. S is then recalculated and checked. The process is continued till ^t < MTIL. As the economic UC table has some inherent spinning reserve,rarely more than one iteration is found to be necessary. For illustration,reconsider fbur unit exampleof Sec.7.3. Let the daily the load curve for the systembe as indicated in Fig. 7.7. The economicallyoptimal UC fbr this load curve is immediately obtained by use of the previously prepared UC table (see Table 7.3) and is given in TabIe 7.4. Let us now check if the above optimal UC table is securein every period of the ioaci curve. For the minimum load of 5 MW (period E of Fig. 7.7) accordingto optimal UC Table 7.4., only unit 1 is to be operated.Assuming identical failure rate ) of l/year and repair rate pr, 99/year for all the four units, let us check if the of systemis securefcrrthe period E. Further assume systemMTIL to be 0.005. the Unit I can be only in two possiblestatesoperating or on forced outage.
Time in hours > Flg.7.7 Dailyloadcurye
Therefore,
2 S = ,? piti : p1r1* p2r2
where
= Pr= P(up)
ffi
= 0.99, rr = 0 (unit = 12MW> 5 MW) I
p+^
= 0.01, rz = t (with unit I down load demand cannot be met)
pz= p( down) =
Hence
S= 0.99x 0 + 0.01x 1 = 0.01> 0.005(MTIL) Thus unit I alone supplying5 MW load fails to satisfytne prescribed security criterion.In orderto obtainoptimalandyet secure UC, it is necessary to run the next mosteconomical i.e. unit 2 (Table unit, 7.3) along with unit 1.
ifl5"#Wl
power Mooern System Rnarysis
E#nft#
= Rs 4,463.216 (startup cost = 0) Clearly Caseb resultsin overall economy.Therefore,the optimal and secure UC table for this load cycle is modified as under,with due considerationto the overall cost.
With both units I and 2 operating, security function is contributed only by the state when both the units are on forced outage.The stateswith both units operatingor either one failed can meet the load demand of 5 MW and so do not contributeto the security function. Therefore, S = p (down; x p (down) x 1 = 0.0001 This combination (units 1 and 2both committed) does meet the prescribed MTIL of 0.005, i.e. ^S MTIL. < Proceedingsimilarly and checking securityfunctions for periodsA, B, c, D and F, we obtain the following optimal and secure UC table for the sample systemfor the load curve given in Fig. 7.7.
Table 7.5 Optimal and secure UC table Unit number Period
Unit number Period
A B C D E F
I I I I I I
I I I I I I
1 1
l'l'
l 0
0
l 0 0
0 0 0
A B C D E F
I
I t<
1 0 1 0 0
I 0 0 0 0 0
*Unit was started due to startup considerations.
7.5
OPTIMUM
GENERATION
SCHEDUTING
* Unit was started due to security considerations.
Startup
Considerations
The UC table as obtainedabove is secureand economically optimal over each individual period of the load curve. Such a table may require that certainunits have to be startedand stoppedmore than once.Therefore,startupcost must be taken into consideration from the point of view of overall economy. For example, unit 3 has to be stopped and restartedtwice during the cycle. We must, therefore, examine whether or not it will be more economical to avoid one restarting by continuing to run the unit in period C. Casea When unit 3 is not operatingin period C. Total fuel cost for periods B, c and D as obtained by most economic load sharing are as under (detailed computation is avoided) = 1,690.756 1,075.356 1.690.756 Rs 4.456.869 = + + Startup cost of unit 3 = Rs 50.000 (say)
From the unit commitment table of a given plant, the fuel cost curve of the plant can be determinedin the form of a polynomial of suitabledegreeby the method of least squaresfit. If the transmissionlossesare neglected,the total system load can be optimally divided among the various generatingplants using the equal incremental cost criterion of Eq. (2.10). It is, howrurr, on."alistic to neglect transmission losses particularly when long distance transmission of power is involved. A modern electric utility serves over a vast area of relatively low load density.The transmissionlossesmay vary from 5 to ISVoof the total load, 4'd therefore,it is essentialto accountfor lcsseswhile developingan economic load dispatch policy. lt is obviousthat when losses present, can no longer use are we the simple 'equal incrementalcost' criterion. To illustrate the point, consider a twobus system with identical generatorsat eachbus (i.e. the sameIC curves). Assume that the load is located near plant I and plant 2 has to deliver power via a lossy line. Equal incrementalcost criterion would dictate that each plant should carry half the total load; while it is obvious in this casethat the ptunt 1 should carry a greatershareof the load demandthereby reducing transmissio' losses. In this section, we shall investigate how the load should be shared among
rr$rrvsu rqrirrrra nlqnfc
Iotal operatrng cost= Rs 4,506.868 Caseb When all threeunits are running in period C, i.e. unit 3 is not stoppecl at the end of period B. Total operating costs= I,690.756+ 1,0g1.704+ 1,690.756
l.rquLor
.rrlro
;^ vYuvrr luuv
l^^^ rvirDsr
4rE .1uuuurrttrg
L^ s
n^,
t()f,
ry
Ine
the overall cost of generation
ODJgCtfVg fS tO mirufiUze
c =
,\rci(Pc')
(7.7)
at any time under equality constraint of meeting the load demand with transmission loss, i.e.
 PI= DP", P, i:l
where
k
0
(7.re)
Equation (7.23) can also be written in the alternative form ( I C) i An Q TL) if i = 1, 2, . . . , k
(7.2s)
k = tatalrnumber of generating plants Pci= generationof lth plant Pp = sum of load demand bt all buses (system load demand) Pr= total systemtransmission loss To solve the problem, we write the Lagrangianas
t=tr,(Pc)^[t""Po".] ,i:l
(7.20)
This equation is referred to as the exact coordination equation. Thus it is elear that to solve the optimum lead seheduling problem; it is necessaryto compute ITL for each plant, and therefore we must determine the functional dependence transmissionloss on real powers of generatingplants. of There are several methods, approximate and exact, for developing a transmission loss model. A full treatmentof these is beyond the scopeof this book. One of the most important, simple but approximate, methods of expressing transmissionloss as a function of generatorpowers is through Bcoeffrcients. This method is reasonably adequatefor treatment of loss coordination in economic scheduling of load between plants. The general form of the loss formula (derived later in this section) using Bcoefficients is
It will be shown later in this section that, if the power factor of load at each bus is assumedto remain constant, the systemloss P, can be shown to be a function of active power generation at each plant, i.e.
Pr= Pt(Pcp P52, ..., P51r)
P. = II
where
pG^B*,pGn
(7.26)
m:7 n:I
(7.2r)
Thus in the optimization problem posedabove,Pci Q = I,2, ..., k) are the only control variables. For optimum real power dispatch,
PG^, PGr= real power generationat m, nth plants B^n= loss coefficients which are constantsunder certain assumed operatingconditions If P6"sare in megawatts,B*n are in reciprocal of megawatts*. Cemputations,of ,: course,may be carried out in per unit. Also, B*r= Bn^. Equation (7.26) for transmissionloss may be written in the rnatrix form as
')
AL = dC, \r ,1Pt ^+ A d PGto Po , # *:O'
i = r' 2' " " k
(7.22)
Rearranging Eq. (7 .22) and recognizing that changing the output of only one plant can affect the cost at only that plant, we have
Pr= PIBPI Where
(7.27)
= ) or
where
Li=
#Li=
) , i = r , 2 ,. . . , k
(7.23)
QAPL iAPGi)
(7.24)
It may be noted that B is a symmetric matrix. For a three plant system, we can write the expressionfor loss as
is called the penalty factor of the ith plant. The Lagrangianmultiplier ) is in rupeesper megawatthour,when fuel cost is in rupees per hour. Equation (7.23) impiies that minimuqr ftrei eost is obtained,when the incrementalfuel cost of eachplant multiplied by its penalty factor is the same for all the plants. The (k + 1) variables(P6r, P62,..., Pct, )) canbe obtainedfrom k optimal dispatchF,q. (7.23) together with the power balance Eq,. (1.19). The parrial derivative)PLIAPGi is referred to as the incrementaltransmission loss (ITL),, associated with the lth generatingplant.
PL= Bn4, + Bzz4, + Bzz4, + 28rrpcrpcz + ZBnpGzpG3 + 2BrrPorpo, e.ZS) wiih the system Dowerioss morieias per Eq. e.z6), we c.an now write ^ f t o_, Ap I
*
= ftl?lPo,"a^^'o')
x Base MVA
*B^,
{in pu) = B^n (in Mwt)
ffif
Modern Po*ersystem Anatysis
k k
p, 4. Calculate =If
j:l J=l
pciBijpcj.
It may be noted that in the aboveexpressionother termsare independentof Po, and are, therefore, left out. Simplifying Eq. (7.29) and recognizing that B, = Bir, we can write
5. Check if power balanceequation (?. t o) is satistied
l
lLPot
 PD pr.l' t
s
l
(a specified value)
a k + P=D zBijpcj
)Fo, j:l
(7.30a)
ln*t I If yes, stop. Otherwise, go to step 6. 6. Increase) by A) (a suitable step size)' tt [* pc, ^ ") <0or
Assuming quadraticplant cost curves as C i (P o i )= * o ,4 ,+ b ,p .,+ d,
We obtain the incremenlalcost as dP'
decrease ) by AA (a suitable stepsize);if
(8",
 po p,
) > 0 ,
dc, = atPo'+ b'
(7.30b)
from above in the coordination
repeatfrom step 3.
Substinrting APL&G' and dcildPci Eq. (7.22), we have
k
Example7,4
'
a i P c i +b , + S l z n , , r o , : ^
J:l
(1.3I)
Collecting all terrnsof P, and solving for Po, we obtain
k
A twobus system is shown in Fig. 7.8. If 100 Mw is transmited from plant 1 to the load, a transmission loss of 10 MW is incurred. Fin( the required generationfor each plant and the power received by load when tie system ,\ is Rs 25llvlWh. The incrementalfuel costs of the two plants are given below: dC
(ai + 2M,,) Pci=  )D
'l
bi + ^ ZBijpGj
J*T
r t&
A*
'28,,P ) L""ii'Gi
]:I
 o'ozPcr 16'oRs,Mwh +  o'o4PG2+ Rs/lvIWh 2o.o
^ la)
n ' t 
:f:
i7'2'"''k (7'32)
Pc,
,, *rB=; )
For any particularvalue of \, Eq. (7.32) can be solved iteratively by assuminginitial valuesof P6,s(aconvenientchoiceis P", = 0; i = l, 2, ..., k). Iterations are stoppedwhen Po,s converge within specified accuracy. Equation (7.32) along with the Dower balance F,q,. (7.19) for a partieular ioaci ciemanci are soiveciiteratively on the following lines: Po 1. Initially choose) = )0. 2 . A s s u m eI t G i =0 ; I = 1 , 2 , . . . , k . 3. Solve Eq. (7.32)iterativelyfor Po,s.
Fig. 7.8 A twobus systemfor Example 7.4
outuuon Therefore
6^rr:
since the ioad is at bus z a\one,p", v,.r,!not have any effect ort Fr. B z z =0 a n d B n = 0 = B z ,
Hence Pl= For Po, 
B nPbr 100MW, Pr = 10 MW, i.e.
(i)
"
l
Mod"rnPo*"r.Syrt"r An"lyri, 10= Bn (100)2
At plant 2 the load increases from 37.59 MW to 125 MW due to loss coordination.The saving at plant 2 is (ii) (iii) =  Rs 2,032.431hr The net saving achieved by coordinatinglosseswhile schedulingthe received r' load of 237.04 MW is = 2,937.69 2,032.43 Rs g,OtS.Ztm, Derivation of Transmission Loss Forrnula
I
Brr= 0'001MWr (7.31) for plant 1 becomes Equation
A.A2P;1+2^;B1P;1 +2^aBr2P62= ^76 and for plant 2 ) 0.04Pc2+2ABrrP"r+Z)BuPcr= 20 Substituting the values of Bcoefficientsand )  25, we get Pc t = 1 2 8 ' 5 7M W Pc z = 1 2 5 M w The transmissionpower loss is Pr = 0.001 x (128.57)' = 16.53 MW and the load is Po = Pct * Pcr Pt = 128.57+ I25  16.53  237.04 MW
An accurate method of obtaining a general formula for transmission loss has been given by Kron [4]. This, however,is quite complicated.The aim of this article is to give a simpler derivation by making certain assumptions. Figure 7.9 (c) depicts the case of two generating plants connected to an arbitrary number of loads through a transmission network. One line within the network is designatedas branch p. Im,agine that the total load current 1, is supplied by plant 1 only, as in Fig. 7.9a. Let the current in line p & Irr.Define
Considerthe system of Example 7.4 with a load of 237.04 MW at bus 2. Find the optimum load distribution between the two plants for (a) when losses are included but not coordinated,and (b) when losses are also coordinated. Also find the savings in rupeesper hour when losses are coordinated. Solution Case a If the transmission loss is not coordinated, the optimum are schedules obtainedby equatingthe incremental fuel costsat the two plants. Thus (i) 0 . 0 2 P + 1 6 = 0 . O 4 P c z + 2 0
\
(7.33)
to The powerdelivered the loadis + Pct * Pcz= 0.001P4r 237.04 we P6.2, get Bqs.(i) and(ii) for P61nncl Solving Pcr = 275.18 MW; and Prr2= 37.59MW
(ii )
Caseb This caseis alreadysolvedin Example 7.4. Optimum plant loadings with loss coordination are MW; Pcz = 125 MW Pct= 128,57 the Loss coordinationcauses load on plant I to reducefrom 275.18 MW to 128.57MW. Therefore,saving at plant I due to loss coordination is
Sirniinriv
(c) Flg. 7.9 Schematic diagram showing two plants connectedthrough a power network to a number of loads
with nient r^_^, ) qinrnc ctrnnirrino the fntel lnqd n s ^ r vn t r v n r a rr t ' E irc . \r 5 ? ok\ ,.rv), r'a wv ^av(ltl
define
Kgi:",+
: o.MPl,r 16)dPc, +r6Pctli),'r',
= Rs 2,937.691hr
Mrz= i=
tD
I^n
e.34)
Mo1 wrd Mp2 are called curcent distribution factors. The values of current distribution factors depend upon the impedances of the lines and their interconnection and are independentof the current Ip.
fW
t
powerSygtem Modern Anatysis
When both generators t and 2 are supplying current into the network as in Fig.7.9(c), applying the principle of superpositionthe currentin the line p can be expressed as where 1ot and Io2 are the currentssupplied by plants I and 2, respectively. At this stage let us make certain simplifying assumptions outlined below: (1) All load currentshave the same phase angle with respectto a common refere{ce. To understandthe implication of this assumptionconsider the load current at the ith bus. It can be written as VDil I (6t d) = lloil l1i where {. is the phaseangleof the bus voltage and /, is the lagging phase angle of the load. Since { and divary only through a narrow range at various buses, it is reasonable assumethat 0,is the same for all load currentsat all times. to (2) Ratio X/R is the same for all network branches. Thesetwo assumptions lead us to the conclusion that Ip1 andI, fFig.7 .9(a)) have the samephase angle and so have Ioz and 1o [Fig. 7.9(b)], such that the current distribution factors Mr, and Mr, are real rather than complex. Let, Ict = llcrl lo, and lcz = 1162l lo2 . where a, and 02 are phaseanglesof 1", and lor, respectiveiywith respect to the common reference. From Eq, (7.35), we can write llrl2  (Moll6l cos a1 + Mpzllo2lcos oz)2a (Mrll6lstn Mr2lls2lsn oz)2 Expanding the simplifying the aboveequation, we get ll,,l2= Mzrrllorlz + tutf,zllczlz 2MolMrzllctl llGzlcos  oz) (a1 +
Now
e, = lstrp( Re
p
Substitutingfor llrl2 fromEq. (7.37), and l1o,l and llurl from Eq. (7.38),'we obtain
D
"
rLurn, 1v31"*fr) p
D2 r Gr
*Wl,r,rMpzRp
lVllv2lcos /, cosQ, 7
P
*'
(7 Equation 3e) "T,o:
8 'n.. 
P3' M32RP tvzP("o, dritT
+ 4,8,,
(7'3s)
;:;::::;:,PczB,z
WGoshfDmS,no p
Brz Bn =
ffiDM"M"Ro
lV,l' (cos )' 6
(7.40)
o; (7.36)
(7.37) (7.38)
T . M ' , rL Y Y R. 7 The terms Bs, Bp and 82, are called loss cofficients or Bcofficients.lf voltagesare line to line kV with resistances ohms, the units of Bcoefficients in are in MWI. Further, with Po, and Po, expressed MW, P, will also be in in
l s.rr l = = ' ? ' l
$ l v tl c o s /,
:II.J' vL
Jl l vrl cosf"
&z
MK
where Pot and Po, are the threephase real power outputs of plants I and 2 at power factors of cos (t, and cos Q2, and yl and V2 are the bus voltages at the plants. ff Ro is the resistanceof branchp, the total transmissionloss is given by*
'The general expression for the power system with t plants is expressedas
The\,bove results can be extendedto the general case of ft plants with transrnishur loss expressedas
k k
P, =Df
m:l n:l
PG^B^.PG.
(7.41)
where 8 r,,, = cos (a,,  on)
P'r
'nt"P' ' .r4rlu*ry ' L' 1yf 1*r6f T,*3,R,t. tVr,l2 1cosffiLu'
Pc^ Pcn cos(a,  on) lV* ilV, l c o s Q * c o s Q ,
F3'
l v  l l v , l c o s Q c o s Q 
(7.42)
It can be recognized as
Pr=ftrB, * ...* P'o4oo * zDpG^B^npGn
m,n:l
;!l
i?ffil
I
..#
power nnatysis ruodern system
r !nrr!na! i i
ffi
i
alreadyare necessary, includingthosementioned The following assumptions if Bcoefficients are to be treated as constantsas total load and load sharing plants vary. Theseassumptions are: between 1, All load currents maintain a constantratio to the total current. 2. Voltage magnitudes at all plants remain constant. 3. Ratio of reactive to real power, i.e. power factor at each plant remains constant. 4. Voltage phase angles at plant buses remain fixed. This is equivalent to assuming that the plant currents maintain constant phase angle with respectto the common reference,since sourcepower factors are assumed constant as per assumption3 above. made, it is fortunate that treating BIn spite of the number of assumptions coefficients as constants,yields reasonablyaccurateresults, when the coefficients are calculated for some average operating conditions. Major system changes require recalculation of the coefficients. by Lossesas a function of plant outputscan be expressed other methods*,but the simplicity of loss equationsis the chief advantageof the Bcoefficients method. Accounting for transmission losses results in considerable operating economy.Furthermore,this considerationis equally important in future system planning and, in particular, with regard to the location of plants and building of new transmissionlines:
Y
lr"
t''
lb1 ' '

b
Refbus v =110"pu
Y
l,o
I ro"o r
Fig. 7.10 Samplesystem Example of 7.6 Solution As all load currentsmaintain a constantratio to the total current, we have _ 3. 6_ jo. g _ o. t 8z6 I, + Id 4. 6 jl. l5 I" I,+ld rd
r:  "i0.25  0 . 2 1 7 4 j1.1s 4.6
Mor= L, Mtr   0.2174, Mrr = 0.2774,Mu = 0.7826 plants 1 and 2 connectedto buses 1 and Figure7.10 shows a systemhaving trrvo 2, respectively. There are two loads and a network of four branches. The bus reference with a voltageof l.0l0o pu is shownon the diagram.The branch cunents and impedancesare: I,=7  j0.25Pu I o = 2  7 0 . 5P u I u = 1 . 6 j 0 . 4 P u Zo = 0.015 + 70.06pu Id = 3.6  70.9 Pu Z, = 0.OI + 70.04 pu M,,2= 0, Mnz = 0.7826, Mrz = 0.2174,Mrtz= 0,7826 Since the sourcecurrentsare known, the voltagesat the sourcebusescan be calculated.However, in a practical size network a load flow study has to be made to find power factors at the buses,bus voltagesand phase angles. The bus voltages at the plants are Vr = 1.0 + (2  j0.5) (0.015+ 70.06) = 1.06+ jO.I725 = 1.066 16.05" pu Vz= 7 + ( 1. 6  jO . 4) ( 0. 015 70. 06) + = 1.048+ jO.O9= 1.051 14.9" pu The current phaseanglesat the plants are (1, = Io, 12= 16r Ir)  ^O :  l4oi o2: t an r ot = t an t+! . 2 2 . 6 9t :  l4o cos (or ot) = cos 0o = 1
more accuratemethodsand exactexpressionfor 0P,./0P6i,references 122,231 may be consulted. *For
Za = 0.Ol + 70.04pu Zo = 0.015 + 70.06 pu Calculatethe loss formula coefficients of the systemin pu and in reciprocal
mesawatfs if the hase is 100 MVA
^^^D
The plant power factors are pfi = cos (6.05" + l4') = 0.9393
ffiffi
Mociern PowerSvstem Ana[,sis
p
OPtimal System Operation
Pfz = cos (4.9 + 14") 0.946 The losscoefficients lBq. Q.a\l are
ffi ffi$ffi l
c =f
ci(Pci)
(7.7)
subject to the load flow equations [see
 0.02224 pu Bzz q2 0.0 5 x (0.7 82q2 + 0.0 x (0.217 + 0.01 (0.7 1 x 82q2 (1.051)2 x(0.946)2
and
e f
j:1
tUilvjily,,tcos(0,,
(7.43)
3 Q, + Ltvinvjlllzulsin (0ul
j:r
(7.44)
= 0.01597 pu x(0.7826)2 + 0.01x(0.217a)2 0.01 + D = Dt2 _ e0.2174)L0.7826X0.015) rJ66 . Lotl x 0.9393 0.946 x = 0.00406 pu For a baseof 100 MVA, theseloss coefficients must be dividedby 100 to obtaintheir values units of reciprocal in megawatts, i.e.
0.02224 LLL+ h  0.02224 x lo2 Mwl Drr =
A 4  Llvllvjlly,,lcos(9,, * 6i  4)= 0 for each pv bus j:7 It is to be noteclthat at the ith bus Pt= Pci Poi Qi= Qci Qu where Po, and ep; areload demands at bus i. Equarions(7.43), (7.44) and (7.45) can be expressed vector form in (7.a3)l [Eq. I
(7.4s)
(7.46)
100
0'01597 = 8.t.,= 0.01597x 1o2Mwl 100 Br.t = 0'0M06 = 0.00406 x loz Mwl 100 LOAD FLOW SOLUTION
'
f (x, y) =
 _tn. \7qql lEq.Q.a, lty,tl
where the vector of dependentvariables is
foreach bus r0  pV foreach busi I
;
\
(7.47)
7.6
OPTIMAL
The problem of optimal real power dispatch has been treated in the earlier section using the approximate loss formula. This section presentsthe more general problem clf real attd reactive povrer flow so as to minirnize the instantaneous operatingcosts.It is a static optimization problem with a scalar objective function (also called cost function). The solutiontechniquegiven here was first given by Dommel and Tinney on [34]. It is based load flow solutionby the NR method,a first order gradient adjustmentalgorithm for minimizing the objective function and use of penalty functions to accountfor inequality constraintson clepenclent variables.The problem of unconstrained optirnalload flow is first tackJed.Later the inequality constraints are introduced, first on control variables and then on dependent variables. Optimal Power Flow without Inequality Constraints
Ld, for eachpV bus_J andthe vectorof independent variables is
.= lr, jforeachrouusf t r 4I n
(7.48a)
a. slack bus 4i 4l pe foreach bus ]= O]
(7.48b)
t"each bus PV ,1,,,j
fn p o w fha e ql ^.r q vv v vo r . r r f^*,,.t^+:^rtlrurulalLlull, r u D l l l r [fle l w ODleQtlVe l L l L l tltnefinn g L l l mrrcf E S r i_^1,A^ a c K d^^ D u _r S
The objective function to be minimized is ihe operating cost
The vector of independent variabresy can be partitioned into two parts_a vector u of control variableswhich are to be variea to achieve optimum value of the objective function and a vector p of fixed or disturbangeor unconhollable
ffi@
Analvsis PowerSvstem Modern
feasible solution point (a set of valuesof x which satisfiesEq. (7.5a) for given u andp; it indeed is the load flow solution) in the direction of steepestdesceht (negative gradient) to a new feasible solution point with a lower value of objeetlve funstion. By repeating the.se moves in *rc dkestisn +f the negadve gradient, the minimum will finally be reached. The computational procedurefor the gradientmethod with relevant details is given below: Step I Make an initial guess for u, the control variables.
P6t on magnitudes PV buses, may parameters* be voltage puru"t"rs. Control
at buseswith controllablepower, etc. The optimization problem** can now be restatedas min C (x' u) subject to equalitYconstraints .f (x, u, p) = 0
(7.4e)
(7'50)
To solve the optimization problem, define the Lagrangian function as (7'51) L (x, u, p)= C (x, u7+ Arf (x, u, P) where ) is the vector of Lagrange multipliers of same dimension asf (x, u, p) Lagrangianfunction conditionsto minimize the unconstrained The necessary (see Appendix A for differentiation of matrix functions). are
Stey 2  Jind a feasible load flow solution from Eq. (7.54) by the NR iterative method. The method successively improves the Solution x as follows. (r +r) * ,(r) + Ax where Ar is obtained by solving the set of linear equations (6.56b) reproduced below:
a f ,= 0 c* l y 1 ' ) _ o
0x 0x L}x J
(7.s2) (7.s3) (7.s4)
0 L = 0 c* l y 1 ' ) _ o
0u 0u LduJ
y) l#r,"',rl]4" r (*('),
rhe end resurts J,;; $; Iti.[l]'*u,,on "^,
Step 3 Solve Eq. (7.52) for
r,
,
a r = (x,u,P) = o f u;
Equation (7.54) is obviously the same as the equality constraints.'The expressionsfor tS as neededin Eqs. (i.52) and (7.53) are rather ^a L ; 0u by involved***.It may howeverbe observed comparisonwith Eq. (6.56a)that Y= Jacobian matrix [same as employed in the NR method of load flow 0x for solution; the expressions the elementsof Jacobianare given in Eqs. (6.64) and (6.65)1. Equations(7.52),0.53) and (7.54) are nonlinearalgebraicequationsand can only be solvediteratively. A simple yet efficient iteration scheme,that can by employed, is the steepestdescentmethod (also called gradient method).
or x and racobian the matri,...
\ =  i( { \ ' l
I
^ rrl
ac
0x
L\dxl J
(7.s5)
Step 4
Insert ) from Eq. (7.55) into Eq. (7.53), and compute the gradient
= Y.c, oc *l 9L1' t
0u L 0 u)
(7.s6)
bus voltage and regulating transformer tap setting may be employed as buses' additional control variables.Dopazo et all26ltuse Qo, as control variable on with reactive Power control'
**rr *L^ ..,.*^Dlvrrl LrMJ
lI lvsr
Slack
It may be noted that for computing the gradient,the JacobianJ  + is already 0x known from the load flow solution (step 2 above). step 5 rf v c equals zero within prescribedtolerance,the minimum has been reached. Otherwise, Step 6 Find a new set of control variables
unew= l.l.^6* L,il
.pal nnrrrcr lncc ic to he minimizedrv
the obiective
function
is
( 7. 57\
C = Pr(lVl, 6) the real Since in this case the net injected real powers are fixed, the minimization of at the slack bus is equivalent to minimization of total system loss, injected power P, This is known as optimal reactive power flow problem' ***The original pup". of Dommel and Tinney t34l may be consulted for details'
where L,u =  o"VC,
(7.s8)
Here A,u is a step in the negative direction of the gradient. The step size is adjustedby the positive scalar o..
#*ff"
f
uooern Power Analvsis Svstem
to the constraint limits, when these limits are violated. The penalty function method is valid in this case, becausethese constraints are seldom rigid limits in the strict sense,but are in fact, soft limits (e.g. lvl < 1.0 on a pebus really
Steps 1 through 5 are straightforward and pose no computational problems. Step 6 is the critical part of the algorithm, where the choice of a is very important.Too small a value of a guarantees convergencebut slows down the the rate of convergence; too high a value causes oscillations around the Inequality Constraints on Control Variables
ry
vshould not exceed 1.0 too much and lvl = 1.01 may still be
The penalty method calls for augmentationof the objective function so that the new objective function becomes
Though in the earlier discussion, the control variables are assumedto be unconstrai""o, are, in fact, always contrained, 1.j":T.1",::tutt e.g.
C t = C ( x , u )f*U t
(7.63)
(7.se)
Pc,, ,nin1 Po, S Pct, ** These inequality constraintson control variablescan be easily handled.If the correctionAu,inBq. (7,57) causes uito exceed one of the limits, a, is setequal to the correspondinglimit, i.e. if u,,oro Au, ) ui,^^ * 7f u,,oro Au, 1ui,^in * otherwise After a control variable reaches any of the limits, its component in the gradient should continue to be computed in later iterations, as the variable may come within limits at some later stage. In accordance with the KuhnTucker theorem (see Appendix E), the necessaryconditionsfor minimization of I, under constraint (7.59) arc:
where the penalty W, is introduced for each violated inequality constraint. A suitable penalty function is defined as  xi,^o)2 i w, = {7i@i ' [ 71G ,  xi, ^i) zi whenever ) xi,rnax xi whenever xr ( r y, m in Q '64)
where Tiis Treal positive number which controls degreeof penalty and is called the penalty factor.
(7.60)
Xmln
0L :0 0u,
7 f u ,,* n < ui < ui ,^^, if u,  ui,^* u i : u i .^u*
Fig. 7.11 Penalty function
of .o ou,
0r,
(7.6r)
o r , o
A plot of the proposedpenalty function is shown in Fig. 7.11, which clearly indicates how the rigid limits are replaced by soft limits. The necessaryconditions (7.52) and (7.53) would now be modified as given below, while the conditions (7.54), i.e. load flow equations,remain unchanged.
T
Thereforer now, in step 5 of the computationalalgorithm, the gradient vector has to satisfy the optimality condition (7.61). Inequality Constraints on Dependent Variables
ax_ac,\awj ,fafl', = __l_ )
dx
o x 4l a r * L a " i )  o
)
(7.6s) (7.66)
:
ax_AC,sdw; ,fAf1',
= L
Often, the upper and lower limits on dependentvariables are specifiedas rmir,SxSr*u^
e.g. lUn,in < lVl < lYl .o ofl a PQ bus
du ou'+ a"*La"l'r:o
'Ihe
(7.62)
vecto
UW:
^trZ
0x
obtained from Eq. (7.64)would containonly one nonzero.
Such inequality constraints can be conveniently handled by the penalty function method. The objective function is augmented by penalties for inequality constraintsviolations. This forces the solution to lie sufficiently close
= 0 asthepenalty # functions dependent on variables independent the control variables. are of
termcorresponding the dependent to variable while x;;
'
ModernPowerSystemAnalysis Mathematical Formulation
By choosinga higher value fot 1,,the penalty function can be made steeper however,will that the solutionlies closer to the rigiA fimits; the convergence, so becomepoorer.A good schemeis to start with a low value of 7j and to increase imization process,if the solution exceedsa certain tolerance limit. This sectionhas shown that the NR method of load flow can be extendedto yield the optimal load flow solution that is feasible with respectto all relevant inequality constraints.These solutions are often required for system planning and operation. 7.7 OPTIMAL SCHEDULING OF HYDROTHERMAL SYSTEM
For a certainperiod of operation7 (one year,one month or one day, depending. upon the requirement),it is assumedthat (i) storageof hydro reservoir at the reservoir (after accounting for irrigation use) and load demandon the system are known as functions of time with complete certainty (deterministic case). The problem is to determine q(t),,the water discharge(rate) so as to minimize the cost of thermal generation. Cr =
rT
JoC
( Por ( t ) ) dt
(7.67)
The previous sectionshave dealt with the problem of optimal schedulingof a Dower system with thermal plants only. Optimal operating policy in this case can be completely determined at any instant without reference to operation at other times. This, indeed, is the static optimization problem. Operation of a system having both hydro and thermal plants is, however, far more complex as hydro plants have negligible operating cost, but are required to operate under constraintsof water available for hydro generationin a given period of time. The problem thus belongs to the realm of dynamic optimization. The problem of minimizingthe operating cost of a hydrothermal sYstemcan be viewed as one of minimizing the fuel cost of thermal plants under the constraint of water availability (storage and inflow) for hydro generation over a given period of operation. J (waterinflow)
under the following constraints: (i) Meeting the load demand
Pcr(r) * Pca(t)  Pr(t)  PoG) = 0; te 10,71
This is called the power balance equation. (ii) Water availability
(7.68)
X (T) x, (o) l' t@ at+ lrqgl =o dt JO JO_
(7.6e)
where J(t) is the water inflow (rate), X'(t) water storage,and X/(0) , Y (T) arc specified water storages at the beginning and at the end of the optimization interval. (iii) The hydro generation Pcrlt) is a function of hydro dischaige and water storage(or head), i.e.
Pcn(r)=f(X'(t),q(t))
(7.70)
The problem can be handled conveniently by discretization. The optimization interval Z is subdivided into M subintervalseach of time length 47. Over each subinterval it is assumed that all the variables remain fixed in value. The problem is now posed as
n^(EL : u^r^4r:....*r"$^r'(pt),rfrrrt,
system hydrothermal Fig. 7.12 Fundamental the problem formulation and For the sake of simplicity and understanding, solution technique are illustrated through a simplified hydrothermal system of Fig. 7 .I2. This systemconsistsof one hydro and one thermai piant suppiying power to a centralized load and is referred to as a fundamental system. Optimization will be carried out with real power generation as control for loss accounted by the loss formula of Eq. (7.26), variable,with transmission
__^_1_ Q
\t.lr)
under the following constraints: (i) Power balance equation
Pt *PtrPIPt =0
where
(7.72t
Ptr = thermal generation in the mth interval Ptn = hydro generation in the zth interval PI =transmission loss in the rnth interval
ffiffifjn
= D77\r57
n tnm
Modern PowerSystem Analysis
t2
)
t

.rn
LDTHTGH
nm
D tn fI Dpy\r6p m
t2
)
Optimal System Operation
PI = load demand in the mth interval (ii) Water continuity equation y^ _ yt(m_r)_ f AT + q^ AT = 0
l e = water head correction factor to account for head variation with storage p = non:effective discharge (water discharge neededto run h dro In the aboveproblem formulation, it is convenient to choosewater discharges in all subintervalsexcept one as independentvariables, while hydro generations, thermal generations and water storagesin all subintervalsare treated as dependentvariables. The fact, that water discharge in one of the subintervals is a dependentvariable, is shown below: Adding Eq. (7.73) for m = l, 2, ..., M leads to the following equation, known as water availability equation
ffiffi
X ^ = water storage at the end of the mth interval J*  water inflow (rate) in the mth interval q^ = water discharge (rate) in the ruth interval The above equation can be written as (7,73) Y  X^ r  J * + q ^ = 0 ; m = I,2, ,.., M = X/*IAT = storagein dischargeunits. where Y InEqs. (7.73), Xo and XM are the specified storagesat the beginning and end of the optimization interval. (iii) Hydro generation in any subinterval can be expressed*as
xM
"oD
m
J^ +la^ = o
m
(7.7s)
Ptn = ho{I + o.5e(Y + Y)l
where
(q*  p)
(7.74)
Becauseof this equation, only (M  l) qs can be specifiedindependently and the remaining one can then be determined from this equationand is, therefore, a dependent variable.For convenience, is chosenas a dependent ql variable, for which we can write
ho=9.8rx rorhto
fto = basic water head (head corresponding to dead storage)
qt = xo  xM* DJ^ Dn^
Solution Technique
M
(7.76)
* where
P3, = 9.81 x 1o2hk@^ ,p ) Mw
(q^  p) = effective discharge in m3ls hry, = average head in the mth interval Now
The problem is solved here using nonlinear programming technique in conjunction with the first order gradient method. The Lagrangian L is formulated by augmenting the cost function of Eq. (7.7L) with equaliry constraintsof Eqs. (7 .72) (7.74) throughLagrangemultipliers (dual variables)
\i \i'and )i. Thus,
L T (X^ + X^ r) l ffi = 7 ro * 2A
where A = draa. crosssectionof the reservoir at the given storage of h' o = basic water head (head correspondingto dead storage, hk= hLll + o.Se(x'+ X"t)l where
=D tc(%r) xT(4r+ 4,  ry ffi + M (y  ytr* + .c, qr)* ^Ttp1, h,(r + 0.5e(y it11* @^ p)rj (7 * .77)
The dual variables are obtainedby equating to zero the partial derivatives of the Lagrangian with respectto the dependentvariables yielding the following equations
AP rJ)e = Ar,rDmt ut/\I CT) /  A ,zl I nr l
AT
Ahto Now
.
7Pt
dPt
'[
rt. Uf t
\
7Pt )
 ^ lu
0 /78)
4n= ho {! + o.Se(x^+ x^t)l @^  P) where h o = 9 .8 7x l 0 3 h to
[The reader may compare this equation with Eq. (7.23)]
#r,M^r['ffi)='
(7.7e)
ffil
a
Analysis Modern Power system
(+) = )7, Ar*' \r p.sh"r(q*p)l )i*t 1o.5ho e I ax^ /)^**t \ *U
*'(q^ DI  o
(7.80)
and using Eq. (7.73) in Eq. (7.77), we get
iently by augmenting cost functioniith p"nulty functionrur air";;;i" the Sec. 7.6. The methodoutlinedaboveis quitegeneral and canbe directlyextended to
a slzstemhaving multi hydro and multither+nal plan+s. The method, however, has the disadvantage of large memory requirement, since the independent variables,dependent variables and gradientsneed to be storedsimultaneously. A modified techniqueknown as decomposition[24) overcomes this difficulty. In this technique optimization is carried out over each subinterval and the complete cycle of iteration is repeated, the water availability equation does. if not check at the end of the cvcle.
(pt) = )rr t\n" fl + 0.5e (zy + Jt  zqt +d)= 0
laq' )
(7.81)
The dual variables for any subinterval may be obtained as follows: (i) Obtain { from Eq. (7.78). (ii) Obtain )! from Eq. (7.79). (iii) Obtain )1, from Eq. (7.81) and other values of ry (m * 1) from Eq. ' ( 7.8 0 ). The gradient vector is given by the partial derivatives of the Lagrangian with variables. Thus respectto the independent
(+)
\oq )m+r
=\f. ^Zh"{r + 0.5e(2Yt + J^  2q^+ p)}
(7.82)
For optimality the gradient vector should be zero if there are no inequality constraintson the control variables. Algorithm Assume an initial set of independent variables q* (m*I) subintervalsexcept the first. for all
Consider the fundamental hydrothermal system shown in Fig. 7.I2. The objective is to find the optimal generationschedulefor a typical dav, wherein load varies in three steps of eight hours each as 7 Mw, 10 Mw and 5 Mw, respectively. There is no water inflow into the reservoir of the hydro plant. The initial water storage in the reservoir is 100 m3/s and the final water storage should be 60m3/s,i.e. the total water available for hydro generationduring the day is 40 m3/s. Basic head is 20 m. Water head correction factor e is given to.be 0.005. Assume for simplicity that the reservoir is rectangular so that e doesnot ehange with water storage. Let the noneffective water discharge be assumed as 2 m3/s.Incremental fuel cost of the thermalplant is dc r.opcr + 25.0 Rsft' dPcr Further, transmissionlosses may be neglected. The aboveproblem has been speciallyconstructed(ratherovelsimplified) to illustrate the optimal hydrothermal schedulingalgorithm, which is otherwise computationally involved and the solution has to be worked on the digital computer. Stepsof one complete iteration will be given here. Since there are three subintervals,the control variablesre q2 and q3. Let us assumetheir initial values to be q2 = 75 m 3ls 15 m2ls The value of water diseharge in the first subinterval can be immediateiy found out using Eq. (7.76), i.e. er = LOO 60  (15 + 15) = 10 m3ls It is given that X0 = 100 m3/s and X3 = 60 m3/s. From Eq. (7.73)
Obtain the values of dependent variables Y, Ptu, F\r, qt using Eqs. (7.73), (7.74), (7.72) and (7.76). 3. Obtain the dual variables )f, \f, )i @ = 1) and )rr using Eqs. (7.78), (739), (7.80) and (7.81). 4. Obtain the gradientvector using Eq. (7.82) and check if all its elements are equal to zero within a specified accuracy.If so, optimum is reached. If not, go to step 5. 5. Obtain new values of control variables using the first order gradient method,i.e.
ek*=q;a++);m=r {\oq* )
(7.83)
where cr is a positive scalar. Repeat from step 2 In the solution techniquepresentedabove, if some of the control variables (water discharges)cross the upper or lower bounds, these are made equal to their respective bounded values. For these control variables, step 4 above is checked in accordancewith the KuhnTucker conditions (7.61) given in S ec .7. 6 .
W
po Modern Xt = f + Jr  gt = 90 m3ls
f=Xr+12q2=75m3/s The valuesof hydro generations the subintervals be obtained in can using Eq. (7.7q as follows: x P b , t = 9 , 8 1 1 0  3 2 0 [ l + 0 . 5 x 0 . 0 0 5 x r+ X 0 ; ) { q '  p ) x 1 = 0.1962 + 25 x 104 x 190} x 8 {I = 2.315 W M { 4 n = 0 . 1 9 6 2 I + 2 5 x l O a x 1 6 5 }x 1 3 = 3.602MW P Z a = 0 . 1 9 6 2l + 2 5 x 1 0 a x 1 3 5 }x 1 3 { = 3.411 MW The thermalgenerations the threeintervalsare then in pLr = pL pl, = J  2.515= 4.685MW 4r = Pto PL, = 10  3.602= 6.398MW F c r = p  ,  4 " = J  3 . 4 1 I= 1 . 5 8 9 W M From Eq. (7.78),we havevaluesof )i as dc(P€)
AGT V D m
Substituting various values, we get
xt,  x3 )l {o.Shoe  p)l _ S! 1O.Sn"e  p)l = 0 (qt @, x?, ll  ^Z {0.5 hoe(q, _ p)l _ )3, 1o.5ho,  p)] = o (qt
) t r = 8 . q 1 4  2 9 . 6 8 5( 0 . 5 x 0 . 1 9 6 2 x0 . 0 0 5 8 l _ 3 1 . 3 9 8 x {0.5x 0.1962 0.005x 13) x = 8.1574 )t, = 9.1574 31.398 (0.5x 0.tg6} x 0.005x t3)  26.589(0.5 x 0.1962 0.005x t3) = x 7.7877 UsingEq. (7.82),the gradientvectoris
(#)=
^ 7  f t n " { 1 + 0 . 5 0 . 0 0(5 x e o 2 x 1 5+ z ) l x 2
= 8.1574 31.398 0.1962 + 25 x 10a x x ,{l l52l =  0.3437
( ar.l  r3 A:2 strn,
lrf )
(2X2 fl + 0.5e + f  zqt+ p)I
_ \m
 ' t l
or
)T=P[,+25
 7.7877 26.589x 0.1962 + 25 x 10a x t22} {I = 0.9799 If the tolerance gradient vectoris 0.1, then optimal conditions fbr are not yet satisfied, since the gradient vector is not zero,i.i. (< 0.1); hence the second iteration will have to be carried out starting with the following new values of the control variablesobtained from Eq. (7g3)
Calculating ), for all the three subintervals, we have
[^i I
lr?l=lrr.lsal j Lr? fzo.ssrJ
Also from Eq. (7.79), we can write
685] [2e
lril
From Eq. (7.81 )
= case  ^3  ^?l: I I r:ea I for ttrelossless
hil lzs.68s1
lnk_l=la[,oLl#l Lq:"*J= L;i;l1+l
Loq" )
Let us take a = 0.5, then
.sag Lri Ld Lze J I I
\l'
^ l = A\hnii + o.5e(?)(0+ il  z q ' + o l  29.685 0.1962{l + 25 x loa (2oo 2o + 2)l x
and from Eq. {7.76)
:fli lf:i= ,'l:::Ijllil [[]
,8.474 FromEq. (7.E0)for m = 1 and2, we have
4'r*= 100  60  (15.172+ 14.510)= 10.31gm3ls The above computation brings us to the starting point of the next iteration. Iterations are carried out till the gradient vector U"comes zero within specified tolerance.
ffil
uodern Power Slrstem Analvsis
Hffiffi
miiiions of iriiocaiories per hour can be expressedas a function of power output Poin megawattsby the equation
PROB iEii/iS
7.1 For Example 7.1 calculate the extra cost incurred in Rsftr, if a load of 220 MW is scheduled Pct= Pcz = 110 MW. as 7.2 A constant load of 300 Mw is supplied by two 200 Mw generators,I and 2, for which the respective incremental fuel costs are dcr = o ' l O P G l+ 2 0 ' 0 Po, dcz o'lzPc2 + 15'o dPo, with powers Pc in MW and costs c in Rsar. Determine (a) the most economicaldivision of load betweenthe generators, and (b) the saving in Rs/day thereby obtained compared to equal load sharing between machines. 7.3 Figure P7.3 shows the incrementalfuel cost curves of generatorsA and B. How would a load (i) more than ZPo, (ii) equal to 2p6, and (iii) less than ZPo be shared between A and B if both generatorsare running.
0.00014+ O.$ft + r2.0po+150
Find the expressionfor inerementalfuel eost in rupeesper megawatt hour as a function of power output in megawaffs. AIso find a good linear approximation to the incremental fuel cost as a function of Fo. Given: Fuel cost is Rs Zhmltion kilocalories. 7.6 For the system of Example 7.4, the system ) is Rs 26a4wh. Assume further the fuel costs at no load to be Rs 250 and Rs 350 per hr, respectively for plants I and 2. (a) For this value of system ),, what are the values of p61, po, and, received load for optimum operation. (b) For the above value of received load, what are the optimum values of Pot and Por, if system losses are accounted for but not coordinated. (c) Total fuel costs in RsArr for parrs (a) and (b). 7.7 FigureP7.7 showsa systemhaving two plants I and 2 connected buses to 1 and 2, respectively. There are two loads and a network of three branches. The bus 1 is the reference bus with voltage of 1.0 I 0" pu. The branch currents and impedancesare Io=2 70.5 pu
L  o= 1 6 i O 4 n r r
1, = 1.8 i0.45 pu Zo= 0.06+ j0..24 pu Zt = 0.03+ J0.12pu
(MW)mtn P6
Flg. P7.3 7.4 Consider the following three IC curves PGr=100+50(IQt2Aqi Pcz=  150 + 60 (IQz  2.5 AqZ Pct=  8.0+ 4a Qq3  1.8 Aqi where ICs are in Rs/IVIWh and P6s are in MW. The total load at a certain hour of the day is 400 MW. Neglect transmission loss and develop a computer programme for optimum generationschedulingwithin and accuracyof + 0.05 MW. Note: All P6s must be real positive.
Z, = 0.03+ /0.I2 pu Calculate loss formula coefficients the system per unit and in the 6f in reciprocal megawatts, the baseis 100MVA if
Ref bus
Flg. P7.7
Samplesystemfor probtemp7.7
W
uoo"rnpo*", syrt"r Anutyri,
/. r\cuenswanoer, J.}(, Modern power systems, International rext York, 1971.
7.8 Fot the power plant of the illustrative exampleusedin Section 7.3. obtain the economically optimum unit commitment for the daily load cycle given in Fig. P7.8. Correct the scheduleto meet security requirements.
Book co., New
w
E,o
i"
0
8
1
2
16
20
24
Timein hours' Flg. P7.8 Dailyload curve for problemp7.9 7.9 Repeat Example 7.3 with a load of 220 Mw from 6 a.m. to 6 p.m. and 40 MW from 6 p.. to 6 a.m. 7.10 Reformulate the optimat hydrothermal schedulingproblem considering the inequality constraints on the thermal generation and water storage employing penaity functions. Find out the necessary equations and gradient vector to solve the problem.
8' singh, c. and R. Billinton, system Reliabitity, Modelling and Evaluation, Hutchinson London,1977. of 9. sullvan, R.L., power systempianning, McGrawHil, New york, 1977. 10' Wood, A'J' and B.F' Wollenberg,Power Generation,operation and Control, Znd edn., Wiley, New york, 1996 11' Mahalanabis, A.K., D.p. Kothari and s.I. Ahson, computer Aided power system Analysis and control, Tata McGrawHill, New Delhi. r9gg. 12. Bergen, A.R., power system Anarysis, prenticeHall, rnc., Ncw Jersey,19g6. 13' Billinton, R. and R.N. Allan, Reliability Evaluation of power System,plenum Press,New york, 1984. 14. Sterling, M.J.H., power SystemControl,I.E.E., England, 197g. 15. Khatib, H., "Economics power systemsReriability,', of Technicopy. r9g0. 16. warwick, K. A.E. Kwue and R. Aggarwar (Eds), A.r. Techniquesin power System.s, IEE, UK, 1997. 17. Momoh, J.A., Electic power System Applicationsof Optimization,Marcel Dekker, Inc., New York., 2001. 18' YongHuaSong (Ed.), Modern optimization Techniques power in Systems,Kluwer Academic Publishers,London, 1999. 19' Debs, A.S., Modern power systemscontror and operation, KAp, New york, 1988. 20. Berrie,T,W., Power System Economics,IEE,London, 19g3. 2r. Berrie, T.w., Electricity, Economicsand pranning, rEE, London, rp92.
D o n l, o ;, o r q Er
NCES REFERE
Books 1. Billinton,R., Power System Reliability Evaluation,Gqrdon and Breach,New york, t970. 2 . Billirrton, R., R.J. Ringlee and A.J. Wood. Power System Reliabitity Calculations, The MIT Press,Boston,Mass, 1973. 5 . Kusic, G.L., computer Aided Power system Analysis, prenticeHall,Nerv Jersey, 1986. 4 . Kirchmayer,L.K., Economicoperation of Power systems, John wiley, New york,
I9)6.
22. Meyer, w.s. and v.D. Albertson, ..Improved Loss Formula compuration by optimally ordered EriminationTechniques',, IEEE TransmpAs, 1971,90: 716. 23. Hill, E.F. and w.D. stevenson,J.R., ..A New Method of Determining Loss Coefficients", IEEE Trans. pAS, July 196g, g7: 154g. 24' Agarwal, S'K. and I.J' Nagrath,"optimal Schcduling of Hydrothermal Systems,,, Proc. IEEE, 1972, 199: 169. 25' Aytb, A'K' and A.D. Patton, "Optimal Thermal Generating Unit Commitment,,, IEEE Trans.,JulyAug 1971,pAS_90: 1752. 26' Dopazo, J'F. et al., "An optimization Technique for Real and Reactive power Allocation", Proc, IEEE, Nov 1967. 1g77. 27. Happ, H.H., "optimal power DispatchA Comprehensive survey,,, IEEE Trans. 1977,PAS96: 841. 28. Harker, 8.c., "A primer on Loss Formula", AIEE Trans, r95g, pt ilr,77: 1434. 29' IEEE CommitfeeReport, "Economysecurity Functions power in systemopoutionS". IEEE ,snprinl Ptthlirntinn 15 ,J f \r r v v rJn oao ( 7w7.w
5 . Kirchmayer, L.K., Economic control of Interconnectedsystems,wiley, New york, 1959. pergamonpress,New Knight, u.G., Power systems Engineeringand Mathematics, York. 1972.
30. Kothari, D.p., "optimal HydrothermalSchedulingand Unit commitment,. ph. D. Thesis,B.I.T.S, Pilani, 1975. 31' Kothari, D.P. and I.J. Nagrath,"security Constrained EconomicThermal Generating Unit Commitment,, J.I.E. (India), Dec. 197g, 59: 156. 32' Nagrath,I.J. and D.P. Kothari, "optimal Stochastic Scheduling Cascaded of Hydrothermal Systems", J.I.E. (India), June 1976, 56: 264.
D\rrD LrE YYr\, l\ew
IorK,
Iyl).
ffi
22 JJ. D^"lL^rvDwrrwtrt I J. ^r vr ^l cl.t
powersvstemAnalvsis Modern
../\r:^l r.,tPtlrrr.u t1^t^l \.VtlLfUl ^f Ul D ^ri^ I\jAL;IYtr nfUWEf Fr rr fIUW , rnnh IEDL ar lfQnS, <  IyO6, Qan vvrr. Q u.r ^i qrrs n D v.l . If ^+L^j nuluatl l,
mal Thermai Generadng'tjnir CommitmentA Review, "Int. J. EPES, Vol. 20, No. 7, Oct. 1998,pp. 443451. 52. Kulkarni, P.S., A.G. Kothari and D.p. Kothari, "Combined Econonic and Emission Dispatchusing ImprovedBpNN", /nf. J. of EMPS, Vol. 28, No. l, Jan 2000,pp. 3t  4_7 53. Aryu, L.D., S.C. Chaudeand D.P. Kothari, "EconomicDespatchAccounting Line Flow Constraintsusing Functional Link Network," Int. J. of Electrical Machine & Power Systems, 28, l, Jan 2000, pp. 556g. 54. Ahmad, A. and D.P. Kothari, "A practical Model for Generator Maintenance Scheduling with rransmission constraints", Int. J. of EMps, vol. 2g, No. 6, June 2000, pp. 501 514. 55. Dhillon, J.s. and D.P. Kothari, "The surrogate worth rrade off Approach for Mutliobjective Thermal power Dispatch hobelm". EpsR, vol. 56, No. 2, Nov. 2000, pp. 103110. 56. Son, S. and D.P. Kothari, "Large ScaleThermal GeneratingUnit Commitment:A New Model", in The Next Genera\ion of Electric Power (Jnit CommitmentModels, edited by B.F. Hobbs et. al. KAp, ,Boston,2001, pp. Zll22557. Dhillon, J.s., s.c. Parti and D.p. \ Kothari, ,,Fuzzy Decision Making in Multi objective Long;term scheduling of Hydrothermal system,,, rnt. t. oy erns, vol. 23, No. l, Jan 2001,pp. lg29. 58. Brar, Y.s., J.s. Dhillon and D.p. Kothari, "Multiobjective Load Dispatch by Fuzzy Logic based Searching Weightage Pattern," Electric power Systems Research,Vol. 63, 2002, pp. 149160. 59. Dhillion, J.s., S.c. Parti and D.p. Kothari, "Fuzzy Decisionmgkingin stochastic Multiobjective shortterm Hydrothermal scheduling," Ip,B proc.tcTD, vol. l4g, z, March 2fi02, pp l9i200. 60' Kothari, D.P., Application of Neural Netwdrks to Power Systems(Invited paper), Proc. Int. Conf., ICIT 2000, Jan. 2000, pp. 62l_626.
../\.:t \_rpl,llllal
P A S .8 7 : 4 0 . 34. Dommel, H.w. and w.F. Trinney, "optimal power Flow solution", IEEE Trans., October1968,PAS. 87: 1866. 35. Sasson,A.M. and H.M. Itlerrill, "SonneApplications of Optimization Techniques to Power SystemProblems", Proc. IEEE, July 1974,62: 959. 36. wu, F. et al., "A Twostage Approach to solving optimal power Flows", proc. 1979 PICA Conf. pp. 126136. 37. Nanda, J., P.R. Bijwe and D.P. Kothari, "Application of Progressive Optimality Algorithm to Optimal Hydrothermal Scheduling ConsideringDeterministic and StochasticData", International Jountal of Electrical Power and Energy Systems, January1986, 8: 61. 38. Kothari, D.P. et al., "Some Aspects of Optimal Maintenance Scheduling of GeneratinglJnits", "/./.E (India), August 1985, 66: 41. 39' Kothari, D.P. and R.K. Gupta, "Optimal StochasticLoad FIow Studies", J.I.E. (India), August 1978,p. 34. 40. "Description and Bibliography of Major Economysecurity FunctionsPart I, il, and III", IEEE committeeReport,IEEE Trans. Jan 1981. pASr00, zlr235. 41. Bijwe, P.R., D.P., Kothari, J. Nanda, and K.S. Lingamurthy, .,Optimal Voltage Control using ConstantSensiiivity Matrix", Electric Power SystemResearch, Vol. II, No. 3, Dec. 1986,pp. 195203. 42. Nanda,J., D.P. Kothari and K.S. Lingamurthy, "Economicemission Load Dispatch through coal Programming Techniques",IEEE Trans.on Energy conversion, vol. 3, No. 1, March 1988,pp. 2632. 43. Nanda, J.D.P. Kothari and s.c. srivastava, "A New optimal power Dispatch Algorithm using Fletcher's QP Method", Proc. IEE, pte, vol. 136. no. 3, May p 1989, p. 153161. 44. Dhillon, J.S., S.C. Parti and D.P. Kothari, "stochasticEconomicEmissionLoad Dispatch", Int. J. of Electric Power system Research,Yol. 26, No. 3, 1993, pp. 179 183. 45. Dhillon, J.S.,S.C. Parti and D.P. Kothari,"MultiobjectiveOptimalThermalPower Dispatch",Int. J. of EPES, Vol. 16, No.6, Dec. L994,pp.383389. 46. Kothari, D.P. and Aijaz Ahmad, "An Expert system Approach to the unit commitment hoblem", Energy conversion and Management",vol. 36, No. 4, April 1995, pp. 257261. 47  Sen, Subir, D.P. Kothari and F.A Talukdar, "EnvironmentallyFriendly Thermal Power Dispatch An Approach", Int. J. of Energy Sources,Vol. 19, no. 4, May 1 9 9 7 ,p p . 3 9 7  4 0 8 . 48. Kothari, D.P. and A. Ahmad, "Fuzzy Dynamic Programming Based optimal GeneratorMaintenance SchedulingIncorporatingLoad Forecasting",in Advances in Intelligent systems,edited by F.c. Morabito, IoS press, ohmsha, 1997, pp.
/.J5  /4U. Aijaz Ahmad and D.P. Kothari, "A Review of Recent Advances in Generator Maintenance scheduling", Electric Machines and Power systems, yol 26, No. 4, 1998, pp. 373387. 50. Sen S., and D.P. Kothari, "Evaluation of Benefit of InterArea Energy Exchange of Indian Power System Based on MultiArea Unit Commitment Approach", Int. J. of EMPS, Vol.26, No. 8, Oct. 1998, pp. 801813. 49.
1.,
R
caused momentary by chargein generafor speecl, tI'r.r.tnr*,i;;?t;qffi; ;; excitation voltagecontrolsare noninteractive smallchanges for and can be modelled analysed and independently. Furthermore, excitation voltage eontrolis
F:tcl :tcfinrr ir rrh vr r nr rv v r rirn h r t h cv  r r c^ r v r f i r r r o Lrr r,ri rJ rrlttw n,rn..r,rhr ^6^,rri+^^.1 vrJrrJr(lrrr urlLUultLtrlcu : rL^e ^$rL^^^r5 llla! ul ulc; ggirtcfalor
field; while the power frequencycontrol is slow actingwith major time constant contributedby the turbine and generatormomentof inertiathis time constant is much larger than that of the generator tield. Thus,the transients excitation in voltage control vanish much faster and do not affect the dynamics of power frequencycontrol.
8.T
.INTRODUCTION
Powersystemoperation considcrcd far was underconditions stcady so of load. However, both active and reactive power demandsare never steady and they continually change with the rising or falling trend. Steam input to turbogenerators water input to hydrogenerators) (or must, therefore,be continuously regulatedto match the active power demand,failing which the machinespeed will vary with consequent change in frequency whieh may be highly (maximum permissiblechangein power fiequency is t 0.5 Hz). undesirable* Also the excitation of generatorsmust be continuouslyregulated to match the reactive power demand with reuctive generation, otherwise the voltagesat various system buses may go beyond the prescribed limits. In modern large interconnected systems, manual regulation is not feasible and therefore automaticgenerationand voltage regulation equipment is installed on each generator. Figure 8.1 gives the schematic diagram of load frequency and excitation voltage regulatorsof a turbogenerator. The controllers are set for a particularoperatirrgcondition and they take care of small changesin load denrand without fiequency and voltageexceedingthe prescribedlimits. With the passage of time, as the change in lcad demand becomes large, the contrcllers must be reseteithernianuallyor automatically. It has been shown in previous chaptersthat for small changesactive power is dependent internalmachineangle 6 and is inderrendent bus voltage: on of whiie bus voitage is dependenton machine excitation (therefore on reactive  " Change frequency in causes change speed the consumers' in of plant affecting production processes. Further, is necessary maintain it to networkfrequency constant so that the powerstations satisfactorily parallel, variousmotorsoperating run in the on the systemrun at the desiredspeed, correcttime is obtainedfrom synchronous clocksin the system, the entertaining and devices functionproperly.
I
P+JQ
Fig. 8.1 schematic diagramof loadfrequency excitation and voltageregulators a turbogenerator of Change in load demand can be identified as: (i) slow varying changes in meandemand,and (ii) fast random variations aroundthe mean. The regulators mustbe dusigned be insensitive thst randomchanges, to to otherwisethe system will be prone to hunting resulting in excessivewear and tear of rotatins machinesand control equipment. 8.2 LOAD FREOUENCY CONTROL (STNGLE AREA CASE)
Let us considerthe problemof controlling the power output of the generators of a closely knit electric areaso as to maintz,inthe scheduled frequency. All the generators such an areaconstitutea coherent group so that all the generators in speeo ancisiow riowii togetiier iip rnarntarnrng thelr reiarrve power angies.Such an area is defined as a control area. Tire boundariesof a coqtrol area will generallycoincide with that of an individual Electricity Board Company. To understandthe load fiequency control problem, let us consider a single turbogeneratorsystem supplying an isolated load.
^  ^ ^ l  I ^ l     .  l ^     ^ L   . r ^ . _ _ _ :  , . r .
W Turbine
Modern power system Analys,s Speed Governing System turbine. Its downward movement opens the upper pilot valve so that more steem is admitted to the turbine under steady conditions (hence more steady power . The reverse Model of Speed Governing System
Figure 8.2 shows schematicallythe speedgoverningsystem of a steamturbine. The systemconsists the following components: of
Steam
Speed changer
Assume that the system is initially operating under steady conditionsthe linkage mechanism stationary and pilot valve closed, stearnvalve opened by a definite magnitude, turbine running at constant speedwith turbin" po*"r output balancing the generator load. Let the operating conditions be characteizedby = "f" systemfrequency (speed) P'c = generator output = turbine output (neglecting generator loss) .IE = steam valve setting We shall obtain a linear incremental model around these operating conditions. Let the point A on the linkage mechanism be moved downwards by a small amount Aye.It is a commandwhich causesthe turbine power output to change and can therefore be written as
Main piston
A
I
Pilot value High pressure oil
t\
Aye= kcAPc
(8.1)
rHydraulic amplifier (speed control mechanism)
Fig.8,2 Turbinespeedgoverning system Reprinted permission McGrawHilt with of BookCo., New York,from Olle l. Elgerd: Electric Energy System Theory: An lntroduction, 1g71, p. 322. (i) FIy ball speedgovernor: This is the heart of the system which senses the changein speed(frequency).As the speedincreases fly balls move outwards the and the point B on linkage mechanism moves downwards. The reversehappens when the speeddecreases. G) Hydraulic amplifier: It comprises a pilot valve and main piston alrangement.Low power level pilot valve movement is converted into high power level piston valve movement. This is necessary order to open or close in the steamvalve againsthigh pressure steam. (xl) Lintcage mechanism: ABC is a rigid link pivoted at B and cDE is anotherrigid link pivoted at D. This link mechanismprovides a movementto the control valve in proportion to change in speed.It also provides a feedback the steamvalve movement (link 4). ,,fr9rn
where APc is the commanded increase in power. \ The command signal AP, (i.e. Ayi sets into rnotion a bequenceof eventsthe pilot valve moves upwards,high pressureoil flows on to the top of the main piston moving it downwards; the steam valve opening consequently increases, the turbine generatorspeedincreases, the frequencygoes up. Let us model i.e. these events mathematically. Two factors contribute to the movement of C: (i) Ayecontributer [?Jl \rl Aya or  krAyo(i.e. upwards) of  ktKcApc
(ii) Increase in frequency ff causes the fly balls to move outwards so that B moves downwards by a proportional amount k'z Af. The consequent movemen of Cwith A remaining fixed at Ayo  . (+) t (i.e. downwards) The net movement of C is therefore orO,  + kAf
AYc= ktkcAPc+kAf
(8.2)
The movement of D, Ayp, is the amount by which the pilot valve opens. It is contributedby Ayg and AyB and can be written as
Ayo=(h)
Ayc+(;h) *,
= ktayc + koAys (g.3) The movement ay.od,epending upon its sign opensone of the ports of the pilot valve admitting high pressure'o' into thJ moving the main piston and opening the steam valve "ynnJ.ithereby by ayr. certain justifiable simprifying assumptions,which ean be rnade at this .tugl, ur", (i) Inertial reaction forces of main pistoi and steam valve are negligible compared to the forces exerteclon the iirton by high pressureoil. (ii) Because of (i) above, the rate of oil admitted to the cylinder is proportional to port opening Ayo. The volume of oil admitted to the cylinder is thus proportional to the time integral o,f ayo. The movementay"i.s obtainedby dividing the oil volume by the area of the crosssection thepiston. of Thus
controt
1 E^,,^ri^/o o\ :. r . 
E
t
riyLr.Lru' \o.o., rs rcpfesenleo ln tne ronn of a block diagram in Fig. 9.3. Ks9 1+ fsss 4Y5(s)
4F(s)
Flg. 8.3 ,Blockdiagramrepresentation speed governorsystem of The speed governing system of a hydroturbine is more involved. An additional feedback loop provides temporary droop compensation to prevent instability. This is necessitated the targe inertia or the penstoct gut" by which regulates the rate of water input to the turbine. Modelling of a hyjroturbine regulating system is beyond the scopeof this book. Turbine Model
Avn= krfoeayrlat
(8.4)
It can be verified from the schematicdiagram that a positive movemen ayo, t causesnegative (upward)movement ayulccounting for the n"gutiu" ,ign used in Eq. (8.4).
Taking the Laplace transform Eqs. (g.2),(g.3) and (g.4), we ger of AYr(s)= k&cApc(") + krAF(s) (8.5) Ayp(s)= kzAyd,s) koAyug) + (8.6)
a y u ( g =  k so r U n l
EliminatingAyr(s) andAyo(s), we can write AYu(s)k'ktk'AP' (s) k,krAF(s) ') (oo '' t "'tr ,/ \
(8.7)
lor,<,r*^or",].i#)
where
Let us now relate the dynamic responseof a steam turbine in tenns of changes in power ouFut to changesin steamvalve opening ^4yr. Figure g.4a shows a two stage steam turbine with a reheat unit. The dynamic *ponr" is targely influenced by two factors, (i) entrained steambetwein the inlet stbam valve and first stageof the turbine, (ii) the storageaction in the reheaterwhich causesthe output of the low pressurestageto lag behind that of the high pressure stage. 'fttus, the turbine transfer function is characterized two time constants. by For easeof analysisit will be assumedhere that the turbinl can be modelled to have Ssingle equivalent time constant.Figure 8.4b shows the transfer function model of a sreamturbine. Typicaly the time constant lies'in the range o.i ro z.s { sec.
(8.8)
=&
Steam valve
n= klc
K2
t_
= speedregulationof the governor
K.,=
.r.
+y
 gainof speed governor
(a) Twostage steamturbine
, rs = ;;l " = tlme constant of speedgovernor rKqkS
AYg(s)FAPds)
(b) Turbine transfer functionmodel Flg. 8.4
#phSi
I
po*", s),rt"r An"ly.i, rrrroarrn
Generator Load Model The increment power in inputto the generatbrload system is
APG _ APD whele AP6 = AP,, incremental turbine incremental loss to be negligible) and App is the load increment. This increment in power input to the syrtem is accountedfor in two ways: (i) Rate of increase of stored kinetic energy in the generator rotor. At scheduledfrequency (fo ), the stored energy is Wk, = H x p, kW = sec (kilojoules) where P, is the kW rating of the turbogenerator andH is defined as its inertia constant. The kinetic energy being proportional to square of speed(frequency), the kinetic energy at a frequency of (f " + Arf ) is given by
Automatlc Generationand Voltage Control
I
=tAP6g)_ aPo(,)r.[#j
2H = pow€r systemtime constant Bf"
(s.13)
Kp,=
+
=powersystem gain
Equation (8.13) can be representedin block diagram form as in Fig. g.5.
^Po(s) 16ffioro,
Flg.8.5 Block diagram representation of generatorload model
laeo(s)
=nr,(r.T)
Rate of change of kinetic energy is therefore
(8.e)
complete Block Diagrram Representation of Load Frequenry Control of an Isolated Power System
=fffrr"n $rr*"r
(8.10)
(ii) As the frequency changes,the motor load changesbeing sensitive to speed,the rate of changeof load with respectto frequ"n.y, i.e. arot\ycan be regarded as nearly constant for small changes in frequency Af ard can be expressed as
AP(s)=trPn15;
@PDl?flAf=BAf (8.11) wherethe constantB can be determined empirically, B is positivo a for predominantly motorload. Writing the powerbalance equation, have we A P c  a Pr^ = T H P*' d ( o f l +B A f , = f.] < Dividingthroughoutbyp, andrearanging,we get
/A' ' n'7' \ AP;q;u)= 1d (Afi^ + B(ptt)af dt f Taking the l,aplace transforrn, we can write AF(s) as AP6$u)(8.i2)
AP6(s)
Flg. 8.6 Blockdiagrammodelof load frequency control powersystem) (isolated Steady States Analysis The model of Fig. 8.6 shows that there are two important incrementalinputs to the load frequency control system  APc, the change in speedchangersetting; situatiqn in and APo, the change in load demand. Let us consider,,.4,.simple
4Fis; 
AP,G) 4PoG) B*' s
Modern
which
; ;;;*;:r;;ffi; *' #ff:Tiil: :,:il?: in system ; steady 3l i: T::: frequencya sudd.n change for .hung", ffi;ffi"ffi;ti'l;
r.han.,oo c demand
the sneerl .hqnrro' 't::; ;::::^
a2e g * ;, 2 rr e o,,
hoo .. g..^) ^.. '(rr cr rr^tr(r ucttrng
\7'e'
af
= o) and the load
anaount *,
(, e.Apog):+)is
obtained follows: as AP^
^f
aF@)l*,(s):o :
I ^L^^ r'E .1uuy' cquauon glves tne steadystate changes in frequencycausedby changes in load demand. Speed regulation R isnaturally so adjusted that changesin frequency are small (of the order of 5vo from no load to ruu load). Therefore,the linear incrementalrelation (g.16)ican be appliedfrom no load to full load' with this understanding,Fig. 8.7 shows the linear relationship betweenfrequency and load for free governor operation with speedchanger set to give a scheduled frequency of r00% at full toao. The .droop, or slopeof this 'l ( I relationship is l \ B+(t/R) ) Power system parameter B is generaily much smalrer* than r/R (a typical value is B = 0.01 pu Mwalz and l/R = U3) so that B canbe neglected in comparison.Equation (8.16) then simplifies to
.rL^
rhe droop isthus by "r,,fl", fjfli;], curve mainly determined R, the speedgovernorregulation.
(8.17)
r^sorr,
K
I ( = 1I .
ap,=_ *"r: (r^;)o",
Decreasesystem = BAf= in load (uffi)*,
Of course,the contribution of decrease system load in is much less than the increase in generation. For typical values of B and R quoted earlier APo = 0.971 APo Decreasein system load = 0.029 ApD consider now the steady effect of changing speed changer setting load demandremainingfixed (i.e. Apo= 0). The sready +)with state changein frequency is obtained as follows. (Or"<rlat
dA^t   \r,, ruuloLoao li\
It is also rccognized that Ko, = 7 / B , w h e r e B  Y ^ /P' (in PuMWunit change ai in frequency). Now
4=(#6)o,.
(8.16)
fi roa
(J
L
8.rog
.c
102 101 100 0
(ii) 60% Load
250 MW machine with an operating load of 125 MW. let the change in load be i%o for IVo change in frequency (scheduledfrequency = 50 Hz). Then
*For
Percent Load
a:?:r?: :2.5 NNVtHz af 0.s
Flg. 8.7 Steady qharacteristic a speed of governorsystem "*l?39frequency
'=(#)b
: #:
o'ol Pu Mwgz
Analysis uodernPower system W
I
Autor"tic G"n"r"tionand Volt"g" Conttol
AD
F
AF@lap,{s):o:
ttsgttf
r t v (t ^ p s
( 1 +T , r s ) ( l * 4 sxl* zors)+ KseKt p,/R K ) \ _ KreKrKp,  lIAP, 4,flr*uoyro,": t / r l ( 1 + K .sK tK ps R I AP',:g
xu'c s
(8.18)
I
(8.1e)
If K rrK,= l
Ar=( "

\ B+llR)
\rc"
(8.20)
If the speed changer setting is changedby AP, while the load demand i.e. changesby APo, the steadyfrequencychangeis obtainedby superposition,
Ar = (
". According to Eq. (8.2I) the frequency changecausedby load demandcan be compensated changingthe settingof the speedchanger,i.e. by APc APo, for Af = Q
Figure 8,7 depicts two load frequency plotsone to give scheduled frequencyat I00Vorated load and the other to give the samefrequencyat 6O7o rated load.
 APo) ru) 'o"
(8.21)
Two generators rated 200 MW and 400 MW are operating in parallel. The of droop characteristics their governors are4Vo and 5Vo,respectivelyfrom no load to full load. Assuming that the generatorsare operating at 50 Hz at no load, how would a load of 600 MW be sharedbetweenthem?What will be the system frequency at this load? Assume free governor operation. Repeat the problem if both governorshave a droop of 4Vo. Solution Since the generators are in parallel, they will operate at the same frequency at steadyload. Let load on generator 1 (200 MW) = x MW and load on generator 2 (400 MW) = (600  x) MW Reduction in frequency = Af Now
af_
x
0.04x 50 200
(i)
(ii)
x af 0.05 50 400 600x EquatingAf in (i) and (ii), we get r) 231 MW (loadon generator
v 
600x=
/A
JOy
trltf
lvlw
(IUau
/1 ^l
ull
^
Btrrltrriltur
L)
on A 100 MVA synchronousgeneratoroperates full load at at frequencyof 50 Hz. The load is suddenly reducedto 50 MW. Due to time lag in governor Determinethe change system,the steamvalve beginsto closeafter 0.4 seconds. in frequencythat occurs in this time. Given H = 5 kWsec/kVA of generatorcapacity. Solution Kinetic energy stored in rotating parts of generatorand turbine = 5 x 100 x 1.000= 5 x 105 kWsec Excesspower input to generatorbefore the steam valve beginsto close= 50 MW Excess energy input to rotating parts in 0.4 sec = 50 x 1,000 x 0.4 = 20,000 kWsec Stored kinetic energy oo (frequency)2 Frequency at the end of 0.4 sec
0'01150 x 231 = 47.69 Hz 200 It is observed here that due to difference in droop characteristics of 2 while generator is underloaded. governors,generatorI gets overloaded It easily follows from above that if both governorshave a droop of.4Vo,they will share the load as 200 MW and 400 MW respectively,i.e. they are loaded corresponding to their ratings. This indeed is desirable from operational considerations. Systemfrequency= 50 ' Dynamic Response
giving the change in frequency as function of To obtain the dynamic response the time for a step changein load, we must obtain the Laplace inverse of Eq. (8.14). The characteristic equationbeing of third order, dynamic responsecan
r' Onfy r Dg  1!  I ODIalneU f,luf A ^^^tC: ^ SPtrUfffU *^:^^1 ll'llll('llua1'I ^^^^ Ua1DE. tI^.,^,,^II(rwsYsIr +L^ LfIs ^L^^^+^i^+in r,Il<ll4ivLsllDrlv
+ zo,ooo 5r rfz = 5ox I soo,ooo )t"=
\ 500,000 )
equation can be approximated as first order by examining the relative involved. Typical valuesof the time constants magnitudesof the time constants of load frequency control system are rdlated as
Trr4T, <To, Typically* t, = 0.4 sec, Tt = 0.5 sec and
I I o
Time (sec)>
t
1
Firstorderapproximatiorl
Flg' 8.8 Firstorderapproximate brockdiagramof road frequency controt an isolated of area Irning Tro = T, = 0: =1), the block diagram of Fig. 8.6 K*\ is Iuld reduced to thlt of F'ig. 8.8, from which we can write AF(s)l*r(s):o = to, ..APo (1+ KpslR)+ Zp.s" s
Dynamic response_of changein frequency a stepchangein load for (APo= 0.01pu, 4s = 0.4 sec,  = 0.5 sLc, Io. = 2b sec, (" = 100, R= 3) The plot of change in frequency versus time for first order approximadon ^rirst given above and the exact response are shown in Fig. a.g. order approximation is obviously a poor approximation. Gontrol Area Concept So far we have considered the simplified case of a single turbogenerator supplying an isolated load. Consider now a practical system with number of e generating stations aird loads. It is possible to divide an extendedpower system (say, national grid) into subareas (may be, State Electricity Boards) in which the generatorsare tightly coupled together so as to form a coherent group, i.e. all the generators respond in unison to changes in load o, ,p"rJ changer settings.Such a coherentareais called a control area in which the frequency is assumedto be the same throughout in static as well as dynamic conditions. For purposes of developing a suitable control strategy, a control area can be reduced to a single speed governor, turbogenerator and load system. All the control strategies discussedso far are, therefore, applibable to an independent control area. Proportional Plus fntegral Control
  "o{1: =xaP,
, l, + ^ + r o ' 1
L R4,J
(,)= Ar ft{'  *,[,,a[n#)]] g 22) *,
TakingR = 3, Kp, = llB = 100, = 20, Apo = 0.01 pu e, Af (t) =  0.029(I  ,t:tt',
Aflrt"udystare  0.029 Hz =
(8.23a) (8.23b)
"For a 250 MW machine quoted earlier, inertiaconstanrIl = SkWseclkVA =2osec , = 4 :o 2 * 5 . ' Bf 0.01x = 50
It is seen from the above discussionthat with the speed governing sysrem installed on each machine, the steady load frequency charartitirti" fi agiven speedchanger setting has considerabledroop, e.g. for the system being used for the illustration above, the steadystatedroop in fieo=ueney will be 2.9 Hz [see Eq. (8.23b)l from no load to tull load (l pu load). System frequency specifications are rather stringent and, therefore, so much change in frequency cannot be tolerated. In fact, it is expected that the steady change in frequency will be zero. While steadystatefrequency can be brought back io the scheduled
ffil
I
Analys Power system Modern
AutomaticGenerationand VoltageControl tin ihe above scheme ACE being zero uncier steaciyconditions*, 4 logical design criterion is the minimization of II,CZ dr for a step disturbance. This integral is indeed the time error of a synchronous electric clock run from the power supply. Infact, modern powersystems keep Eaekofintegra+e4tinae errsr all the time. A corrective action (manual adjustment apc, the speed changer setting) is taken by a large (preassigned)station in the area as soon as the time error exceeds a prescribed value. The dynamics of the proportional plus integral controller can be studied numerically only, the systembeing of fourth orderthe order of the system has increasedby one with the addition of the integral loop. The dynamic response of the proportional plus integral controller with Ki = 0.09 for a step load disturbance of 0.01 pu obtained through digital computer are plotted in Fig. 8.11. For the sake of comparisonthe dynamic responsewithout integral control action is also plotted on the samefigure.
speedchangersetting,the systemcould under go intolerable vaiue by adjus'ring dynamic frequency changes with changes in load. It leads to the natural suggestion that the speed changer setting be adjusted automatically by monitoring the frequency changes.For this purpose, a signal from Af is fed diagram througfan integrator to the s configuration shown in Fig. 8.10. The system now modifies to a proportional plus integral controller, which, as is well known from control theory, gives zero steady state error, i.e. Af lrt""d" ,,ut,= 0.
Integral controller APe(s)
APp(s)
l+tr8I I t  +
AF(s)
t
l
AP6(s)
sensor Frequency control plus Fig. 8.10 Proportional integralload frequency by The signal APr(s) generated the integral control must be of oppositesign to /F(s) which accounts for negative sign in the block for integral controller. Now
+ I
I t o
r x
1
AF(s1=
Kn, Ko, (r %"s). +). ( l * f , r s ) ( l + 4 s ) + (**
RKo,s(l+{rs)(l+ 4s) * + {'s)(1 + 4sXl f zo's)R Ko'(KiR f s)
"+
actiOn
(8.24)
obviousry
Flg. 8.11 Dynamic response loadfrequency of controller with and without integral controlaction(APo = 0.01pu, 4s = 0.4 sec, Ir = 0.5 sec, Ips= 20 sec, Kp.= 100,B  B, Ki= 0.09) 8.3 IOAD FREOUENCY CONTROL AND ECONOMIC DESPATCH CONTROL
freouencv_ J _ I control ________ with ,.___ inteorel
__O
(8.25) Af l"t"^dy = so/F(s) : o state , In contrast to Eq. (8.16) we find that the steady state changein frequency has been reduced to zero by the additio4 of the integral controller. This can be argued out physically as well. Af reaches steady state (a constant value) only
rr.,lrsrr uAp^ rcwlMl HrD
Ap
=

.onsfant
vvuulqr!.
Becarrse of
fhe
intes!'atins
Of the
Load
eonfrnller
qnhierrAe 'vu
lvrv
?a?^
otvsuJ
craolrr
Dl4lg
ora+o
controller, this is only possible if Af = 0. In central load frequency control of a given control area, the change (error) in frequencyis known as Area Contol Error (ACE). The additional signal fed back in the modified control schemepresented above is the integral of ACE.
frequencyelTor and a fast dynamic response,but it exercisesno control over the relative loadings of various generating stations (i.e. economic despatch) of the control area. For example, if a sudden small increasein load (say, 17o) occurs
a control is known as isochronous control, but it has its time (integral of frequency) error though steady frequency error is zero. 'Such
f area, the road conrior ,changesthe speed changer 1i..l1r_::ltrol frequency Dcrurgsor tne governors of all generating units of the area so that, together, theseunits match the load and the frequenry returns tp the scheduled value (this action takesplace in a few seconds).However, in the,process of this change the Ioadings of u@units change in a manner independent of economi@ In fact, some units in the pro""r, may even get overloaded. Some control over loading of individual units cafi be Lxercised by adjustingthe gain factors (K,) includeJin the signal representing integral of the area cogtrol error as fed to individual unitr. However, this is not satisfactory. _Tcommand signai generated'oythe centrai economic despatch computer.Figure 8'12 gives the schematicdiagram of both thesecontrolsior two typi.ut units of a control area.The signal to changethe speedchan3ersetting is lonstructed in accordance with economic despatcherror, [po (desired) pJactual)]. suitabry modified by the signal representingintegral ncg at that instant of time. The signal P6 (desired) is computed by the central economic despatch computer (CEDC) and is transmitted to the local econornic despatch controller (EDC) installed at each station. The system thus operateswith economic desfatch error only for very short periods of time beforJ it is readjusted. 8.4 TWOAREA LOAD FREOUENCY CONTROL Automatic
"fnceot
An extendedpower system can be divided into a number of load frequency control areasinterconnectedby meansof tie lines. Without loss of generality we shall consider a twoarea case connectedby a single tie line aslilusnated in Fi g. 8.13.
Speed
Fig.B.i3 Two interconnected contror areas(singre rine) tie The control objective now is to regulatethe frequency of each area and to srnnultaneously regulatethe tie line power as per interareapower contracts. As in the caseof frequency, proportional plus integral controller will be installed so as to give zero steady state error in tie line power flow as compared to the contractedpower, It is convenientlyhssumedthat each control area canbe represented by an equivalentturbine, generatorand governorsystem.Symbols used with suffix I refer to area 7 and those with suffi x 2 refer to area 2. In an isolated control area case the incremental power (apc _ apo) was accountedfor by the rate of increaseof stored kinetic energy and increase in areaload causedby increasein fregueircy.since a tie line t *rport, power in or out of an area, this fact must be accounted for in the incremental power balanceequation of each area. Power transported out of area 1 is .eivenbv
lr.
EDC  Economic despatch controller CEDC  Central economic despatch computer Flg. 812 Control area load frequency and economic despatch control
Reprinted (with modification) with permission of McGrawHill Book Company, New York from Olle I. Elgerd: Electric Energy SystemsTheory: An Introd.uction, I971,p. 345.
Ptie, = r where q'q
''rrl''l X,,
sin ({  q
(8.26)
 poweranglesof equivalent machines thetwo areas. of
308 
I
I
Modefn Power SystemAnalysis
Automatic Generation Vortage and contror
I APti".r(s)
Fil
in tie For incrementalchanges { and 6r, the incre.mental line power can be
expressed as AP,i,,r(pu)= Tp(Afi  462) where
T, = 'Y:t'Yf PrrXrz coefficient cos(f  E)  synchronizing
(8.27)
we Since incrementalpower anglesare integrals of incrementalfrequencies, can write Eq. (8.27) as
Fig. 8.14 The corresponding block diagram is shown in Fig. g.15.
+
AP,i,,r = 2*.(l
Afrdt
I Urat)
(8.28)
APti",r(s)
where Afi nd Af,, arc incremental frequency changes of areas 1 and 2, respectively. tie Similarly the incremental line power out of area2 is given by aPt;", z = 2ilzr([ yrat where [ ayrat)
AF1(s)
iE= Fig. 8. 15
n7ri"l
(8.2e)
For the control area 2, Ap6", r(s) is given by tEq. (g.Zg)l  : gr r r , apt i", z( s)= [ AFr ( s) 4F, ( s) ]
tYr:J ({ cos L  E): [S]ti z: ar2rrz rzr= LL "
Przxzr \Prr)
(s.30)
( g: 35)
With referenceto Eq. (8.12), the incrementalpower balance equationfor area 1 can be written as
= APo, APor + *w)+ Jr" or
nrzr* AP,,",t
(8.31)
It rnay be noted that all quantitiesother than fiequencyare in per unit in Eq.(8.3l). we of Taking the laplacetransf'orm Eq. (8.31) and reorganizing, get A F (s ) = IAP 6 1 G  A P r,(s) ) where as definedearlier [seeEq. (8.13)]
= Kp31 I/81 Tpil = LHr/BJ"
A P ti " ,,1r;] $.32) " t$ 4,,t,! I+
which is also indicated Uy ,i. block diagramof Fig. 8.15. \ Let us now turn our attentionto ACE (areacontrol error; in the presence of a tie line. In the case of an isolated control area, ACE is the change in area frequency which when used in integral control loop forced the steady state frequencyelror to zero. In order that the steadystatetie line power error in a twoarea control be made zeroanotherintegralcontrol loop (one for each area) must be introducedto integratethe incremental line power signal and feed tie it backto the speed changer. This is aeeomplished a singleintegrating by bloek by redef ining ACE as a linearcombination incremental of frequenryand tie line power. Thus, fbr control area I ACEI = APu".r+ brAf, ( 8. 36) where the constant b, is called areafrequency bias. Equation (8.36) can be expressed the Laplace transform as in ACEl(s) = APo., r(s) + b1AF1g) Similarly, for the control are a 2, ACE2 is expressedas ACEr(s) = APti".z(s)+ b2AF,(s) (8.37)
(8.33)
Comparedto Eq. (8.13) of the isolated control areacase,the only changeis ol the appearance the signal APri"J(s) as shown in Fig. 8.14.
'l'^LiI4ArrrS AS fho T ^l^a rrrw lsl/l4vv fL lo nl o frn rrm r * 4l D u rr ^ fr u E^ LY. /a ta\ \v.L9), tl h a lrv c i o nro l^ or6r s ,4P /"\ tie.I\.r/ ", ic nlrfoinerl
 /4 (s)l = AP,i.,1(s) ffroor(s)
(8.34)
( 8. 38) Combining the basic block diagramsof the two control areascorresponding to Fig. 8.6, with AP5rg) and Apr2(s) generatedby integrals of respective ACEs (obtainedthrough signalsrepresentingchangesin tie line power and local frequencybias) and employing the block diagramsof Figs. g.t+ to g.15, we easily obtain the composite block diagram of Fig. g.16.
WIU&
Modern Power svstem Analvsis
Let the step changes loads APo, and APrrbe simultaneously in appliedin control areas 1 and 2, respectively.When steady conditions are reached,the output signals of all integratingblocks will become constantand in order for this to be so, their input signalsmust become zero. We have, therefore, from
F i e .8 . 1 6 KtL) APu",, + b rAfr= O finputof integrating block\ ,r) K'z) APti",, + brAfr= o finpot of integrating block\ r l Afr  Afz =o finpurot integrating block '4'\ \ s FromEqs.(8.28)and(8.29) APn",, =Tr, . AP.i",z, Tzt; I.=constant ar2 (8.41) ) (8.39a)
trJ
o
(8.3eb) (8.40)
ra
S
I il
oi
g
A
oo
N
o (U o !t o oy,
.Y
EF
a (\
.g q
<.1
:pE
o 6 c) =a d  (
8 6
Hence Eqs.(8.39) (8.41)aresimultaneously satisfied only for
A P r i " , r =A P , : " , 2 = 0 and
v'it
(8.42)
lr
*li
ol. a
g t u Q c l e.> (g()
E 9
o o
o o o o
E o
Afi=Afz=0
Thus, under steady condition change in the tie line power and frequency of each area is zero. This has been achieved by integration of ACEs in the feedbackloops of each area. Dynamic responseis difficult to obtain by the transfer function approach (as used in the single area case)becauseof the complexity of blocksand multiinput (APop APor) and multioutput(APri",1, Ap6",2, Afr Afr) situation. A more organizedand more convenientlycarried out analysisis through the state spaceapproach(a tirne domainapproach). Formulation of the statespacemodel for the twoarea system will be illustrated in Sec. 8.5. The results of the twoareasystem(APri", changein tie line power and,Af, changein frequency) obtainedthrough digital computerstudy are shown in the form of a dotted line in Figs. 8.18 and 8.19. The two areasare assumedto be identical with system parameters given by Trs= 0.4 sec, 7r = 0.5 sec, ?r, = 20 sec K o r =1 0 0 , = 3 , b = 0 . 4 2 5 , & = 0 . 0 9 , 2 f l r 2 = 0 . 0 5 R 8.5 OPTTMAL (TWOAREA) LOAD FREOUENCY CONTROL
5 * * o
O E
F A
G'
6
o o o
E (D () (U
b 6 tr(')
E
H'.s
tt)
E g
.9
d
* 3 5 u = o a'9 '85
oo. *,n EOo o o o
@
.n o o o r\ ai cit lr
CL
d <;
l!
t
l+
ld
5l
u I f:
Modern control theory is applied in this section to design an optimal load frequencycontroller for a twoa3ea system.In accordance with modern control terminology APcr arrdAP62 will be referred to as control inputs q and u2.ln the conventional approachul anduzwere provided by the integral of ACEs. In
AutomaticGeneration and Voltagecontrot
itZ I
Analysis System Power rrrrodern 
f*
ComparingFigs. 8.16 and 8.17, xt = Aft .r2 AP,;1 tt1= APg, w1= AP" For block 1 xq = Af. x5 = AP52
u2 = /)Pa w2 = APp,
M&
moderncontrol theory approach ur and u2 wtll be createdby a linear the For (full state feedback). formt'lating states of combination all the system feedbackloops are this purposethe conventional statevariablernodel for
block as shown in bv resented a se Statevariablesare definedas the outputsof all blocks baving either Fig. 8.17. We an integratoror artirne,constanf.. immediately notice that the systernhas nine state variables.
XS= JACEit t, = JACE, dt
1+r+i,
+ I
I
I l
f.'f\
Optimal case (full state feedback)
x1 + T. r r i, = K ^ t ( x z L P
h
 w) (8.43)
.
h l 
1
4 l
o
o X
( t'2
3
I
, *f
Kprt
t psl
Kprt , Kprt ,z ;wt x  ;  x t
t ptl t ptl
'psl
1
,I
r;+.1
/ /' ' I' 8
t t
For block 2 x.2+ Tiliz=
' With integralcontrolaction
xt
or
*z=+r**n
t r + { , s r i ': = R L r, r + r , r
( 8. 445
due to step load(0.01pu) changein area1 F i g . 8 .1 8 change in tie linepower '
For block 3
r
21.+_'';;;71=a.11=1
12 14 20 18 16 Time(sec)
A
L I
N
or
* t=
X+ * n
^h
r,
t* ,* * ,,
wz) Ko*2
(8.45)
o x
IL
I
with integral control action
For block 4
Torz*+= I Krrz(xs + ar2x7 Knrz
'\A'1.{<
Optimalcase (full state feedback)
or
iq=
Tprz

Tps2
''
 T   y '  a  _  W ^
at?K or2
Tps2
'
Tpsz
z
( 8. 46)
For block 5
Fig. 8.19 Change in frequencyof area 1 due to step load (0.01 pu) 1 change in a.rea Before presenting the optimal design, we must formulate the state model. This is achieved below by writing the differential equations ciescribing each individual block of Fig. 8.17 in terms of state variables (note that differential equations are written by replacing s UV *1. ' dt'
x s t 7,2i5  x6 or is=
l
Ttz
Y 4
r<
I  V t
1
4 l
T,z
u
( 8. 47)
For block 6 . l x s * I ,szx6 ; x4 + u2
I\2
or
io=#*o**u
'2t sg2 t sg2
( 8. 48)
'3i4',"1
T
ModernPower SystemAnatysis ' co","".t"" " constructedas under from the state variables x, and rnonly. ut=is= brx, + x.i Kirxs= Kir IeCn,Ar Kiz la.Cerar
For block 7 it=2iTtzxt2iTr2xa For block 8
(8.4e) (8.5O) (8.s ) 1
For block 9
i9= b2xa anxt
uz= Ki{s=
(8.43) to (8.51) can be organizedin the following vector The nine equations matrix form (8.s2) *=Ax+Bu+Fw
where
ln the optimal control schemethe control inputs u, anduz aregeneratedby means of feedbacks from all the nine states with feedback constants to be determinedin accordance with an optimality criterion. Examination of Eq. (8.52) reveralsthat our model is not in the standard form employed in optimal control theory. The standardform is i=Ax+Bu which does not contain the disturbance term Fw present in Eq. (g.52). Furthermore,a constantdisturbance vector p would drive some of the system states and the control vector z to constantsteadyvalues;while the cost function employedin optimal control requiresthat the systemstateand control vectors have zero steady state values for the cost function to have a minimum. For a constant disturbancevector w, the steady state is reachedwhen
x _ l x r x2 ... xg)r = state vector u = f u t u2fT= control vector vector w = l w t w2fT = clisturbance while the matrices A, B and F are defined below:
I I Tpst, 0 1 Rr4er
A _
2
Y 'tPsl
3
0
4
0
5
0
7
Tprt
8
9
bLoo
0 0
atzKprz
Tprt
 1
Tt
1
Ttr
o
o
0 0 0 0
0 0
o
0
 1
Trst 0 1
o
K p Tprz '
o
z Tprz Ttz
* = 0 in Eq. (8.52); whichthengives 0 = A . r r "+ B u r r + F w (8.s3) Definingx and z as the sumof transient steady and state terms, can write we = ,
x x' * Ir" (8.54) n = ut * z', (8.55) Substituting and z from Eqs. (8.54) and (8.55) in Eq. (8.52), we have r i' = A (r/ + x"r) + B(at + usr)+ Fw By virtue of relationship(8.53),we get *' = Axt + But ( g. 56) This represents system model in terms of excursion of state and conhol vectors fiom their respectivesteadystate values. For full state feedback, the control vector z is constructed by a linear combination of all states.i.e. u=Kx (8.57a) where K is the feedback matrix. Now ttt+ Itrr= l( (r/+ rr") For a stable system both r/ and ut go to zero, therefore ur, = _ Kx* Hence
tt /= Ikl
0 0 6 7 8 9 0 2 irrz
bL
Tprz
o o
0 0 0
o o
0 0 0
i 
o 1
10
1 7,, I
TreZ
0 0 0 1
atz
0 0 0 0 0 0 0 0 0 0
0
RzTrsz 2ilr2 0 0 0 b 2 0
0 0 0
[ o o TI I
Br = 
I L
ss1
o o
o
aco) 'O
00.l
I
I J
l o oo o o +
Kprt 0 Tprt
0 0l
I
,;T
(8.s7b)
Modern Power SystemAnalysis Examinationof Fig. 8.17 easily reveals the steady state values of state and control variablesfor constantvaluesof disturbance inputs w, andwr. These are
Ilrr=X4"r= /7r" = 0
Automatic Generation
b ? o0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0
0 0 0 0 0 0
4 0 0
0 0 0 0 0 0
ulr, = wl r5rr= x6rr= lv2
(8.s8)
arzbz 0 0 0 0 0 0 0 0 0 0 0 0
uzr, = wz
Igr, = COnstant I9r, = Constant The values of xr* and xe* depend upon the feedback constants and can be determined from the following steady state equations: utrr= kttxtr, + ... + ftt8r8", * kt*sr, = wl = r,t2ss k2txlr, + ... + kzgxgr., kz*gr, = wz *
4 0 0  a n b z 0 0 Q+a?)o o 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1
= symmetric matrix R  kI = symmetric matrix
(8.se)
K = RrBrS
The feedbackrnatrix K in Eq. (8.57b) is to be determined so that a certain performance index (PI) is minimized in transferring the system from an arbitrary initial state x' (0) to origin in infinitie tirne (i.e. x' () = 0). A convenient PI has the quadratic form
(8.63)
'
Pr=
+ u'r Ru'dt ;ll '.'' Qx'
(8.60)
The acceptablesolution of K is that for which the systemremainsstable. SubstitutingEq. (8.57b) in Eq. (8.56), the system dynamics with foedback is definedbv i' = (A  BIgx, (g.64) Fol stability all thc cigenvalues the matrix (A  Bn shouldhave of negative real parts. For illustration we considertwo identical control areaswith the following systellparameters: 4r* = 0'4 scc; T'r= 0.5 sec; 7'r* = 20 sec /l = 3: (n* = l/lJ = 100 b = O . 425;Ki = 0. 09; up = I ; 2iln = 0. 05
The manices Q arrd R are defined for the problem in hand through the following design consiclerations: (i) Excursions of ACEs about the steadyvalues (r,t + brx\;  arrxt, + bzx,q) are minimized. The steady values of ACEs are of course zero. (ii) Excursions JnCg dr about the steadyvalues (xts, xte) nrinimized. of are The steacly valuesof JeCg dt are,of course,constants. (iii) Excursionso1'the contt'ol vector (ut1,ut2) about the steadyvalue are rninirnized. The steadyvalueof the controlvectoris, of course, constant. a ' This nrinimization intended indirectlylimit the controleffbrt within is to the physical capability of components. For example, the steam valve catmot be openedmore than a certain value without causingthe boiler presisure drop severely. to With the abovereasoning, can write the PI as we
pr= * fU+ + h,.r,,)2 ( tt,2xt, brxta)z (.r,?+ ,,]) + + + 2Jtt'
+ kfu'l u,11 + at
From the PI of Eq. (8.51), Q md R can be recognizedas
0.021 0.0100 0.52tt6 l.l4l9 0.68  0.0046 0.743 0.gggg0.00001 l3 7 f, ^ = [ Lo.tl0460.o2tl0.0100 0.5286 t.t4rg 0.6813 0.74370.0000 0.gggsl
(8.61)
'*Refer Nagrath and Gopal [5].
iiii'f:l
power Modern rystem in4gs
c= vR.f vr '_
As the control areasextend over vast geographical regions, there are two ways of obtaining full state information in each area for control purposes. (i) Transport the state information of the distant area over communication channels. This is, of course,expensive.
The error initiates the corrective action of adjusting the alternator excitation. Error wave form is suppressedcarrier modulated, tt" carrier frequency being the system frequency of 50 Hz.
Change in voltage caused by load Load change
tG
1+Iers
skrt
8.6
AUTOMATIC
VOLTAGE
CONTROL
Fig. 8.21 Brockdiagramof arternator vortageregurator scheme Error amplifier: It demodulatesand amplifies the error signal. Its gain is Kr. scR power amplffier and exciter fierd: It provides the n"."rriry power amplification to the signal for controlling thl exciter n"ro. arr*;"g ,rr" amplifier time constant to be small enoughio be neglected, the ovelail fansfer function of these two is K, l* T"rs where T"yis the exciter field time constant.
L
Figure 8.20 gives the schematic diagram of an automatic voltage regulator of a generator.It basicallyconsistsof a main exciter which excites the alternator field to control the output voltage. The exciter field is automaticallycontrolled through error e = vr"r  vr, suitably amplified through voltage and power amplifiers. It is a type0 system which requiresa constant error e for aspecified voltage at generatorterminals. The block diagram of the systemis given in
o
A D
Alternator; Its field is excited by the main exciter voltage vu. Under no road it producesa voltage proportional to field current.The no load transferfunction is Ks 7*T*s where
Potential
Fig. 8.20 Schematic diagramof alternator voltageregulator scheme Fig. 8.21. The function of important components and their transferfunctions is given below: Potential transformer: It gives a sample of terminal voltage v.. Dffirencing device; It gives the actuating error
T*= generatorfield time constant. The load causes a voltage drop which is a complex function of direct and quadratureaxis currents.The effect is only schematically reBresented hlock hv G.. The exact load model of the alternatoris beyond ,t" ,iop" ;rhtJ;;: stabitizing transformer: T4*d are large enough time constantsto impair the system's dynamic response.lq weil known that the dynami. Itjs r"rpoor" of a control system can be improved by the internal derivative feedback loop. The derivative feedback in this system is provided by means of a stabiyzing transformer excited by the exciter output voltage vE. The output of the
320' l
I
Modern Power SystemAnalysis
Automatic Generation Voltage_l and Control
Jffif
E
I
is transformer fccl ncgativclyat the input terminalsof thc SCR power stabiliz,ing amplifier. The transferfunction of the stabilizing transfo"meris derived below. is Since the secondary connectedat the input ternfnals of an amplifier, it can be assumedto draw zero current. Now
dt vr = Rr i., + LrJilL ' d t
valuesimposedhy the limitersareselected resffictthe generation The banded to rate by l}Vo per minute. I
I.g 9t", u't _+( +/
A l
'rr= MY
dt
t*9r"'l
Taking the Laplace transform, we get
(s) %, _ sM VuG) R, * s,Lt
sK",
sMlRt l*Irs Fig.8.22 Governor modelwith GRC The GRCs result in larger deviationsin ACEs as the rate at which generation can change the area is constrainedby the limits imposed. Therefore, the in duration for which the power needsto be imported increasesconsiderably as cornparedto the case where generationrate is not constrained.With GRCs, R should be selected with care so as to give the best dynamic response. hydroIn thennal system, generationrate in the hydro area norrnallyremainsbelow the the safe limit and therefore GRCs for all the hydro plants can be.ignored. 8.8 SPEED GOVERNOR DEADBAND ON AGC AND ITS EFFECT
1 +{ , s
Accurate staterrariablemodels of loaded alternator around an operating point can are available in literature using which optimal voltage regulation schemes devised.This is, of course,beyondthe scopeof this book. be 8.7 LOAD FREOUENCY CONTROL WITH GENERATION RATE CONSTRAINTS (GRCs)
the so controlproblcmdiscussed far doesnot consicler effect frcquency The l<racl of the restrictionson the rate of changeof power generation.In power systems maximum only at a specified can powergeneration change plants, havingsteam (fiom saf'etyconsiderations the equipment)for o1 rate. The generationrate reheat units is quit low. Most of the reheatunits have a generatiol rate around 3%olmin. Some have a generation rate between 5 to 7jo/o/min.If these largc tttottrclttrry systerttis likely to c:ha.sc constraintsarc not consirlcrcd, undue wear and tear of the controller. Several Thrs results in disturbances, of the to havebeenproposecl consider effect of GRCs for the clesign methocls dynamic the When GRC is considered, systeln generation controllers. automatic cannotbe appliedfor and linearcontroltechniques nonlinear rnodelbecomes of the optimization the controllersetting. the ratesdenotedby P", are included in the statevec:tor, If the generation systerm order will be altered.Insteadof augntentingthem, while solving the stare equations,it may be verified at each step if the GRCs are viclated. Another way of consicieringGRCs for both areas is to arjri iinriiers io ihe governors[15, 17] as shown in Fig. 8.22, r.e., the maximum rate of valve by openingor closingspeedis restricted the limiters.Here 2", tr,r,, iS the power rate limit irnposedby valve or gate control. In this model
l A Y E l .  g u , n r
(8.6s)
The eff'ectof the speed governor deadband that for a given position of the is governor control valves, an increase/decrease speed can occur before the in positionof the valve changes. The governor deadband materiallyaffect the can system response.ln AGC studies, the deadband eff'ect indeed can be significant,sincerelativcly small signalsare under considerations. governor TlLespeed characterristic. though nonlirrear, beenapproxinraaed has by linear characteristicsin earlier analysis. Further, there is another noniinearity introducedby the deadband the governor operation. in Mechanical f'riction and backlashand also valve overlapsin hydraulic relays cause the governor deadband. Dur to this, though the input signal increases, speed the governor may not irnmediately react until the input reachesa particular value. Similar a.ction takesplace when the input signal decreases. Thus the governor deadbandis defined as the total rnagnitudeof sustainedspeedchangewithin which there is no change in valve position. The limiting value of deadbandis specifiedas 0.06Vo. was shown by Concordia et. al [18] that one of the It effects of governor deadbandis to increasethe apparentsteadystate speed regulation R.
lFFf
Modrrn Po*., svrt.t Analuri,
The effect of the deadbandmay be included in the speed governor control as loop block diagram shown in Fig. 8.23.Consideringthe worst caseforthe deadband,(i.e., the system starts responding after the whole deadband is block in Fig. 8.23,the following set of and traversed) examiningthe deadband ly define the behaviourolthe dead.band [9]
DiscreteTime Control Model
The continuoustimedynamic systemis describedby a set of linear differential equations
x=Ax+Bu+ fp
(8.67)
where f u, P are state, conhol and disturbance vectors respectively and A,B and f are constantmatrices associated with the above vectors. The discretetime behaviourof the continuoustime systemis modelled by the system of first order linear differenceequations:
Speed governor Deadband
x(k+1)=Qx(k)+Vu(k)+jp&)
(8.68)
control loop in Flg. 8.23 Deadband speedgovernor u(r+1)= 7(r) 1:
"(r+l)
_ x, 1 deadband "(r+1) (8.66)
where x(k), u(k) and p(k) are the state,control and disturbancevectors and are specified t= kr, ft = 0, 1,2,... etc.and ris the sampling at period. 6, tl,nd 7 Te the state, control and disturbance transition matrices and they are evaluatedusing the following relations. d= eAT
_ deadband; if x('+l)  ,(r) I g
tf Xr*l _ xt < 0 "(r+1). (r is the step in the computation)
{=({r_lntr
j=(eArDAtf where A, B and, are the constantmatrices associatedwith r, ,,LO p vectors I in the conespondingcontinuoustimedynamic system. The matrix y'r can be evaluatedusing various welldocumentedapproacheslike Sylvestor's expansion theorem, series expansion techniqueetc. The optimal digital load frequency controller designproblem is discussed detail in Ref [7]. in 8.10 DECENTRALIZED CONTROL
Reference[20] considersthe effect of governor deadbandnonlinearity by using the describingfunction approach[11] and including the linearised equationsin the state spacemodel. oscillatory. makesthe dynamicresponse of The presence governordeadband It has been seen [9J that the governor deadbanddoes not intluence the selectionof integral controller gain settingsin the presenceof GRCs. In the the system presence GRC and deadband even for small load perturbation, of highly nonlinear and hencethe optimization problem becomesrather becomes complex. 8.9 DIGITAL LF CONTROLLERS
In view of the large size of a modern power system, it is virtually impossible to implement either the classicalor the modern LFC algorithm in a centralized manner.ln Fig. 8.24, a decentralized control schemeis shown.x, is usedto find out the vector u, while x, aloneis employed to find out u". Thus.
In recent years,increasinglymore attentionis being paid to the questionof digital implementation of the automatic generationcontrol algorithrns. This is mainly due to the facts that digital control turns out to be more accurate and
rcliqhlc
r v^rEv^vt
nnrnnaef
in qize
less censifive
to nnise
end drift
nnd more
flexihle
Tt
may also be implemented in a time shared fashion by using the computer in systems load despatchcentre,if so desired.The ACE, a signal which is used for AGC is availablein the discreteform, i.e., there occurssampling operation ; betweenthe systemand the controller. Unlike the continuoustimesystem,the control vector in the discretemode is constrainedto remain constant between Flg. 8.24 Decentralized control
i,i2[,:*.1
4
Power Analysis Modern System
x  (x1 x2)' ut=ktxt u2kzxz

and Automatic Generation voltageControl
ffi
'  aF(s) 1' af (t)dr: liq, * '4F(s) hm/F(")] : I n,n,, tf^461dv , t JO JO s0 s+0 S
L
been shown possible using the modal control principle. Decentralized or hierarchicalimplementationof the optimal LFC algorithmsseemsto have been case since the real load disturbancesare studied more widely for the stochastic A truely stochastic. simple approachis discussedin Ref. [7]. It may by noted that other techniquesof model simplification are available in the literatureon alternativetools to decentralizedcontrol. Theseinclude the method of "aggregation", "singular perturbation", "moment matching" and other techniques[9] for finding lower order models of a given large scale system.
8.4 For the two area load frequencycontrol of Fie. 8.16 assumethat inte controller blocks are replacedby gain blocks, i.e. ACEI and ACE are fed to the respective speedchangersthrough gains  K, and  Ko. Derive an expressionfor the steadyvalues of changein frequency and tie line power for simultaneouslyapplied unit step load disturbanceinputs in the two areas. 8.5 For the two area load frequencycontrol employing integral of area control error in each area (Fig. 8.16), obtain an expressionfor AP6"$) for unit step disturbance in one of the areas.Assume both areas to be identical. Comment upon the stability of the system for parameter values given below: 4e = 0'4 sec; Z, = 0'5 sec; Zp. = 20 sec K p r = 1 0 0 ;R = 3 ; K i = l ; b = 0.425 ar2= I;2tTr, = 0.05 lHint: Apply Routh's stability criterion to the characteristicequation of the system.l
IE PROBI/IS
rated 200 MW and 400 MW are operating in parallel. 8 . 1 Two generators of The droop characteristics their governors are47o and 5Vorespectively from no load to full load. The speedchangersare so setthat the generators operate at 50 Hz sharing the full load of 600 MW in the ratio of their to ratings.If the load reduces 400 MW, how will it be sharedamong the generators and what will the s)/stemfrequency be? Assumefree governor operatlon. of The speedchangers the governorsare resetso that the load of 400 MW at is sharedamong the generators 50 Hz in the ratio of their ratings. What of are the no load frequencies the generators? 8 . 2 Consider the block diagrammodel of lcad frequencycontrol given in Fig. 8.6. Make the following approximatron. (1 + Z.rs) (1 + Z,s) = t + (7rg + T,),s= 1 + Z"c.r giveu below. Given AP,  0.01 pu Solve for Af (l) with parameters T"q= 0.4 + 0.5 = 0.9 sec; 70, = 20 sec ; K r r K , = 1 ;K p r = 1 0 0 R = 3 given in Fig. 8.9. Coinparewith the exact response
NCES REFERE
Books
l. Elgcrd, O.1., Elccu'ic Energv.Sv,s/clrT'lrcorv: An ltttnxlut'lion. 2nd cdn. McCrawHill, New York, 1982. 2. Weedy, B.M. and B.J. Cory Electric Pow'er Systems,4th edn, Wiley, New York, I998.
a 1
Cohn, N.,
Control of Generation and Power Flou, on Interconnected Systents,
Wiley, New York, i971.
4 . Wood, A.J., and B.F. Woolenberg, Power Generation, Operation and Control,2nd
edn Wiley, New York, 1996.
5 . Nagarth, I.J. and M. Gopal, Control Systems Engineering, 3rd edn. New Delhi,
2 0 0l .
8 . 3 For the load frequency control with proportional plus integral controller
clJ
oc
olrn'rn.iorlvYYll
ll(
T i iLc . 6
L
e
v.
1n
rvt
nhfain
vuLarrr
en
AsnrAccinn
fnr
tha
cfenrlrr
cfrfp
errnr
in
6 . Handschin, E. (Ed.), Real Time Control of Electric Power Systems, Elsevier, New
York 1972.
time APr. What is the corresponding cycles,i.". f'41t)d r; for a urrit step t ^ "
, 1 ,
7 . Mahalanabis, A.K., D.P. Kothari and S.I Ahson, Computer Aided Power Systent
Analysis and Control, Tata McGrawHill, t959. New Delhi, 1988.
lirnl*m
of (with respect 50 Hz).lCommenton the dependence to error in seconds error in cycles upon the integral controller gain K,.
8 . Kirclrrnayer, L.K., Economic Control of lnterconnected Systems,Wiley, New York, 9 . Jamshidi, M., Inrge Scale System.s:Modelling and Control, North Holland, N.Y.,
1983.
10. Singh, M.G. and A. Titli, SystemsDecomposition, Optimization and Control PergamonPress, Oxford, 197g. I I' Siljak, D'D., NonLinearSystems: The Parametcr Analysis antl Design, Wiley, N.Y. 1969. .t,apers 12. Elgerd, o.I. and c.E..Fosha,"The Megawatt Frequency control problem: A New Approachvia optimal control Theory", IEEE Trans.,April 1970,No. 4, pAS g9: 556. 13' Bhatti, T'S., C.S Indulkar and D.P. Kothari, "Parameter optimization of power Systemsfor Stochastic Load Demands" Proc. IFAC. Bangalore,December 19g6. l4' Kothari, M'L., P.S. Satsangi and J. Nanda,"sampledData AutomaticGeneration Control of Interconnected ReheatThermal Systems ConsideringGenerationRate Constraints", IEEE Trans.,May 19g1, pAS_100;2334. 15' Nanda, J', M.L. Kothari and P.S. Satsangi,"Automatic GenerationControl of an Interconnected Hydrothermalsystem in continuous and DiscreteModes considering Generation Rateconstraints' IEE proc., prD, No. l, January19g3,130 : 17. 16' IEEE committee Report,'DynamicModels for Steamand Hydroturbines power in system studies" IEEE Trans., Nov/Dec. rg73, pAS92, 1904. l7' Hiyama, T', "Optimization of Discretetype Load FrequencyRegulatorsConsidering GenerationRate constraints" proc.lE4 Nov. g2, r2g, pt c, 2g5. I8. concordia,c., L.K. Kirchmayer ..Effect and E.A. Szyonanski, of speed Governor Deadbandon Tie Line Power and Frequency Control performan ce,, AIEE Trans. A u g . 1 9 5 7 ,7 6 , 4 2 9 . 19' Nanda,J', M.L. Kothari and P.S. Satsangi, "Automatic Control of ReheatThermal SystemConsidering Generation Rate Constraint and Covernor Deadband,,. J.I.E. (India),June 1983, 63,245. 20. Tripathy, s.9., G.s. Hope and o.p. Marik, ,,optimisatiorr of Loadfrcqucncy C<lntrolParameters Power systemswith ReheatSteamTurbines for and Governor Deadband Nonlinearity", proc. IEE, January rgg2, rzg, pt c, No. r, r0. 21. Kothari, M.L., J. Nanda, D.p. Kothariand D. Das,.,Discretemode of a.two_ AGC area Reheat Thermal system with New Area control Error,,, IEEE Trans. on Power System,Vol. 4, May 19g9, 730 22' Daq D. J. Nanda, M.L. Kothari and D.p. Kothari, ,.AGC of a Hydro_Thermal systemwith New ACE considering GRC", Int. J. EMps,1g, No. 5, rggo, 46r. 23' Das, D', M.L' Kothari, D.P. Kothari and J. Nanda, "Variable Structure Control strategy to AGC of an Intcrconncctcd Rcheat Thermal systcm,,, prctc. IEE, r3g, p t D , 1 9 9 1 ,5 7 9 . 24. Jalleli, Van Slycik et. al.. "lJndersfandingAutonnaticGeneration Control,,, IEEE Trans.on P.S., Vol 07, 3 Aug. 92, 1106_1122. 2 5 . Kothari, M.L., J. Nanda,D.p. Kothari and D. Das, ,,Discrete Mode AGC of a two Area ReheatThermal Systemwith a NACE consideringGRC,,, J.LE. (rndia), vol. 72, Feb. 1992,pp Zg7303. 2 6 . Bakken, B.H. and e.s. Grande,"AGC in a Deregulatedpower system,,, IEEE Trans.on Power Systems,13, 4, Nov. 199g,pp. 1401_1406.
9.1
INTRODUCTION
So far we have dealt with the steadystate behaviour of power system under normal operating conditions and its dynamic behaviour under small scale perturbations.This chapter is devoted to abnormal system behaviour under conditions of symmetrical short circuit (symmetricalthreephase.fault*).Such conditions are caused in the system accidentally through insulation failure of equipment or flashover of lines initiated by a lightning stroke or through accidentalfaulty operation.The systemmust be protectedagainst flow of heavy short circuit currents(which can causepeffnanentdamageto major equipment) by disconnectingthe faulty part of the system by means of circuit breakers operated by protective relaying. For proper choice of circuit breakers and protective relaying, we must estimatethe magnitude of currents that would flow under short circuit conditionsthis is the scope of fault analysis (study). The majority of systemfaults are not threephase faults but faults involving one line to ground or occasionallytwo lines to ground.These are unsymmetrical faults requiring special tools like symmetrical componentsand form the subject of study of the next two chapters.Though the symmetrical faults are rare, the symmetrical fault analysis must be carried out, as this type of fault generally leads to most severe fault current flow against which the system must be protected. Symmetrical fault analysisis, of course, simpler to carry out. A power network comprises synchronousgenerators,ffansfonners, lines and loads. Though the operating conditions at the time of fault are important, the loads can be neglectedduring fault, as voltages dip very low so that currents drawn by loads can be neglectedin comparisonto fault currents.
*Symmetrical fault may be a solid threephase short circuit or may involve are impedance.
325 
t
ModernPowerSystemAnalysis
tffiffi
I
The synchronousgenerator during short circuit has a characteristictimevarying behaviour.In the event of a short circuit, the flux per pole undergoes dynamic changewith associated transients damper and field windings.The in reactanceof the circuit model of the machine changesin the first few cycles from a low subtransientreaetanecto a higher transient value, finally settling at a s'iitt higher synchronous (steady state) value. Depending upon the arc intemrption time of circuit breakers,a suitable reactancevalue is used for the circuit model of synchronousgeneratorsfor short circuit analysis. In a modern large interconnectedpower system, heavy currents flowing during a fault must be interruptedmuch before the steady state conditions are established. of Furthermore,from the considerations mechanicalforcesthat act on circuit breaker components,the maximum current that a breaker has to carry For selecting a circuit breakerwe must, momentarily must also be determined. therefore, determine the initial current that flows on occulTenceof a short circuit and also the current in the transient that flows at the time of circuit intemrption. 9.2 TRANSIENT ON A TRANSMISSION LINE
42V = *sin
lzl
(cr,rf a_ +
A
z = (Rz Jr\tt"(t: +
tanl +)
ir = transient current [it is such that t(0) = t(0) + L(0) = 0 being an inductive circuit; it decayscorrespondingiothe tim6 constantiRl. =  i,(6)e$tL)t = 9Y s i n( d  a ) g  . ( R t D t
tzl
Thus short circuit current is given by
(e.1)
Synrnretrical short circuit current DC otT set curnent
Let us consider the short circuit transient on a transmission line. Certain simplifying assumptions madeat this stage. are (tlte casewhcn the line is (i) The line is led I'rorna constant voltagcsoLrrcc fed from a realisticsynchronons ma.chrne tre treatedin Sec.9.3). will (ii) Short circuit takes place when the line is unloaded(the caseof short circuit on a loaded line will be treatedlater in this chapter). (iii) Line capacitance negligibleand the line can be represented a lumped by is RZ seriescircuit. L , ,F.
+ v = JI vsin (o,t *) rV)
A plot of i* i, and'i = i, + i, is shown in Fig. 9.2.rnpower systemterrninology, the sinusoidal steady state current is called the symmetrical short circuit current and the unidirectional transient component is called the DC offset current, which causes total short circurit the currentto be unsymmetrical the till transient decays. It easily follows fiom F'ig. 9.2 that the maximum momenro) short circuit currcnt i,,,,,, corresponds the firstpenk. theclecay trnnsient to If of current this in short time is neglected,
.l
I
i
r+V\'\'
 Jrv'sin c) * E' (dlzl tzl
Sincetransmission line resistance small. 0  9C,. is
(e.2)
Im*=
_
.
Jiv
rzr
cosa+
JTv
rzl
(e.3)
F i g .9.1 by With the aboveassumptions line can be representecl the circuit rnoclel the to of Fig. 9.1. The short circuit is assumed take place at t = 0. The parameter It <rcontrolsthe instanton the voltagewavewhen shortcircuit occLrrs. is known of from circuit theory that the currentafter short circuit is composed two parts,
1.tr. t I"+ I.t
This has the maximum possible value for o. = 0, i.e. short circuit occurring when the voltage wave is going through zero.Thus = '# i,n,n possible) lrnu* e.4)
= twice the maxirnum of symmetricalshort circuit current (doubling effect) For the selectionof circuit breakers.momentaryshort circuit currentis taken corresponding its maxirnumpossible value(a sat'e to choice).
whcre i, = steadystatecurrent
.w
ffiffif
The nevf ntrecfinn ic
Modern Power System Analysis
turhqf ic fhc r.rrrrcnf fn hc i n f e r r r r n f/ p s l ? t r el lv Aa hqo haan
b,a#&
reactancewhen combined with the leakagereactance of the machine is called Xi synchronousreactance X4 (direct axis synchronousreactancein the case of salient pole machines).Armature resistance being small can be neglected.The ne ls snownln rrg. on per phasebasis.
pointed out earlier, modern day circuit breakers are designed to intemrpt the cunent in the first few cycles (five cycles or less). With referenceto Fig, 9.2 it meansthat when the current is intemrpted,the DC offset (i,) has not yet died the value of the DC offset at the time of intemrption (this would be highly complex in a network of even moderately large size), the symmetrical short circuit current alone is calculated.This figure is then increasedby an empirical rnultiplying factor to account for the DC offset current. Details are given in S ec .9. 5 .
(a) Steady stateshortcircuitmodel of a synchronous machine
(b) Approximate circuit modelduring period shortcircuit subtransient of X1
(c)Approximate model circuit during period short transient of circuit Fig. 9.3 Consider now the sudden short circuit (threephase)of a synchronous generator initially operating under open circuit conditions. The machine undergoesa transientin all the three phasefinally ending up in steady state conditions describedabove. The circuit breakermust, of course,intemrpt the current much before steady conditions are reached. Immediately upon short circuit, the DC offset currents appear in all the three phases,each with a different magnitude since the point on the voltage wave at which short circuit occurs is different for each phase.These DC offset currents are accounted for separately on an empirical basis and, therefore, for short circuit studies, we need to concentrate our attention on syimmetrical (sinusoidal) short circuit current only.Immediately in the event of a short circuit, the symmetrical.short circuit current is limited only by the leakagereaitance of the machine. Since the (theorem of constant air gap flux cannotchangeinstantaneously flux linkages), to counter the demagnetization of the armature short circuit current, currents appearin the field winding as well as in the damper winding in a direction to help the main flux. These currents decay in accordancewith the winding time constants.The time constant of the damper winding which has low leakage inductanceis much less than that of the field winding, which has high leakage
Fig.9.2 Waveform a short of circuit current a transmission on line 9.3 SHORT CTRCUTT A SYNCHRONOUS MACHTNE (ON OF NO LOAD)
Under steady state short circuit conditions, the armature reaction of a generator produces demagnetizing In termsof a circuitthis a synchronous flux.
MocjernPower Sysiem nnaiysis Thus during the initial part of the short circuit, the damper and field inductance. model currentsinducedin them so that in the circr,rit windingshavetransfurnrer thcir reactancesX,of field winding and Xa* of damper windingappear in parallelx with Xo as shorvn in Fig. 9.3b. As the danqpqlruadlag cullqqls 4!q first to die out, Xr* effectively becomesopen circuited and at a later stage X1 from the parallel thus chauges reactance becomes open circuited.The rnachine combinationof Xo, Xy and Xu. during the initial period of the short circuit to of X,,and Xrinparallel (Fig.9.3c) in the rniddleper:iod the short circuit, and presented the machine by finally to X,,in steadystate(Fig. 9.3a).The reactance in the initial period of the short circuit, i.e.
1_: X. rL (11x,,+UXJ+llxd,,) X'j "
FaultAnalysis symmerrical
I'lt,5ffit
I
I
0)
Steady state current amPlitude Tlme
(e.5)
Fig.9. 4
current short symmetrical circuit machine (b)Envelope synchronous of
<>f is called the subtrunsientreoctutxc:e the nrachine.While the reactance effective after the darnperwinding currents have died out, i.e. (e.6) X' ,t= X , + (X,,l l X,) of is called the transientreactance the machirre.Of course,the leactanceunder of reactance the machine.Obviousiy Xf7< steacly conditionsis the synchronous which changesfront < X'd Xu.The machinethus offers a timevaryingreactance Xttoto Xtaandfinally to Xn.
I
Subtransient oeriod
b
l I
I I
I
Steady state period
E a
q)
'o0
g
()
f,
If we examinethe oscillograrnof the short circuit currentof a synchronous trom it, we will tind machineafter the DC ottset cuitentshave beenrettroved the current wave shapeas given in Fig. 9.4a. The envelopeof the current wave shapeis plottedin Fig. 9.4b, The shortcircuit currentcan be divided irtto three periodsinitial subtransientperiod when the current is large as tire tnachine reactance,the middle transient period where the machine offers subtransient and finally the steadystateperiod w\n the machine offers transientreactance, reactance. ofters synchronous : If the transientenvelope is extrapolatedbackwards in tinre, the difference envelopesis the cunent Ai/' (correanclsubtransiertt betwecnthe tlansicrrt sponding to the clamperwinding current) which decays fast according to the clamperwinding time constant.Similarly, the difference Ai/ betweenthe steady with the field time constant. in decays accordance envelopes state1nd transicnt above,we discussed the oscillogram,the cunentsand reactances In termsof can wrlte lIl =
oa t; \t z. lEsl Y,
t
a
Time
(9.7a)
o
E E a
o
i
of Extrapolation steady valrre of Extrapolation transientenvelope
lll =
Actual envelope
\IL
ob _ l E 8 l t; xtd
(e.7b)
tltt= 32.: Y+
J2
where
X,J
(9.7c)
(a) Symmetricalshort circuit armature current in synchronousmachine
Fig. 9.4 (Contd.) *Unity turn ratio is assumedhere.
l1l = steady state current (rms) !//l = transientcurrent (rms) excluding DC component lltl = subtransientcurrent (rms) excluding DC component reactance Xa = direct axis synchronous
'ii#I:
Modern Power Svstenn Anelveie
Fault
n
Xtd= direct axis transientreactance X'j = direct axis subtransientreactance lErl = per phaseno load voltage(rms) Oa,Ob,Oc = interceptsshorvn iLEigs 9Aa and, b_ The intercept Ob for finding transient reactance can be determined accurately by means of a logarithmic plot. Both Ai, and al decav exponentially as Aitt = Ai( exP ( t/q,) Ait = Ai6 exp 1_ t/r7) where r4, and rf arerespectively damper, and field winding time constants with Td* 4 ry At time / 2' r4*, Aitt practicalry dies out and we can write log (Aitt+ At,)1, ,
Trl*
,  log Ai' =  Aint ,/ ry
t l c a b
*+
different types of machines. Tabie 9.r gives typical valuesof machinereactances which can be userjin fault calculations and in stability studies. Normally both generator and motor subtransient reactancesare used to determine the momentary current flowing on occurrence of a short circuit. To decide the intemrpting capacity of circuit breakers, except those which open instantarreously,subtransientreactanceis used for generatorsand transient reactancefor synchronousmotors. As we shall see later the transient reactances are used for stability studies. The machinemodel to be employed when the short circuit takesplace from loaded conditionswill be explainedin Sec. 9.4. The method of computing short circuit currents is illustrated through examplesgiven below.
rhe values reactances tf j? normallyrie within f:T:jr:lltlq certain predictable",l"ilirion), limits for
pon ii'ragntir rutlrarion
s
g
O)
o
o
Fig. 9.5
r f
For the radial network shown in Fig. 9.6, a threephase fault occurs at F. Determine the fault current and the line voltage at l l kv bus under fault conditions.
1OMVA 15%reactance 1 1k V 10 MVA 12.5oh reactance \ \
ne : 3o km,z = (0.27 jo.3e a/ km +
Ai,l,:o : Aito exp(r/ ,t )1,:o: Ai,o : ob Table 9.1 Typicalvaluesof synchronous machine reactances (All valuesexpressed pu of ratedMVA) in
Synchronous
r NO 2: 5 MVA,8ohreaclance riO.0B) / km o F z xn caote / Fig. 9.6 Radial networkfor Example g.1 Solution Select a system base of 100 MVA. V6ltage bases are: I I kVin generators, 33 kV for overhead line and 6.6 kV for cable.
Type of machine
Turboalternator Salient pole (Turbine (Hydroelectric) generator)
compensator (Condenser/ capacitor)
Synchronous motors*
X, (or X,) X^ .t
xd xti x2 xo
ru
l .0 0 2 .0 0.91.s 0 .1 2 0 .3 5 0.r0.25 _ x,d 0.040.14 0.0030.008
0.61.5 0 .4  1.0 0.20.5 0 .1 3 0.35 _ x,d 0.020.2 0 .0 0 30.01s
r.5r2.5
0.951.5 0.30.6 0.180.38 0.170.37 0.0250.16 0.0040.01
0.81.10 0.650.8 0.30.35 0.180.2 0.190.35 0.050.07 0.0030.012
Reactance G, = of Reactanceof G2 Reactanceof Z, =
ro = AC resistanceof the armature winding per phase. * Highspeed units tend to have low reactanceand low speedunits high reactance.
PowerSystemAnalysis
I
Fault Analysis Symmetrical
 I
337
Reactance Tr= ;WY4 of )
= 71.6pu
= (0.93+ j05s) + (71.6) (0.744 i0.99) + (t1.0) + + = I.674+ j4.14= 4.43176.8"pu at Voltage 11 kV bus= 4.43 167.8"x 0.196l 708" = 0;88 I :T ptt = ft88 x 11 = 9;68 kV
x  Z (in ohms) MVA""'" line impedance overhead (kvBur" )2 30x(0.27+j0.36)x100
Q'2
 (0.744 70.99) pu + q21tlq = (0.93 = 3(91!t,rJr0,q, pu + 70.55) cableimpedanc"
(6 .6 )' Circuit model of the systemfor fault calculationsis shownin Fig. 9.7. Since the systemis on no load prior to occurrenceof the fault, the voltagesof the two generatorsare identical (in phaseand magnitude) and are equal to 1 pu. The generatorcircuit can thus be replacedby a single voltage sourcein serieswith the parallel combination of generatorreactancesas shown.
11kV bus +70. +i0.99) i1.6 (0.93 j.0 (o.744 55) ,666' I I  66d. I Cable I t T2 Line T1
A 25 MVA, 11 kV generator with Xl = 20Vo is connected through a transformer, line and a transfbrmer to a bus that suppliesthree identical motors as shown in Fig. 9.8. Each motor has Xj = 25Voand Xl = 3OVoon a base of rating of the stepuptransformeris 25 MVA, 5 MVA, 6.6 kV. The threephase 11/66 kV with a leakage reactance of l0o/o and that of the stepdown of transformer is 25 MVA, 6616.6kV with a leakagereactance l0%o.The bus fault occurs at the point F. voltage at the motors is 6.6 kV when a threepha.se For the specified fault, calculate (a) the subtransientcurrent in the fault, (b) the subtransientcurrent jn the breaker .8, (c) the momentary current in breaker B, and by (d) the current to be interrupted breakerB in five cycles. on line = l5%o a baseof 25 MVA, 66 of Reactance the transmission Given: kV. Assurne that the systemis operating on no load when the fdul" occurs.
i
I l
Flg. 9.8
Fig. 9.7
= + + Totalimpedance (j1.5ll j1.25)+ (t1.0) (0.744 i0.99)+ (i1.6)+ (0.93 70.55) +  1 . 6 7 4 j 4 . 8 2= 5 . 1 1 7 0 . 8 " p u + ''? Isc= tt = 0'196I  70'8"Pu 5.r170.8" = to*.10; = 8,750 A 18u." J3 x6.6 = A Isc= 0.196x 8,750 1,715 F between and11 kV bus Total irnpedance
Sotution Choosea systembaseof 25 MVA. For a generatorvoltage baseof 11 kV, line voltagebaseis 66 kV and motor voltage base is 6.6 kV. (a) For each motor
X',j*= j0.25 x
+
= i1.25 pu
are Line, transtbrmersand generatorreactances already given on proper base values. The circuit model of the systemfor fault calculationsis given in Fig. 9.9a. and motor inducedemfs are The systembeing initially on no load, the generator The circuit can thereforebe reducedto that of Fig. 9.9b and then to identical. Fis. 9.9c. Now
fiffirel
!
power Modern System Anatysis I s c = 3 > < l  + = +
j1.25 j0.55
momentarycurreni ihrough breaker B  1.6 x 7,4i9.5  17,967A (d) To compute the current to be intemrpted by the breaker, motor subtransient reactance (X!j = j0.25) is now replaced by transient reactance (X a = /0.3O).
 j4.22pu r
Basecun'entin 6.5 kV circui, 
25 x 1,000 = 2.187 A
Issc= * r4rrlt]frf* o o,;;
(b) From Fig. 9.9c, current through circuit breaker B is
XI (motor)= 70.3x
25 = Jr.) pu T
I s c (' B ) 2 x  + + . ] _
j1.25
j0.55
:  i 3. 42 r .
110"
= 3.42x 2,187 7,479.5 = A
1 1 0+ '
of The reactances the circuit of Fig. 9.9c now modify to that of Fig. 9.9d. Current (symmetrical) to be intemrpted by the breaker(as shown by arrow) 1 1 ^ " =3.1515pu =2x +
jl.s
jO.ss
jo.2
j0.15
j0.1
110'
F tll?9
;< to"
Allowance is made for the DC offset value by multiplying with a factor of 1.1 (Sec. 9.5). Therefore, the current to be interrupted is x 1 . 1 x 3 . 1 5 1 5 2 . 1 8 7= 7 . 5 8 1A 9.4 SHORT CIRCUIT OF A LOADED SYNCHRONOUS MACHINE
machine, it was In the previous article on the short circuit of a synchronous that the machine was operating at no load prior to the occurrence of aBsumed machine is short circuit. The analysisof short circuit on a loadedsynchronous complicatedand is beyond the scope of this book. We shall, howevbr, present here the methods of computing short circuit current when short circuit occurs under loaded conditions.
lr rBtlc
rl:^ f\
>. Lv
1/.
Slluws
^L^.,^
+L^
Llrg
urrUurt
^i^.,i+
lrlu(lEl
*^A^l
ur
^f
a
i0.55
(b)
i0.55
rcuitbreaker)
(c)
generatoroperatingunder steadyconsynchronous ditions supplying a load current /" to the bus at a terminal voltage of V ". E, is the induced emf under loadedcondition andXa is the direct axis synchronous reactanceof the machine.When short circuit occurs at the terminals of this machine, the circuit model to be used for computing short circuit Fig. 9.10 Circuitmodelof current is given in Fig. 9.11a for subtransient a loaded current,and in Fig. 9.1lb for transientcurrent. The m achine inducedemfs to be used in thesemodels are given bY E,l= v" + ilTtj
Fig. 9.9 (c) For finding momentary current through the breaker, we must add the DC offset current to the symmetricalsubtransient current obtainedin part (b). Rather than calculating the DC offset current, allowance is made for it on an empiricalbasis.As explainedin Sec.9.5,
EL V'+ il"Xto The voltage E!is known as the voltage behind the subtransientreactance and the voltage E!is known as the voltage behind the transient reactance.Infact, if 1o is zero (no load case),EJ= Etr= Er, the no load voltage,in which case the circuit model reducesto that discussedin Sec. 9.3.
(e.8) (e.e)
340

I
Modernpo*s1_qqe!l inslygs
_Symmetrical
Fault Arralysis
solution Aii reactancesare given on a base of 25 MVA and appropriatgi voiiages.
:t
t/o
Prefaultvoltage V" = J'9 l1
= 0.9636 l0
pu
Load = 15 NfW, 0.8 pflEading
( a ) C i r c u i tm o d e l f o r c o m p u t i n g subtransient current (b) Circuit model for computing transient current
F i g .9 . 1 1 Synchronous motors have internal emfs and reactances similar to that of a generatorexcept that the current direction is reversed. During short circuit conditions these can be replaced by similar circuit moclels eicept that the voltage behind subtransient/transient reactanceis eiven bv
= l: = 0.6 pu, 0.8 pf leading 25 prefault _ 136.9. 0.77g3I 36.9"pu currenlI" = _9{__ = 0.9636 0.8 x Voltagebehindsubtransient reactance (generator) E", 0.9636I tr + j0.45 x 0.1783I 36.9"  0.7536 70.28 + pu Voltage behindsubtransient reactance (motor) El,,  0.9636/_ tr i0.15 x 0.7783/_ 36.9" = 1.0336 .70.0933 pu Theplefurult.equivalent is shown Fig.9.l2b. circuit in Uncler fhrrltecl c.nclit i o n( l r i g . .l 2 c ) 9
. , , 0. 7536+"i0. 21t 00 . 6 2 2 6 _ j 1 . 6 7 4 6 p u =0 I';," i 0.45 I',l,= Current in fault 1.03i6 rooc)? 1 i0.1s
E'lr= v"  jI"xU E'*= v"  jI"4
(e.10) (e.11)
Wheneverwe are dealing with shortcircuit of an interconnected system,the synchronousmachines (generatorsand motors) are replaced by their corres po n d i n g i rc u i t m o c l e l s a v i n g v o l tagebehi ncl c h (transi ent) srrhtransi ent reactancein serieswith subtransient (transient) reactance. The rest of the network beingpassive rentains unchanged.
I
9.3 I Example . r^_._ . ... . ._ , .:
".._
A synchronousgenerator and a synchronousmotor each rated 25 MVA, I I kV having l5Vo subtransient reactance are connected through transfbrmers and a line as shown in F'ig. 9.12a. The transfbrmers are ratecl 25 MVA. lll66kV and 66lll kV with leakage reactance of l\Vo each. The line has a reactance of lTTo on a base of 25 MVA, 66 kv. The motor is drawing 15 Mw at 0.9 power factor leading and a terminal voltage of 10.6 kV when a symmetrical threephasefault occurs at the motor terminals. Find the subtransient culrent in the generator, motor and fault.
Gen 
I J = I : i + 1 , , , , = _g . 5 6 5 3 u j p
Basecttrre (gen/moto 44q1 nt I =
) lr;
j0.1 j0.1 j0.1
t Tt'  '!
Line
J3xll
= 1.3 12.2 A
Now
( a ) O n e  l i n e i a g r a mf o r t h o s y s t o m f E x a m p l e 3 d o 9 t" F t ; .6'f,1,'. 'dtI.  'ltrd. t I
i
I'J  1,312.0 (0.6226 j1.6746) (816.4 jL,tgt.4) A _ = I'J= 1,312.2 0.6226 j6.8906)= ( 816.2 jg,O4L8) (._ A 1tjtt,23gA short circuit (sc) current computation through the Thevenin Theorem
, '. )1 i 0 . 1 5 r +l
I I
I
(b) Prefault equivalent circuit (c) Equivalent circuitduringfault
Fi9.9"12
An alternate method ol' cornputing short circuit currents is through the application of the Thevenin theorem.This method is faster and easily adopted
342 
I
I
Modern Po*er SystemAnalysis
to ,yrtlatic computation for large networks. While the method is perfectly gcncrerl, rs illustratcd it hcrc tlrrougha sinrplccxanrplc. generatorfeeding a synchronousmotor over a line. Consider a synchronous Figure 9.I3a showsthe circuit model of the systemunder conditions of steady As a first step the circuit model is replacedby the one shown in Fig. 9.13b, wherein the synchronous machinesare represented their transientreactances by (or subtransient reactances subtransient currents of interest) serieswith are in if voltagesbehind transientreactances. This changedoes not disturb the prefault current I" and prefault voltage V" (at F). As seenfrom FG the Thevenin equivalent circuit of Fig. 9.13b is drawn in prefault voltage V" rn serieswith the passive Fig. 9.13c.It comprises Thevenin impedancenetwork. It is noticed that the prefault current 1" doesnot appearin that the passivcThcvcninirnpcdancc nctwork.It is thcretore t<lbe rcmcnrbcrcd this current must be accountedfor by superposition after the SC solution is obtainedthroughuse of the Thevenin equivalent. now a lault at 1,'thloughan irnpedance .liigure 9.13dshowsthe Zl Consider Thevenin equivalent of the system feeding the fault impedance. We can im m e d i a te lw ri te y ', l '  V" jXrn + Zt
AI^
Xle +X (xlh,+x + xi
(e.14)
Postfault currents and voltages are I{= I" + alr I{=Postfault voltage vf where Av = ix^tf
obtained follows by superposition: as
I" + AI^ (in rhedirection of AI^)
(9.15)
fffl:liiwith
respect thereference ct;;;..d to bus by theflow
vo + (  , r xn, I f ) = v" + Av = eJ6) is the voltage of thefaultpoint F/ on the Theveninpassive of
faurt
incethe prefault current flowing out of t curent out of F.is independeniof load on is summarizedin the following four Step I: Obtain steady state solution of loadedsystem (load flow study). step 2: Replace reactances of synchronous machines by their subtransienU transient values. Short circuit uit sources.The result is the passive Thevenin network. "rr 'Step 2 at the fault point by negarive of ies with the fault impedanc". 6o_pur" rterest. r areobtainedby adding results of Steps The following assumptions be can safery madein SC computations reading :ation: ragnitudes are I pu. tre zero. to actual conditions as under normal nity.
(e.r2)
Current causedby fault in generatorcircuit
AI,  
x'd^
(xhs+x + xl^
f
(e.13)
(a)
G
(b)
G
ruc cnanges ln current caused by short circuit are quite large, of the order of 10_20
(c)
(d)
equivalent Fig.9.13 Computation SC current the Thevenin of by
tion 2. Fi9'e't+ Let us illustrate the above method b, recalculating results of Example the 9.3.
F is the faultpointon the passive Thevenin network
Modern Power SystemAnatysts 9'3 for computationof postfault The circuit model for the systemof Example conditionis shownin Fig' 9'14'
,Qrrrnmalrinal
E^..t+
A r.
!
70.60 =  78.565 pu _ 0.9636x
current due to fault' Change in generator
If sc MVA (explainedbelow) is more than 500, the above multiplyingiactors are increasedby 0.1 each. The multiplying factor for air breakers rated 600 v or lower is 1.25. The current that a circuit breakercan intemr
rng voltage over a certain range, i.e.
Amperes at operating voltage
io'+ Pu , , B_ AI^= rl8.s6s j 0 . 6 0   i2'141 "." "_ t
Change in motor current due to fault'
x Ptt At^= 78.565 j.$*i  i6'424
current to obtain the subtransient To these changeswe add ttre prefault current in machines.Thus I'l= I" + AIr  (0.623 j1.67$ Pu P In =  I" + AI^= ( 0.623 76.891) u calculated already which are the same (and shoutd be) as through th9 Thevenin we have thus solved Example 9.3 alternatively in4eecl,is a powerful method for large theorem ond ,up.rposition. This, networks. 9.5 SELECTION OF CIRCUIT BREAKERS Rated intemrpting MVA (threephase) capacity
= '6ty(tifle)lrated x 11(line)lrated inremrpting cunent
where V(line) is in kV and 1 (line) is kA. Thus, instead of_computing the sc current to be intemrpted, ' r ' we cbmpute threephase MVA to be intemrpted, where SC SC MVA (3phase)_ Jt x prefault line voltage in kV x SC currentin kA. If voltage and current are in per unit values on a threephase basis
SC MVA (3phase) = lylp,..roul, 11116 (MVA)uur. x x
O hL i' o r r ul'ru J r J ' irc if i o A l \ / \ / '' v Y rs r q uu
(e.r7)
. I w o o t t h c c i r c u i t b r c a k c r r a t i n g s w h i c l r r c c ; u i r c t h c c t r n l p t l t t t t i o n o f S C c u r r e n t rnurcthln (or cclualto) thc sc MVA required to be intemupted. symmetrical interruptirtg c:nrrent' For the selectionof a circuit breakerfor a particular are: rated momentarycurrent and rated location,we must find. using subtransient reactancesfor the maximum possible SC MVA to be intemrpted with Syrnmetrical SC current is obtained by respectto type and by location of fault and generating capacity (also Momcntary current irms) is then calculated synchronous synchronousrnachines. rnotorl load) by a factor of 1'6 to accountfor connectedto the system. A threephase fault though rare is generally the one multiplying the symmetricalo"nory current which gives the highest SC MVA and a circuit breaker of the presence DC offset current' must be capable of subtransient interrurptingit. An exception is an LG (linetoground) is computedby r'rsing currentto be intcrrupted Symmetrical f.ault close to a for synchronous and generator*.In a simple systemthe fault locationwhich gives generators transientreactances synchronous tor reactances synchronous the to be addedto The DC offset value highest sc MVA may be obvious but in a large system motorsinduction motors are neglected*. various possible multiplying the accounted for by locations must be tried our to obtain the highestst nava obtain the current to be interrupted is requiring Lp"ur"a SC computations.This is ilustrated by the examplesthat by a factor as tabulated below: symmetrical SC current follow.
Circuit Breaker SPeed 8 cyclesor slower 5 cycles 3 cyclcs 2 cycles Multiplying Factor 1.0 1.1 1.2 1.4
 l v t v AA ii  +i^ *i,r u. pi i i l g  c a p a c l t y o f a c i r c u i t n e  : breaker is to be
r.; Iii"'n" I
Three6.6 kv generators B and c, eachof I0o/o A, leakagereactance MVA and ratings 40, 50 and 25, respectively are interconnected electrically, as shown in
tThis will be explained in Chapter 1l.
during a short currentscontributedby induction motors t'In some recent attempts, circuit have been accountedfor.
15:!"'l
.";: I
T
{d46..t
AnalysisPowerSystem Modern
currentscan then be calculatedby the circuit model of Fig. g.l6acorresponding to Fig. 9.13d.The circuit is easilyreduced rhat of Fig. 9.16b, ro where 7 ( 0. 069 + j0. 138) + j0. r 2s il 00. 15+ jo. 22 j0. 44)
reactance Fig. f.i5, by a tie bar through curent timiting reactors,each of I2Vo A to which it is connected. threcphase baieclupon the rating ofthe machine of 6'6 kV' feeder is supplied from the bus bar of generatorA at a line voltage
^ f n 1 a
short Q/phase.Estimatethe maximum MVA that can be fed into a symmetrical far end of the feeder. circuit at the
= 0.069 j0.226 = 0.236173 + SC MVA = Volf = V"('+') = + pu (sinceVo = 1 pu) \Z) Z I Z
(MVA)Ba."
50 = 2 1 2M V A 0.236
Tie bar
Fi g. 9.15 Sotution Chooseas base50 MVA, 6.6 kV' Feeder imPedance Considerthe 4bus systemof Fig. 9.17. Buses1 and 2 aregenerator busesand 3 and 4 are load buses. The generators are rated l l kv, 100 MVA, with transientreactance l07o each. Both the transformersare 1ll110 of kV, 100 MVA with a leakagereactanceof 5Vo.The reactances the lines of to a base of 100 MVA, 110 kv are indicated on the figure. obtain the short circuit solution for a threephase solid fault on bus 4 (load bus). Assume prefault voltages to be 1 pu and prefault currentsto be zero.
(G)
=
%
=(o.o6e+/0.138)pu
 o'1[50 = 0.125Pu Gen A reactance 40 = 0.1 Pu GenB reactance * 4 G e n C r e a c t a n c e = 0 . 1 = 0.2 pu 25 = o't''I tn = 0 . 1 5p u A Reactor reactan." 40 = B Reactor reactance 0.12Pu x 0.12 50  0.24 pu Reactor C reactance =
j0.2
1
Fig. 9.17 Fourbus g.5 system Example of Solution Changesin voltages and currentscausedby a short circuit can be calculatedfrom the circuit model of Fig. 9.18. Fault current 1/ is calculated by systematic network reduction as in Fig. 9.19,
(b)
j0.12 +i0.138) (0.069
yo = 1Z0o(
(a)
jo.24
Fig. 9.16
,^o
I
I
Moclern Power Svstem AnalYsis
=  jt.37463 pu tt = *= j0.13s60
lo*
14
!o;E
_ jo.2
jo 1,), 7 I 4+
D':
r  rnTl
[ 0t ,
i0.1
A, l,zs
I
t,= rsx i: i3;:: =  j3 837or pu j0.37638
12=r, x {: i:::: =  j3.53 pu 762 '' j0.37638
Let us now computethe voltagechanges busesl,2and 3. From Fig. fclr 9 .l 9 b , w c g i v c AVr  0  ( / 0. 15)(  j3. 8370r ) =  0. 57555 pu
ib.rs
Izt
2
t, ti, ,
10.15
i ro'rs
o)
\
AV, = 0  (iO.l.s)( .i3.53762)  0.53064pu = Now
'AtiA'
r fI l, i0 1'':,
t 
i.0.11
5 V , ' = l + l V t = 0 . 4 2 4 4 .p u
F'tt r l r ' ' r),} o' rs i o.t\P I I (
2 i I 0 6 ' )I
V ) ,= l + . / lt=
N ow
AVz=0.46936pu
, , t , . ( .^l
4T ( tvl=to * irt
t l
iots
(a) (b)
(. )vl = t.o ir
l
v , J v J = J0'17964 pu ,of tr.o.t
fr'
AVy= 0  [(/0.15) j3.83701) Q0.15) (+ (/0.17964)l =  0.54fi(r0 pu V t t = I  0 . 5 4 8 6= 0 . 4 5 1 4 u 0 p
(e)
(c)
l ' ' l
+
I ( ) v?= t.o
vlo= o
The determination currentsin the remaining of lines is left as an exerciseto tltcrrr'ldcr. Short circuit study is completewith the computation SC MVA at bus 4. of (SC MVA)^ = 7 .37463 x 100 = 737.463MVA It is obvious that the heuristic networkreductionprocedureadoptedabove is not practical for a real power network of even moderate size. It is, therefore, essentialto adopt a suitablealgorithmfbr carryingout short circuit study on a digital computer.This is discussed Sec.9.6. in 9.6 ALGORITHM FOR SHORT CIRCUIT STUDIES
l
'' ,j;0. uuuo
(
, ,l J 10.04166
Tr
i
(
t
tt*)
\.. ) V ?
.
= 1.0
Ill I
reductionof the network of Fig' 9'18 Fig. 9.19 Systentatic
So far we have carriedout short circuit calculations simple systemswhose for t can I wc l l i tssi vc t ct wor ks be easilyr educed.n t his scct ion ext cnd our st udy t o
i*.3€0l
Modern Power System Analysis _ Symmetr.ical  , . . . . . , *", , ; FaultAnat lr sis
large ,tyrt"r. In orcler to apply the four steps of short circuit computation it to developedearlier to large systems, is necessary evolve a systematicgeneral algorithm so that a digital computercan be used.
_.rr.vv\, network.
rrrv u(rr vurraegcs oI mls
Now
4V = Z"urJf
where
otn Gen 2
.7 'l '
(e.20)
I
i = busimpedance matrix of the I Znn passive Thevenin network )
u// = bus current injection vector Since the network is injected with current 
e.2D
Fig. 9.20 nbussystemundersteadyload in Consideran nbus systemshown schematically Fig.9.20 operatingat steady load. The first step towards short ciicuit computation is to obtain prefault voltages at all buses and currentsin all lines through a load flow study. Let us indicate the prefault bus voltagevector as
0 0 : /rf I, : I', f
1/ only at the rth bus, we have
(e.22)
(e.18)
:
Substituting Eq. (9.22) in Eq. (g.20), we have for rhe rth bus AV, =  ZrJf By step 4, the voltage at the nh bus under fault is
Let us assume that the rth bus is faulted through a fault impedance Zf . The postfault bus voltage vector will be given by
V{ur= VBus AV +
(e.1e)
where AV is the vector of changes bus voltagescausedby the fault. in As'step 2, we drawn the passiveThevenin network of the system with generators reactances with their emfs shorted replacedby transient/subtransient ( F ie.9 .2 1 ).
v!= vor+avo, vor Z,Jf
However, this voltage must equal Vd = 7f 1f We have from Eqs. (9.23) and (g.24) zftf  vo,_ z,Jf
(e.23) (e.24)
or
f=
V: Zr, + Zf
(e.2s)
At the rth bus (from Eqs (9.20)and(g.22)) AV, =  Z,Jf
Fig. 9.21 Network of the system of Fig. 9.20 for computingchanges in bus voltages caused by the fault
v{= v? Z,Jf, i = 1,2, ..., substituting // from Eq. (9.25).we have for
(e.26)
vI= vf  : zl';rv! z*+L
(e.27)
Fault ,Analysis Siimmetrica! 352 i Power System Analysis rtllodern Fori=rinEq.(9.27)
I ?53 rca 
matrix for the network of Fig. 9.18 is formed First of all the bus admittance as fol l ows:
(e.28)
In the above relationship V,o'r, the prefault bus voltagesare assumedto be known from a load flow study. Zuu, matix of the shortcircuit study network of Fig. 9.21 canbe obtainedby the inversionof its furr5matrix as in Example in 9.6 or the Zru, building algorithm presented Section9.7.It should be observed here that the SC study network of Fig. 9.2I is different from the correspondingload flow study network by the fact'that the shunt branches do reactances not appearin the load flow study to corresponding the generator are network. Further, in formulating the SC study network, the load impedances much larger than the impedances of lines and ignored, these being very ginerators. Of course synchronousmotors must be included in Zuur tormulation for the SC study. Postfault currents in lines are given by For calculation of postfault generatorcurrent, examine Figs. 9.22(a) and (b). From the load flow study (Fie. 9.22(a)) Prefault generator output = PGr+ iQci
r I
I
t
1
r
I
  j28.333
l Y trL. > = Yt r t = _ L
l
j0.2
175.000
Y3= Yy =
;r* 1 Yrq= Yq= J,3.l
 i6. 667
= i10.000
Y z z = :  + +j0.1s + jj.r . j0.15  + + : ^
" Y z t =Y n = + =
Yzq= Yn=
l
j0.2
= j28.333
f u= yu (vri vt)
Q.29)
j0. 1
j10.000
 i6. 667
F; 'Y 3 3 .=
I I =i16.667 + ' j0.15 jo.1
Ytq= Yqt= 0.000 I v Yqq= ,(a)
*
I ,.""

i76'667
(b)
Fi1.9.22 f", = &#; = (prefault generator output Pci + iQci) (9.30)
(9'31) By inversion we get Z",tt as
v,u
E'Gi = V, + jXt"/t)",
From the SC study, Vf ,is obtained. It then follows from Fig. 9.22(b) that
rrc,=tfr! iir;rnr;;.; I
(e.32)
j0.0s97 j0.0719 j0.0903 j0.0780 j0.0780 j0.13s6 jo.o719 j0.0743
Now, the postfault bus voltagescan be obtainedusing Eq. (9.27) as V { = r V Z  Z t oo VaP o o r
above,we shall recomputethe short circuit To illustrate the algorithm discussed solution for Example 9.5 which was solved earlier using the network reduction technioue.
,2.tr!A
JJ'
I

^/lndarnl lYlVVVll
Pnrrrar I Vltvl
Qrratarn Vtvlvlll
, rlrsrtvre
Analrrcic
malriaal   tglt tvql
E qa t u
I
The prlfault conditionbeing no load, V0r= Voz= V03 Voo= 1pu
r
, J

_
9 r . o  i 9 9 1 9 x 1 0= 0 . 4 2 4 8 p u
i0.1356
r
1.000 2,,{or Zzz)

1.00 =  j I L 0 7 4 I 9 7 pu j0.0903
vrz=v:  o?': L L z o
= 1.0 
r,',
By Inventing Y"u" /nus = Yrus Vsus or
or
oi 0 .0 7:1 0 x 1 . 0 = 0 . 4 6 9 8 p u :
j0.1356
7
v{= v\  1* vf Zoo
 :: _: = 1 . 0 ' i0.0743 1 . 0= 0 . 4 5 2 1 u p j0.r3s6
= Vsus [Y"ur]t /eus = Znus/eus
(e.33)
Zsvs= [Yuur]t The sparsity of fsu, may be retained by using an efficient inversion technique [1] and nodal impedance matrix can then be calculated directly from the factorized admittancematrix. This is beyond the scope of this book. Current Iniection Technique
vi = o'o
Using Eq. (9.25)we can obtainthe fault currentas =, j7.3j463pu 7r= .r^0.9^o=j0.13s6 earlierin Example9.5.Let us also with thoseobtained agree Thesevalues in the calculate shortcircuitcurrent lines13,12, l4,24 and23. v,r v{ [,r 2 = : 1 z.n 0.42480.4521 j o ' 1 8 2 Pu i0.15
Equation (9.33) can be written in the expandedform
V 1: 2 1 1 \ * Z t z l z 1 . . . * Z n I n v2:22111+ 22212 ... + zznln + V, : Zntll * 2,,21r* .. .+ Z,,nl,, It immediately follows from Eq. (9.34) that
(e.34)
 r r z = v,r v{2  :  0.42480.4698_ j 0 . 2 2 5 u '_ ' tr p zrz .i0.2 r ,r t4 yr v{ 4q zz+ _ o.4z4vo jo'l  i4.248pu
 j3.132pu
"ii
zv'l 
I r l , r : r z: . . . : r n = o
I1r0
(e.35)
q . irq = ' u'l  v{ I 
0.46e8o =
i0'15

Also Zi,  Ziii (Znvs is a symmetrical matrix). As per Eq. (9.35) if a unit currentis injectedat bus (node)7, while the other ere buses kept oponcircuited,the bus voltagesyield the valuesof theTth column of Zuur. However, no organized computerizabletechniquesare possible for finding the bus voltages.The technique had utility in AC Network Analyzers 'where the bus voltagescould be read by a voltmeter.
0.4521 r v{ v^ = 0.4698 =  ./0.177 IIzt=tiutt Pu TOFor the exampleon hand this method may appearmore involved compared to the heuristicnetworkreductionmethodemployed in Example9.5. This, however, is a systematicmethod and can be easily adopted on the digital computer for practical networks of large size. Further, another important feature of the method is that having computed Zu's, we can at once obtain all the required short circuit data for a fault on any bus. For example,in this particular system, the fault current for a fault on bus I (or bus 2) will be Consider the network of Fig. 9.23(a) with three buses one of which is a Evaluate Zsus. reference. Sotution Inject a unit current at bus I keepingbus 2 open circuit, i.e., Ir = I. andIr= 0 as in Fig. 9.22(b).Calculatingvoltagesat busesI and 2, wehave Ztt=Vt=7 Zzt=Vz=4
356 
.f
Modern PowerSystemAnalysis
Fault Analysis Symmetrical
!, f5l
Now let It = 0 and 12= 1. It similarly fbllows that Ztz=Vt=4= Zn
Zvus= 6l l+
l7
41
Becauseof the above computationalprocedure,the Zru, matrix is referredto as the 'opencircuitimpedance matrix'. Z"vs Building Algo4ithm
the dimension of Zsu5 goes up by one). This is type2 modification. branch(i.e., a new loop is formed an 3. Zuconnects old bus to the reference This is type3modification. dimensionof 2o,," doesnot change). (i.e., new loop is formed but the dimensionof two old buses 4. Zuconnects Zuu, does not change).This rs type4 rnodittcation. 5. Zu connectstwo new buses(Zeus remains unaffectedin this case). This situation can be avoided by suitable numbering of busesand from now onwardswill be ignored. Notation: i, jold buses; rlsfelence bus; knew bus. Type 1 Modification
, 4r,
It is a stepbystep programmabletechniquewhich proceedsbranch by branch. It has the advantagethat any modification of the network does not require complcte rebuilding Z"ur. of Consider that Zrur has been formulated upto a certain stage and another branchis now added.Then Zrur (old) Zo:branch impedance Zsus (new)
Figure 9.24 shows a passive (linear) nbus network in which branch with impedance2,, is addedto the new bus k and the referencebus r. Now
V*= ZJ*
Z r i = Z * = 0 ; i = 1 , 2 . . . . .t t Zm = Zu
Hence
Upon adding a new branch, one of the following situations is presented.
Zsvs (new) 
(old) s,,"
(e.36)
Passive linear nbus network
Fig. 9.24 , Type1modification Fig. 9.23 Current injection method computing of Zru. Type2 Modification
t . 26 is added from a new bus to the referencebus (i.e. a new branch is
added and the dimension of Zry5 goes up by one). This is typeI modificution.
Zo is addedfrom new bus ft to the old bus7 as in Fig. 9.25.It follows from this figure that
35S;,1
MorJern
Powcr
Srrctarn
anatrroio
1O_
Zrj
lr
n
Passivelinear nbus network
zzj
Zsvs(old) I :
(e.38)
(e'I )' 3 Eriminate { inthe 3":';,,iiy';;;?::"i"t i.I" a;y>;i:ation
Fig. 9.25 TypeZ modification
oI
lzyziz...zi,I zii + z, )lI*
'o= 12
1
or
Now
(\1Ir+ Zi2Iz+ "' + ZlnI)
(9'39)
Vo= Zdo + V, = Zr,I*+ ZiJr+ Zlzlz+... + Zii ei* I) + ... + Zinln
Rearranging, V*=4lt+ Consequently Zti Zzj :
Znj
V; = 2; 111+Z, r I , + "' + Z'nI n+ Z; / r SubstitutingEq. (9.40) in Eq. (9.39)
(e.40)
Z l z l z+ . . . + 2 , , 1 , + . . . +Z i , l n + ( Z i i + Z ) l k
,,= lr^ h(z*
* 1rr]a lr,,
rz uz,,>)r, "h
2,,1)r, (s.4t)
+ + lr^ ^i{zu
Equation (9.37) can be written in matrix form as
(e.37)
zjj + zb
IYpe3 Modification
: ltz't"Zv) = (new) znvs(old ;+;l ZsLr5 "jjT"ulz,, l
TypeModification
.
t"f
(e.42)
zu connectsan old bus (l) to the referencebus (r) as in Fig. 9.26. This case follows from Fig.9.25 by connectingbus ft to the referencebus r, i.e. by setting
can . as two zo connects old buses in Fig. 9.27 Equations be writtenas follows for all the networkbuses.
v*=o'
(l;+ lp
J
Passive linear nbuS network
modification Fig. 9.27 TYPe4
Fig. 9.26 Type3 modification
"' + Zr i ( I i + / ) + ZU Q i /o)+ ...+ ZiJ"(9.43) Similar equationsfollow for other buses. + Vi= Z; 11, Z, lr +
360 
Modern power System Analysis
Fault Analysis Symmetrical Step I: Add branch 2,,  0.25 (from bus I (new) to bus r) = ZBUS t}.25l Step
I fOt
I he voltages of the buses i and j are, however, constrained by the equation (Fig. 9.27)
V,= Z ,rl o + V i or 2 1 1 1 1 \ 2 1 2 + . . . + Z i t ( 1 , + I ) * Z i i ( I i  I ) + . . . + zi,In + = Z t l * + Z , r l , + Z , r l "+ . . . + Z l U , + I _I earTangtng 0 =(Zit 21) Ir+ ... + (Zti Z) Ii+ eu Z) Ii +...+ (Z*  21,) In + (Zu * Zii * Zii  Zi  Z) 11,
(e.44)
(i)
Add branch Zzt = 0.1 (from bus 2 (new) to bus I (old)); typeZ
e.45)
zsvs=',lZ:11, 31i]
= zsvs o.zs o.3s o.2s   lo.2s 0.2s 0.3s..l
Step 4:
(ii)
(iii)
Collecting equations similar to Eq. (9.43) and Eq. (9.45) we can write vl
Step 3: Add branch ,rt = 0.1 (from bus 3 (new) to bus I (old)); typeZ modification
V
vn 0
Eliminating 1o in Eq. (9.46) on lines similar to whar was done modification,it follows that
(new) = Zsuts (old) Z131,s
lo.2s o.zs 0.2s1
(9.46)
Add branch zz, (from bus 2 (old) to bus r); type3 modification
10.2s o.zs 0.25'l =  Zet)s o.zs 0.3s 0.2s  o'3so +
[0.25l 0.3s  o.tsI to.zs 0.251
zb+zii*Ziizzij
l'"
lZ,,
Lo.2so.2s o3sl
" Ln.trj
[0.14s8 0.1042 0.14s8l = I o.ro+z o.r4s8 o.lo42 
lzit  z1t)... (zi,  zi))
fo.tott o.ro42o.24s8l
Step5: Acldbranchizr= 0.1 (from bus 2 (olct)to brrs31old)): type4 modification [o.l4s8 0.1042 0.l4s8l
with the useof fbur relarionships Eqs (9.36), (9.37),(9.42) and (9.47) bus impedancematrix can be built by a stepbystep procedure(bringing in ng hci one bra n c h a t a ti n l e ) a s i l l trs tra tc 'itl l i xrrrnpl e9.8. Thi s pror.cdrrrc n a mechanicalone can be easily computerized. When the network undergoes changes, the modiflcationproceclures be can cttrployccl rcviscthc bLrs to irtrpcdance rnatrixul'thc nctwol'k. Thc opeling o1 a line (Ztt) is equivalentto adding a branch in parallel to it with impedance  2,, (see Example9.8).
= zsvs o.'oo, 0.1458o.to42l8 l  0 . 1 0 . 1 4 5+ 0 . 2 4 5 8  2 x O . l O 4 2 +
) 10.l4s8 O.tO42 0.24s8 1042.l [0 = 0.0417 [o.to+z 0.0411 0.0417] I 
j Example9.8
For the 3busnetworkshown Fig. 9.28build Zsus in I I
I
[o.o+tzl 0.1103 0.12501 lo.r3e7 = 0.1ro3 o.I3e7 0.1250  I  0.1250 0.1250 0.17sol
an to o Oltenirtg tinu (line 32): This is equivalent connecting intpeclance 0.I betweenbus 3 (old) and bus 2 (old) i.e. type4 modification.
!,0.25 l,'
(I l
I
r 9 z 'irl l I l L, ili ,  6"11. I tr 3
0.1 o.t
I i,J
or.
Ref bus r Fig. 9.28
I
Zsus=Zuur(olcl) (01)+ol?5 #
filf.lid,
ModernPower System Anatysis
rr^r :)ymmelrlual r^..tr raull Ana*raio rurqryo's ffir r**
L o.osoo_J
I
[ 0.0147l t 0.0147 l 10.0147 0.01470.0s001  
Now
  0.1042 0.14s8 0.1042l; (sameas in step 4) 0.1458 0.1042 0.2458 
and
i' li = ( F^ E  ) = 0 . 2 8 6 2[ Z
zu)
symmetry of the given power These two voltages are equal becauseof the network (c) From Eq. (9.29)
Iri = Y,j(vl  vl)
For the power system shown in Fig. 9.29 the pu reactances shown therein. are For a solid 3phasefault on bus 3, calculate the following (a) Fault current
= Irtz=+(0.286  0.286; g ' j0.1
and
(b) V\ and vt (c) It,r, I'\, uxl Il, (d) 1fr, andIf, Assume prefault voltage be I pu. to
0.2
JIc)  i,,
I\t=tt= ro.186o) fr
  i2.86 As per Eq. (9.32) ,r r GrE' orvrf ixtic + ixr
ffr
F
0.09 1
(d) 0.1
\, o '' \ \ / , ,/
'l
// ,/'0., \.. .
l 3 '
But
no = E'Gr 1 Pu (Prefault load)
IL.= GtSimilarlY
10'286= j2'86 io.2+io.os
Fig. 9.29
I f cz= i2. 86
Solution The Thevenin passivenetwork for this systemis drawn in Fig. 9.28 with its Zru, given in Eq. (iv) of Example 9.8. (a) As per Eq. (9.25) Jt 
LE PROB MS
5 0.1 anclresistance ohms is suddenly line 9 . 1 A transmission of inductance H as shown in Fig' P9'1' Write the shortcircuitedat t =0 at the bar end approximatelythe value of for expression short circuit current i(r). Find (maximum momentary current)' the first current maximum currentntaximum occurs at the sametime as [Hint: Assumethat the first ttrcuit current') the first current maximum of the symmgtricl.:non
V: 2,, + Zl
or
1f 
zn
Y o  : j0.175  j s . j l ._1
(b) As per Eq. (9.26) rr.fv
i 
vl'  Ztyu
zrr+zl'r
564,'l
I
power System Analvsis rr]logern
vyrrrrlvt"vE'
errrnrnorrinal Farrlt Analvsis "' " J
I
I
geS
I
fl
'i
64A' v^/v\
0.1H
SO
i
u = 1 OOn(100 f + 15.) si n
Fig. p9.1 9 '2 (a) What should the instantof short circuit be in Fig. p9.1 so that the DC offset current is zero? (b) what should the instantof short circuit be in Fig. p9.1 so rhar the DC offset current is maximum? 9 . 3 For the system of Fig. 9.8 (Example 9.2) find the symmetrical currents to be interrupted by circuit breakersA and B for a fault ar (i) p and (ii) e. 9 . 4 For the system in Fig. p9.4 the ratings of the various componenrs are: Generator: 25 MVA, 12.4 kV, l\Vo subtransient reactance Motor: TransformerT,: Transfbrmer Tr: Line: 20 MVA, 3.8 kV, l5%o subtransient reactance 25 MVA, 11/33 ky, gVoreactance 20 MVA, 33/3.3 kV, I \Vo reactance 20 ohms reactance
Fig. P9.5 ansi 9 . 6 A s y n c h r o n o u s g e n e r a t o f r a t e d 5 0 0 k V A , 4 4 0 v , 0 . l p u s u b t rpowele n t 0.8 lagging ."u"'tun". is supptying a passiveload of 400 kW at for a threephasefault factor. Calculate tfr" in'itiui symmetrical rms current at generator terminals. unit is connected to a line through a circuit 9 . 7 A generatortransformer breaker.The unit ratings are: pu, Xi= o,2o pu and Generator: 10 MVA, 6.6 kV; X,,d= 0.1 = 0'80 Ptt X,r
The systemis loadedso that the motor is drawing 15 Mw at 0.9 loading power factor, the motor terminal voltage being 3.1 kv. Find the subtransientcurrent in generatorand motor for a fault at generator bus. lHint: Assumea suitablevoltagebasefbr the generator. The voltagebase for transformers,line and motor would then be given by the transforma_ ti 0 n ri l ti < l sF rl r c x a n rp l ci,f' w c chooscgcrrcrator. tagcbasc . vol as I I kv, the line voltage baseis 33 kv and motor voltagebaseis 3.3 kv. per unit reactances calculatedaccordingly.] are
Transformer:10MVA,6.9133kV,reactance0.08pu of 30 kv, when a threeThe systemis operatingno load at a line voltage breaker'Find phasetault occurson ih" tin" just beyondthe circuit symmetrical rms current in the breaker' (a) the initial offset currenr in rhe breaker, iui tt. maximum possible DC current rating of the breaker' (c) the momentary kVA' (d) the current,o U. intemrpted by the breakerand the intemrpting and short circuit current in the breaker' (e) the sustained 50 MVA at 1 1 k v , 0 . 8 9 . 8 The systemshown in Fig' P9'8 is delivering ils infinite. laggirrgp0w0rl.itctttrillttlltbrrswhichrtriryberegirrdecl are: of Particulars various systemcomponents 60 MVA, 12 kV' X,/ = 0'35 Pu Generator; kV' X = 0'08 pu (each): 80 MVA, 12166 Transtbrmers 12 Reactance ohms' resistanc:negligible' Line: that the circuit breakersA and B will Calculatethe syrnmetricalcurrent fault occurring at be called upon to interrupt in the event of a threephase B' F near the circuit breaker
Line
T
,1
T2
Fig. p9.4 9 . 5 Two synchronous motors are connectedto the bus of a large system through a short transmission line as shown in Fig. p9.5. The ratings of vanouscomponents are: Motors (each): 1 MVA, 440 v,0.1 pu transientreactance Line: 0.05 ohm reacrance Large system: Shortcircuit MVA at its bus at 440 V is g. when the motors are operating at 440 v, calculate the short circuit cuffent (symmetrical) fed into a threephase fault at motor bus.
Fig. P9.8
*Sf,S
tgdern power Svstem Anatvsis t . 9'9 A two generatorstaticln suppliesa f'eeder througha bus as shown in Fig. P9'9' Additional power is fed to the bus throJgh a transfoilner from a large system which may be regardedas infinite. A reactor X is included
bus 2 of the system.Buses 1 and 2 are connectedthrough a transformer and a transmissionline. Per unit reactancesof the various components are: to Generator(connected bus bar 1) 0.25 line Transmission 0.28
and,1"!T to limit theSCrupruring capacity of :,'jT:^"1^:t::1u1.',f"'Ter 4v<rl\er
rr ru JJJ rvtyA (Iault close to breaker). Find the
inductive reactanceof the reactor required. system data are: Generator Gr: 25 MVA, 757o reactance Generator Gr: 50 MVA, 20Vo reactance Transformer Tr: 100 MVA;8Vo reacrance Transformer T2: 40 MVA; l\Vo reactance. Assumethat all reactances given on appropriate are voltagebases.choose a baseof 100 MVA.
ti Tz
The power network can be representedbi a gene ator with a reactance (unknown) in series. on With the generator no load and with 1.0 pu voltage at each bus under operating condition, a threephaseshort circuit occurring on bus 1 causesa current of 5.0 pu to flow into the fault. Determinethe equivalent reactanceof the power network. are 9.12 Considerthe 3bus systemof Fig. P9.I2. The generators 100 MVA, with transientreactanceIjVo each. Both the transforners are 100 MVA of with a leakage reactance 5%t. The reactanceof each of the lines to a base of 100 MVA, 110 KV is 707o.Obtain the short circuit solution for solid short circuit on bus 3. a threephase Assume prefault voltagesto be 1 pu and prefault currents to be zero.
Fig. p_9.9
T1
G)
k 1 1 /1 1 0 V f l rn kv\xxx"
^T" T '2
9'10 For the threephase power network shown in Fig. p9.10, the ratings of the various componentsare:
j0.1
Fault
Fig. p9.10 Fig. P 9. 12 Generators Gr: 100 MVA, 0.30 pu reactance Gz: 60 MVA, 0.18 pu reactance Transformers(each): 50 MVA, 0.10 pu reactance Inductive reactor X: 0.20 pu on a base of 100 MVA Lines (each): 80 ohms (reactive); neglect resistance. with the network initiaily unroaded and a rine vortageof 110 kv, a symmetrical short circuit occurs at mid point F of rine r. calculate the short circuit MvA to be intemrpted by the circuit breakers and B at the ends of the line. what A would these values be, if the reactor X were eliminated?Comment. impedancedata are of 9.13 In the systemconfiguration Fig. P9.12,the system given below: Transient reactanceof each generator= 0.15 pu Leakage reactanceof each transformer = 0.05 pu = Ztz = i0.1, zp  i0.12, 223 70.08 Pu For a solid 3phasefault on bus 3, find all bus voltagesand sc currents in each component.
365 j
t
ModernPower SystemAnalysis
9.14 For the fault (solid) location shown in Fig. P9.I4, find the sc currentsin lines 1.2 and 1.3. Prefaultsystemis on noloadwith 1 pu voltageand prefault currents are zero. Use Zuu, method and compute its elementsby the current injection technique.
ll
I
'r[bl Q9 Fig. P9.14
110 111 kV 0.5 pu reactance 0.1 pu reactance
10.1
INTRODUCTION
ERE S REF NCE
B oo k s
l. B I r r w t t . H . L ) . . S o l t t t i o t rt t . f I . a r , g , e e t w r t r k l t y M u t r i r M c t l u t d s , W i l c y , N c w Y o r k , N 'i'cxtbook M t t t i t : t ' t t i ' t n v ' u ' 5 ' v s ' / e l l , l ,i r r t c r r t a t i t t r t a j eorrrplny,
In our work so far, we have considered both normal and abnormal (short circuit) operettionsof power systern Llncler cornpletely balanced (symmetrical) c o l t t l i t i o t t s . L J l r r l c rs u c l t o l t c t ' a t i o t t l t c s y s t c r l r i r u l t c d a n c c si n c a c l r p h a s c a r e identical and the thrcephasevoltages and currents throughout the system are c o l l t p l c t c l y h l t l i t n c c t l .i . c . t h c y h u v c c r p r i r l n u r g r r i t u t l c sn c l r c h p h u s c u n d t r c i prtrgrcs.sivcly displaced in tirne phase by 120' (phase u leads/laesphase b by 1 2 0 ' a n d p h a s e i r l e a d s / l l g sp h i t s e c b y 1 2 0 " ) . I n a b : r l l n c e d s y s t e n r , l n a l y s i s can proceed on a singlephase basis. The knowledge of voltage and current in one phase is sufficient to completely determine voltages and currents in the other two phases. Real and reetctive powers are simply three times the corresponding per phase values. Unbalanced system operation can result in an otherwise balanced system due to unsymmetrical fault, e.g. linetoground fault or linetoline fault. These f'aultsatrc,itt lhct, ol' lnot'econlnroll occurrence+thalt the syrnntetric:al(threephase) fault. Systern operation may also become unbalanced when loads are unbalanced as in the presence of laree singlephase loads. Analysis under unbalancetl conditions has to be carried out on a threephase basis. Alternatively, a nlore convenient method of analyzing unbalanced operation is through symtnetrical components where the threephasevoltages (and currents) which may be unbalanced are transfbrmed into three sets of balanced voltages (and * Typical relative frequencies occurrenceof different kinds of faults in a power of (irr syst(:llt ordcr ol' dccrcasing scvcrity)are: Threephase(3L) faults 5Vo Double linetoground(LLG) faults lj%o I)outrlc lirrc (l1,) lirults  5t/o (LG) faults Single linetoground 7jVo
tL)'7 5. 2 . f . l c t r c t t s w a r t d uj'.,i t . , New York, 1971. '3.
Stagg, G.W. and A.H. ElAbiad. Computer Methods in Power SystemsAnalysis, McCrawHill Book Co., Ncw York, 1968.
4. Anderson, P.M., Analysis of Faulted Power Systems,Iowa State Press,Ames, Iowa, 1973. ( ' f r r t ' k c .l  , . . ( ' i n ' t t i t A r r t t l . t , , toi.,ft' A l t t r n u l i t t , qC t r r r c r t l l ' t t w r r S \ , . r ' / r , r l .V.o l . I , ' r New York. 1943. 6 . S t c v c r r s o n , . D . J r . , E l e t r r c n t .o f P o w c r S y s t e m sA n u l y  s i s , 4 t hc d n , M c ( i r z W s Ncw York, l9tl2.
r.
P ap e r
"7. Brown. H.E. et al. "Digital Calculation o[ ThreePhase Short Circuits by Matrix Methods", AIEE Trani., 1960, 79 : 1277.
compo1gnlg symmetrica!
Fortunately, in sucha transformation currents)called symmetricalcomponents. the impedancespresentedby various power system elements (synchronous generators, transformers, lines) to symmetricalcomponentsare decoupledfrom each other resulting in independentsystem networks for each component anced set). This is the basic reason for the simolicitv of the svmmetrical component method of analysis. TO.2 SYMMETRICAL COMPONENT TRANSFORMATION
! .a?.,{fii
Thus above.
Vo Vot *
bYbL
Voz I
vbz
Voo
(10.s) ( r0.7)
Vr= VrI *
Vrz *
Vro
by A set of three balancedvoltages (phasors)Vo, V6, V" is charactertzed equal and interphasedifferencesof 120'. The set is said to have a phase magnitudes if sequenceabc (positivesequence) Vulags Voby l2O" and V. lags Vuby I20". in be The three phasorscan^then expressed terms of the referencephasor Vo as Vo = Vo, V6 = a"Va, V, = aVo where the complex number operator cr is defined as sL  air20" It has the following properties
,*2:ei24o':eilAo" (o')* : o a3:l l+ala2:0 _*
The three phasor sequences(positive, negative and zero) are called the of components the original phasorset Vo, V6,V,. The addition of symmetrical symmetricalcomponentsas per Eqs. (10.5) to (10.7) to generate Vo, Vr, V, is indicatedby the phasordiagramof Fig. 10.1.
V61=crV61
V6fo?V11
(10.1)
is sequence),then If the phase sequence acb (.negative Vr= & V o V o = V o , V u = tu Vo , Thus a set of balancedphasorsis fu'lly characterizedby its referencephasor (positive or negative). (say V,) and its phasesequence A Suffix 1 is commonly usedto indicatepositive sequence. set of (balanced) positive sequencephasorsis written as
Vo1, V61 &Vu1, Vrr = aVot
(r0.2)
components obtain to of addition the symmetrical Fig. 10.1 Graphical in the set of phasorsV", V6,V" (unbalanced general) phasors Voy Eqs. (10.5) to (10.7) in termsof reference Let us now express Vo2and Voe.Thus (10.8) Vo= Vot* Vozl Voo
Vu= a.2Vor+ aVor* V"= cVol+ o2vor* Voo Voo
Similarly, suffix 2 is used to indicate negativesequence.A set of (balanced) negative sequencephasorsis written as
Vo2, V62= dVn2, Vrz= Q'Voz
(10.3)
A set of three voltages(phasors)equal in magnitudeand having the samephase Thus a set of zero sequencephasorsis written is said to have zero sequence.
AS
Vng, V6g =
Vo1, Vr1 = Vo1
(10.4)
Consider now a set of three voltages (phasors)Vo, V6, V, which in generalmay be unbalanced.According to Fortesque'stheorem* the three phasorscan be
* The theorem is a general one and applies to the case of n phasors [6],
(10.e) (10.10)
These equations can be expressed in the matrix form
ffi#d __r_:
ModernPo*et sytttt Analysi,
Io = il'
(10.19) (10.20)
(10.11)
and I' = ArI' where
Vp = AV,
(10.12)
v, =
l'u 
1,"1 of originalPhasors = vector
Ir=lIu I,=lIo, l;ana I LI,) Lr,o_i
Of course A and Ar are the same as given earlier. In expanded form the relations (10.19) and (10.20) can be expressed as follows: (i) Constructionof current phasorsfrom their symmetricalcomponents:
Io= Iot * Ioz * Ia= o2lott Ioo loo dozr
l ''l
[r,' I
Lv, )
% =  Voz  = vector of symmetrical components
Iu'l
Lu"'.1
Ia
[' A 1 "
V, = 4'V,
[t
o2 tl
I rl o nt I
Ir= dot+ azlor* Ioo (ii) Obtaining symmetrical components of current phasors:
(10.21) (r0.22) (r0.23) (ro.24)
(10.13)
Iot= * t,r.=
1
e"+ du+ o2lr)
We can wrire Eq. (10.12) as
(10.14)
d o']
ComputinE ,{r
and utilizing relations (10.1), we get
; (Io+ azlu+ aI,) i ; I r o = , Q o +I u + I r )
(ro.2s)
.
(r0.26)
o  ' = * l r c r 2a I 'Lt
(r0.1s)
r rl
Certain observationscan now be made regarding a threephase system with neutral return as shown in Fig. 10.2.
In expandedform we can write Eq. (10.14) as
Vor=
I
,<V"+
cVu+ a2VS
(10.16) (10.17) (10.18)
Vao
V""
voz= ! fr"+ ozvu+o%) 3
Vuo=
I
;
V,+ Vr,* r,)
Equations (10.16) to (10.18) give the necessary relationshipsfor obtaining symmetricalcomponents the original phasors,while Eqs. (10.5) to (10.7) of give the relationships for obtaining original phasors frorn the symmetrical components. The symmetrical component transformationsthough given above in terms of voltageshold for any set of phasorsand thereforeautomatically apply for a set of currents.Thus
ln
Fig. 10.2 Threephase systemwith neutral return The sum of the three line voltages will always be zero. Therefore, the zero sequencecomponent of line voltages is always zero, i.e.
= vobo
trr**
= vu,+ v"o) o
(10.27)
On the other hand, the sum of phasevoltages(line to neutral)'may not be zero so that their zero sequence component Vn, may exist. Since the sum of the three line currentsequals the current in the neutral wire. we have
_\
B _____}_
Ioo=!U"+16+r")=!+
(10.28)
i.e. the currentin the neutralis three times the zerc sequence line current.If the neutral connectionis severed.
Flg. 10. 3 Solution or Io + I" * Is = Q 10130" + l5l 60o+ Ic = O I c=  16. 2+ j8. 0 = 18 1154" A
Ioo=lr,=o 3
(ro.2e)
t.e. in the absence of a neutral connection the zero sequence line current is always zero.
Power Invariance We shall now show that the symmetrical componenttransformationis power invariant, which means that the sum of powers of the three symmetrical componentsequalsthe threephase power. Total complex power in a threephasecircuit is given by
to From Eqs.(10.24) (10.26) Itr= :0Ol3O" + 751(60"+ l2O" ) + I8l(154" + 240"))
5 1
= 10.35 j9.3 = 14142"A + Iez= ;Q0130' + 151( 60o+ 240") + I8l(154' + 120"))
J
(i)
I
S = f p $ = 4 1 , + V u t t +V , t !
or s = [Av,,]'t'lAl,l.
(10.30)
= 1.7 j4.3 = 4.651248" A
Iao= I t^ (lo + IR + J
(ii) (iii)
/c) = 0
= v! 'qr'q. tI
Now
(10.31)
FromEq. (10.2) Im= 141282"A Inz= 4'6518"A Iao=oA
Check: Ia= Iat * I,rz* I,to= 8'65 + j5 = 10130" Convertingdelta load into equivalentstar,we can redraw Fig. 10.3 as in Fig. 10.4.
la
..
r rl l1 0 0l [r o? 'Tt o16.=lta a n' .rl:ro r olt, (10.32) "tll I JLa, c  l l L 0 0 l l [r r s=31yti,=341.,
= 3V^1, + 3V"r!), + 3V"oI), = sumof symmetrical powers component (10.33)
Icr = 141162" A Icz = 4.651128"A Ico=oA
I Example10.1  T
A delta connectedbalancedresistive load is connectedacross an unbalanced threephasesupply as shown in Fig. 10.3. With currents in lines A and B specified, find the symmetrical componentsof line currents. Also find the symmetrical componentsof delta currents. Do you notice any relationship betweerisymmetricalcomponentsof line and delta currents ? Comment.
Modernpower system anatysis $,$i.M I Delta currents obtained follows are as
Qrrmmatriaal
lamnnnanla
tffiffi
veB= U^ Ir) Io
ItB=zAB)R
Similarly,
tUo
ru)
rnc=er Ir) I
5
Positive and negativesequencevoltagesand currents undergo a phaseshift in passing through a stardelta transformer which depends upon the labelling of terminals.Before considering this phaseshift, we need to discussthe standard polarity marking of a singlephasetransformer as shown in Fig. 10.5. The transformer ends marked with a dot have the same polarity. Therefore, voltage Vun, is in phase with voltage V..,. Assuming that the small amount of magnetizingcurrent can be neglected,the prirnary current 1r, enteringthe dotted end cancelsthe demagnetizing ampereturns the secondarycurrent 1, so that of I, and12with directions of flow as indicatedin the diagram are in phase.If the direction of 1, is reversed, 1, and 1, will be in phase opposition.
Ice=!frr5
Io)
Substituting the values of Io, Iu and Ir, we have
r5l 60")= 6186 A Ien= Oozzo" ! rnc= 60" rllr54") = lO.5l 41.5"A lOsz rol3o") 8 . 3 1 1 7 3A = rcn= " !W2154"
The symmetricalcomponentsof delta currents are Ie m= ;(6 1 8 6 " + IO.5 l (  4I.5' +
J
Flg. 10.5 Polaritymarking a singlephase of transformer\ Consider now a star/delta transformer with terminal labelling as indicated in Fig. 10.6 (a). Windings shown parallelto eachother are magneticallycoupled. Assume that the transformer is excited with positive sequencevoltages and carries positive sequencecurrents. With the polarity marks shown, we can immediatelydraw the phasor diagramof Fig. 10.7.The following interrelationship betweenthe voltages on the two sidesof the transformer is immediately observedfrom the phasor diagram VtBt= x Vabrl3V, r  phasetransformation ratio (10.34) As per Eq. (10.34), the positive sequence line voltages on star side lead the corresponding voltageson the delta side by 30" (The same result wo,'ld apply to linetoneutral voltages on the two sides). The same also applies for line currents. If the delta side is connectedas in Fig. 10.6(b) the phaseshrft reverses(the readershould draw the phasor diagram);the delta side quantitieslead the star side quantities by 30".
1
l 2O" ) + 8.31(173"+ 240" ))(i v)
= 8172" A
(v) I,qaz=:6186" + 1051( 41.5"+ 240")+ 8.31(173" 120")) +
J
I
= 2 . 7 1 2 1 8A "
Ie n o = 0 Incr, Ircz, IBC,,lsn1, Iga2 and 1.oo can be found by using Eq. (10.2). Comparing Eqs. (i) and (iv), and (ii) and (v), the following relationship between symmetrical componentsof line and delta currents are immediately observed: (vi)
IeBr=+130" V J Ienz= \ zzo"
t
'
(vii) (viii)
! J
The reader should verify theseby calculatrng Io', andl*2from Eqs. (vii) and (viii) and comparing the results with Eqs. (iv) and (v).
ar System Analys'is
A ________
Componentg Symmetricat
o a
 3?9
I
l
/\
correspotxding positive sequence quantities on the LV side by 30'. The reverse is the case for negative sequence quantities wherein HV quantities lag the corresponding LV quantities by 30". vsc2
(a) Star side quantitieslead delta side quantitiesby 30o
(b) Delta side quantities lead star side quantities by 30o
Fig. 10.6 Labelling star/delta of transformer
vcea
Vcrcz
Vesz
voltages a star/delta on Fig. 10.8 Negative sequence transformer
IO.4
SEOUENCE IMPEDANCES
OF TRANSMISSION
LINES
Figure 10.9 shows the circuit of a fully transposedline carrying unUalancecl currents. The return path for 1,,is sufficiently away for the mutual effect to be ignored. Let X" = sell'reactance o1'each line X. = mutual reactanceof any line pair The fbllowing KVL equationscan be written down from Fig. 10.9. Vo  V'o= jXJo + jX*Iu + .ixmlc
Vec'l
Fig. 10.7 Positive sequence voltageson a star/delta transformer Instead, if the transformerof Fig. 10.6(a) is now excited by negative sequencevoltages and currents,the voltage phasor diagram will be as in Fig. 10.8.The phaseshift in comparison the positive sequence to casenow reverses, i.e., the star side quantitieslag the delta side quantities by 30'. The result for Fig. 10.6(b)also correspondingly reverses. It shall from now onwards be assumed that a star/delta transformer is so labelled that the po,sitive sequencequantities on the HV side lead their
V6
vc
lr= l"+ lo+1"
(
Fig.10. 9
Modern PgwgfSy$gln Snalysis
Svmmetrienl cnmnonent(
tflIlFffi
T
4  r[: jxJo + jxh + jx*[" l',  V!: iXJ" + jxmlb + jxJ"
or in rnatrix form
^a
(10.3s)
t, o{__]Z1 o 12 "f}22 o
ls Zs o)f_l.
vb
v"
V, r
vI vi

= J
x^x,x. x_x*x,
Ib I"
(10.36) (10.37) (10.38) (10.3e)
(a) Positive sequence network
(b) Nagativesequence network
@)Zero sequence network
I// / n
zIo
AIZAI,
F i g .1 0 . 1 0 The decouplingbetween sequence networks of a fully transposed transmission holds also in 3phase synchronousmachines and 3phasetransformers. This fact leads to considerable simplications in the use of symmetrical rnethod unsymmetrical in components fault analysis. In case of three static unbalancedimpedances,coupling appearsbetween sequence networks and the method is no more helpful than a straight forward unalysis. 3phaso
or or Now
A (I/, v,

v ! ) : zuIs
r// ,'s
jx,
AI ZA :
JX^
jx^
jx^ jx, jx^
(10.40)
l*, ', :J L:
Thus Eq. (10.37) can be written as
0 X,X* 0
10.5 SEQUENCE IMPEDANCES AND SEOUENCE NETWORK OF POWER SYSTEM
Power system slernsnfstransmission lines, transformers and. synchronous nlachineshave a threephasesymmetrybecause which when'currentsof a of particular sequenceare passed through these elements, voltage drops of the same sequenceappear, i.e. the elementspossessonly self impedancosto sequencecurrcnts. Each eleurent can therelbre be representedby tlree decoupledsequencenetworks (on singlephase basis) pertaining to positive, negative and zerosequences, respectively. EMFs are involvedonly in a positive sequence network of synchronousmachines. For finding a particular sequence impedance, elementin questionis subjected currentsand voltagesof that the to only. With the elementoperating sequence undertheseconditions, sequence the impedance can be determined analytically or through experimentaltest results. With the knowledge of sequence networks of elements,complete positivel' negativeand zero sequencenetworks of any power systerncan be assembled. As will be explained in the next chapter. these networks are suitably interconnectedto simulate different unsymmetrical faults. The sequence currentsand voltagesduring the fault are then calculatedfrom which actual currentsand voltases can be found. far"rlt 10.6 sEQuENCg I;IpTDANCES AND NETWoRKs SYNCHRONOUS MACHINE oF
l " rlo Jl l'rltr:t 'JL o I o Lr
z 0 0
wherein Zr:
 41 l'tr( I
I x" x^
t 0 2 2 0 z
llr, I x,  x,,, o ll r,  o x, *zx^ )j,l
0 l[ /' 0 o
0
o
(l o.4l)
.l
lLr, i
ll,,I
(r0.42)
j(X,  X,r) : positive sequence impedance
(10.43) (r0.44) (10.4s)
Zz: j(X,  X^) : negative sequence impedance Zo: i(X, + 2X)
/t;\ \r,, o^rrol vyuar ^^o.i+i"o pvrrlryu ^ ( tr ll r l
: zero ,tequence impedance
^:^^)^^^strqu{rrruE rrup€uallutis.
We conclude that a fully transposed transmission has:
^^^+i ut/ts4Lrvg
(ii) zero sequenceimpedancemuch larger than the positive (or negative) sequenceimpedance(it is approximately 2.5 times). It is further observedthat the sequencecircuit equations (10.42) are in decoupleclfbnn, i.c. thcrrearc no rnutual scqucncc inducternccs. F)quation (10.42) can be represented network form as in Fig. 10.10. in
rlachine (generator motor) or Figure 10.11depictsan unloadedsynchronous groundedthrougha reactor (impedanceZ).8o, E6andE, are the inducedemfs
3S2 ] I
_
powerSystem Modern Anatysis
Symmetrigel_qg4pe!e$s _
behindappropriltc thc conditiorts, voltagt: lirrnr kradccl the shortcircuit occurs constitutesthe positive (subtransient,transient or synchr:onous) reactance
uence voltage. Figure lO.IZa shows the threephasepositive sequencenetwork ntodel of a
of the three phases.when a fault (not shown in the figure) takes place at machine terminals, currents I,,and /. flow in the lines.whenever the 1,,, fault involvesground,currentIn= In+ Iu + ^I"flows to ne'tral from ground "i^I^'.
Hrvvvvs
know the equivalentcircuits presentecl the *u.hin" to by the flow of positive, negative and zero sequence culrents, respectively. Because of winding symmetry currents of a particular sequence produce voltage drops of that sequence only. Therefore,thereis a no coupling betweenthe equivalent circuits of varioussequences*.
la
yyrrrr r(rLrll 4rr.1IyJIs (\_napler
i l),
we
must
machine. Z, doesnot appearin the model as Iu = 0 for positive synchronous by currents. Since it is a balancednetwork it can be represented the sequence network model of Fig. 10.12b for purposesof analysis. The singlephase bus for a positive sequencenetwork is at neutral potential. Further, reference no since current flows from ground to neutral,the neutral is at ground potential.
lat > a
* )e"
.n
Referencebus
'( t + \=\'..
t:.6 dn>_ u \ t6 > b
____l_" Fig' 10'11 Threephase synchronous generator withgrounded neutral Positive Sequence Impedance and Network Sincea synchronous machineis clesigned with symmetricalwindings, it induces emfsof positivesequence only, i.e. no negatrveor zerosequence voltagesare inc lt rc e n i t' Wh c n th c tttl tc l ti nc u rr i cs id posi ti vcscqucncc gl curr.cnts l y, thi s ntodeof operationis the balancedmoclediscussedailength in Chapter 9. The armature reaction field caused by positive sequence currents rotates at 'synchronous speedin the salneclirection the ,otu., i.e., it is stationary as with respect to field excitation. The machine equivalently offers a direct axis reactance whose value reducesfrom subtransicntreactance(X,a) to transient reactance (Xtr) and finally to steadystate (synchronous)reactanJe(Xa), as the short circuit transient progressesin time. If armature resistance is assumed negligible,the positive sequence impedanceof the machine is 21= jXtj (if I cycle transientis of interest) (10.46) = jX'a Gf 34 cycle transientis of interest)
L
(
I
ln't , lc't m (a)Threephase odel
b
I I
(b) Singlephase odel m
machine of network synchronous sequence Fig. 10.12 Positive voltageof terminal c With referenceto Fig. 10.12b,the positive sequence to the referencebus is given by with respect
V,,l= E,, Zll,,l
(10.4e)
and Network
Negative Sequence Impedance
(10.47)
sequence has zeronegative rnachine It hasalreadybeensaidthat a synchronous currentsin the stator a voltages. With the flow of negative sequence induced rotating field is createdwhich rotates in the opposite direction to that of the speedwith respect field and, therefore,at double synchronous positivesequence to rotor. Currents at double the stator frequency are thereforeinduced in rotor field and damper winding. In sweeping over the rotor surface, the negative of with reluctances direct and quadrature mmf is alternatelypresented sequence axes. The negative sequence impedance presented by the machine with considerationgiven to the damper windings, is often defined as
= jXa (if steady state value is of interest) (10.48) If the machine short circuit takes place from unloaded conditions, the terminaivoltageconstitutes positivesequence the voltage;on the other hand.if
*'fhis can be shown to be so by synchronousmachine theory,[5].
(10.s0) ;lZ2l<lZrl 2 machine,on a threenetwork models of a synchronous Negative sequence The andb, respectively. basisare shownin Figs. 10.13a phase and singlephase referencebus is of course at neutral potential which is the same as ground potential.
Z . t =j
xt: + x,!
fl$4.:rl
t
powLr Mooern System nnarysis

symmetricat Components
Zs= 3Zra Zos (10.53) in orderfor it to havethe samevoltagefrom a to reference bus.The reference bushereis, of course, groundpotential. at
From Fig. 10.14b zero sequencevoltage of point c with respect to the
Voo =  Z{oo
From Fig' 10.13bthe negative sequence voltage of terminal a with respecr to referencebus is
Voz=  Zzloz
FjSiffi
(10.s1)
(10.54).
Order of Values of Sequence Impedances of a Synchronous Generator
laz a
Typical valuesof sequence impedances a turbogenerator of rated 5 MVA, 6.6 kV, 3;000 rpm are: Zr = lZ%o(subtransient) Zr = 20Vo(transient) Zr = 7l0Vo (synchronous) Zz= I2Vo Zo= 5Vo For typical values of positive, negative and zero sequencereactances of a synchronousmachine refer to Tablu 9.1. IO.7 SEOUENCE IMPEDANCES OF TRANSMISSIO\ LINES
(b) Singlephase model
Flg. 10.13 Negative sequence network a synchronous of machine Zero Sequence Impedance and Network
we state once again that no zero sequence voltages are induced in a synchronous machine. The flow of zero sequencecurrents creates three mmfs which are in time phase but are distributed in space phase by 120". The tesultant air gap field caused by zero sequence currents is therefore zero. Hence, the rotor windings present leakage reactanceonly to the flow of zero sequence currents (Zos < Zz < Z).
/66
Reference bus
a Iao
(a)Threephase model
(b)Singlephase model
1o.14 Zerosequence network a synchronous of machine Zero sequence network models on a threeand singlephase basis are shown in Figs. 10.14aand b. In Fig. r0.l4a, the cu'ent flowing in the inpedance zn betweenneutral and ground is In = 3lno.The zero sequence voltage of t"rinul a with respect to ground, the referencebus. is therefore Vr1=  3znioo Zorlno=,r, (32, + Z0)loo e0.52) wh11eZo, is the zero sequence impedanceper phase of the machine. Sincethe singlephase zero sequence networkof Fig. 10.14bcarriesonly per phasezero sequence current, its total zero sequence impedancemust be
A fully transposedthreephase line is completely symmetrical and therefore the per phase impedanceoffered by it is independentof the phase sequenceof a balanced of currents. other words,the impedances set In offired by ii to positive and negative sequence culrents are identical. The expressionfoi its per phase inductive reactanceaccounting for both self and mutual linkages ias been derived in Chapter 2. When only zero sequence currents flow in a transmissionline, the currents in each phase are identical in both magnitude and phase angle. part of these currents return via the ground, while the rest return throu h the overhead ground wires. The ground wires being grounded at severaltowers, the return currentsin the ground wires are not necessarily uniform along the entire length. The flow of zero sequence currentsthroughthe transrnission iln"r, ground wires and ground createsa magnetic field pattern which is very different from that causedby the flow of positive or negativesequence currentswhere the currents
h,r.ro rrqYv ^ (r
Prrd'Ds' urrrsrtrIlutr
^L^^^
J:ff^^^^
UI
^f
ILV
t ^no
an(l
tne
fetufn
Cuffent
lS
sequence impedance of a transmission line also accounts for the ground impedance(zo = z.to+ 3zri.since the ground impedance heavily dependson soil conditions, it is esseniialto make some simplifying assumptions obtain io analytical results. The zero sequenceimpedanceof transmission lines usuallv
ZefO.
T\e
ZefO
3Sd I
ModernPower SystemAnatysis
Symmetri
a
ranges fromZ to 3.5 timesthe positivesequence impedance*. This ratio is on the higher side for double circuit lines without ground wires.
ter" "*"
"tt "
transformers
10.8 SEOUENCEIMPEDANCES AND NETWORKS OF
It is well known that ahnosl all present day installationshave threephase transformerssince they entail lower initial cost, have smaller spacerequirements and higher efficiency. The positive sequence seriesimpedanceof a transformerequalsits leakage impedance.Since a transformeris a static device, the leakageimpedancedoes not changewith alterationof phasesequence balancedapplied voltages.The of transformernegativesequence impedanceis also thereforeequal to its leakage reactance.Thus. for a transformer
Zt= Zz= Zr"ukug"
networks of various types of transformer Beforeconsideringthe zero'sequence are three importantobservations made: connections, side. current only if there is cunent florv on the secondary currentscan flow in the legs of a star connectiononly if (ii) Zero sequence the star point is groundeclwhich proviciesthe necessaryreturn path for culrents.This fact is illustratedby Figs. 10.15aand b. zero sequence
/ao= 0

':o
__,
a
{> .
r'l
( r0.55)
Assuming suchtransformerconnectionsthat zero sequence cunents can flow on both sides,a transformeroffers a zerosequence impedancewhich may differ slightly from the corresponding positive and negativesequence values.It is, however, normal practiceto assumethat the seriesimpedances all sequences of are equal regardlessof the type of transformer. The zero sequence magnetizingcurrent is somewhathigher in a core type than in a shell type transformer. This difference does not matter as the magnetizing curent of a transformer is always neglectedin short circuit analysis. Above a certain rating (1,000 kvA) the reactanceand impedance of a transformerare almostequal and are thereforenot distinguished.
/co= 0 f' /oo=o
star (a) Ungrounded
(a) Grounded star
in currents a starconnection Fig. 10.15 Flowof zerosequence (iii) No zero sequencecurrents can flow in the lines connecte.dto a delta culrents.Zero sequence connectionas no returnpath is availablefor these currents can, howevet, flow in the legs of a deltasuch currents are in voltages the delta connection. of causedby the presence zerosequence This f act is illt r st r at ed Fig. 10. 16' bY
/ao= 0 a
*We can easily compare the forward path positive and zero sequenceimpedances of a transmission line with groundreturnpath infinitely away.Assumethat eachline has a self inductance,L and mutual inductance M between any two lines (completely symmetrical case).The voltage drop in line a causedby positive sequencecurrents is VAnt= uLlnr+ uMI,,r+ aMI,.1 [uL+ (& + a) aMllo, = a(Ltu)Ior
ii', currents a deltaconnection Fig. 10.16 Flowof zeroSequence Let us now consider various types of transformerconnections. Case I; YY transformerbank with any one neutrttl grounded. zero sequence is If any one of the two neutralsof a YY transformer ungrounded, thesecannotflow star and consequently, currentscannot flow in the ungrounded network in the grounded star.Hence,an open circuit existsin the zero sequence i.e. betweenthe two parts of the system connectedby the between H and L, as transformer shown in Fig. 10.17.
Positive sequencereactance= a(L fuI) The voltage drop in line a causedby zero sequencecurrents is VAno= aLloo+ utMluo+ uMlrs = a(L + 2M)Ino Zero sequence reactance= w(L + 2W Obviously,zerosequence reactance much more than positive sequence is reactance. This result has alreadybeenderivedin Eq. (10.45).
ffi
(seeFig. 10.19). If the star neutral is groundedthrough 2,, an impedance3Zn appearsin series with Z, in the sequencenetworkCasb
Zg
YA tran
rmer bank with ungroundedstar
tll
d
L
YYtransformer bankwithone neutral grounded its zero and sequence network Case 2: YY transforrner bank both neutrals grounded When both the neutralsof a YY transformerare grounded, a path through the transformer exists for zero sequencecurrents in both windings via the two groundedneutrals.Hence,in the zero sequence network 11andL areconnected by the zero sequence irnpedanceof the transformeras shown in Fig. 10.1g. Case 3: YA transformer bank with grounded y neutral
This is the special case of Case 3 where the neutral is grounded through Zn = oo. Thereiore no zero sequence curent can flow in the transformer network then modifies to that shown in Fig. 10.20. windings. The zero sequence
{ L
bus Reference
"ifr6L H 2 ;
o
L
Fig. 10.20
star bankwith ungrounded and its Yl transformer network zerosequence
Case5: AA transformer bank
Reference bus
o4TdL=o H Z s
Since a delta circuit provides no return path, the zero sequencecurrents cannot flow in or out of AA transformer; however, it can circulate in the delta windings*. Theretore, there is an open circuit between H anE L and Zo is connectedto the referencebus on both endsto accountfor any circulating zero sequencecurrent in the two deltas (see Frg. 10.2I).
L
Fig. 10.18 YYtransformer bankwith neutrals groundedand its zero sequence network
dE;1
^r
a _. I
Reference bus
network bankand its zerosequence Fig. 10.21 AAtranstormer
Fig. 10.19 Y1 transformer yneutral and its zero bankwith grounded sequence network If the neutral of star side is grounded,zero sequence currentscan flow in star becausea path is availableto ground and the balancing zero sequence currents can flow in delta.Of courseno zero sequence currentscan flow in the line on the delta side. The zero sequence network must therefbre have a path from the line 1{ on the star side through the zero sequence impedance of the transformer
10.9 CONSTRUCTION OF SEOUENCENETWORKS OF A POWER SYSTEM
In the previous sections the sequencenetworks for various power system elementssynchronous machines,transformersand lineshave been given. Using thesq, complete sequencenetworks of a power system can be easily constructed.To start with, the positive sequencenetwork is constructed by
*Such circulating currents would exist only if zero sequencevoltages are somehow induced in either delta winding.
Modern Power SystemAnalysis examination the onelinediagrarnof the system.It is to be noted that positive of sequence (gcnerators motors) voltages present synchronous are in machines and only. The transitionli'orrr positil'e sequence network to negative sequence network is straightforward, Since the positive and negative sequence irnped,) \rrrrvD qrtu Lt otrJltrt ttlvt Jrr, rrrv LrltlJ vrrarr
ffi
MVA base in all other circuits and the following voltagebases. line voltase base= 11 x Transmission 121 123.2kY 10.8
neclessary positive sequence in network to obtain negatit'e sequence network in respect of synchronous machines. Each machine is representedby neg;ttive sequence impedance, negativesequence the voltage being zero. The referencebus for positive and negativesequence networks is the system neutral.Any impedance connectedbetweena neutraland ground is not included in these sequence networks as neither of these sequencecurrents can flow in suchan impedance. Zero sequencesubnetworks for various parts of a system can be easily combinedto form completezero sequence network. No voltage sourcesare presentin the zero sequence network. Any impedanceincluded in generatoror transformerneutralbecomes three tirnesits valuein a zero sequence network. Special care needs to be taken of transforners in respect of zero sequence network. Zero sequence networks of all possibletransformer connectionshave
been dealt with in the preceding section. The procedure for drawing sequence networks is illustrated through the following examples. f":'
= x Motor voltagebase 123.2 ++ = 11 kV 121 to line of The reactances transformers, andmotorsareconverted pu values as bases follows: on appropriate
= 0.1 pu reactance x =. f++) = 0.0s05 Transformer 3 0 \ 1 1/
loo x 25 : Line reacl a n c e =
ffi
=o.164pu
2s /10\2 =  0.345pu of Reactance motor I = 0.25 x  x t 15 \lll
Reactance motor 2  0.25 x ?1, rf Lo)t = 0.69 pu of 7.5 \ ll / in network is presented Fig. 10.23. The requiredpositive sequence
Referencebus
r Example 10.2 1
t _ _ _ _ _ _ : _  l
A 25 MVA, 1l kV, threephase generator has a subtransient reactance 20Vo. of T lr cgcrrc ri tto ru p p l i c s o rrto to rs v 0r' r,r s tw o transnri ssi on w i Lhtral tsl ol l rrcl ' s l i nc at both ends as shownin the oneline diagramof Fig. I0.22. The motorshave ratedinptrtsof 15 and 7.5 MVA, both 10 kV with 25Vosubtransient reactance. Thethreephase transformers both rated30 MVA, I0.8/L2I kV, connection are AY with leakagereactance 70Voeach.The seriesreactanceof the line is of 100 ohms. Draw the positive and negative sequencenetworks of the system with reactances markedin per unit.
e
Es( ) +l
I
I Enrt( ) +l I t! .r: i
,1
) Eara l +
ioz:),
d ,66XLg ,.I6XU j0'164 i oo8o5
.
i o 3 4 5 : r
iio6e
9
,
r "xfi i 00805
P r'
In
network Example for 10.3 sequence Fig. 10.23 Positive
Reference bus
Ir t s ' t o
p ' u910' (t, 'r' )
.
q \  / r \ '
t
Fig. 10.22
(2) ,
IYA
Motor ,,1
I
Assume that the negative sequence reactance of each machine is equalto its subtransientreactance.Omit resistances. Select generator rating as basein the generatorcircuit.
i 0.0805
j4164
loo8o5
Fig. 10.24 Negative sequencenetworkfor Example 10.3 Since all the negative sequence reactances of the system are equal to the positive sequence reactances,the negative sequencenetwork is identical to the
' rti..;;:1 . ::i.l
3,42'l
l4odern Pov"'ei' SvstemAnalvsis
symmetricat compblents _1 Zerosequence reactance motor  0.06 of 2 " +. = 0.164 pu
= Reactance currentlimiting reactors of
ffi
positive sequence network but for the omission of voltage sources.The negative network is drawn in Fig. 10.24. sequence
(i+)'
''.1.i3t = 0.516pu (11)',
For the power system whose oneline diagram is shown in Fig. 10.25,sketch the zero sequence network.
T1 T2
Reactance current of limiting reactor included zero sequence in network' = 3 x 0.516  1.548 u p Zerosequence reactance transmission = ry:21 of line (r23.D2 = 0.494pu The zero sequence networkis shownin Fig. 10.27.
Fig.10.25 Solution The zero sequence network is drawn in Fig. 10.26.
j1548
Reference bus 32n
Reference bus
j1.il8 jo164
io06
Zost
Fig. 1O.27 Zero sequencenetworkof Example 10.5
Zrt Fig. 10.26

Z6(line) 26(line)
272
Zero sequence network of the system presented in Fig. 10.25
IEIvIS PROB
i   i    _ *  ;
1Example10.4 *"
1
Draw the zero sequencenetrvork for the system describedin Example 10.2. Assumezero sequence reactances the generator for and motors of 0.06 per unit. Current limiting reactorsof 2.5 ohms eachare connected the neutral of the in generatorand motor No. 2. The zero sequence reactanceof the transmission line is 300 ohms. Solution The zero sequencereactance of the transfonner is equal to its positive sequence reactance. Hence : Transformer zero sequence reactance 0.0805 pu Generator zero seqllence reactances: 0.06 pu Zero sequence reactance motor 1 : 0.06 x of
10.1 Computethe following in polar form (i) o2t 1ii; I a (iii) 3 d + 4cy+ 2 (iv) ja "? 10.2 Thlee identical resistorsare star connectedand rated 2,500 V, 750 kVA. This thre.ephase unit of resistors is connected to the I side of a AY transformer.The following are the voltages at the resistor load = lVo6l 2,000 Y; lVu,l = 2,900 V; lV,ol = 2,500 V Choosebaseas 2,500 V, 750 kVA and determinethe line voltagesand currents in per unit on the delta side of the transforrner.It may be assumed that the load neutral is not connectedto the neutral of the transformer secondary. 10.3 Determine the symmelrical components three voltages of Vo= 2001ff, Vt = 2001245" and V, = 2001105' y
: 0.082
Power SystemAnalysis acrossiines bc of resistiveload of 100 kVA is connected 10.4 A singlephase of components the the symmetrical supply of 3 kV. Compute a balanced line currents. 10.5 A delta connectedresistive load is connectedacross a balanced threepnasesupply
'
""*' 2: 25 Generator " MVA, 11 kV, Xtt = ZOVo
transformer(each):20 MVA, ll Y1220Y kV, X = I5Vo Threephase The negative sequencereactanceof each synchronousmachine is equal of machine is 8Vo.Assume that the zero sequencereactances lines are reactances. of 25OVo their positive sequence
250 f) B
ABC Fig. P10.5 Phasesequence of 400 V as shown in Fig. P10.5. Find the symmetrical componentsof line cunents and delta currents. of 10.6 Three resistances 10, 15 and 2O ohms are connectedin star acrossa threephasesupply of 200 V per phase as shown in Fig. P10.6. The supply neutral is earthedwhile the load neutral is isolated. Find the currentsin each load branch and the voltage of load neutral above earth. Use the method of symmetrical components'
la
X = 5o/o at machine,l rating
at machine 2 rating
Fig. P10.8
r0.9For the power system of Fig. P10.9draw the positive, negative and zero
and transformersare rated as follows: networks. The generators sequence l: G ener at or 25 M VA, 11 kV, Xt t =0. 2, X2 = 0. 15,Xo = 0. 03 pu 2: G ener at or 15 M VA, 11 kV, Xt =0. 2, Xz= 0. 15, X0 = 0. 05 pu Motor 3: 25 MVA, 11 kV, Xt = 02, Xz= 02, Xo = 0.1 pu Synchronous Transformerl: 25 MVA, I 1 Lll20 Y kV, X = IIVo 2: 12.5 MVA, 11 LlI20 Y kV, X = l07o 3: 10 MVA, I2O Ylll Y kV, X  IjVo Choosea base of 50 MVA, I I kV in the circuit of generatorl.
T1
I r
'0.:"t
I
tB
te
ls Fi g. P 10.6
c1
1
_L__
_l
tb
10.7 The voltagesat the terminalsof a balancedload consistingof three 20 V. 100 1255.5'and 200 llsf arc resistors 2OO4O", ohm Xconnected Find the line currents from the symmetrical components of the line What relationexistsbetween voltages the neutralof the load is isolated. if the syrnrnetricalcomponentsof the line and phase voltages'/ Find the power expanded in three 20 ohm resistors from the symmetrical componentsof currentsand voitages. impedancenetworks for 10.8. Draw the positive, negative and zero sequence the power systemof Fig. P10.8. lines, and Choosea baseof 50 MVA, 220 kV in the 50 0 transmission and transformers in mark all reactances pu. The ratings of the generators are:
\7
Fig. P 10. 9 Note: Zero sequencereactanceof each line is 250Voof its positive reactance. sequence
j,,W
voo"inpo*",,syrt"r Rn"tu"i,
10.10 Considerthe circuit shownin Fig. P10.10.Suppose Vo, = L00 l0 Xr=12 Q = 60 160 vbn Xob=Xbr=X*=5d)
F i g . P 10.10
(a) Calculate Io, 16,and 1, without using symmetrical component. (b) Calculate Io, 16,and 1" using symmetrical component.
II.I
INTRODUCTION
NCES REFERE
Books l. wagncr, c.F. and R.D. Evans,symmetrit:al componenfs, McGrawHill,Ncw york , 1933. 2. Clarke,E., Circuit Analysisof Alternating Current Power Systems, Vol. 1. Wiley, Ncw York, 1943. 3. Austin Stigant,5., Master Equationsand Tablesfor Symmetrical ComponentFault Studies, Macdonald,London, 1964. 4. stcvenson, w.D., Elcments Power Sy.stem oJ' Analysis,4thedn,McGrawHill, New York, 1982. 5. Nagrath, I.J. and D.P. Kothai, Electric Machines,2nd edn., Tata McGrawHill, New Delhi, 1997. Paper 6. Fortescue,C.L., "Method of SymmetricalCoordinatesAonlies to the Solution of Polyphase Networks', AIEE, 1918,37: 1O27.
Chapter 9 was devoted to the treatment of symmetrical (threephase)faults in during suchfaults,analysis a power system. Sincethe systemremainsbalanced basis. In this chapter, we shall could conveniently proceed on a singlephase deal with unsymmetrical faults. Various types of unsymmetrical faults that occur in power systemsare: Shunt Wpe Faults
(i) Single linetoground (LG) fault (ii) Linetoline (LL) fault (iii) Double linetoground (LLG) fault Series Type Faults (i) Open conductor (one or two conductorsopen) fault. (3L) fault being the most severe It was statedin Chapter 9, that a threephase the rupturing capacity of circuit breakers,even though must be usedto calculate this type of fault has a low frequency of occurrence, when compared to the unsymmetricalfaults listed above. There are, however, situationswhen an LG fault can cause greater fault current than a threephasefault (this may be so when the fault location is close to large generating units). Apart fiom tliS, unsymmetrical fault analysis is important for relay setting, singlephase switching and system stability studies(Chapter 12). The probability of two or more simultaneousfaults (crosscountryfaults) on a power system is remote and is therefore ignored in system design for abnormalconditions.
*g9,S.I
Modernpower System Anatysis
Fault Analysis Unsymmetrical
The method of symmetricarcclmponents presentedin chapter 10, is a powerful tool for study of unsymmetricalfaults and witl be fully exploited in this chapter.
Vat (a)
UNSYMMETRICAL
FAULTS
Consider a general power network shown in Fig. 11.1. It is assumedthat a shundtype fault occursat point F in the system, a result of which currenm is Io, 16, flow out of the system, and vo, v6, v"are voltagesof lines a, b, /, c with respectto ground.
(b)
(b)
Vao
(c)
(c)
_l
lao
Fig. 11.1 A general powernetwork Let us also assume that the system is operating at no load before the occurrence of a fault. Therefore, the positive sequence voltages of all synchronous machineswill bc identicaland will equalthe prefaultvoltageat F. Let this voltage be labelled as Eo. As seenfrom F, the power systemwill presentpositive, negative and zero sequence networks, which are schematicallyrepresented Figs. Il.2a, b and by c. The referencebus is indicatedby a thick line and the point F is identified on each sequence network. Sequence voltagesat F and sequence currentsflowing out of the networks at F are also shown on the sequencenetworks. Figures 11.3a,b, and c respectively,give the Thevenin equivalents the three sequence of networks. Recognizingthat voltageEo is presentonly in the positive sequence network and that thereis no coupling betweensequence networks,the sequence voltages at F can be expressedin terms of sequencecurrents and Thevenin sequence impedances as
networks seen as Fig. 11.2 Sequence fromthe fault point F
Fig. 11.3 Thevenin equivalents the of sequence netWorks seen as fromthe faultpoint F
cunents and voltages are Dependingupon the type of fault, the sequence constrained, leading to a particular connection of sequencenetworks. The sequencecurents and voltages and fault currents and voltages can then be easily computed.We shall now consider the varioustypesof faults enumerated earlier. 11.3 STNGIE LINETOGROUND (tG) FAULT
fault at F in a power systemthrougha fault Figure 11.4showsa linetoground impedanceZI. The phasesare so labelled that the fault occurs on phase a.
1r",1lt"1 lt, 0 olr",'l l r " , l = l o lo l 2 2 o l l , * l Lv"o)Lol Io o zo)lr"o)
(11.1)
(LG) fault at F Fig. 11.4 Single linetoground
Unsvmmetrical Anatvsis Fault
At the fault point F, the currentsout of the power systemand the line to ground voltagesare constrainedas follows:
fffiH
r 'at =
(zr * zz * z) +3zr
E
o
.
(11.7)
I u =o
Ir= 0 vo = zllo The symmetrical components of the fault currents are
(rr.2)
11 . 3 ) (11.4)
Fault current /o is then given by
(zr * zz * zo)+32
The above resultscan also be obtained directly from Eqs. (11.5) and (11.6) by using Vop Vo2and Voefrom Eq. (11.1). Thus (Eo IolZ) + ( IorZr) + ( I"&d = 3Zf lot I(4+ 4+ Zo)+ 3zfllor= Eo
Eo
(11.5)
Ior =
Expressine ,r':l; ,';',.*1;,ij;.*.ar Eq
Vot * Voz *
(zr * zz t Z) +3zr
componenrs, wehave
(11.6)
Voo = ZfIo = 3zflor
The voltage of line b to ground under fault condition is Vu= &Vor+ aVo2* Voo
As per Eqs. (11.5) and (11.6) all sequence currentsare equal and the sum of sequence voltagesequals3zf lot Theiefore, theseequationssuggest a series connection of sequencenetworks through an impedanie 3zf ut rrio"*n in Figs. 11.5 aa n c ib .
Substituting for /o from Eq. (11.8) and reorganizing,we ger
Va= Eo
=o,2^+)* (n' 4',+). +) ("
3rl2 + Z2(&  a)+ zo(&  D zr (Zr*22+Zo)+3Zt
(11.e)
The expression for V, can be similarly obtained.
Fault Occurring
Under Loaded Conditions
la,t= = t"o= h laz t
l"s=I,,1=t, t
/a\
(b)
Fig' 11'5 Connection sequence of network a singlelinetoground fault for (LG) In terms of the Thevenin equivalent of sequence networks, we can write from Fig. 11.5b.
When a fault occursunderbalancedload conditions,positive sequence currents alone flow in power system before the occurrence of the fault. Therefore, negativeand zero sequence networksare the sanleas without loacl.The positive sequencenetwork must of course carry the load current.To account for load current, the synchronous machines in the positive sequence network are replacedby subtransient, transientor synchronous (dependingupon reactances the time after the occulTenceof fault, when currents are to be determined) and voltages behind appropriate reactances.This change does not disturb the flow of prefault positive sequence currents (see Chapter9). This positive sequence network would then be usedin the sequence network conhectionof Fig. 11.5a for computing sequence currentsunder fault In casethe positive sequence network is replacedby its Thevenin equivalent as in Fig. 11.5b,the Theveninvoltage equalsthe prefaultvoltage Vi atthe fault point F (under loaded conditions).The Thevenin impedanceis th6 impedance betweenF and the reference bus of the passivepositive sequence network (with voltage generatorsshort circuited).
i'4ot',,1
ModernPo@is
fni, i, illustratecl a two machinesystemin Fig. 11.6.It is seenfiom this by figure that while the prefault currents flow in the actual positive sequence network <lfFig. 11.6a, samcdo not exist in its Theveninequivalent the network of Fig' 11.6b.Therefore,when the Theveninequivalentof positive sequence network is used fot calculatlng fault eurre+ts, the positive sequeneeeurrents within the network are those due to fault alone and we must superimpose on these the prefault currents. Of course, the positive sequencecurrent into the fault is directly the correct answer,the prefault current into the fault being zero.
from whlch we get Io2=  Iol 1.,0= 0 The symmetrical components of voltages at F under fault are
(11.11) (11.12)
I v,,f , [t a o'fl'' I l r " , l = i l a 2 " l l r , z r r a ) t1 I
LV"o) Lr
(b) (c)
(11.13)
lL",
we Writing the first two equations, have 3vot = vo + ( a+ &) vu  &zf t u
Fig. 11.6 Positive sequence network and its Thevenin equivalent before occurrence a fault of The aboveremarksare valid for the positive sequenienetwork, independent of the type of fault. t7.4 LrNETOLrNE (LL) FAULT
3voz= v, + (tr + ,l) vo  ull Iu
from which we get
3(va voz) (a  &1zrtu iJi y' to
Now 16= (& a) Iot (: I o z =  I o t , 1 o o= 0 ) .r
( 11 . 1 4 )
=  jJt I"r
16 Substituting from Eq. (11.15)in Eq. (11.14),we get Vot  Voz= Zf I ot
( 11.15)
(11.16)
Figure I 1.7 showsa linetolinefault at F in a power systemon phases and b c through a fault impedanceZf . The phases can always be relabelled,such that the fault is on phasesb md c.
b
(11.11) and (11.16)suggest parallelconnection positive and Equations of negative V sequencenetworksthrougha seriesintpedance as shown in Figs. 11.8a and b. Since loo= 0 as per Eq.(11.12),the zero sequence network is unconnected.
l6
Vaz
Fig. 11.7 Linetoline(Lt) fault through impedance Z/
F lat
_1
laz
The currents and voltages at the fault can be expressedas
[1":o I Io=
I
l I u Ir t ' lI, : _i
l ,V o 
Vr= IoZf
(11.10)
Fig. 11.8
(LL) fault Connectionof sequencenetworksfor a linetoline
The symmetrical components the fault currents are of
In terms of the Thvenin equivalents, we get from Fig. 11.8b
( 11 . 1 7 )
t
Fault Analysis Unsymmetrical _1 or
Voo= Vot * 3Zf Ino
I SQt*
F ro m Eq . (1 1 .1 5 ), e g e t w
I u ' !
be found.
Z t * 22 + zZ l " + 4 ,++2 22 +f Knowing l,trwecarlcalsqld{e and, v,,, vorfromwhich vsltages thefault, at
L

 jJi u n, J '  o
(l 1.23)
( 11 . 1 8 ) \.
From Eqs. (11.19),(ll.22a) and (11.23),we can draw the connection of networksas shown in Figs. 11.10aand b. The readermay verify this sequence by writing mesh and nodal equationsfor these figures.
If the fault occursfrom loadedconditions,the positive sequence network can be modified on the lines of the later portion of Sec. 1r.3. 11.5 DOUBLE LrNETOcRouND (Lrc) FAULT
Figure 11.9 showsa doubre linetoground fault at F in a power system.The fault may in generalhave an impedanceZf as shown. ab
tr a
Y/u=9 ,..\.
Z1 Ea
_.__
r"l t'o  ___ z! t'"o _,__r:]tr"
(b)

I :1
Vaz
laz
II
+
Vss I
Fig. 11.9 Double linetoground. (LLG) fault through impedanceZl The current and voltage (to ground) conditions at the fault are expressed as
Fig. 11.10 Connectionof sequence networksfor a double linetoground (LLG) fault ln terms of the Thevenin equivalents, we can write from Fig. 11.10b I ,. = l
ot
Io=o i", + Io, + I"o =o
l
( 11 . 1 9 ) (r 1.20)
V,,= V, = Zf (lt, + Ir) = 37rf 1,,r, The symmetricalcomponentsof voltagesare given by
@
Eo
l ! : ' l =* ,l_l 1"v,,,)l I
fiom which it follows rhat
a l!:'1 [r a 2 "']lr.l ' o
rll,ur)
En
zt + z2(zo *3zI ) I (22+ zo+ 3zt )
(11.24)
l l vI u
(rr.21)
The above result can be obtainedanalyticallyas follows: f or Vut , Vuz and V, * in t er m sof E, , in Eq. ( 11. 1) and S ubst ir ut ing impedance premultiplyingboth sidesby Zt (inverseof sequence matrix), we get
v,,t= V.z= Llv,, + (a + r11V,,1 3"'
(Il .22a)
v,,o= + 2vu) 1""
From Eqs. (11.22a) and (t I.Z2b)
(rr.22b\
lt;' I o z;'
I o o
voo vot=
tr,
 ,r &1 vu= vt = 3zfloo
'=17' ;, ilt?l l';"1
z;' ll E,, 2,r,,, I,,nl +3zf Io o z o t ) L o )L r , o _ J
ol[r,,ztrut o lln,2,r",
I I
(rr.2s)
406t'l
Modern powe1_syglgll 4!g!yg!s
both sides by row marrix tl 1 1l and using Eqs. (11.19) and 1.gltiitiUying (1I.20), we ger
be more than the threephase fault current.
From Eq. (11.22a), have we Eo Ztlot= Zzloz Substituting loz=  (Ior + Io) fseeEq. (11.19)]
O<
\
?
lx,
i
l
.
E o  Z rl o t= Z z (Io r+ Io s) or E ^o 'I u =  o  ( ,l zt+z'\'
Negative
Fig. 11.12 Sequence networks synchronous of generator grounded through neutral impedance
," 1"'
Substituting this value of Iooin Eq. (rr.26) and simplifying, we finally get
f "l =
If the fault takesplacefrom loadedconditions,the positive sequence network will be modified as discussed Sec. 11.3. in
r   l i e+tflP.f ;;!utrl,,r
Figure 11.11 showsa synchronous generator whoseneutral is groundedthrough a reactancaXn.The generatorhas balancedemfs and sequence reactances X1, X., and Xu such that X, = Xz > Xo.
(d) Write expressionfor neutral grounding reactance,such that the LG fault current is less than the threephase fault current. Solution (a) Figure 11.12 gives the sequencenetworks of the generator.As stated earlier voltage source is included in the positive sequence network only. (b) Connectionof sequence networks for a solid LG fault (ZI = 0) is shown in Fig. 11.13, from which we can write the fault current as ll)rc 
3 lE , l
zxt+&,*3X,,
3lE"l
(i)
(c) If the neutral is solidly grounded llolLG= =
2\+xc
3lE"l
(ii)
For a solid threephase fault (see Fig. 11 . 1 4 )
_ l I ) r r = ;E:o:l t
xn
l'FJf[\:  : """ /
I
Ea
(iii)
lqs= 131. 1 Fig. 11.13 LG fautt
.\"/ ) +))
Comparing (ii) and (iii), it is easy to seethat llol Lc> llol3L
b c
Fig' i i.i i
Syne hronous generator grounded through neutralreactance
(a) Draw the sequence networks of the generatoras seen from the terminals. (b) Derive expression for fault current for a solid linetogroundfault on phase a.
An important observationis made here that, when the generator neutral is solidly grounded,LG fault is more severethan a 3I_ fault. It is so because, Xo * Xr = X, in generator.However, for a line Xo D Xt = Xz, so that for a fault on a line sufficiently away frorn generator, fault will be ntore severe 3L than an LG fault.
E"()
ft
n
q
I
l
\
l
l
I
Xt'4
7 _ _ li L
Fig. 11.14 Threephase fauft
 a . 1 .  . 4 s ^ & g . . .  ' r } . } . * ' * . 3 a    } l !  }  } . } ' Q t p t 
408
usqelrl@
(a) Ir=
If J  ' A A A r : A A 4 E ,
Unsymmetrical Fault Analysis
l'i
(d) with generatorneu,tralgrounded through reactance,comparing Eqs. (i) and (iii), we have for LG faurt current to be less than 3L fault 3 lE , l
3x1
: . n r
tltrffi
f
J
j 0.0e j0.07 j0.1+ 0.9e 0.99 j0.26s s+ + +
24 + Xo+3X,
= 2.827  j0.756 (b) Cunent in the grounding resisror= If = 2.g27 _ j0.756 llrl = 2.926 x :L" J3 xll (c) Voltageacrossgrounding resisror= = 3.07 kA *1 2 r e.g2i  j0.756)
2X, + Xo + 3Xn> 3Xl
*^, I(xr  xo)
J
(iv)
 0.932 j0.249
= 0.965
= 6 . 1 3k V
i;;;ili",i:rl
Two 11 kV, 20 MVA, threephase, connectedgenerators star operatein parallel a s s h trw ni n F i g ' l l ' 1 5 , th c p o s i ti ve, negati vc andzero sequence reactances of eachbeing,respectively, j0.1g, j0.r5,7o.10 pu. The star point of one of the generators isolatedand that of the otheris earthed is through a 2.0 ohm resistor. A single linetogroundfault occurs at the terminals of one of the generators. Estimate (i) the faurt current, (ii) current in grounding resistor, and (iii) the voltage acrossgrounding resistor.
For the systemof Example 10.3 the onelinediagramis redrawnin Fig. I 1.16. On a baseof 25 MVA and I 1 kV in generator circuit, the positive,nega"tiue and zerosequence networks of the systemhave been drawn alreadyin Figs. 10.23, 10.24and 10.27Before the occurrenceof a solid LG at bus g, the motors are loadedto draw 15 and 7.5 MW at 10 kV, 0.8 leading power factor. Ifiprefault currentis neglected,calculate the fault current and subtransientcurrent in all partsof the system.
What voltage behind subtransient reactancesmust be used in a positive sequencenetwork il prefault current is to be accounted fbr'/ Xo = 0.06pu Tz e
F i g .1 1 . 1 5 Solution (Note: All values are given in per unit.) Since the two identicar generatorsoperatein paralrer, Xr.o = j0.18
2.5A
(H+d )( 
Xo = 300 fl
= i0.0g,Xr"q=t'#
= j0.075
Fig. 11. 16 one r ine diagr am t hesyst em Exam pr 11. 3 of of e solution The sequence networks given in Figs. 10.23, 10.24 and r0.2i are connected Fig. 11.17to simulatea solid LG fault at bus g (seeFig. in 11.16).
[f nreforrlt n urrfrfv rn f D vr a rlo ara qrv nanlo^+^.i rrvSlvvL(/\I
Sincethe star point cf the second generator isolated, zero sequence is its reactance doesnot comeinto picture.Therefore, Zo"q=7O.10 n=j0.10+3x +3R
For an LG fault, using Eq. (11.1g), we ger
4:+
(ll)',
= 0 . 9 9 +ri0 . 1
= E'l= E',!,t= E',1,2 vj (prefault voltage at g)
Iy (fault current for LG fault)  Io = 3lo, =
3E
X t" o+ X z" ql Zo" ,
=
= pu 1+ o.eoe
410 I
I
Analysis Modern PowerSystem
.
Unsymmetrical Anatysir Fautt
.l''. E"o( )
I
rb6T___
'>
Z z = Z r = j 0 . 1 6p u Fiom thescquence network connection ' o rf _ J
T
mbl
l02:l (
dt
f
Vf
737;4
j2.032
io'szs
0.99+i0.607
= q'901 =  jo.Mlpu
Ioz= Ioo= Ior =  j0.447 pu
l io.t+z
+ i0.136 i0.1 36 *_  j0.311 .447
Fault current= 3loo= 3 x ( j0.447) =  jL341 pu The componentof Io, flowing towards g from the generator side is
j0.447 !: ?:= =  70.136 x pu j0.7ss
and its component flowing towards g from the motors side is
i 0.608
11.712
 jo.Ml* i,?s?s= j0.311 =pu j0.7ss
Similarly, the component of Io2from the generatorside is  j0.136 pu and its component from the motors side is 70.311. All of Iosflowstowardsg from tnotor2. Fault currentsfrom the generatortowards g are
I p.oo,
Fig. 11.17 Connectionof the sequence networksof Example 11.3. are currents shownon the diagramin pu for a solid Subtransient fault linetoground at g The positive sequencenetwork can now be easily replaced by its Thevenin a e q u i v a l e n ts s h o w ni n F i g . 1 1 . 1 8 .
Reference bus
I rl [ro.r30lli0.2721 I r,1l [ rl l . l l l l I r u l = l a z G r l l  i o . r r = l 7 0 . 1l 3 6 l6 pu Lr.J lo u2 rJL o j L jo.r36j
and to g fiom motors are
T
r
Now
L/,J la
I [t..] [t I t o" l  la 2l a l ^
e"
l
llfro.ltt1 [.rl,06el r l l  r o . s u: l  i o . r 3 6 l p u l
l " l l ' l '
I)Lj0.447J Lj0.136_l
The positive and negative sequence componentsof the transmission line cunents are shifted 90" and +90o respectively, from the corresponding componentson the generatorside of Tr, i.e. Positivesequence current =  j(jO.136)   0.136 pu Negative sequence current  j( j0.I36) = 0.136 pu Zero sequence current = 0 ('.' there are no zero secluence currents on the transmission line, see Frg. ll.17) Line a current on the transmission line =Q.136+0.136+0=0 Iu and I, can be similarly calculated.
4l:iN
Modern power System Anatvsis
t tal u
Unsymmetricat Anglysis Fault lll4#hF
F a
Let us now calculatethe voltages behind subtransient reactancesto be used if the load currents are accountedfor. The per unit motor currents are:
x*q* Xr",
i0.09+ i0.075
= _ j6.06
25x0.909x0.8 25x0.909x0.8
136.86 = 0.66+ j0.495pu = 0.4125136.86'= 0.33+ j0.248pu
UsingEq. (11.15),we have 1y(fault current) = Io = i Now
Vot = Voz = Eo Iorxleq = 1'0  ( j6.06)
3Xj6.06):10.496
Motor 2: E!"2 =l.l:t, l';:::iir:tfi il r"
Generator,E'{ = 0.909 + j0.525 x I.2375136.86"
Motor 1: tr'^' =llltr:';:: ="ri;Yilii.il r"
Total cuffent drawn by both motors = 0.99 + j0.743 pu The voltages behind subtransient reactancesare calculatedbelow:
u0.0e)
("' /oo = 0)
= 0.455
Voo= I^/'og Voltage of the healthy phase, Vo= Vot * V n z * V o o= 0 . 9 1
= 0.52 + j0.52 = 0.735145. pu It may be noted that with thesevoltages behind subtransient reactailces, the Thevcnin equivalent circuit will still be the same as that of Fig. 11.19. Therefore, in calculating fault currents taking into account prefault loading condition, we need not calculate EIy E/ft and E(. Using the Thevenin eqtrivalentapproach, can first calculate currentscauied by fault to which the we load currents can then be added. Thus, the actual value of positive sequencecurrent from the generator
.^.^l^
3orExample 11.2,assume that the groundedgenerator solidly grounded.Find is .hefault current in each phase and voltage of the healthy phase for a double inetogroundfault on terminals of the generator.Assume solid fault (Zf  01. )olution Using Eq. (1I.24) and substitutingthe values of Zp* Zr"rand Z*, rom E xam ple1l. 2, we get ( not e Zf =0, Z0"q=j0. 1)
I _ ral 
t '+r 7 0 ' 
ruw:lrus ule Iault ls
rl
f^lr
!
io.oe. Yj:+ J:: {: l: j0.075 j0.i0
10. 075 i 0. 10 ^ x
  ; 7 < ?
0.99 + j0.743 j0.136 = 0.99 + j0.607 andtheactualvalueof positive sequence current from the motorsto thefault
l s '
Vot= Vo2= Vo1= Eo
Ior Zr.q= 1  ( j7.53) (/O.09)
= 0.323
I _ ra2
0.99 j0.743 j0.311=  0.99 jr.054 In this problem, becauseof large zero sequence reactance,load current is comparable with (in fact, more than) the fault current. In a large practical system,however, the reversewill be the case, so that it is normal practice to neglect load current without causing an appreciableerror.
_Voz 
Zr.,
Voo Zo.,
j0.07s
0.323 j0. 10
=
0.323  j4.306
.^ A^ / 1
tta 0 
t 1
Iu= rllot
+ alo, + Ioo
For Example I1.2, assume that the grounded generatoris solidly grounded. Find
f}ro ^"1+ rrrv frcrlrrL n r r * a * # ttursrrl
auLr vurLil$tr
^J
,,^lr^^
ul
^f
eL^
ttl9
l^^trr^
lr€artlly
r
pllase
  
r
IOf
a lfnetollne
r.
laUlt
r
a.
On
terminals of the generators. Assume solid fault (Zf  0). Solution For the LL fault, using Eq. (11.I7) and substituting the values of X,"u and Xr"u from Example 11.2,we get
= (0.5 /0.866) (j7.53)+ (O.5+ _i0.866) + U4.306) j3.23 =  10.248 j4.842  11.3341154.74' +
Ir= el,r, + ozlor* Ino
= (0.s + 70.866) (j7.s3) + (0.5 j0.866) (j4.306)+ j3.23
4!4" 1
I
ModernPower SystemAnalysis  10.248+ j4.842  71.334125.28"
Fault Analysis Unsymmetrical 
Positivesequence network
,415
Voltage of the healthy phase, V o = 3 Va t = 3 x 0 .3 2 3= 0.969 11.6 OPEN CONDUCTOR FAULTS
An open conductor fault is in series with the line. Line currents and series voltagesbetween broken ends of the conductorsare required to be determined,
Negativesequence network
lc
cic'
Zerosequence network
fault in and Fig. 11.19 Currents voltages openconductor Figure 11.19 shows currentsand voltages in an open conductor fault. The ends of the system on the sides of the fault are identified as F, F', while the conductorends are identifiedas ua /, bb / and cc'. The set of seriescurrentsand voltagesat the fault are ll I f v .l '  l lt ' ,, I : v : l l '. ". " l l l v oo'l , 'r p l,1''P I
Fig. 11.20 Sequence networks for open conductor laull at FF/ [n terms of symtnctrical cotnponents, we can wrlte
Vonl * Voo,2* Vno,g = O
I o t = I o 2 = I n o + 1 "
L/I ,
L V ,), ,
of The symmetrical components currents and voltagesare
l " ' l  L/.,,J lVou'o ) networks can be drawn for the power system as seen from FF/ The sequence and are schematically shown in Fig. ll.2O. These are to be suitably connected dependingon the type of fault (one or two conductorsopen). Two Conductors Open
F IF'
t. =  I^. l: v":l v^," I " " " ' l
[ /.'I
1v,,,,,,1
Figure lI.2l representsthe fault at FF /with conductors b and c open. The currents and voltages due to this fault are expressedas
Voo'= 0 1 6 = I r  Q
'tD
I I I
(rr.27)
(11.28)
c " c '
I
o Fig 11.21 Two conductors pen
Fig.11.22 Connection sequence of networks two conductors for open
tt'6 il
t
Modern Power System Anatysis
prefault voltage Vf_. of bus r in series with the passive positive sequence and zero sequence short circuited).Sincenegative network (all voltagesources prefault voltages are zero, both these are passivenetworks only.
Equations(11.29) and (11.30) suggesta seriesconnectionof sequence networksas shown in Fig. II.22. Sequence currelrtsand voltagescan now be computed.
Onc Conluctor Open
For one conductoropen as in Fig. 11.23,the circuit conditionsrequire Vbb,=Vrr,=O Io = O
1 Voor2= Voory= *Voo,
J
Reference bus
(11.31) (11.32)
for passive positive sequence network
In terms of symmetricalcomponents theseconditionscan be expressed as
Vool=
(11.33) (11.34)
Iot* In + Ioo=0
Equations(11.33) and (1I.34) suggesta parallel connection of sequence networksas shown in Fie. 1I.24.
t
I
[l
V""'z I
for networks LG fault of Fig. 11.25 Connection sequence network sequence on the r th bus (positive equivalent) by represented its Thevenin
l
L'i ' ryl
F
la
It may be noted that subscripta has been droppedin sequencecurrents and voltages, while integer subscript is introduced for bus identification. Superscripts o and /respectively, indicate prefault and postfault values. network For the passivepositivesequence
Vr"us= ZrnusJr"ut
where
(11.35)
a
= Vtuus
t ,c+c', c/ i
positivesequence voltagevector (1 1.36) bus
lao
Fig. 11.23 One conductoropen
Fig. 11.24 Connectionof sequence networks for one conductor open
Ztrl
Zrnus
Ztnn )
: l positive sequencebus impedance matrix
/1 1 2?\ t
\ L r . J )
II.7
BUS IMPEDANCE MATRIX METHOD UNSYMMETRICAL SHUNT FAULTS
FOR ANALYSIS
OF
and
Bus impedance method of fault analysis, given for symmetrical faults in Chapter9, canbe easilyextended the caseof unsymmetrical to faults.Consider fbr example an LG fault on the rth bus of a nbus system. The connectionof sequence networks to simulatethe fault is shown in Fig. I1.25. The positive sequencenetwork has been replacedhere by its Thevenin equivalent,i,e.
[/'' I
e bus Jrsus= l t t . '  = p o s i t i vs e q u e n c e cunent injectionvector l : l
r l
(l1.38)
t4l8 
I
todern power SvstemAnalvsis network connectionof Fig. 11.25,we can now write From the Sequence
t2r
As per the sequence network connection,current  IJr_, is injected only at the faulted rth bus of the positive sequence network, we have therefore
rf
vro, 2r,, * zz,, I Zor,+3zf
(rr.47)
( 11 . 3 e )
other types of faults can be simila*Seomputed using Zlrr, Zrr, and Zon in placeof Zr , Zr and Zoin Eqs. ( 1I . 7) , ( 11. 17) and ue
(11.24) with E,  Vi,. by at voltages anybuscannow be computed superposing sequence Postfault
on prefault bus voltage, the voltage developed owing to the injection of appropriatesequencecurrent at bus r'. at network,the voltagedeveloped bus i owing Foi passivepositive^sequence to the injection of  IIr, at bus r is (11.48) Vtr= Zr,Jfr,
s ub s ti tu ti n g q .(1 1 .3 9 ) i n E q . (11.35),w e can w ri re the posi ti vesequence E voltage at the rth bus of the passivepositive sequencenetwork as V,,' =  Zrrrlfr, ( 11.40) Thus the passivepositive seguence network presentsan impedanceZr_ r, to the positivesequence currentI{_r. For the negativesequence network = Zz_sus Vzuus (11.41) Jz_nus The negativesequence network is injected with current lfr_, at the rth bus only. Therefore,
voltageat bus I is givenby postfaultpositivesequence Hence V l  , = V i  ,  Z r  , , f r  , ;i = l ' 7 ' " ' ' t t where
voltageat bus i prefaultpositive sequence
( l r.49)
0 0 = Jzsus
of Zr,, = irth component Zt"ut bus voltages are zero, the postfault Since the prefault negativesequence bus voltagesare given by sequence negative Vf''=0+ Vzr   zr,rl'fr, where
l r  , , = i r t l t c o l l l p o l l c l t to l ' Z t  t , , t Similarly, the postfault zero sequence bus voltages are given by
,i{ ,
0
zr rrlf, I
(1r.42)
(l r . s 0 )
The negative sequence voltage at the rth bus is then given by Yr,=
(l 1.43)
where
Thus, the negative sequence network offers an impedance Zr_rrto the negative sequence currentltr_, Sirnilarly,fbr the zero scqucncenetwork Vuuu,= ZosusJ,,",r, (11 44) .
Vd' =  Zu''lfu''; j = l' 2' "'' tI
(i l . s1 )
= irth component tlf Z9 sLr.
0 0
Josus=
 r^{  r u
0
( l l.4s)
Vo_, =  Zs_,.rlfs_,., (r 1.46) That is, the zero s.equencenetwork off'ers zrn intpeclance Zrr_,.,. the zero to sequence cuffent l'*r. and
currentsin voltagesknown at the buses,sequence With postfault sequence as: linescan be comPr'rted yl adrnittarlces ur, Jz',u,and yor, For line uv, having sequence ( V f t  u v [  r ) f ,rr= ltu, (rr.s2) I f r  r r = ! 2  , , r ( V t  , , V 5  r , ) , I i, r  ur = Jo u,( Vio ,  Vf o r l andcurrentscanbe voltages phase and currents, voltages Knowing sequence transformation easilycomputed by the useof the symmetricalcomponent Vr, = AV" Ir, = AI,
'i[di;,l
Modern powSr Jy$ern_4nglysis
Fault Unsymmetricalnnarvsis l_+ Yraa= _+^ =_ jtj.422 io.z io.o'os Yra"= Yrreto;;t : jr24zz
It appears first, as if this methodis more laboriousthan computingfault at currents from Thevenin impedancesof the sequencenetworks, as it requires computation of bus impedancematrices of all the three sequence networks.It must, however,be pointed out herethat once the bus impedance matriceq have been assembled, fault analysiscan be conveniently carried out for all the buses, which, in fact, is the aim of a fault study. Moreover, bus impedancematrices can be easilymodified to accountfor changes power networkconfiguration. in
liffiffi
Yt_,,= Yrn= #*
Yr"f *h:
j0.085 *
.#
 i78.s1e
i6.0s1
I I =i16.769 + j0.345 j0.69
For Example 10.3, positive, negative and zero sequencenetworks have been drawn in Figs. 70.23, 10.24 and 10:27. Using the bus impedancemethod of fault analysis,find fault currents for a solid LG fault at (i) bus e and (ii) bus I Also find bus voltages and line currents in case (i). Assume the prefault currents to be zero and the prefault voltages to be 1 pu. Solution Figure 11.26 shows the connection of the sequencenetworks of Figs. 10.23, 10:24 and 10.27 for a solid LG fault at bus e.
Y.  t88
d t7.422
v v r IBUS _ I2_BUS_
e f I 0 0 12.422  18.519 0 6.097 t2.422 12.422 19 6.097  18.5 0 16.769 12.422 0 0
I
E"* Positive sequence
network
a
10.345
v r odd 
I = _ j0.62I j1.608
1 j0.0805 ' =i14.446 j0.494
@
@
'vo  e e  'Y t, t t    t 1
Yrrr=
Negative sequence network iO'345
jth=io'584
i 0.6e
o Yoar= 'o
.1 = j2.024 Yo,r= jofu
@
@
i0.0805
Yofs= 0.0
Zero sequence network
{ @
@
JU.4v+ /U.UUUs
v I O _ B U S_
v/
orl 0 " r r0l
0
d 0.621
e 0
f
0 2.024 14.446 0
o 6
r4.M6 2.024 0
Fig. 11.26 Connection the sequence of networks Example 1.6 for an LG of 1 faultat bus e Refer to Fig. 11.26to find the elements the bus admittancematricesof the of three sequence nefworks.as follows:
0 0 0 0.584
lnverting the three matricesaboverendersthe fbllowing threebus impedance matrices
Analysis
  0.417pu VL" =  Zu,,Ifo, =  j0.0706x ( j2.362)   0.167pu 0 0.07061 0.00989 0.00989 0.07061 0
The fault currentwith LG fault on busc is
0
0 0 0 r.71233
Vfzr= 4r"llr1 =j0.08558 x( j2.362) =  0.202pl
Vfus=ZurJfu"=O Using Eq. (11.52), the currents in various parts of Fig. 1I.26 can be as cornputed follows: (i)
0
3x1 =  j7.086pu j 0J7636+ j 0.r7 + j 0.07 636 061 The fault curent with LG faulton bus /'is
il=
e
II+= Yrp UI+ vI+,)
  j6.097 (0.728 0.s84) = _ 70.88
If ,0"= Yr4" (Vf ,a Vf ,r)
Irr
__3x1 + .i0.t82gg iOlAZgg +/O"O?Gi
) j0.436s9
a
=  7 6 . 8 7 1p u (ii) Bus voltagesand line currentsin case(i) can easily be computcdusing Eqr. (11.49)(tI.sZ). Given below is a sample calculation for computing voltage at bus f and current in line ef FrorrrEq. (11.49) VIa= Vi_a Zr_0" Ifr_"
= =  j12.422 (0.703 0.584)  jL482
I,,t= If tf" * If t,t, =  70.88 + ( ,tl .482)
  i2.362 which is the same as obtainedearlier [seeEq. (i)l where If, = 3lut. IJrsf = YFcf (vf ',  vf ')  j12.422 (0.798  0.728) =  i0.88 Notice that as per Fig. 11.26,it was requiredto be the sameas llrs". Izf" = Yrr, (Vfrr  Vf,r) =  j6. 097 (  0. 212+ 0. 417)=  7O . 884 IL tr.= nt1a(Vtt1 Vfrr,) =  j2.024 ( 0.023+ 0.167) =  jO.29I pu Ittn (a) = IJrf" * It)r, + {* =  j0.88 + ( 70.88)+ ( j0.291)
= t.0  j0.t2s7s(' 7'0ttb = 0.703 pu \ " 3 )) Vftt= Vit  Z, tu If , .,
= r.o jo.ns4j(r, orru = o.zzs p, )
V J r  "= V " r  "  Z r _ " o  l J r_"
= 1.0 i0.17638 j2.363)= 0.584pu (vIr= vis  z,r"I{, = 1.0 j0.08558 j2.363)= 0.798 (pu vfzf=  zr.fJfr" =  ./0.11547 ( j2.362)=  0.272pu x  Zof"Ifo, j0.00989 ( j2.362) Vfor= =x = _ 0.023pu
  jz.os
Sirnilarly,other currentscan be computed.
A single line to ground fault (on phasea) occurs on the bus I of the system of Fig. 11.27.Find
I
qU;l
Modern Power System Anatysis
Fault Unsymmetrlcal AnalysisT
f,r'4l#
( i)
= zrws rLo.o+s
matnx ls
i t0.105 0.0451
Zz.sus o.1o5.J:
Tero sequencenetwork of the system is drawn in Fig. II.29 and its bus
ted below.
bus Reference
Ftg. 11.27
0.15
0.15
Current in the fault. (b) sc current on the transmissionline in all the three phases. : (c) SC current in phase a of the generator. (d) Voltage of the healrhy phasesof the bus 1. Given: Rating of each machine 1200 kvA, 600 v with x, = x, = rTvo, xo = 5vo. Each threephase transformeris rated rz0o kvA, 600 v  nlgroo VY with leakage reactance SVo, of The reactances the transmission of line are xr = Xz = 20vo and Xo = 40vo on a baseof 1200 kVA, 3300 V. The reactances of the neutral grounding reactors are 5Vo on the kVA and voltage base of the machine. Note: Use Z"u, method. Solution Figure 11.28 shows the passive positive sequencenetwork of the systemof Fig.l1.27. This also represents negativesequence the network for the system.Bus impedancematrices are computedbelow:
Reference bus
(a)
0.05
0.05
0.05
0.4 Fig. 11. 29
0.05
bus Bus 1 to ref'erence Zosus= i [0.05] Bus 2 to bus I
l0.05 = /Lo.os 0.051 Zonus 0.451
Bus 2 to referencebus
7 zOBUS 
l 'Lo.os _ l 0.45 0.0s _ lt o . o so . 4 s l + 0.4s1 10.451
(ii)
l0.0s 0.051
;
l0.051
or
0.0051 = .1[0.04s ZosusJZIO.OO5 0.0451
0.05
0.2 Flg. 11.28 0.05
As per Eq. (11.47) IIt = VO
, Z;tr * Zz_tr Zo_r+3Zl *
Bus I to referencebus Zt_svs= j[0.15] Bus 2 to Bus I
unloaded before fault) But V", = I Pu (sYstem Then j1.0
fr_,
Z r  8 t ,. =,lfO '.1S ,r r o t 5 Bus2 to reference bus
;;;l
0.1s1 r
o l5t
=  j3.92 pu ,0.105+0.105+0.045
Ifrt=ltr;=f
(a) Fault current, I\ = 3If ,, =
^2Pu
z L Brs. ="/Lo.rs r, ,[o'15
o35J**
""
[o'ls o'3s] Lo.rt.l
lO.tst
O) Vfvr = Vo r, = Ztrr lfrt = 1. 0 70. 105x  j3. 92 = 0. 588; Vot ; = 
Vf tr= Vorr ZrrJfrri Vora = 1.0 (system unloaded beforefault)  j0.M5 x  j3.92 = 0.g24 = 1.0
(c) 4c =
r
l
(1  o's88)t33" 1.37i2.38
,o;  ( 0.412))t3o" .:= j0.1s to r.37 i2.38
vtr;=  zrtrfrt
= 70.105  j3.92 = 0.412 x  Zrr, Ifr, V{z=   j0.045x  j3.92 = _ 0.176 = vfot  zurrlL,   j0.045x  j3.92 = _ 0.176
rLc=
vtz= Zuy IL,
=  7O.005  j3.92 =  0.02 x I{rz= yvrz (VI_r Vfr_r) =  1  ( 0 . 5 8 8 0 . 8 2 4 ) =j l . r 8 j0.2 Ifrrr= !zn U{t  Vfrr) = *.r0 . 4 1 2 0 . 1 7 6= i 1 . 1 8 + ) j0.2' If o_r,= lo_tz(Vfu,  Vfur)
IIoc= 0 (seeFig' 1I'29) If o_c= et37  j2.38)+ (1.37 i2.38) = _ j4.76 can of b Currentin phases andt:c the generator be similarly calculated. (d) Vfor= ZVfvr + Vf,, + Vfur = 0.588 1240" 0.4L2 1120" 0.176 =  0.264 j0.866= 0.905 l107" VIrt= Vf,t + Vfr, + Vfu' = 0.588ll20'  0.412 1240" 0.176   0.264+ i0.866= 0.905 1107"
= ;,
( 0'176 o.o2o) io.3s = +
LEIVIS PROB
has a x"o= 0'2 pu' Its negative and zero 11.1 A 25 MVA, 11 kv geaerator 0.3 and 0.1 pu. The neutral of the are reactances respectively sequence current in the generatoiis solidly grounded.Determine the subtransient conditions when generatorand the linetoline voltagesfor subtransient an LG f'ault occurs at the generatorterminals. Assume that before the occuffence of the fault, the generatoris operating at no load at rated voltage. Ignore resistances. 11.2 RepeatProblem 11.1 for (a) an LL fault; and (b) an LLG fault. generatoris rated 25 MVA, 11 kV. It is starconnected 11.3 A synchronous is with the neutral point solidly grounded.The generator operating at no are voltage. Its reactances Xt' = Xz = 0.20 and Xo = 0'08 load at rated
pu. rr^r^ir^+^ \aruulalLE +l^^ Llls .,m,o+einol DJrluuvlrrv4r o r rv r feoa c i c n t ou l u r r n line etrrre.nfs for (i) S in o   Sle
Iro _ tz =j l .1 8 + 7 1 .1 8 j 0.39_ j 2.75 +
7 . _  f D _,t. z  Ji lr . t a ru  ro ./1A r 4  a v I r o _ r J. 'r1 r o . /7^fr z_tLv . .n ^^ + Jv.Jy
= _ j07g If,rz= jl.18 lI20 + 71.1glZ4V + il.3g = j0.79
ll.4
linetoground fault; (ii) double line fault; (iii) double linetoground fault; and (iv) symmetrical threephasefault. Compare these currents and comment. to of For the generator ProblemI 1.3,calculatethe valueof reactance be in the generatorneutraland ground, so thatlinetogroundfault included
lW:il
Mqdernpower Syst€m Anatysis current equals the threephase fault current. What will be the value of the groundingresistance achicvcthc samcconclition,l to
with the reactance value (as calculated above) included between neutral and ground, calculate the double line fault current ancl rlso double linetoground faul
sub^ t*t. ignored. current prelault
7.5MVA kv 3.3/0,6
X = 10o/o
11'5 Two 25 MVA, 11 kv synchronousgenerators are connected to a common bus bar which suppliesa feeder. The star point of one of the generatorsis groundedthrough a resistanceof 1.0 ohm, while that of the other generatoris isolated.A linetoground fault occurs at the far end of the feeder. Determine:(a) the fault current; (b) the voltage to ground of the sound phasesof the feeder at the fault point; and bi r.l "orilg" the star point of the groundedgenerator with iespect to ground. The impedances sequence to currentsof eachgeneratorand feeder are given below: Generator (per unit) jo.2 i0.15 j 0.08
Feeder (ohms/phase) j0.4 j0.4 j0.8
fauft X'= Xz= 2oo/o
Xo = 5o/o Xn = 2'5o/o
Yr6i* F i g .P  l 1 . 8 n.g the A double linetoground fault occurs on lines b and c at point F in in phase c of current system of Fig. itf.q. Find the subffansient are machine 1, assuming prefault currents to be zero. Both machines of x//= xz=lUvo and xo= rated 7,200kvA,600 v with reactances trtnslorutcris rutcd 1.200kVA. 600 VA/3.300 5%. llac5 thrccphirsc of Vts with leakage reactanceof 5Vo.The reactances the transmission 4oToon a baseof 1,200kVA, 3,300V' line areX,=X,=20vo andXo= The reactancesof the neutral grounding reactors are 5Vo on the kVA base of the machines.
Positive sequence Negative sequence Zero sequence
rI'6
Determine the fault currentsin each phase following a double linetoground short circuit at the terminals of a starconnected synchronous generator operating initially on an open circuit voltage or i.o pu. The positive, negative and zero sequencereactance of the generator are, respectively,70.35,j0.25 andj0.20, and its star point is isolated from ground.
11'7 A threephase synchronous generatorhas positive, negative and zero sequence reactances phase per respectively, 1.0,0.g and 0.4 ohm. The of winding resistances negligible. The phase sequence are of the generator is RYB with a no load voltageof I I kV betweenlines. A short circuit occurs between lines I and B and earth at the generatorterminals. Calculate sequence currentsin phaseR and currentin the earth return circuit, (a) if the gencrator neutralis solidly earthecl; ancl (b) il the generator neutral is isolated. Use R phase voltage as reference. 11.8 A generator supplies a group of identical motors as shown in Fig. P11'8. The motors are rated 600 V, 9O%o efficiency at full load unity power factor with sum of their output ratings being i rrrrW.The motors
afe sharino enrrqllrr
_r_rr* "1_
F l g .P ' 1 1 . 9
,_L
i^
vn6l
:
lagging and 90vo efficiency whea an LG fault occurson the low voltage side of the transformer. Specify completely the sequence networks to simulatethe fault so as to include the effect of prefault current. The group of motors can be treatedas a singleequivalent motor.
o l v .l r ^ ^r s r^ os u ^ ^ '. r t ll \vt r r rv i^ltr l a.L € u v 'r' .' a g e , u . 6 p o w e r \€ / v tactor
1 machine1 generating pu voltageis connectedthrough 11.10 A synchronous 0.1 pu to two transmissionlines in transformerof reactance a Y/Y parallel. The other ends of the lines are connected through a YN 0.1 transformerof reactance pu to a machine2 generating 1 pu voltage' For both transformers X, = Xz = Xo' Calculatethe current fed into a double linetogroundfault on the line '2. 'side terminals of the transformer fed from machine The star point of The machine I and of the two transiormers.are soiiriiy grourrded. base are reactancesof the machines and lines referred to a common xo Xz xl 0.05 0.25 0. 35 Machine 1 0.04 0.20 0.30 2 Machine 0. 80 0.40 0.40 Line (each)
Modernpower
'
ll'll
I lncrrrnmatrinal
Farrlt
Anahraio
lL.'A+\l
Thepot+: q"geEye zero uno seque;r1:#;:"# fl:f{",,":t' viltr'om vv'rv\,rerr,"in pEr fiit *:"
Positive GeneratorI Generaror2 Each rransformer Infinite bus Line 0.15 0.25 0.15 0.15 0.20 Negative 0.15 0.25 0.15 0.15 0.20
i,:il".5.T1",ffi,1,iJl.:^::l:l;,i""i;ffi*"#,J'#::,irfr
""*"t;;;il#,
Zero 0.0g oo (i.e. neutral isolated) 0.15 0.05 0.40
:T"tl' "."11'::::1.r: :
i.H ,:i il"lili ": :i::,{:,il.:ni fr;;il;; ;ffi# Hffi flffi 1:' f:"9,,.* *'' ; ;".il:i ff un ri te u il; in s j.
#,f
nerwork twogenerators with connecred iltr"::,i"il"ti .1T:Toow.er
Xr = Xz = 0.1 pui Xo = 0.05 pu X, (Brounding reactance)= 0.02 pu = 0J?u Transformer: Xr: Xr:Xt Generator:
X, (grounding reactance)= 0.04 pu Form the positive, negative and zero sequence bus impedancematrices. For a solid LG fault at bus 1, calculate the fault current and its contributionsfrom the generatorand ffansformer.
1 _f6lY €fff
,
L
2
sequence networks of the power l.] ;;_ (b) P_....*,the with borh generarors and infinite bus op"rutinf r.o pu voltage on no road,a rinetogrouncr faurt occursat .ne of "a terminars the of the starconnected winding of the transformerA. caiculate the currents flowing (i) in the fauli; and (ii) through rhe transformer A. ,q 6L_'
\7  )r I I ]L  x lI_ J  l , Itz2,H / " . .Y \ '
rTftYA
F i g .P  1 1 . 1 3 Hint: Notice that the line reactancesare not given. Therefore it is convenientto obtain Zt, svs directly rather than by inverting Ir, sus. Also ro, it singular and zs, BUS cannot be obtained from it. In such "us situations the method of unit current injection outlined below can be used. For a twobus case
_t
F i g .p  1 1 . 1 1
rl'r2 A star connectedsynchronous generatorfeeds bus bar r of a power system. Bus bar I rs connected to bus bar 2 thro'gh a star/crerta lt'itnsl0t'ttlcl' scrics with a ilt transmissionline. The power network connectedto bus bar 2 can be, representedby a star_ connectedgeneratorwith equar "quiuutently' positive incr ncg.tivc sc(rr.r0rccs rcactances. star porntsare Alr solidry connected ground.The to per unit sequence re:lct*nces v'rious corrlponents of are given berow:
pol;itive 0.20 OJZ 0.30 Nc,,gutivt: 0.l5 0.12 0.30 Generator Transfbrmer Transmission Line PowerNerwork Zt:ro 0.05 0.12 0.50
li;,1=17,',7,',)l';,1
Injectingunit currentat bus 1 (i.e. Ir = tr, !2= 0), we get Zn= Vt Zzt = Vz S ir nilar ly injcct ingr r uitcur r cutut bus 2 1i. c./ r = 0, lz = l) , we get Ztz = Vl 7:tz= Vz ll.l4 Zou5could thus bc dircctly obtained this technique. by Considerthe 2bus systemof Example 11.3.Assumethat a solid LL fault occurson busf Determinethe fault currentand voltage(to ground) of the healtlry phase.
X X 0.10 Under no load condition with 1.0 go voltage at each bus bar, a current of 4'0 pu is fed to a threephase short circuit on bus bar Z.Deitrmhe the positive sequence reactance X of the equivarent generator of the ,, . power network. For the same initial conditions, find the faurt current for single line_ toground fault on bus bar l.
11.15 Write a computer programme to be employed for studying a solid LG fault on bus 2 of the sy'stemshown in Fig. 9.17. our aim is to find the fault current and all bus voltages and the line currents following the fault. use the impedance data given in Example 9.5. Assume all transformers to be YlA type with their neutrals (on HV side) solidly grounded.
il
reactances of the Assume that the positive and negative sequence, generatorsare equal, while their zero sequencereactanceis onefourth of their positive sequence reactance. The zero sequence reactances the of reactances. lines are to be taken as 2.5 times their positive sequence Set all prefault voltages = 1 pu.
12
NCES REFERE
Books
1. 2. 3. 4. 5. 6. W.D., Elementsof Power SystemAnalysis,4th edn., McGrawHill, Stevenson, New York, 1982. Elgerd, O.I., Electric Energy Systems Theory: An Introduction, 2nd edn., McGrawHill, New York, 1982. Analysis, Wiley, New York, 1979. Gross,C.A., Power System J.R., Modern Power Systems, Ncuenswander, International Textbook Co., Ncw York, 1971. Bergan,A.R. and V. Vittal, Power System Analysis,2nd edn.,PearsonEducation Asia, Delhi, 2000. Soman, S.A, S.A. Khapardeand Shubha Pandit, ComputationalMethods for Analysis, KAP, Boston, 2002. Large SparsePower Systems
T2.T
INTRODUCTION
Papers
7. 8. Brown, H.E. and C.E. Person,"Short Circuit Studiesof Large Systemsby the Impedance Matrix Method", Proc. PICA, 1967,p. 335. Involving Open Smith, D.R., "Digital Simulation of SimultaneousUrrbalances IEEE Trans.PnS, 1970,1826. and FaultcdConductors",
power systemis its ability to returnto normal The stabilityof an interconnected or stable operation after having been subjectedto some form of disturbance. Conversely, instability means a condition denoting loss of synchronism or have beenrecognizedas an essential falling out of step. Stability considerations part of power system planning for a long time. With interconnectedsystems continually growing in size and extending over vast geographicalregions, it is bdtween various becomingincreasinglymore difficult to maintain synchrortism '. parts of a power system. by The dynamicsof a power system are characterised its basic featuresgiven
tjElt w.
1. Synchronoustie exhibits the typical behaviourthat as power transfer is gradually increaseda maximum limit is reachedbeyond which the system i.e., it falls out of step. cannot stay in synchronism, 2. The systemis basically a springinertiaoscillatory system with inertia on the mechanicalside and spring action providedby the synchronoustie wherein power transfer is proportional to sin d or d (for small E, 6 being the relative internal angle of machines). 3. Because of power transfer being proportional to sin d, the equation determining system dynamics is nonlinear for disturbances causing large variations in angle d, Stability phenomenonpeculiar to nonlinear systemsas distinguished from linear systems is therefore exhibited by power systems (stable up to a certain magnitude of disturbance and unstable for larger disturbances). Accordi^rglypower system stability problemsare classified into three basic types*steady state,dynamic and transient.
*There are no universally accepted precise definitions of this terminology. For a definition of some important tenns related to power system stability, refer to IEEE Standard Dictionary of Electrical and Electronic Terms, IEEE, New York, 19i2.
434 .'l
'
l
Modernpower SystemAnalysis
Dnrrrn'
Qrra+^
Or^L:l:r.
I
^
Th"t study of steady state stability is basically concerned with the determination the upperlimit of machineloadingsbeiore losing of synchronism, providedthe loading is increased gradually. Dynamic instability is more probable than steady state instability. Small disturbances are esntinuaHy oeeurring irr a po*.. system r"#"ti"* i" loadings,changesin turbine speeds,etc.) which are small enough not to cause the systemto lose synchronismbut do excite the systeminto the"itate of natural oscillations.The system is said to be dynamically stable if the oscillations do not acquiremore than certainamplitudeand die out quickly (i.e., the systemis welldamped).In a dynamically unstablesystem, the oscillation amplitude is large and thesepersist for a long time (i.e., the system is underdap"a;. rni, kind of instability behaviour constitutesa seriousthreat to system security and createsvery difficult operating conditions. Dynamic stability can be significantly improved through the use of power system stabilizers.Dynamic system study has to be carried out for 510 s and sometimesup to 30 s. computer simulationis the only effective meansof studying dynamic stability problems. The samesimulationprogrammesare, of course,appiicableto transient stability s t udie s . Following a sudden disturbance on a power system rotor speeds, rotor angulardifferencesand power transferundergofast changeswhose magnitudes aredependent upon the severityof disturbance. a large disturbanc", For in angulardifferencesmay be so large as to .:ausethe machines "hung.. to fall out of step' This type of instability is known as transient instability and is a fast phenomenon usuallyoccurringwithin I s fbr a generator close to the causeof disturbance. There is a large rangeof disturbanceiwhich may occur on a power system, a fault on a heavilyloaclecl which requires but line openingthc lipc t<l clear the fault is usually of greatestconcern.The tripping of a loadJd generator or the abrupt dropping of a large load may also causeinstability. The effect of short circuits (faults), the most severetype of disturbance to which a power systemis subjected,must be determinedin nearly all stability studies' During a fault, electrical power from nearby generators is reduced drastically,while power from remote generators scarcelyaf1'ecte4. is ln some cases'the system may be stable even with a sustainedfault, whereas other systems will be stable only if the fault is cleared with sufficient rapidity. Whetherthe systemis stableon occurrence a fault depends of not only on the systemitself, but also on the type of fault, location of fauit, rapidity of clearing and method of clearing,i.e., whether clearedby the sequential opening of two or more breakersor by simultaneousopening and whether or not the faulted line is reclosed.The transientstability limit is atmost always lower than the steady statelimit, but unlike the latter,it may exhibit different valuesdepending on the nature,location and magnitudeof disturbance. Modern power systemshave many interconnected generatingstations,each with severalgenerators and many loads.The machineJlocatedatany one point ln a systemnormally act in unison.It is, therefore, common practice in stability
machines which are not separatedby lines of high reactance are lumped togetherand consideredas one equivalent machine. Thus a multimachine system can often be reduced to aq equlyalg4t fery lq4qhrue system If Synchronlsm lost, ttrern'achinei eaitr gioup stay togetheralthough they go ii of out of step with other groups.Qualitative behaviour of machinesin an actual system usually that of a two machine system.Because its simplicity, is of the two machine system is extremelyuseful in describingthe general concepts of power system stability and the influence of varioust'actors stability. on It will be seenin this chaptertbata two machine systemcan be regardedas a single machinesystemconnected.to infinite system. Stability study of a multimachine systemmust necessarily carried out on be a digital computer. I2.2 DYNAMICS OF A SYNCHRONOUS MACHINE
The kinetic energy of the rotor at synchronous machineis
K E = 1 JJ,^ x 106 MJ 2
where / = rotor moment of inertia in kgm2 aro, = synchronous speedin rad (mech)/s But whorc
u.r,n= rotor speed in rad (elect)/s
P = nuurbelo1'rnaclrine poles
= KE +(t(?)'c.,. xro.)*
L M,
2 ' ,
M = J ( ? \ ' u , x ,1 0  6 ' \P/ = moment of inertia in MJs/elect rad We shall detine the inertia constantH such that
G H = K E =u % M J ! 2
G = machinerating (base) in MVA (3phase) H = inertict constant in MJ/I4VA or MWs/MVA
{36 i
I
'.: : il
Power SystemStability MociernPower Sysiemnnaiysis e Swing Equation
I I
437
It immediatelyfollows that 2GH = GH MJs/elect rad M = (ts lt f = ffi 14Js/electdegree 180f
(r2.r)
and electrical ;ure 12.1 shows the torque, speedand flow of mechanical windage, triction and wersin a synchronousmachine.It is assumedthat the nloss torque is negligible. The differential equation governing the rotor namicscan then be written as
M is also calledthe inertia constant. Taking G as base, the inertia constantin pu is M (pu) = +
 d ' o ^ =T^J :; t ot(r2.2)
r " Nt
(r2.3)
nf
s2lelectrad ),, degree s'lelect
'here 0* = angle in rad (mech) T* = turbine torque in Nm; it acquiresa negative value for a motoring T, = electromagnetictorque developedin Nm; it acquiresnegativevalue
for a motoring machine machine
= H ffi
The inertia constant H has a characteristicvalue or a range of values for Table 12.1 lists some typical inertia constants. each classof machines. machines* of , Table 12.1 Typicalinertiaconstants synchronous
Type oJ Mat:hine Turbine Generator Condensing
Intertia Con.slunt H Stored Energy in MW Sec per MVA**
t m
o_>
machine powers a synchronous in and Fig.12.1 Flowof mechanical electrical While the rotor undergoesdynamics as per Eq. (12'3), the rotor' speed (1s) [Sec. by changes insignificant magnitude for the time period of interest 0.i. Equation (12.3) can therefore be converted into its more convenient
lnrltar ,vvrvr Tnrm tvrrlt
NonCondensing Water wheel Gencrzttor (< Slowspeed 200 rpm) Highspeed(> 200 rpm) +' C S y n c h l o r r o u so r r d c r r s c r4+ Large Small Motor with load varying li'ortr Synclrrcnous 1.0 to 5.0 and highcr lor hcavy l'lywhccls
1,800 rpm 3,000rpm 3,000rpni
96 74 43
LJ
.
*  Lv . 2. \,\.l .ll l,l =l y : * * n l l
vJ
' e rl !l c a *t nl r r^ r ( U n U Ur r . l t n lrvcl nqrr al i n : l r t l l r t l J U Lv
I
cv nr nrs rt r:rt n t a t t h g S V n C h f O n O U S ! r u q
24 t.25 l.00
peed(ur,.). Multiplying both sidesof Eq. ( 12.3)by u),,^'we can write t':t);" x lo('  P^ J6"n,
dtt where
P,. Mw
(t2.4)
2.00
'ttt
p

higher for from Table 12.1 th'atthe value of H is considerably It is observed Thirty to sixty per cent of wheelgenerator. thzrn steamturbogenerator tbr watc:r unit is that of the priine mover, the total inertia of a steam turbogenerator whereasonly 4 I5Vo of the inertia of a hydroelectric generatingunit is that of the waterwheel,including water.
D t 
,

mechanical ptlwer inPut l n electrical power outPttt t n
MW lossis assumed copPer MW; stator
negligible. Rewriting Eq. (12.4)
/ )\2 x ul;) u.r, 1o6) ff s2a
 P^ P,Mw
where
* Rcprinted with permission of the Westinghous Electric Corporation from Electrical Transmission and Distribution Reference Book. +* Where range is given, the first figure applies to the smaller MVA sizes. *tc+ HydrogenCooled,25 cent less. per
0, = angle in rad (elect) nod = Mi; ' \ , = ,P,,pP" (12.5)
ol
43S I
Modernpower SvstemAnalvsis
t  
_
Power System Stabitity
Grnu"h= machine rating (base) Gryrt",o= system base
It is more convenient measure angularposition of the rotor with respect to the to a synchronously rotating frame of reference.Let o= (a;,?;[' ;"':?: f# (called torque angle/power angle) ul ar di sP1 ement fro m sy nchronou sl y ac

_
I +Sl
Equation (12.11) can then be written as
(r2.6)
From Eq. (12.6) d rg , _ d 2 6 ,Jt2 dt2 Hence Eq. (12.5) can be written in terms of d as ' , d26 M,  : = P ^  P e M w otWith M as definedin Eq. (12.1), we can write GH d2d
 g'"" +" (Yry*l =(p p.)
Gsystem \/ dt' \ ) "t c' Grrr,.*
(r2.7) or
(12.8)
{ " " " ' o ' ! = 'P* P  P"PU sYstem in base dl "f = **" Hryr,"* Hu"ht'+*l
\..Gt""'n /
(t:.r2) (r2.r3)
where
= machine inertia constantin system base
fVP*P"Mw
Dividing throughoutby G, the MVA rating of the machine,
a
(r2.9)
Machines
Swinging
Coherently
Consider the swing equationsof two machinesor a common systembase. Hr d261
M(pu) = P* P,; + dt'
in pu of machine rating as base where
"f (12.r0)
a;
A;
= P*r  P"t Pu
= P*2 P"zPu
(r2.14) . (12.rs)
H2 d262
=+ M(Pu) lrj
or H dzb = P*^nf dt" P, pu (IZ.1I)
"f Since the machinerotors swing together(coherentlyor in unison) 6, = [r= [ A ddi ng Eqs ( 12. 14)and ( 12. 15) H"q d26 = P*trf dtz whcrc Prr=P*r+ Pn Pr=P"l * Prz H"q=Hr+ H, P"
(r2.16"
(Eq. (12.1D)l}q.(12.1 is called thestvingecluatioy This equation 1)), and it describes rotor dynamicsfbr a synchronous the machine(generating/motoring). It is a secondorder differential equationwhere the darnpingterm (proportional to d6ldt)is absent because the assumption a lossless of of machineanclthe fact that the torqueof damper winding has been ignored. This assumptionleads to pessimisticresultsin transientstability analysisdampinghelps to stabilizethe system.Dampingmust of coursebe considered a dynamic stability study. in Sincethe electrical power P, depends uponthe sine of angled(seeEq. (12.29)), the swing equationis a nonlinear secondorderdifferential equation. Multimachine System
(r2.r7)
The two machinesswinging coherentlyare thus reducedto a single machine as in Eq. (12.16).The equivalentinertia in Eq. (12.17) can be wrirten as
H"q, = Hl ,nu.h Gl ^ach/Grystem* Hz ^u"h G2 u"h/G.yr,.. (12.18)
The aboveresultsare easily extendable any numberof machinesswinging to coherently.
In a multimachine systema common systembase must be chosen. Let
ffimel"L2rl
1
A 50 Hz, four pole turbogenerator rated 100 MVA, 11 kv has an inertia
g6nsf anf nf R O MI/I\/IVA

t4/s' I torJern Power Srrctarn Anarrraio
mqtsqllvteryerSvele
3. Effect of voltage regulating loop during the transientis ignored, as a also This assumption constant. machineemf remains the consequence generated the helps to stabilize system. resultsvoltage regulator to leads pessimistic to Before the swing equation can be solved,it is necessary determinethe of dependence the electrical powel otttput(P,,)upon the rotor angle. Simplified Machine Model
(a) Find the stored energy in the rotor at synchronous speed. (b) If the mechanical input is suclcJenly raisedto
80 MW fbr an electricalload of 50 MW, tind rotor acceleration,neglecting rrlechanicaland electrical losses.
(c) If the acceleration calculated part (b) is maintainedfor 10 cycles,find in the changein torque angle and rotor speedin revolutionsper minute at the end of this period. )olution (a) Storedenergy= GH = 100 x g = g00 MJ
For a nonsalientpole machine, the per phaseinduced emfterminalvoltage understeadyconditionsis equation (r2.Ie) E =V + jXolu+ jXol,,;X,r) X,r where I=Ia+ Is
( b )P , = 8 0  5 0 = 3 0 M W = M 4
r,
. 1  
(r2.20)
dt'
CH 180/
800 1 8 0x 5 0
4
MJr/elect deg
andusual symbolsare used. condition Undertransient XaX'a1Xa but X'o = Xn since the main fleld is on the daxis Xtd< Xo ; but the differenceis less than in Eq' (I2.I9) Equation(12.19)during the transientmodifiesto (12.2r) Et =V + jxt lo+ jXnln  I) + jXotlo =V + jXq(I ' (12.22) = (Y + jxp + j(X'a  Xq)Id The phasor diagram colrespondingto Eql (12.21) and (12.22) is drawn in Fig. 12.2. Since under transient condition, X'a 1X, but Xn remainsalmost unaffected, that it is fairly valid to assttme
4 d26 = 3o +'j 6,2 or
r ) c
 337 elect '5 deg/s2 (c) 10cycres =
Change d= !{ZZI.S) x (0.2)2= 6.75 elect degrees in
= 60 x 337'5 2x360J . ! = zo ' 11 ^ E 6 2r {PnVs
"lT
.'. Rotor speedat the end of l0 cvcles
x'a = xq
(r2.23)
 r2ox5o 'z8.tz5 0.2 + x
4 = 1505.625 rpm I2.3 POWER ANGLE EQUATION
In solvingthe s'uving equation(Eq (12.10)),certain simplifying assumpticns are usually made. These are: 1. Mechanicalpower input to the machine (P*) remainsconstantduring the period of electromechanical transientof interest. In other words, it meansthat the effect of the turbine governing loop is ignored being much slower than the speedof the transient.This assumptionleads to pessimisticresultgoverning loop helpsto stabilizethe sysrem. '2. Rotor speedchangesare insignificantthese have alreadybeen ignored in formulating the swing equation.
Fig. 12.2 Phasor diagramsalient pole machine
,firZlll
r"orrr r
E,=V+jXnI = V + jXotl
atysis _
\
Power Svstem Stabilitv
t 2
equation (12.22)now becomes
T*
[ECAC,+'
(r2.24)
D'
o
O f
@
r
synchronous reactance)
The machine mod also applies to a cylindrical rotor machine where
X,l = X*/(transient
Fig. 12.5 Twobusstability study network For the 2bus system of Fig. 12.5
f Y,, Y"1 = Y n u s I  j '  : ' l ; YI , z =Y z r
LY^ Yr,[
Fig. 12.3 Simplifiedmachine model
(r2.2s)
The simplified machine of Fig. 12.3 will be used in all stabilitv studies. P.ower Angle'Cunre For the purposesof stability studies lEl1, transient emf of generatormotor, remains constant or is the independent variable determined by the voltage regulatingloop but v, the generatordeterminedterminal voltage is a dependent variable'Therefore, nodes(buses) the stability the of study network pertainto the ernf terminalin the machinemodel as shown in rig. 12.4, while the machine reactance (Xu) is absorbed the systemnetwork as in clifferent from a load flow sttlcly' Fttrther, loirtls (othcrthrrn lhe liu'gcs;ynchronous riitittir.s) bc r.cpiacetl will by equivalent static admittances (connected in shunt between transmission network buses and the referencebus). This is so becauseload voltages vary cltlringa stabilitystutly (in a loaclllow stucly, the.se remain constant within a nalrow bancl).
Complex power into bus is given by Pi+ jQiElf At bus I Pr + jQr  Er' (YyE1)* + E, (YpEil* But E't =lEil l6; E / = l E ' z l1 4 . Yr z= lYr zl l0 p
(r2.26)
Y, u = G r t + jBt i
Since in solutionof the swing equationonly real power is involved, we have from Eq. (12.26) I r t = l E l i 2 G t r + l E r tI i E i i i y , r l c o s ( , t i , h  0 , 2 ) u 2 . 2 7 ) A similar equationwill hold at bus 2. Let lEllzG, = P, l E t' l lEztI lY7l = Pn, *
,/
and
4 6=6
Qn = x/2 1 Then Eq. (12.27) can be written as Pr = P, * P,n"* sin (6  1); Power Angle Equation For a purely reactive network Gtt = 0 (.'. P. = 0); lossless network } t z = n 1 2 ,: . J = 0 P,=P^*sin 6
Systemnetwork
(IZ.ZB)
Fig. 12.4
(12.29a)
j0.5
I
I
where
I lE''  ' P^^, = lE'' x ' '
simplified power angle equation
(r2.29b)
i0.5
X = transfer reactancebetween nodes (i.e., between E{ andEi) graphical plot of power angle equation (Eq.(12.29)) is shown in Fig. 12.6. where 'Ihe
p"l
I
D I max (Ps6+APe).....t*
Peo
Generator
lE'lt6
1lo"
(b)
Fg.12.7Asimpesystemwithitsreactancediagram
Fig. 12.6 Poweranglecurve The swing equation (Eq. (12.10)) can now be written as
Yl aZt
0.5 X r z= 0 ' 2 5+ 0 ' 1 + (12.30) = 0.6 at faultoccurs the a case now a morecomplicated wherein 3phase Consider that becomes diagram the case reactance midpointof oneof thelinesin which of Fig. 12.8(a). StarDelta Conversion to that of the Converting star at the bus 3 to delta,the networktransforms wherein Fig. 12.8(b)
+=
 P^
sin Pn'u* dpu
7rI dt
differential equation with which, as already stated,is a nonlinear secondorder no damping. T2.4 NODE ELIMINATION TECHNIOUE
In stability studies, it has been indicated that the busesto be considered are thosewhich are excitedby the internal machinevoltages(transientemf's) and not the load buseswhich are excited by the terminal voltagesof the generators. Therefore, in Y"u, formulation for the stability study, the load buses must be eliminated. Three methods are available for bus elimination. These are illustrated by the simple system of Fig. 12.7(a) whose reactancediagram is by bus 3 getseasilyeliminated drawn in Fig. I2.7(b).In this simplesituation, parallel combination of the lines. Thus
o

(,
j0.25
at( U
lE/lt6
1t t r
(a)
446 L
power System Modern Anatysis
l u t _ , This method obviously is cumbersometo apply for a network of even small complexity and cannot be computenzed.
Power SystemStability Node Elimination Technique
lE/lt6
Formulate the bus admittances for the 3bus system of Fig. 12.8(a).This network is redrawn in Fig. 12.9 wherein instead of reactance branch. admittancesare shown. For this network.
CD
@
o
(r
ra
(c) Fig. 12.8
Fig. 12. 9
v
_ 0.25x 0.35+ 0.35x 0.5 + 0.5 x0.25 = 1.55
ruus
The bus 3 is to be eliminated. In general for a 3bus system
This method for a complex network, however, cannot be mechainzed for preparing a computerprogramme. Thevenin's Equivalent
With ref'crcncc Fig. l2.ti(a), tho Thcvcnirr'sccluivalcntlbr lo thc network portion to the left of terminals a b as drawn in Fig. 12.8(c)wherein bus t has been modified to 1/. , rr h = v 0.25 lEtl l5 025+0,5
lh  lY^ llv,I Llrl Lv,, Yn rrrJL%l
or
Y"t"lIu'I f''I=fr' Y, vu
(r2.31)
is at Sinceno source connected the bus3 =o It Y r r V r +Y r r V r + Y r r V r = O
= 0 .4 1 7 Et I 16 xrh =
Itlow
or
vz=?rr ?r, Yrt Y.,
(r2.32)
0 .3 5 x 0 .2 5 035+025
 o'146
Substituting this value of V3 in the remaining two equations of Eq. (12.31), thereby eliminating Vy It =Y,Vt * YrzVz+ YttVt
Xt2 =0.146+ 0.5 = 0.646*
*This value is different from that obtained by star delta transformation as V* is no longer lEtl I { in fact it is 0.417 lEtl 16.
=(",  Y,rYr, +(ty," ) u ' [ ' ' !+)v" Yr, Yr,
(.^tt
)''
In compactform
to be 1.0 pu.
=lt,i,' (reduced) Ynus
lY'r,
l,l,'1 Y'r,
)
(12.33)
Solution E E l . x =l f fr lil =^m = o ' 4I9ZPI u (2) Equivalent circuit with capacitive reactor is shown in Fig. 12.71 (a). p'e65 j1.o jo.l i0.25 io.1 i1.o
Y ' t l = Y '33 r y "
Y'12= Y'21= Ytz 
Q23aa) (r2.34b) (r2.34c)
lEnl= 1.2
YrtYt, Ytt
vt,  rr t 22= y 'zz  Y r t Y t yT In general, in eliminating node n
i1.0
=
lEml= 1.0
Yo,(new) = Yry(old)
Yo^(old)Y,,(old)
(r2.3s)
(a)
(b)
Applying Eq. ( 12.34) to the example in hand
Flg.12.11 Converting star to delta, the network of Fig. 12.11(a)is reduced to that of Fig. 12.11(b)where 7 X ( t r a n s f e r )  / 1 . 3 5 X / 1 . 1 + / 1 . 1j1.0 J 1 . 0 ) + (  / 1 . 0 ) X J 1 . 3 5 . . X(_
lt.gzt = (reduced); Ysu5 1 0.646 L
It then follows that X,t=:o.646 = 1 . 5 4 8 (= 1 . 5 5 )
= j0.965 1 . 2 x 1= . pu 1244 statepowerlimit = Steady ffi we by replaced inductivereactance, get the reactance (3) With capacitive ster Converting to delta,we havethe circuitof Fig. 12.12. equivalent of trasferreactance i1.35 i1.1
static capacitive reactorof In the systenrshown in Fig. 12,10,a threephase reactance 1 pu per phase is connectedthrough a switch at motor bus bar. Calculatethe limit of steadystatepower with and without reactorswitch closed. Recalculatethe power limit with capacitive reactor replaced by an inductive reactor of the same value.
Fig.12.12 x.t1.35 x.r1.l+ "11.1x + 11.0 ll.0 _ j1.35 i 1.0  j3.e35
7X(transfer) F i g .1 2 . 1 0
!5o' l
""attt
t"t,
PowerSystemStability in As alreadycalculated this section, Xn = l'55
hiffi
Steadystatepower timit ':'!] = 0.304 pu 3.935
Example12.3
The generatorof Fig. 12.7(a)is delivering 1.0 pu power to the infinite bus (lVl = 1'0 pu), with the generatorterminal voltage of v,r = 1.0 pu. calculate the generatoremf behind transientreactance. Find the maximu io*". that can be transferredunder the following conditions: (a) Systemhealthy (b) One line shorted (3_phase) rhe middle in (c) One line open. . Plot all the three power angle curves. Solution Vt = lV,l la From power angle equation L,et =  la or
= !".!;!! = 0.6e4 pu
P, = 0.694 sin d (ii)
(c) One line open: It easily follows from Fig. 12.7(b) that X r z= 0 . 2 5 + 0 . 1 + 0 . 5 = 0 . 8 5
P*.*=tit^o]t =r.265 0.85
or P" = I.265 sin 6 (iii) The plot of the three power angle curves (Eqs. (i), (ii) and (iii)) is drawn in Fig. 12.13. Under healthy condition, the systemis operatedwith P,, = P, = 1.0 pu and 6o= 33.9",i.e., at the point P on the power anglecurve 1.79 sin d As one line is shortedin the middle, Po,remains fixed at 1.0 pu (governing system act instantaneously)and is further assumedto remain fixed throughout the transient(governing action is slow), while the operatingpoint instantly shifts to of Q on thc curvc 0.694sin dat d= 33.9".Noticcthutbccuusc machineinertiu, suddenly. the rotor angle can not change
t l _v t l v I s t n o = p " X ( t"t ) < t= I
[025+oJJsrn or rr = 2 0 .5 " Current into infinite bus. I _ lV,llalVll0"
1.79
jx  1 2 0 . 5 I l 0
1.265 Pn=1'O 0.694
0 . 6 9 4s i n 6
i0.3s
=l+j0.18=1.016 10.3" 1
Voltage behind transient reactance, Et =tltr + .j0.6 (l + 70.1g) x = 0 .8 9 2+ j 0 .6  1.075133.9"
0
33.90
I 900
A
Fig. 12.13 Poweranglecurue,s
(a) Systemhealrhy p^u* = lv )lEt   1x 1.075= ., t 9 ,,  t ' F,.\ PU x,, c,5 P, = I'79 sin d (b) One line shorredin the middle: (i) I2.5 SIMPLE SYSTEMS to Infinite Bus
Machine
Connected
to Figure 12.14 is the circuit model of a single machineconnected infinite bus Xr.In this simple case through a line of reacthnce
t

Power System Stability
I Xtransf.er =X'a* X,
I lt5'
From Eq. (12.29b)
and
Xt urrrf..
,, ='4U
sin d= p.u*sin d
. ="f( P  r  P " r \ "tlH') it l2?:)= d26"
(rn)
(r2.39b)
(r2.36)
SubtractingEq. (12.39b)from Eq. (12.39a)
The dynamicsof this sysremare describecl Eq. (12.11) as in
d2@,;6)=^r(':jr!,] .J
dtz \ HrH, ) or H"q d26 * ., =Pn7rI (ltP,
,.  P,)
(r2.40) (r2.4r) (r2.42) (r2.43)
#ft= P^P" Pu
where
lE/lt6
6=4
6.
HtH, rr "eq Hl + Hz The electrical power interchange is given by expression
Fi1.12.14 Machine connected infinite to bus Two Machine System The caseof two finite machines connectedthrough a line (X") is illustrated in Fig. 12.15whereone of the machinesmust be generating und th" other must be motoring.Under steadycondition, beforethe system goesinto dynamicsand
p"= ,E!4!' X'0, x, + xd2 ,in6
+
(12.44)
The swing equation Eq. (12.41) and the power angle equation F;q (12.aa) have the same form as for a single machine connectedto infinite bus. Thus a to twomachine systemis equivalentto a single machineconnected infinite bus. bus) qYstemwould be (connected intinite to Becauseof this, the singlemachine studied extensively in this chapter.
Fig. 12.15 Twomachinesvstem P*t=P*z=P.
In the system of Example 12.3,the generatorhas an inertia constant of 4 MJ/ of MVA, write the swing equationupon occurrence the fault. What is the initial to can be assumed remain constantfor If this acceleration angularacceleration? at the end of this time interval and the new Lt = 0.05s, find the rotor angle acceleration. Solution (12.38a) Swing equation upon occurrenceof fault H 1 g 0 fd v 4 d'6 _o  r m  'D e d,26 ,
the mechanicalinput/outputof the two machinesis assumedto remain constant at these values throughout the dynamics (governor action assumed slow). During steadystateor in dynamic condition, the electrical power output of the generatormust be absorbedby the motor (network being lossless).Thus at all time Prl=Prz= P" The swing equationsfor the two machinescan now be written as
(12.38b)
*"t#=l0'694
dt"
s i n6
t4 = z2so  0.6e4 d;. (1 sin
(12.39a)
P ) t L _"_ ir ( P ^ r _ " r_ _ " ( p .  + \  ' dtz ff, ):ut t rr, l
,.4#jirf
I
vooern powersysremAnatysis
Initial rotor angle do= 33.9" (calculatedin Example 12.3) a2 l s = 2 2 5 0 (l  0.694 si n 33.9" ) ;l d t ' l , n +
Linearizing about the operatingpoint Qo (P"0, 4) *" can write
ffi
L ltc..:'
]i;E!!";1&t
LP,=(*).
o,
The excursionsof A d are then describedby
#l l ll r : o* C +
= 0; rotor speed cannotchange suddenly
= + no9i+'  P^ (P,o aP,')  L,P,
d,r
A, (i n A,t = 0 .0 5 s )= I x 1379 (0.05)2 x
2
t .7"
6 t = 6 + 4,6 = 33.9+ 1.7"= 35.6"
M d'+' * dr
(r2.47)
a26l = 2250(l  0.694sin 35.6") ., d r ll r = 0 . 0 5 s
 l34I elect deg/s2 Observethat as the rotor angle increases, electrical power output of the the generatorincreasesand so the accelerationof the rotor reduces. 12.6 STEADY STATE STABILITY
where
P
d dt
The system stability to small changesis determined from the characteristic equation
Mp, +[#], =o
whosc two roots are
f / t \ t r , . 1 ( \ f {
The steady state stability limit of a particularcircuit of a power system is defineclas the maximutnpower that can be transmittedtri fhe receivinoen; v'r6 vrru without loss of synchronism. Consider the sirnplesystemof Fig. 12.14whose clynamicsis describect by equations M* dl = P^ P " M W; Eq. (12.8) ln Pu sYstem ) ,in 6'= p^u*sin d
P = + l  \ u 1 t 0 o ) ol L M I
As long as (0P/0 0o it positive, the roots are purely imaginary and conjugate and damper and the system behaviour is oscillatory about do.Line resistance windings of machine,which have been ignored in the abovemodelling, cause the system oscillations to decay. The system is therefore stable for a small increment in power so long as
M = H 7 and p, = !4)!
(12.4s)
(12.46'
(aP,/aa > o
(12.48)
x,t
For determinationof steadystate stability, the direct axis reactance (X.r) ant, voltagebehind X4 are usedin the above equahons. The plot of Eq. (12.46)is given in Fig. 12.6.Let the system be operaring with steadypower transferof P^ = P^with torque angle as indicated in the d figure. Assume a small increment AP in the electric power with the input from the prime mover remainingfixed at p*(governor r.rforrr" is slow compared to
When (0 P/AD, is negative, the roots are real, one positive and the other without increases The torqueangle therefore negativebut of equal magnitude. (disturbancc) and the bound upon occurrcnccol a small powcr incretrtent synchronismis soon lost. The systemis therefore unstablefor @ Pe/aDo < 0 cofficienr. This is also called stffiess known as synchronizing @p/A[ois (electrical) of synchronousmachine. Assuming lEl and lVl to remain constant,the systemis unstable,if
po*e,systmst"uititv
lEllvl c o sd ^ < o X
b{dffi
r
(12.4e) 4>90" Themaximum powerthatcanbe transmitted withoutlossof stabili
or
Iaee 1 l 
1.8 L 06 Jro" = 0.577MW (pu)/elect rad
M ( pu) = 4 t r x5o r s?/ cr cctat r
L_x _ _ _2_ r c o s 3 0 "
( r2.s0)
and is given by
P^u*
lEnvl
H = o*r o From characteristicequation
(12.sr)
If the system is operating below the limit of steady stability condition (Eq. 12'48), it may continue to oscillate for a long time if the iamping is low. Persistent oscillations are a threat to system security. The study oi system damping is rhe study of dynamicalstability. The above procedure is also applicable for complex systems wherein governor action and excitation control are also accounted for. The describing dif f er e n ti a le q u a ti o ni s l i n e a ri z e c l bout the opcrati ngpoi nt. a C oncl i ti ep fbr steady state stability is then determinedfrom the correspondingcharacteristic equation (which now is of order higher than two) It was assumedin the above account that the internal rnachinc voltage lEl remains constant(i.e., excitation is held constant).The result is that as loading increases,the terminal voltage lv,l dips heavily which cannot be toleratcdin practice. Thereforc,we must consicler steadystatestabilitylimit the by assuming that excitation is adjusted for every load increase to keep lv,l constanr.This is how the system will be operaiedpractically. It may be understocdthat we are still not considering the effect of automatic excitation control. steady state stability limit with lv,l anrt lvl constantis consiclered in Example 12.6.
P=tr[(*),,"1*)'
+i(WiY")u == i4.76
FrequencYof oscillations= 4.76 railsec 4'76 0. 758Hz 2r (ii) For 807oloading
= s i nd o +
Pu*
=0.8or =53.1" 6
rqa)
\ 05 )rr,
 r'Zxrcos 53.1"
1.8 = 0'4 MW (Pu)/elect rad
p =!, (q+k)* =* i3s6
Frequencyof oscillations = 3.96 radlsec
?q6
A synchronous generator reactance1.20pu is connected an infinite bus of to bar (l Vl = 1.0pr) throughtransformers a line of total reactance and of 0.60 pu. The generatorno load voltageis I .20 pu and its inertia constantis H = 4 MWsilvIVA. The resistance and machinedamping may be assumednegligible. The systemfrequencyis 50 Hz. Calculate the frequency of natural oscillations if the generatoris loaded to (1) 50Voand (ii) 80Voof its maximum power limit. Solution (i) For 50Voloading s m d o i t
'
')*
Find the steady state power limit of a system consisting of a generator equivalent reactance0.50 pu connectedto an infinite bus through a series is gf of rcactance 1.0pu. The terminalvoltage the generator held at 1.20pu and the voltage of the infinite bus is 1.0 pu. Solution The systemis shown in Fig. 12.16.Let the voltage of the infinite bus be taken as reference.
P
max
= 0 . 5 o r4 = 3 0 o
m Analysis Then Now
Power SrrstemStahilitu
EsE
V = 7 . 0 /  f f , V t= ! . 2 l 0
I_
1.2107.0 jI
Xa= O'5
lE4t6
Vt=1.219
I
V = 1.0100
Fl g. 12.16 E = Vt + jXdI = 1.2 l0 + j0.5
L
I
rr . 2l g
A knowledgeof steady state stability limit is important for various reasons.A system can be operated above its transient stability limit but not above its steadytatelimit. Nowrwith increased fault el,earing speedsjt is possible to make the transient limit closely approach the steady state limit. As is clear from Eq. (12.51),the methodsof improving steadystate stability limit of a systemare to reduceX and increaseeither or both lEl and I Vl. If the transmissionlines are of sufficiently high reactance, the stability limit can be raisedby using two parallel lines which incidently also increases reliability the of the system.Series capacitorsare sometimesemployed in lines to get better voltage regulation and to raise the stability limit by decreasing the line reactance.Higher excitation voltages and quick excitation system are also employed to improve the stability limit. I2.7 TRANSIENT STABITITY
 1.0
J
I
.E = l.g l0 _ 0.5= (t.g cos e_ 0.5)+ sin 71.g 0 Steady state porver rimit is reached whenE hasan angle 6= 90o, of i.e.,its real part is zero.Thus, 1 . 8 c o0  0 . 5 = 0 s
or Now
It has been shown in Sec. L2.4 that the dynamics of a single synchrono,rs machine connectedto infinite bus bars is governed by the nonlinear differential equation
0 = 73.87" Vt = 1.2/.73.87"= 0.332+ j t . 1 5 2
,, d'6 = P ^  P " M iF
where or P, = P* sin d
r _ 0.332+.ir.rs2_r + j0.669 = t.t52
E 0.332+ jr.r52+ 70.5(1.152 j0.668) +  0.002 j1.728= 1.728 + I90. Steady statepowerlimit is givenby 1.7 L p ^ u  l E l l V l = i5 2 8 x= *= l'152 Pu V;+V If instead, generator is held fixed the emf at a valueof r.2pu, the steady state power limit would be
It is observed regulating generator that the emf to hold the terminalgenerator ','oltage r.2 pu raises at thepowerli.it frorn0.g pr'ro r.r52pu; this is how thevoltageregulating loop helpsin power system stab'ity.
= P*"* i#
=o'8 Pu
_ d26 (r2.s2) M +  P* P** sind otAs said earlier, this equation is known as the swing equation.No closed form solution exists for swing equation except for the simple case P = 0 (not a practicalcase)which involves elliptical integrals.For small disturbance(say, gradual loading), the equation can be linearized (see Sec. 12.6) leading to the concept of steady state stability where a unique criterion of stability (APrlAd>0) could be established.No generalizedcriteria are available* for determining system stability with large disturbances(called transient stability). The practical approach to the transient stability problem is therefore to list all important severe disturbances along with their possible locations to which the systernis likely to be subjected according to the experienceand judgement of the power system analyst. Numerical solution of the swing equation (or equationsfor a multimachine case) is then obtained in the presence of such disturbancesgiving a plot of d vs. r called the swing curve. If d starts to decrease after reaching a maximum value, it is normally assumed that the systemis stableand the oscillation of daround the equilibrium point will decay
tRecent literature gives methods of determining transient stability through Liapunov and Popov's stability criteria, b:rt thesehave not been of partical use so far.
ffiffif
I
power Modern System Anatysis
StabilitY S','stem Power
ffiffi
andfinally die out. As alreadypointed out in the introduction,important severe distulbancesare a short circuit or a sudden loss of load. For easeof analysiscertainassumptions and simplificationsare always made (some of thesehave already been made in arriving at the swing equation (Eq. /1 /l <.t\\ a rr 11consequences upon accuracyof results. 1. Transmission line as well as synchronous machine resistance are ignored. This leads to pessimisticresult as resistance introducesdamping term in the swing equationwhich helps stability. In Example I2.11, line iesistance has been taken into account. 2. Damping term contributedby synchronousmachinedamperwindings is ignored. This also leads to pessimisticresults for the transientstability limit. 3. Rotor speed is assumedto be synchronous.In fact it varies insignificantly during the course of the stability transient. 4. Mechanical input to machineis assumedto remain constantdurins the transient,i.e., regulating action of the generatorloop is ignored. This leais to pessimisticresults. 5. Voltage behind transientreactanceis assumedto remain constant, i.e., action of voltage regulating loop is ignored. It also leadsto pessimisticresults. 6. Shunt capacitances not difficult to account for in a stability study. are Where ignored, no greatly significant error is caused. 7. Loads are modelled as constant admittances. This is a reasonablv accuraterepresentation. Note: Sincerotor speedand hencefrequency vary insignificantly, the network parametersremain fixed during a stability study. A digital computer programme to compute the transient following sudden disturbance aan be suitably modified to include the effect of governlr action and excitation control.
p"rrnon.ntly till cleared manually. Since in the majority of faults the first ieclosure will be successful, the chances of system stability are greatly enhancedby using autoreclosebreakers.
Fig.12.17
In the caseof a perrnanentfault, this system completely falls apart. This will not be the case in a multimachine system.The stepslisted, in fact, apply to a system of any size. 1. From prefault loading, determine the voltage behind transientreactance of and the torque angle 16o the machine with reference to the infinite bus. fault, determine the power transfer equation Pr(A during 2. For the specified ' ^ault. In this system P" = 0 for a threephasefault' ' as obtained in step 1, calculate 5 . From the swing equation starting with fi das a function of time using a numerical technique of solving thetnonlinear differential equation. 4 . After clearanceof the fault, once again determine P, (A and solve further for d (r). In this case,P"(A = 0 as when the fault is cleared,the system getsCisconnected. 5 . After the transmissionline is switched on, again find P" (0 and continue to calculate d (r). 6 . If 6 (t) goes through a maximum value and starts to reduce,the system is Calculation regardedas stable.It is unstableif d(r) continuesto increase. after a suitable length of time. is ceased An important numericalmethod of calculating d(t) from the swing equation will be giurn in Section 12.9.For the single machineinfinite bus bar system, stability can be conveniently determined by the equal area criterion presented in the following section. I2.8 EOUAL AREA CRITERION
Upon occulTenceof a severedisturbance, say a short circuit, the power transfer between machines is greatly reduced, causing the machine torque angles to swing relatively. The circuit breakers near the fault disconnect the
In a systemwhere one machine is swinging with respectto an infinite bus, it is possible to study transient stability by means of a simple criterion, without resorting to the numerical solution of a swing equation.
Considerthe swing equation d26
AF
=
I
*@^
= P ' 1 &rt ,"= accelerating power (r2.s3)
The stability criterion for power systemsstatedabovecan be convertedintc a simple and easily applicableform for a single machine infinite bus system. Multiplying both sides of the swing equation * [t#), we get
M=! rf
ln pu system
2P" d6
M d t
Ifrtegrating,we have
(r2.s6)
where dois the initial rotor angle before it begins to swing due to disturbance. From Eqs. (12.55) and (12.56), the condition for stability can be written as
Fig. 12.1g protof 6 vs tfor stabreand unstabre systems lf the system is unstable dcontinuesto increase indefinitely with time and the machine loses synchronism.on the other hand, if the system is stable, 6(t) performs oscillations (nonsinusoidal)whose amplitude decreasesin actual praetice because darnpingterms (not included of in the swing equation).These two situations are shown in nig. 12.1g. since the system is non_linear, the nature of its response160l is not unique and it may exhibit instability in a fashion different from that indicated in Fig. rz.rg,depending upon the nature and severity of disturbance. However. experienceindicatesihai the response 6!'l j" a power system generallyfalls in the two broad categoriesas shown in the figure' It can easily be visualizednow (this has alsobeenstatedearlier) that for a stable system,indicationof stability will be given by observation the of first swing where dwill go to a maximumand will Jturtto reduce.This fact can be stated as a stability criterion, that the system is stableif at some time d6 =o (r2.s4) dt and is unstable,if d  6 > 0 <lt for a sufticiently long time (more than 1 s will genera'y do).
6
or
[r"ad
6,,
o
(r2.s7)
The condition of stability can therefore be stated as: the system is stable if the power) dcurve reduces zero at some value of to areaunderPo(accelerating areaunder Po 6curve must equal d In other words, the positive (accelerating) 'equal area' criterion of the negative (decelerating)area and hence the name lstability. To illustrate the equal area criterion of stability, we now consider several that may occur in a single machine infinite bus bar types of disturbances system. Sudden Change in Mechanical Input
Figure 12.t9 shows the transientmodel of a single machinetied to infinite bus bar. The electrical power transmittedis given by
Infinite bus bar lvlr0o Fig. 12.19
(12.ss)
>
Pm
ffifftl
Modern Power svstem nnarys,s
lEtllvl P, = u, # sin d P** sind ndT^e
,
Under steady operating condition P.o = Pro= P** sin do
4n Ar=)(PnP")d6
Az=i<r,P^)d6
6l
be possible to find angle d2 such that rg condition is finally reached when 41 own in Fig.l2.2L Under this condition, hat  6t= 6. = 6^o= T
n'sinl +:
(12.58)
6 b f % ?'i
i
6
2 (' s
Q <(4
Flg. 12.20 P" 6 diagram sudden for increase mechanical in inputto g e n e ra to rf F i g .1 2 .19 o This is indicatedby the point ainthe Pr 6 diagram of Fig. 12.20. Let the mechanical input to the rotor be suddenly increased to Pn (by opening the steam valve). The acceleratingpower 1o = P*t  P, causes'the rotor speed to increase (u> a,,r)and so does the rotor angle. At angle 6r,, Po= P*r Pr(= P* sin 4) = O (statepoint atb)but the rotor anglecontinues to increaseas t.,) ur.Po now becomesnegative(decelerating),the rotor speed begins to reducebut the anglecontinuesto increasetill at angle 6., a= ur once again (statepoint at c. At c), thedeceleratingarea A, equals the accelerating
bc
input with mechanical stability case of transient Fig. 12.21 Limiting increased suddenlY for available A, is lessthan A1' e Any turther increas in P^, meansthat the area the kinetic energy causes d to increasebeyond point c and so that the excess power, with the system decelerating power changes over to accelerating shown by use of the equal consequentlybecoming uistable. It has thus been increasein mechanicalinput areacriterion that thereisan upper limit to sudden (P^r Po,s),for the systemin question to remain stable' ' 'ii will remain stableeven ',,uy'ulso be not"i from Fig. 12.21thatthe system so long as the equal area beyond though the rotor may oscillate {^=.90"' for use in steady state = criterion is met. The condition of d 90" is meant apply to the transientstability case' stability only and does not
j.e., areaA, (areasare shaded),
= J ,, Od 0. Since the rotor is decelerating,
6o
the speedreduces below ur andthe rotor angle begins to reduce.The statepoint now traversesthe P,  6 curvein the oppositedirection as indicated by arrows in Fig. 12.20.It is easily seenthat the systemoscillatesabout the new steady statepoint b (6= 4) with angle excursionup to 6 *d 4.on the two sides. These oscillations are similar to the simple harmonic motion of an inertiaspring system except that these are not sinusoidal. As the oscillations decay out because of inherent system damping (not modelled),.the systemsettlesto the new steady state where P^t = P, = Prn.* sin dl
trs#ffi {i..fuu.xdr 
r._r_..._ ruooern A Fower uvslem Anarvsrs
Power a
l':fiffn#;i;
Effect of Clearing Time on Stability Let the system of Fig. 12.22 be operatingwith mechanicalinput P^ at a steady angleof d0 (Pn,= P") as shown by the point a on the Pr 6 cliagramof Fig. 12.23.If a 3phasefault occurs at the point P of the outgoing radial line, the electrical output of the generatorinstantly reducesto zero, i.e,, p, = 0 and the statepoint drops to b. The accelerationareaA, begins to increaseand so does the rotor angle while the statepoint movesalong bc. At time /. corresponding to angle 6, the faulted line is cleared by the opening of the line circuit breaker. The values of /, attd 4 are respectively known as clearing time and,clearing angle. The system once again becomeshealthy and transmits p, = p,ou,. d sin i.e. the state point shifts to d on the original P,  d curve. The rotor now decelerates and the decelerating area A, begins while the state point moves along de.
rresponding to a clearing angle can be t only by numerical integration except in this simple case.The equal established area criterion therefore gives only qualitative answer to system stability as the time whgn the breaker should be opened is hard to establish.
Pe
'D max
Pm

 lt___j
l

l
/'\
\/
co )
\
I
d",
6r"*
+
Criticalclearing angle
angle Fig. 12.24 Criticalclearing F19.12.22 If an angle fi canbe found such that A2= Ap the system is found to be stable. The systernfinally settlesdown to the steadyoperatingpoint a rn an oscillatory mannerbecauseof inherent damping.
Pe
D , max
and so\oes d, to As the clearing of the faulty line is delayed,A, increases find A2 = Ar till 6r = 6^ as shown in Fig. 12.24.For a clearing time (or angle) larger than this value, the systemwould be unstableas A, < Ar The maximum aiiowabie vaiue of the clearing time and angle for the systemto remain stabie are known respectively as critical clearing time and angle. For this simple case (P, = 0 during fault), explicit relationships for 6, (critical) and t" (critical) are establishedbelow. All angles are in radians. It is easily seen from Fig. 12.24 that 4nu*=T d; and Now
uct
(12.59) (r2.60)
P*= Pr* sin 6o
Pm
A t =
J h
(P^ 0) d 6 = P^ (4, 
6)
and
6^^
6 e l f . "
(3phasefault)
61
Az=
P"=o ./
i
Clearing angle
6,,
J (P**
sin d P^) d6
= P.u* (cos d,  cos d*)  P* (6^o  6"i) Fig. 12.23
ffil
uoo"rnpo*"r system Anatvsis cos{. = !^
,.774
,, /__\ tt \l
W
(r2.61)
Pm
For the systemto be stable, A2= A1,which yields
(5,^^" d) + cos 4o"* \ tniu( Prn*
)
l_Lr__l

I
L r l
l
I Infinite
bus
IVVO"
4, = critical clearing angle Substituting Eqs. (1259) and (12.60) in Eq. (12.61), we get 4r = cost [(r, _ Z6l sin do_ cos 6o] During the period the fault is persisting,the swing equation is
where
(a)
(r2.62)
IVVO"
= rf P^: P, = o d,r, 1r:
Integrating twice
d,2 d
(12.63)
6 = ,rf P*tz + $ 2H
(b)
0 " , = # P ; 2 " ,1 6 0
where /cr = critical clearing time 4, = critical clearing angle From Eq. (12.6a)
Fi1.12.25 Singlemachine tiedto infinite bus through par:allel two lines (12.64) Both thesecurves are plotted in Fig. 12.26,wherein Pu*n ( P_u*ras (Yo * Xr) > (Ya + Xr ll X).The systemis operatinginitially with a steadypower transfer Pr= P^ at a torque angle 4 on curve I. Immediately on switching off line 2, the electrical operating point shifts to curve II (point b). Acceleratingenergycorresponding areaA, is put into rotor to followed by decelerating energy for 6 > q. Assuming that an area A2 correspondingfo deceleratingenergy (energy out of rotor) can be found such that At = Az, the system will be stable and will finally operate at c correspondingto a new,rotor angle 6, > 60" This is so becausea single line offers larger reactanceand larger rotor angle is neededto transfer the same steadypower.
," / (both lines in)
2H(6, 4) ,
TrfP*
(r2.6s)
where d, is given by the expressionof Eq, (12.62) An explicitrelationship clctenninirtg i, porisiblc I'rrr r., in this caseas tluring the faulted condition p" = o and so trre ,wing equation can be integrated in closedform. This will not be the casein mosi other situations. Sudden Loss of One of parallel Lines considernow a singlemachinetied to infinitebus throughtwo parallellinesas in Fig. 12.25a. circuit model of the sysremis given in Fig. r2.25b. Let us study the transientstability of the ,yir"rn when one of the lines is suddenly switched off with the system operating at a steady road. Before switchingoff, poweranglecurve is given by sin d= Pm*l sin d xa i xt llx2 Immediately on switching off line 2, power angrecurve is given by P"r= lE'llvl
P"n= g:+
,\d T rt7
sin d= pmaxr d sin
Fig.12.26
Equalareacriterion applied the opening one of the two to of linesin parallel
ffiffi4l
po*rr.surt* Mod"rn
nrryr',
Power Syt St"blllry via the healthy line (through higher line reactance X2 in place of Xl ll Xz)7; with power angle curve
sln
sin d
4=4o*_T_6, which is the samecondition as in the previous example. Sudden Short Circuit on One of parallel Case a: Short circuit at one end of line Lines
Let us now assume disturbanceto be a short circuit at the generatorend of the line 2 of a double circuit line as shown in Fig. 12.27a.We shall assumethe fault to be a threephaseone.
obviously, Po[ ( P"*r. The rotor now starts to decelerate as shown in Fig. 12.28. The system will be stable if a decelerating areaA, can be found equal to accelerating area A, before d reachesthe maximum allowable value 4o*.At areaA, dependsupon clearing time /. (correspondingto clearing angle {), clearing time must be less than a certain value (critical clearing time) for the system to be stable. It is to be observedthat the equal area criterion helps to determine critical clearing angle and not critical clearing time. Critical clearing time can be obtained by numerical solution of the swing equation (discussedin Section 12.8).
P"y, prefault (2 lines)
postfault (1 line) P6n1,
X2 , (b)
? t 6
F19.12.27 Shoftcircuit one end of the line at Before the occurrenceof a fault, the power angle curve is given by
Ffg. 12.28 Equalarea criterion applied the systemof Fig. 12.24a, to I systemnormal,ll faultapplied, faultedline isolated. lll It also easily follows that larger initial loading (P.) increases for a given A, clearing angle (and time) and thereforequicker fault clearing would be needed to maintain stable operation. Case b: Short circuit away from line ends
'
p"t= d ' xi xltx2 sind= p_*, sin ,)4,'rlr,,,a, +
which is plotted in Fig. 12.25. Upon occulrenceof a threephase fault at the generatorend of line 2 (see Fig. I2.24a), the generator getsisolatedfrom the power systemfor purposes of power flow as shown by Fig. 12.27b.Thus during the period the fauit lasts, The rotor therefore accelerate, .i;t:;i"s dincreases. synchronism will be lost unless the fault is cleared in time. The circuit breakers at the two ends of the faulted line open at time tc (corresponding angle 4), the clearing time, disconnectingthe faulted line. to
When the fault occurs away from line ends(say in the middle of a line), there is some power flow during the fault thoughconsiderablyreduced,as different from case a where Pen= 0. Circuit model of the system during fault is now shown in Fig. 12.29a.This circuit reducesto that of Fig. 12.29cthrough one deltastar and one stardelta conversion. Instead,node elimination technique of Section 12.3 could be employed profitably. The power angle curve during fault is therefore given by P"t=  Ellvl
'1r'II
s i n d = P m a xs irn d r
ffiil
+t*F;4l
Modern Power svstem Analvsis
'!'vev" ' ' 'v' v, '" ' ' " '', '
ls#z#i
system operationis shown in Fig. 12.30,wherein it is possibleto find an area A, equalto A, for q. < 4nu*. At the clearing angle d. is increased, area ai increat"t und to nna Az = Ar, 4. increasestill it has a value 4n*' t6" ooi*,, ollnrvohle stahilitv This caseof critical clearineangleis shown fnr in Fig. 12.3L
Pe
I
x,t
X6
Pr1,prefault (2 lines)
postfault (1 line) P6111, G
xc
@
(b)
Xr
Fig. 12.31 Faulton middleof one lineof the systemof Fig. t2.l4a, case of criticalclearingangle
(c) Fig. 12.29 P"rand P,u as in Fig. 12.28 and Per as obtained above are all plotted in Fig. I2.3O. Accelerating area A, corresponding to a given clearing angle d is less
Pe
Annlvins eoual areaeriterion to the caseof critical clearing angle of Fig. 12.31' we can wnte
4, dntn'
j (P^ 4n*u sinfldd= J {r^*r sind P^)d6
6,,
60
where
P"'Prefault (2 lines) ,.a postfault line) (1 P"11;,
 sinr (:t_) 4,,* =T
V maxIII
(r2.66)
./
Integrating, we get l6* cos (P^a + Pmaxrr d)  * (P'*,,1 cos d +
16o
= Q
P"11, duringfault
or
P^ (6",  6) * P.u*u (cos '[.  cos do) I P* (6**  6"r) * Pom (cos fi*  cos 4J = 0 Fig. 12.30 Faulton middleof one line of the systemof Fig. 12.24a with d"< {,
t
cos {r =
4naxtn
 PmaxII
:otd",*
(12.67)
critical clearing anglecanbe calculated from Eq.(12.67) above. The angles in are in radians. equation The mooiiiesas belowif the angJes are :lt::t1on ln oegrees.
cos {. 
ft
r.(6 ^i*  do)  Pmaxrr do * prnu*ru d,ou* cos cos
Pmaxltr  Prnaxn
Give the system of Fig. 12,33 where a threephascfault is applied at rhe point P as shown. i0.s
Infinite bus vFlloo
Case c: Reclosure If the circuit breakersof line 2 arereclosed successfully (i.e., the fault was a transient one and therefore vanishedon clearing the faurty line), the power transfer once again becomes P"N = P"r= p*u*I sin d Since reclosure restores power transfer, the chances of stable operati'n improve. A case of stable operationis indicated by Fig. 12.32. For critical clearing angle
Flg. 12.33 Find the critical clearing angle for clearing the fault with simultaneousopening of the breakers I and 2. T\e reactance values of'various components are indicated on the diagram. The generatoris delivering 1.0 pu power at the instanr preceding the fault. Solution With referenceto Fig. 12.31, three separate power angle curves are involved. f. Normal operation (prefault)
4 = 4r* = 1T sinl 1p_/p*.*r;
t ucr 6rc
60
sin J @r, Pmaxrr 0 dd = J (p.*m sin d_ pm) d,6
6r, dru, t
+ J (P,*r sin d_ p^) d6
6.
Xr=0.2s+ffi+0.05
= 0.522 pu
p,t=rysind:
= 2.3 sin d
ffirino
(il
Prefault operatingpower angle is given by 1.0 = 2.3 sin 6 or IL During 6o =25.8" = 0.45 radians fault
It is clear from Fig. 12.31 that no power is ffansferredduring fault, i.e., = o
(Clearing angle)
(Angle reclosure of )
\
,0.:"o
Fig 1232 Faurtin middre a rineof of the systemof Fig. 12.27a where trrj tr, + r; T = time betw,een clearing anclreclosure.
trin 1n 1A
iffi
powerSystem Modern Anatysis =  1.5 (cos2.41 cos 6,)  (2.41 6") = 1.5cos 6", + 6r, I.293
Setting A = Az and solving
U
E
rrr. Post fault operation (fault cleared by openingr the faulted Iiae)
6r,  0.45 = 1.5cos 6r,+ 6r, 1.293
l.2xl.0 n Perrr=
ff
sin d= 1.5sin 6
(iii)
or
cos {,. = 0.84311.5 0.562
Pe
or 4, = 55.8" poweranglediagrams shownin Fig. 12.35. are The corresponding
Pn=1 'O
Find the critical clearing angle for the system shown in Fig. 12.36 for a threephase fault at the point P. The generator is delivering 1.0 pu power under prefault conditions. 5 i0.1 i 0. 1s
lnfinite bus lvF1.otoo
66=0.45 rad
6^rr=2.41rdd Fig. 12.35
Pu lF,l=1.2
.10.15
afea ^ Al = . A2 (Sge flg.
jo.15 Flg. 12.36
TTto rrrw
ooi*"* urour.rLurr
^*:^^:Ll^ psllluDDlulc
^l^ alilBrtr
C Omax
f^l()f
given by I = 2.4Lradians 1.5 Applying equal area criterion for critical clearing angle { Ar = P^ (6",  6) = 1.0 (6",  0.45) = 6c, 0.45 4ou*=rsinl
dr*
12.35) is
Solution f, Prefault bus is operation
Transfer reactance between generator and infinite
& = 0.25+ 0.17+
0.15+0.28+0.15 = 0.71 2 (i)
Az= !{r,np^)d,6
6,, 2.41 I 6., r2.41
Pc' = r ' Z x l s i n d = 1 . 6 9s i n 6 , 0. 71 power angle is given by The operating
sin 1.0= 1.69 ,fr
or do = 0.633 rad The positive sequencereactancediagram during fault is IL Durtng fault presentedin Fig. 12.37a.
=  (1.5 in6 1) dd s J
=  1 . 5 c od _ d l s
 6",
Power System Stabilitv
J 000 L
Mi#ffi
r*
j0.25 0 '
+
/ 000 Lr \000
Perrr=U!l '
jo.17
sin d = r'2 sin 6
(iii)
j0.15
j0.14
j0.14
j0.'15 V=1.0
With referenceto Fig. 12.30and Eq. (12.66),we have
+
) E1=t.z
To find the cqitical clearing angle, areas A1 and A, arc to be equated.
6",
(a) Positive sequence reactance diagramduringfault j0.25 j0.145 j0.145 ', j0.17
At = l.o (6,, 0.633) ,
and
J
o.+e5 d dd sin
60
lE'l=1.2
V=1.OlOo
Az =
Now
dmax f
J 6cr 
 1 . 2s i n d d d  1 . 0( 2 . 1 5 5 4 )
At =Az
(b) Network afterddltastar conversion
or
6r, = 0.633l9l=1'z
V=1.0100 2.155
o.63 J
J
0.495 d dd sin
=
(c)Network stardelta after conversion . Ftg. 12.32
J 6cr
[ t.Z rin 6 d6  2.t55 + 6,, '
or  0.633 0.495 olo' =  1.2cos ol"tt 2.155 + cos
lo.orr la., or  0.633 + 0.495 cos 6,,  0.399 = 0.661 + 1.2 cos 6",  2.155 or cos 6r, = 0.655 U 6r, = 49.I"
Converting delta to star*, the reactancenetwork is changed to that of Fig. 12.37(b). Further, upon converting star to delta, we obtain the reactance network of Fig. .r2.37(c). The transfer reactanceis given 6y
(0.2s 0.145) 0.072s (0.145 0.17) + + 0.0725 (0.25 0.145) + + + (0.14s 0.17) + Xu=
0.075
_ 2.424 p r eI = fir. Postfault lal!
i.+Z+ uur v  v.a/, sin d = 0.495 sin
6
off)
(ii)
operation
(faulty
line switched
A generatoroperating at 50 Hz delivers 1 pu power to an infinite bus through a transmission circuit in which resistance is ignored. A fault takes place reducingthe maximum power transferable 0.5 pu whereasbefore the fault, to this power was 2.0 pu and after the clearance the fault, it is 1.5 pu. By the of use of equal area criterion, determine the critical clearing angle. Solution All the three power angle curves are shown in Fig. 12.30.
X r l = 0 . 2 5 + 0 . 1 5+ 0 . 2 8 + 0 . 1 5 + 0 . 1 7= 1 . 0
*Node elirnination techniquewould be'used for complex network.
,'ffi
Ilere
Mod"rn Po*.. sEl!"nAn"lytit
= P"*r =2.0 pu, Pmaxl= 0.5 pu and Pmaxrrr 1.5 pu Initial loading P^ = 1.0 pu
2.The angular rotor velocity u= d6ldt (over and above synchronous velocity
( p 6r,ro= I zr sin tffiJ
\
E7sinl
1 :2.4!rad 1.5
r>2
U
n1 solution Discrete
n Continuous solution
t Af
Applying Eq. (r2.67)
 0.523)0.5 0.523 1.5 2.41 cos 1.0(2.41 + cos = o ??? cos {, 1.5 0.5
6r, = 70'3" T2.9 NUMERICAT SOTUTION OF SWING EOUATION
un/2
t
unI/T+tsn4l2
u13/2
In most practical systems,after machine lumping has been done, there are still more than two machines to be considered from the point of view of system stability. Therefore, there is no choice but to solve thp swing equation of each machine by a numerical techniqueon the digital computer. Even in the case of a single machine tied to infinite bus bar, the critical clearing time cannot be obtained from equal area criterion and we have to make this calculation
numerlca[y
! . .rrrr
n2
p3l2
n'l
r>112
n
t Af
mrougn
l
swulg
:

equauulr.
Ll^,
zFL^^
 ^Ll^+i^^+^l *^+L^lt rttrIc aIU ssvtrIilr JuPurDtruilL('(l lllELlluLlD
$ni
now available for the solution of the swing equation including the powerful RungeKutta method. He.rewe shall treat the pointbypoint method of solution which is a conventional, approximatemethod like all numerical methods but a well tried and proven one. We shall illustrate the pointbypoint method for one machine tied to infinite bus bar. The procedure is, however, general and can be appliedto every machine of a multimachine system. Consider the swing equation d26 1 = ;T ;e*P^*sind): PolM;
6nz
Jn2 Fig. 12.38 n1 n Pointbypoint solution of swing equation
Af
In Fig. L2.38, the numbering on tl\t axis pertains to the end of intervals. At the end of the (n l)th interval, the accelerationpower is Pa (n_r)Pm P* sh 4r Q2.68) where d_1 has been previously calculated.The changein velocit! (a= d6ldt), causedby the Pa@r),assumedconstant over At from (n312) to (nll2) is w n '2wn 3t 2=( Lt / M ) Pa@ r )
(*  9H orinpu system = +) M
\ 7t
iTf)
(12.6e) (12.70a) (12.70b)
The solution c(r) is obtained at discrete intervals of time with interval spread of At uniform throughout. Accelerating power and changein speed which are continuous functions of time are discretrzedas below: 1. The acceleratingpower Po computed at the beginning of an interval is assumed to remain constant from the middle of the preceding interval to the middle of the interval being consideredas shown in Fig. t2.38.
The change in d during the (nl)th interval is L6rt= 6r1 6n2= A'tun4'2 and during the nth interval L6r 6n 6n t = / \ t un 112
,ir Yirtvt r t Ar lr.rl b r )
power SystemStabilitv [i{8il;r
\rvtu.,v,t
Q^t,,1;^^
SubtractingEq. (12.70a\ from Eq. (12.70b) and using Eq. (12.69), we get
nsluls
E
^f^^
ws
Lall aPPt.y ultt
^*1,,
+L^
^+^
stEPUyslttP
L..
^+^
^rl.^l
lIIculUU,
Wtr lltrC(l t() Calculate
 ^ t
^
 r
ra
L'6,= A6,t + Using this, we can write
(A r)2 D r
M
a(.nI)
(12.7r)
the inertia constant M and the power angle equations under prefault and postfault conditions. Base MVA = 20 IneRia coflstant, Mepu\ =
G2.72) The processof computation is now repeatedto obtain Pa61, L6r*tand d*t. The time solution in discrete form is thus carried out over the desiredlength of time, normally 0.5 s. Continuous form of solution is obtained by drawing a ;mooth curve through discrete values as shown in Fig. 12.38. Greater accuracy of solution can be achi.evedby reducing the time duration of intervals. The occurrence or removal of a fault or initiation of any switching event power Po.lf such a discontinuityoccurs causesa discontinuity in accelerating at the beginning of an interval, then the averageof the values of Po before and after the discontinuity must be used.Thus, in computing the incrementof angle occurring durirrg the first interval after a fault is applied at t = 0, Eq. (I2.7I) becomes 7,,6, ,= (Ar)t o M * P a2 +
6n= 6nt + L,6n
180 /
x 1.0 L52 x 180 50
= 2.8 x 10+ s2le\ect degree I I Prefault &' = 0 . 3 5 + 0'2 =0.45 2
Pd= Pr.*r sin d !,.lxt r = '.;;sin . d = 2.M sin 5 Prefault power transf'er = + = 0.9 pu 20 Initial power angle is given by 2.44sin4=0.9 or 6o= 21.64" \ (i)
where Pos* is the acceleratingpower immediately after occurrence of fault. Immediately before the fault the system is in steady state, so that Poo= 0 and ds is a known value. If the fault is cleared at the beginning of the nth interval, in calculation for this interval one should use for Pa@r)the value llP"6r>is + Po6_9*), where Pa@_r) the acceleratingpower immediately before clearing and Po6_r)+is that immediately after clearing the fault. If the discontinuity occurs at ihe miciciie of an intervai, no speciai proceciure is neecled.The increment of angle during such an interval is calculated, as usual, from the value of Po at the beginning of the interval. The procedure of calculating solution of swing equation is illustrated in the following example.
II During fault A positive sequence reactancediagram is shown in Fig. 12.39a.Converting star to delta, we obtain the network of Fig. 12.39b, in which 0. 35x 0. i + A. 2x0. i + 0. 35x0. 2 ,, = trtr = 0l
P.u = Pmaxtt sin d
1 AI..Z) pu
 1'1x1 r;n d = 0.88sin 6 1.25
(ii)
A 20 MVA, 50 Hz generator delivers 18 MW over a double circuit line to an infinite bus. The generatorhas kinetic energy of 2.52 MJA4VA at rated speed. The generator transient reactanceis X/o = 0.35 pu. Each transmission circuit lE/l = 1.1 pu and has R = 0 and a reactance of 0.2 pu on a 20 MVA bgqe. infinite bus voltage V = 7.0 10". A threephaseshort circuit occurs at the mid point of one of the transmissionlines. Plot swing curves with fault cleared by simrrltaneous opening of breakersat both ends of the line at2.5 cycles and 6.25 cycles after the occuffence of fault. Also plot the swing curve over the period of 0.5 s if the fault is sustained.
Fig. 12. 39
lil. Postfault
With the faulted line switched off,
055 J;=:::,J;2
which it is obvious that the systernis unstable'
: :t  . I = 1: . 1' x'1s i n d = 2 . 0 s i n d
Let us choose Al = 0.05 s The recursive relationshi'ps for stepbystep swing curve calculation are reproducedbelow. Pa(n_r)=P^  P** sin 4_r (iv) L6n= L6nt* (Lt)z
M 6n= 6nt + A,6n
t
o E
@
;100
o 8 0 o)
c (5 o
' 'oa(nl)
(v) (vi)
Sincethereis a discontinuity P, andhence Po, the average in in valueof po mustbe usedfor the first interval. P"(0)= 0 pu and Po (0*) = 0.9  0.88 sin 2I.64" = 0.576pu = 9t#ZQ Po(ouu.,us"l
L
ioo P
= 0.288pu
I o t l 0.1
fault cleared at 2.5 cycles
l
Sustained Fault Calculations are carried out in Table 12.2 in accordance with the recursive relationship (iv), (v) and (vi) above. The secondcolumn of the table showsP* the maximum power that can be transferredat time r given in the first column. Pn * in the case of a sustainedfault undergoesa sudden change at t = 0* and remains constant thereafter.The procedure of calculations is illustrated below by calculating the row correspondingto t = 0.15 s. = (0'l sec) 31.59" P."* = 0.88 sin d (0.1 s) = 0.524 P, (0.1 s) = P,,,u* 6 (0.1 s) = 0.88 x 0.524  0.46I sin P, (0.1 s) = 0.9  0.46I  0.439
( At\2
t l l 0.3 0.2 f (s) 
l i 0.4
i 0.5
_ r l
0.6
fault and for for for Fig. 12.40 Swingcuryes Example12.10 a sustained in clearing 2.5 and 6.25 cYcles
aaDle !A A aZ.Z n^i. l., ^aia* fUllll'UyPulllt rratinnc t/vlllPutqrrvrro n{ vr c r r r r nV er' rir n n ur' n / a vr fnr qtrctainarl fattlt
/ f = 0. 05s
t sec
P^u pu
sin 6
P"=Prrr,*sin6 P,,= 0'9 P, Pu Pu
a
deg
6
6
deg
YP, M
(0.1 s) = 8.929x 0.439 3.92 6 (0.1ss) = Ad (0.1s) + qL
M
P, (0.1s)
U \
= J.38"+ 3.92"= 11.33" d ( 0 . 1 5s )= d ( 0 . 1s ) + A d ( 0 . 1 5s ) = 31.59" 11.30' 42.89" = +
0+ o^u, 0.05 0.10 0 .t 5 0.20 0.25 0.30 0.35 0.40 0.45 0.50
2.44 0. 88 0. 88 0.88 0.88 0. 88 0. 88 0. 88 0.88 0.88 0.88 0.88
0. 368 0. 368 0. 368 0. 41 0.524 0.680 0. 837 0.953 0.999 0.968 0. 856 0.657
0324 0.361 0.46r 0. 598 0. 736 0. 838 0. 879 0. 852 0.154 0.578
0. 0 0. 576 0. 288 0. 539 0.439 0. 301 0.163 0. 06 0.021 0.048 0.145 0.32r
2.57 4.8r 3. 92 2.68 r . 45 0. 55 0.18 0.426 1.30 2. 87
2.57 7.38 11.30 13.98 15.43 15.98
2r.64 2r . 64 74.21 31.59 42.89 56.87 72. 30 r 6 . r 6 88. 28 1 6 . 5 8 r04.44
r 7 . 8 8 r2r.02 20.75 138. 90
1s9.65
I
f
F . a t t l t su.a
Flaae^'J vtcrete.t :tlt
Modern po@is
o E ntA,i, fryCIeS
ltr
progressively greater clearing time till the torque angle d increases without bound. In this example, however, we can first find the critical clearing angle using Eq. (12.67) and then read the critical clearing time from the swing curve corresponding to the sustainedfault case. The values obtained are: Critical clearing angle = 118.62 Critical clearing time = 0.38 s Table 12.4 Computations swing curve for fault clearedat of , 6. 25cycles( 0. 125s)Af = 0. 05s
sin 6 P"=P^ *sin6 Po= 0.9 P"
Time to clear f'ault= 2.5 = 0.05 s 50 Pu^ suddenly
to 2.0 at t = 0.05. Since the
be assumedto remain constant from 0.025 s to 0.075 s. The rest of the prtrcedureis the sameanclcomplete calculations are shown in Table 12.3.The swing curve is plotted in Fig. 12.40 from which we find that the generator undergoesa maximum swing of 37.5" but is stable as c5finaily begins to decrease. Table 12'3 Computations swing curyesfor fauttcleared of at2.s cvcles (0 .0 5s ), At = 0.05s
P,,,,,^ pu .sin.5 Pr,=P,rr.,*,tin5 pr,= 0.9 pu pu pu
P,no
4 deg
6
6 deg
pu A6 deg 6 deg 21.64 21.64 21.64 24.21 24.21 24.21 29.54 34.10 36.70 37.72 34.16 29.64 24.33 19.73 17.13 0 0+ ouu, 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
A
2.44 0.88 0. 88 0.88 2.O0 2.00 2.00 2. 00 2. 00 2.00
Z.W
' \ N
0 2.44 0. 0 .8 8 ouu, 0. 05 0 . 8 8 0. 05+ 2 .0 0 0.05uus 0 . 1 0 2.00 0 . i 5 Z .U U o. 20 2 .0 0 0.25 2 .0 0 0 . 3 0 2 .0 0 0 . 3 5 2 .0 0 o. 40 2 .0 0 0.45 2 .C 0 0. 50
T
0.368 0 .3 6 8 0 .3 6 8 0.41 0 .4 1 0.493 0 .5 6 0.s91 0.597 0.561 0.494 0.41 0.337
0 .9 0.324 0 .3 6 0 .8 2 0 .9 8 6 I.t2 T.I9 r.t9 I.t2 0 .9 8 9 0.82 0.615
0.0 0 .576 0.288 0 .54 0 .08 0.31  0.086  0.22  0.29  0.29  0 .22  0 .089 0 .08 0 .225
2.57
2.57

2.767 0.767 1.96 2.s8 2.58 1.96 0.79 0.71 2.0
5.33 4.56 2.60 0.02 2.56 4.52 5.31 4.60 2.6
0. 368 0. 368 0. 368 0.41 0.524 0.680 0.767 0. 78 0. 734 0.613 0.430
u. z11
0.9 0.324 0.36r o.46t 1.36 1.53 1.56 1.46 r.22 0. 86
u.4c)0

0.0 0.576 0. 288 0.539 0.439 4.46 0.63 0.66 0. 56 0.327 0.04
u.4J4
.
2.57 2.57 4.81 7. 38 3. 92 11.30 4.10 7.20 5.66 r . 54 5.89  4.35 5. 08  9.43 2.92  12.35 0. 35  12.00  u.t  J r.6 /
2r.&
21.64 zt.U 24. 2r 3t.59 42.89 50. 09 51.63 47.28 37.85 25.50 l_J. ) u 5.37

u.4J
A ?
0.50 2.oo
T2.TO MULTIMACHINE STABITITY
_ _ 
Fault Cleared in 6.25 Cycles Time to clearfault= ua?t = 0.125 s s0
From what has been discussedso far, the following steps easily follow for determiningmultimachinestability. 1. From the prefault load flow data determine E/ovoltage behind transient . reactancefor all generators.This establishesgeneratoremf magnitudes lEll which remain constant during the study and initial rotor angle 6f = lEt. Also record prime mover inputs to generators,P*o  PoGk 2. Augment the load flow network by the generator transient reactances. Shift network busesbehind the transient reactances. 3. End Inus for various network conditions*during fault, post fault (faulted line cleared), after line reclosure. 4. For faulted mode, find generator outputs from power angle equations (generalizedforms of Eq. (12.27)) and solve swing equationsstep,by step (pointbypoint method).
4ffi 
Modern power SystemAnalysis
5. Keep repeating the above step for post fault mode and after line reclosuremode. 6. Examine d(r) plots of all generators and establish the answer to the stability question. The above stgps are illustrated in the following example.
l,,m
220kV, 100M VAbase
Bus to bus Series Z HaIf line charging
A 50 Hz, 220 kV transmissionline has two generatorsand an infinite bus as shown in Fig. 12.4I. The transformer and line data are given in Table I2.5. A threephasefault occurs as shown. The prefault load flow solution is presented in Table 12.6. Find the swing equation fbr each generatorcluring the fault period.
Line 45 Line 51 Line 4I Trans;24 Trans:35
0.018 0.004 0.007
0.1r 0.0235 0.04 o.022 0.04
0.113 0.098 0. 041
valuesin pu on 220 kV, loadflow Tabfe 12.6 Bus data and prefault 100M VAbase
S.No. and Bus No. Voltage Polar Form
Bus Upe
Real e
Voltage Imaginary f
Generation
Load
o
Vz=1.0328.2350
o
i
) v i F1.O217.16o
( )vs=1.ollz [5,
lltI'
Se ).5+j(
22.490
Slack  1.010" PV 2 t.0318.35" 3 t.0217.16 PV 4 1.017414.32"PQ 5 1 . 0 1 1 2 1 2 . 6 9 "P Q
1.00 t.0194 1.0121 1.0146 1.0102
 3. 8083 0.2199 0 0 0. 0 0 0.6986 0 3. 25 0. 1475 0 0.3110 0 0.1271 2.10 1. 0 1. 0 0. 44 0 0.167 0 0.5 0.16 0 0.0439
6) i6,u
Solution Before determining swing equations, we have to find transient internal voltages. The current into the network at bus 2 basecion the <iatain Tabie i2.6 is
'rz _
PriQ, _ 3.25i0.6986 r.o3l 8.23519" v: 3.2sj0.6986 x 0.06719V r.03l8.23519'
+ + E{= (1.0194 j0.1475)
o
Fig. 12.41
Data are given below for the two generatorson a 100 MVA base. Gen 1 500 MVA, 25 kV, XJ = 0.067 pu, H = 12 MJAyIVA Gen 2 300 MVA, 20 kV, X,j = 0.10 pu, H = 9 MJA4VA Plot the swing curves for the machines at buses 2 and 3 for the above fault which is cleared by simultaneous opening of the circuit breakersat the ends of the faulted line at (i) 0.275 s and (ii) 0.0g s.
= 1.0340929 j0.3632368 + = = 1.0960333 119.354398" 1.0960lo.337l tad I El = I.0 l0' (slackbus) r 2 . 1 j 0 . 3 1 x 0.1 l9O" + E4  Q.0I2r + j0.1271) r.021 7.15811'
= = 1.0166979 j0.335177 1.0705 lI8'2459" +  1.071 10.31845 rad by The loads at buses 4 and 5 are represented the admittancescalculated as follows:
yrr.= ''9 .!.:,0!(0.9661jo.4zsr) 6 (1.0174)"
PowerSystem Stability v 0 . 5_ j 0 . 1 6 During Fault Bus Matrix
I a,grr, _ .
Load admittances, al
prefaulr *Trl "r"
rr.off
(0'488e i0'1s647)
therefore' now :::,::::^'l:,::,1,:nt il::H,:T:, designateas buses2 and3, the fictitious internal nodes between the internal voltages and the transient reactancesof the machines.Thus we get
reactanles the machines wlll, of we wlr, we
;;'#;;"J###::
Since the fault is near bus 4, it must be short circuited to ground. The Ynus during the fault conditions would, therefore,be obtained by deleting 4th row prefault Y".r. rnatrix. Reduced fault and 4th column from the above augmented matrix (to the generatorinternal nodes)is obtained by eliminating the new 4th row and column (node 5) using the relationship
Y*iqn"*1= Y*j@tat Yt n(oltt) ynj(old Yrn )/ @ta)
2 = Y^2  @ =  i r r . 2 3 6
Yzq= j11.236= yqz
Ytt =
The reducedfaulted matrix ()'eus during fault) (3 x 3) is given in Table I2.8, from the other busesduring the fault which clearly depicts that bus 2 decouples and that bus 3 is directly connected to bus 1, showing that the fault at bus 4 reducesto zero the power pumped into the system from the generatorat bus 2 and renders the second generator at bus 3 to eive its power radially to bus 1. Table 12.8 Elementsof Yrus (duringfault) and Ysus(post fault) for Ex. 12.11,admittances pu. in
Reduced during fault Yu^
io'04+io]
Yzs= j7.I43 = yst
=  i7'143
Y u = Y t q + y q t * y q s + Bo,
)
Bo,' ) *
L
Yzq
Bus
I 2
a J
= 0.9660877 j0.4250785 + 4.245 j24.2571+ 1.4488 _ j i 8 . 8 5 3 8 +0 . 0 4 1 r O . 1 1 3 _ j r t . 2 3 5 g + _ 6.6598977 j44.6179 _ Ur, Y s s =Y r s * Y s q * Y s r * 9* *' r3 , ' "5 2 2  0.4889 j0.1565 r.4488 j8.8538 + _ + 7.039r j41.335 + /O.1 + j0.098_ j7.t42} 13 _ 8.97695s js7.297202 _ The complete augmented prefaultlzuu,matrix is show n.iri'Table 12.1. Table12.T Theaugmented prefaurt admittance bus matrix Ex. 12.11, for admittancespu in
Bus
s.7986j35.6301 0  0.068175.1661 +
0  jrr.236 0
 0.2214 j7.6289 + 0.s j7.7898 0
 0.0681 j5.166l + 0 j6.2737 0.1362'
Reduced post Jault Yurt
I 2 J
r.3932 jr3.873r  0.2214 j7.6289 +  0.0901 j6.0975 +
 0.0901+ j6.O975 0 0.159 j6.1168 1
Post Fault Bus Matrix Once the fault is cleared by removing the line, simultaneouslyopening the circuit breakersat the either ends of the line betweenbuses4 and 5, the prefault Y"u5 has to be modified again. This is done by substituting Yqs= Ysq 0 and of subtracting the seriesadmittance of line 45 and the capacitive susceptance half the line from elements Yooand Ytt.  Yqs 84512 = Y++lporrfault)Y++(prefault) = 6. 65989 j44. 6179 1. 448+ 78. 853 j0. 113 = 5.2III  j35.8771 = Similarly, Ysr(oo*, rault) 7.528I i48.5563 The reducedpost fault Y"u5 is shown in the lower half to Table 12.8. It may be noted that 0 element appears in 2nd and 3rd rorvs. This shows that,
rr.284_j6s.473
2 3 4 0 _i1r.23sg 0 6 4.245+ j24.257 jll .2359 7.039+ j41.355 0
0 0  j7.1428 0
4.245 + j24.257 7.039 + j4r.35s j11.23s9 0
0 j7.t428 6.6598_j44.617 1.4488 +j8.8538 0 + j7.1428 1.4488+ j8.8538 8.9769 + j57.2972
{92 
L..^:^^lr. rr 
Modern power System Anatysis
Stanitity PowerSystem
I 491,1
tlle lrrryr'ruarry' generarors I an0 Z are not Interconnected when line 45 is removed. During Fault Power Angle Equation Prz= 0
It rnay be noted that in the above swing equations,P,, ntay be written in general as follows: sin Pn= Pr,  Pr. P,r,n* (6 7) Solution of Swing Equation
P,3= Re [YrrEr,El* El* \F(]; sinceyy= 0 * = E{2 Gn + lEil lEil lrr,l cos (6zr_ Lzt) = ( 1 . 0 7 1 )( 0 .1 3 6 2 + I x 1 . 0 7 1 x 5 . 1 6 6 5 o s( d 3_ 9 0 . 7 5 5 " ) 2 ) c = 0 . 1 5 6 1 5 . 5 3 1 i n( h  0 . 7 5 5 ) P"3 + s Postfault Power Angle Equations p"z= lE/P G22+ lElt lEll ly2Ll (dr, _ 0zr) cos = 1.0962 0.5005+ I x 1.096x 7.6321 x cos ({  9I.662") = 0.6012+ 8.365sin (d, _ I.662) p,3 = tE{ 2q3 + Ell tElt t\l cos ( 6r,_ 0rr)  7.0712 0.1591 1 x 1.07rx 6.09gcos (dr  90.g466") x + = 0.1823+ 6.5282 (d, _ 0.9466") sin Swing EquationsDuring Fault
The above swing equations(during fault followed by post fault) can be solved by the poinrbypoint method presented earlier or by the Euler's method presented in the later part of this section. The plots of E and 4 are given in Fig. 12.42 for a clearing time of 0.215 s and in Fig. 12.43for a clearing time of 0.08 s. For the case (i), the machine 2 is unstable, while the machine 3 is stable but it oscillates wherein the oscillations are expected to decay if effect For the case(ii), both machinesare stablebut of damper winding is considered. the machine 2 has large angular swings.
(lnfinite bus) 1 Machine is reference
#=ff
(P^zP,z)=
= t:9/ e.2s  o) elecr deg/sz 12 d,q _ 180 f (PnP"t) a'r= k = t*:/ 9 = Y + e . r  { 0 . 1 s 6 1 5 . 5 3 1s i n ( 6 t  0 . 7 s s ) } l sin l.g43g  5.531 (d,  0.755")l electd,eg/sz
at) (0.275s faultcleared
Swing EquationsPostfault p . 2 s  1 0 . 6 0 r + 8 . 3 6 5 i n ( d )  t . 6 6 2 " ) }e r e c d e g / s 2 z s J t _ + 6.5282 sin(d3 0.g466.)}lelect deg/sz
12.1for 2 F\g.12.42 Swingcurvesfor machines and 3 of Example at clearing 0.275s. If the fault is a transient one and the line is reclosed,power angle and swing equations are neededfor the period after reclosure.Thesecan be computed from the reduced Ysus matrix after line reclosure.
#={f
dtd, 180 f * =*12.10{0.1823
dt' 9
494

Modern power System Analysis
Machine 1 is reference (lnfinite bus)
PowerSystemStabilitv
l,,4PFr (t2.74)
500
.
i * = # ( P " o r  P c o )k, = 1 , 2 ,. . . ,m r7.2
M a c h i n e2
Initial state vector (upon occurrence of fault) is xoLk=fr= lEot x " z k =0 The state form of swing equations (Eq. (12.74)) can be solved by the many available integration algorithms (modified Euler's method is a convenient choice).
c) o
q,
L
at)
o)
!
o (t c
100
Algorithm Computational Modified Euler's Method
i l l t l
for Obtaining
Swing Currzes Using
_L__ I _ 0.8 0.9
 _ 1.0
Faultcleared after4 cycles Fig. 12.43 Swing curves for machines 2 and 3 of
Example 12.11 clearing 0.0gs for at Gonsideration of Automatic voltage Regulator (AVR) and Speed Governor Loops
Carry out a load flow studY (prior to disturbance) using specified voltagesand powers. 2 . Compute voltage behind transient reactancesof generators(Eo*) using and initial rotor angle emf magnitudes Eq. (9.31).This fixes generator. bus voltag" Y?). lreference slack Compute, Ysu5 (during fault, post fault, line reclosed). J. l.
a
4. Set time count r = 0. 5 . Compute generatorpower outputs using appropriatgts"us with the help
of the geniral form of Eq. (12.27). This givet Pg},for / 1 /'). Note: After the occurrenceof the fault, the period is divided into uniform discretetime intervals(At) so that time is countedas /(0),t(l), ......A typical value ol' lt is 0.05 s. 6. Compute t(i[;'},i\7'),k 1 , 2 , . . . , m l f i o m E q s .( 1 2 . 7 4 ) . 'pr r liu' cst thc 7. Corttputc I'ilslstutc ir r t ir lc's t = , {t +l) ) , [ , 0 * r= , t l o + i [ , 0 a t ,
state variable Formulation of swing Equations The swingequation the hh generator for is
I = .2. .... rrt
At ,ff')  *V)+ *$'o)
(r2.73)
of the 8. Cortrpute first estimates E^('+t) = BQ+D E? lcos x,(i*r)+7 sin *\lf')) g. ComputeP8;'); (appropriate Y"u5and Eq. (12.72))' (12.74). k 10. compute[t;{in'),it:o*t'), = r,2, ..., mf fromEqs. valuesof statederivatives the 11. Compute average i[i,), urr= ][iu,cl +;l[*t)]
k=1,2,...,ffi
+ t' d
= + ( p ' o o  p " ) ;k  r , 2 ,. . .f,f i u Hk'
For the multimachine case,it is more convenientto organiseEq. (12.73) is state variableform. Define xrk= 6r= lE*' xz*= 6t Then i t*= xzt
= + l iL'),*,*I*\: ,tt'i"
Power System Stabitity
1't
LL.
Lompute tne lrnal state estimates for  = t\r+t)
A

491;,,
*;Iu" = *([)+ ill) uus, at ,&*r) ,([l + if).^,rat
k = 7'2'"'' frt
factors which affect transient stability and therefrom draw th" .on"llsions, regarding methods of improving the transientstability lirnit of a system and making it as close to the steady state limit as possible. For the case of one machineconnected to infinite bus, it is easily seen from the angle through which it swings in a given time interval offering thereby a method of improving stability but this cannot be employed in practice because of economic reasons and for the reason of slowing down the response of the speed governor loop (which can even become oscillatory) apart from an excessiverotor weight. With referenceto Fig. 12.30,it is easily seenthat for a given clearing angle, the accelerating area decreasesbut the decelerating area increases as the maximum power limit of the various power angle curves is raised, thereby adding to the transient stability limit of the system. The maximum steady power of a system can be increasedby raising the voltage profile of the system and by reducing the transfer reactance. These conclusionsalong with the various transient stability casesstudied,suggestthe following method of improving the transient stability limit of a power system. 1. Increaseof system voltages,use of AVR. 2. Use of high speedexcitation systems. 3. Reduction in systemtransfer reactance. 4. Use of high speedreclosing breakers (see Fig. 12.32).Mo&rn tendency is to employ singlepole operation of reclosing circuit breakers. When a fault takes place on a system, the voltagesat all buses are reduced. At generator terminals, these are sensed by the automatic voltage regulators which help restoregeneratorterminal voltagesby acting within the excitation system.Modern exciter systems having solid statecontrols quickly respondto bus voltage reduction and can achieve from onehalf to one and oneh'alfcycles (l/2l]) gain in critical clearingtimes fbr threephase taults on the HT bus of the generator transformer. Reducing transfer reactance is another important practical method of increasing stability limit. Incidentally this also raises system voltage profile. The reactance of a transmission line can be decreased(i) by reducing the conductor spacing,and (ii) by increasingconductordiameter (see Eq. (2.37)). Usually, however, the conductor spacing is controlled by other features such as lightning protection and minimum clearanceto prevent the arc from one phase moving to another phase.The conductor diametercan be increasedby using material of low conductivity or by hollow cores. However, norrnally, the conductor configuration is fixed by economic considerationsquite apart from stability. The use of bundled conductorsis, of course,an effective means of reducing series reactance. Compensation for line reactance by series capacitors is an effective and economical method of increasing stability limit specially for transmission
Compute the final estimate for Eo at t = r('*l) using BQ+t) = l4llcos xf;+r) + 7 sin *f1r) 14. Print (",9*t),*;:o*D k = I,2, ..., m ); 15. Test for time limit (time for which swing curve is to be plotted), i.e., ch e c k i f r> rn n u r. n o t, r r+ r and repeat If fromstep5 above. Otherwise print results and stop. The swing curves of all the machines are plotted. If the rotor angle of a machine (or a group of machines) with r"rp".t to other machines increases without bound, such a machine (or grouf of machines) is unstable and eventually falls out of step. The computational algorithm given above can be easily modified to include simulation of voltage regulator, field excitation response, saturation of flux paths and governor action Stability Study of Large Systems
To limit the computer memory and the time requirements and for the sake of computationalefficiency, a large multimachine system is divided into a study subsystemand an external system.The study subsystem is modelled in detail whereas approximate modelling is carried out for the external subsystem.The total study is renderedby ihe modern techniqueof dynamic equivalencing.In the externalsubsysteln, nutnber<lfrnachines drasticallyrecluced is using various methodscoherency based rnethodsbeing most popurar and widely used by v anous p o w e r u ti l i ti e si n th e w o rl d . I2.T7 SOME FACTORS AFFECTING TRANSIENT STABILITY
We have seenin this chapterthat the twomachine system can be equivalently reduced to a single machine connected to infinite bus bar. The qualitative conclusions regarding system stability drawn from a twomachine or an equivalent onemachineinfinite bus system can be easily extended to a multimachine system. In the last article we have studied ihe algorithm for determiningthe stability of a multimachrnesystem. It has been seen that transient stability is greatly affected by the type and location of a fault, so that a power systemanalyst must at the very outset of a stability study decide on these two factors. In ou, we have selected "*uples a 3phase fault which is generally more severe from point of view of power transfer' Given the type of fault and its location let us now consider other
'4lI
I
Modern power System Analysis
I I :.^Ir I.1.':/tlltt,
. i::
distancesof more than 350 km. The ciegreeof series compensation,however, accentuatesthe problems of protective relaying, normal voltage profiles, and overvoltagesdrrring linetogroundfaults. Seriescornpensation becomesmore effective and economical if part of it is of compensationupon the occurrencec Switehed series eapaeitorssimultaneot and raise the transient stability limit tc limit. Switching shunt capaciiors on ol stability limits (see Example 12.2) but the MVA rating of shunt capacitors required is three to six times the rating of switched series capacitors for the same increase in stability limit. Thus series capacitors are preferred unless shunt elements are required for olher pu{poses,siy, control of voltage profile. Increasing the number of parallel lines between transmission points is quite often used to reduce transfer reactance.It adds at the same time to reliability of the transmissionsystem. Aclditional line circuits are not likely to prove economical unit I aftet all feasible improvements have been carried out in the first two circuits. As the majority of faults are transientrn nature,rapid switching and isolation of unhealthy lines followed by reclosing has been shown earliei to be a great help in improving the stability marginr. ih. modern circuit breakertechnology has now made it possible fbr line clearing to be done as fast as in two cycles. Further, a great majority of transient faults a'e linetoground in nature. It is natural that methodshave beendevelopedfor selective single pole opening and reclosing which further aid the stability limits. With ,"f"r".,"" to Fi;. lz.r7, if a transient LG fault is assumedto occur on the generator bus, it is immedi ately ol,lnt the fault therewill now be a definite amount of power rransfer, ::.1,:l_., ab uurereni rrom zero power transfer for the case of a threephase fault. Also when the circuit breaker pole corresponding to the faulty line is opened, the other two lines (healthyones)remain intact so that considlrable power transfer continuesto take place via theselines in comparison to the caseof threepole switching when the power transferon fault clearing will be reducedto zero. It is, therelbre, easy to see why the single pole swiiching and reclosing aids in stability problem and is widely adopted.These facts arelilustrated by means of Example 12'12. Even when the stability margins are sufficient, single pole switching is adopted to prevent large swings and consequent vortage dips. Single pole switching and reclosingis, of course,expensiu. in t..*s of relaying and introducesthe as.socjatec! problernsof overvoltagescausedby single pole opening owing to line capacitances.Methods are available to nullify these capacitive coupling effects.
nd dr++^s^^L L
stzeof rotor reducesinertia constant, Iowering thereby the stability margin. The loss in stability margin is made up by such features as lower reactance lines, fastercircuit breakersand faster cxcitation systenrs tliscussetlalreacly,and as a faster system valving to be discussed later in this article. A stage has now been reached in technology whereby the methods of irnprovinE=stability; discussetl above, have been pushed to their limits, e.g., clearing times of circuit breakers have been brought down to virnrally irreducible values of the order of two cycles. With the trend to reduce machine inertias there is a constant need to determine availability, feasibility and applicability of new methodsfor maintaining and/or improving system stability. A brief account of some of the recent methods of maintaining stability is given below: HVDC Links Increased use of HVDC links ernploying thyristors would alleviate stability problems. A dc link is asynchronous,i.e., the two ac system at either end do not have to be controlled in phase or even be at exactly the same frequency as they do for an ac link, and the power transmitted can be readily controlled. There is no risk of a fault in one system causing loss of stability in the other system. Breakingr Resistors
For improving stability where clearing is delayed or a large tn)a i, suddenly lost, a resistive load called a breaking resistor is connected at or near the generatorbus. This load compensates at least sonreof the reduction of load for on the generatorsand so reducesthe acceleration. During a fault,the resistors are applied to the terminals of the generators through circuit breakers by means of an elaboratecontrol scheme.The control schemedeterminesthe amount of resistance be applied and its duration. The breakingresistorsremain on for to a matter of cycles b<lth during fault clearing and after system voltage is restored. Short Circuit Current Limiters
These are generally used to limit the short circuit duty of distribution lines. These may also be used in long transmission lines to modify favourably the transferimpedance during fault conditions so that the voltage profile of the systemis somewhatimproved, therebyraising the systemload level durins the fault. Turbine Fast Valuing or Bypass Valuing
Recent Trends Recent trends in design of large alternators tend towards lower short circuit ratio (scR = r/x), which is achieved by reducing machine air gap with consequentsavings in machine mmf, size, weight and cost. Reduction in the
The two methods just discussedabove are an attempt at replacing the sysrem load so as to increase the electrical output of the generator during fault conditions. Another recent method of improving the stability of a unit is to decrease the mechanical input power to the turbine. This can be accornplished
by rneansof fast valving. where the differenee between mechanicalinput and reduced electrical output of a generator under a fault, as sensedby a control scheme, initiates the closing of a turbine valve to reduce the power input. Briefly, during a fast valving operation, the interceptor valves are rapidly shut (in 0.1 to 0.2 sec) and immediately reopened. This procedure increasesthe critical switching time lons enoush stable for faults with stuckbreakerclearing times. The schemehas been put to use in some stations the USA. in FUII Load Rejection Technique
rJ
,t0.15
 l
0'3 0.1 idT L_r6TT\_
7
Fast valving combined with highspeedclearing time will suffice to maintain stability in most of the cases.However, there are still situations where stabilitv is difficult to maintain.In such cases, normal procedureis to automatically the trip the unit off the line. This, however, causesseveral hours of delay before the unit can be put back into operation.The loss of a major unit for this length of tirne can seriouslyjeopardize the remaining system. To remedy these situations, a full load rejection scheme could be utilized after the unit is separated from the system. To do this, the unit has to be equipped with a large steam bypass system. After the system has recovered from the shock caused by the fault, the unit could be resynchronized and reloaded. The main disadvantageof this method is the extra cost of a large bypasssystem.
0'1 p I L z rfd]uir
I  l
 I
(b) Negative sequence network
t ;
ao
'' .
l]o t
Ti
t
)
l go ''
I_

tI
p 1'o    I L.'  irT60 t;,'r";*u1n""
= C xo 0.0e15 l =
networl
Fig. 12.45 For an LG fault at P the sequence networks will be connectedin seriesas shown in Fig. 12.46. A stardeltatransformationreducesFig. 12.38 to that of Fig. 12.47 from which we have the transfer reactance
= Xr2(LG faulQ 0.4+ 0.4+ nOII''O = 1.45 0.246
The systetnshown in Fig. 12.44is loaded to I pu. Calculate the swing curve and ascertainsysternstability for: (i) LG fault threepole switchingfollowecl reclosure. hy linc lounclhealthy. (ii) LG fault singlepole switchingfbllowed by reclosure,line found healthy. Switching occurs at 3.75 cycles (0.075 sec) and reclosure occllrs at 16.25 cycles (0.325 sec). All values shown in the figure are in pu. H = 4.167
Xt=0'15
I
( lEl=1'2
.t.
) I rrd 60 0'4
I
P
o'4
l . r . l = ?,_ ,.
X r =0 . 1
'.
'L
Fig.12.46 Connection sequence of networks an LG fault for
= Xr o3Oj
xo= o.t
?*f ( / h'Ir!.&f!
AYl
Fig. 12.44
(,r
I I
fl /'  fXr n r r t f\
'lr
___J
t
Fig.12.47 Transferimpedance an LG fault for When the circuit breakerpoles correspondingto the faulted line are opened (it correspondsto a singleline open fault) the connectionof sequence networks is shciwn in Fig. 12.48. From the reduced network of Fig. 12.49 the rransfer reactance with faulted line switched off is
Solution The sequence networks of the systemare drawn and suitably reduced in Figs. I2.45a, b and c.
Y.^ ^ll
(fettlterl
\*
line
nnen\

0.  A t  v
Aa L A1 v.
r r
(l A _
V..+
1 ..
L../L
I .o.ro
Under healthy conditions transfer reactance is easily obtained from the positive sequencenetwork of Fig. 12.45 a as Xrr(line healthy)= 0.8
Pettt = 0
PrN= Pd = 1.5 sin d Now
46, A6n, @LP^,  l ,, + 'a(n )
M
o'1 P
Negative sequence
H = 4.167MJA{VA
 4.63 x 10a sec2lelectricaldegree 1,1= JU1 8 0x 5 0 Taking At = 0.05 sec
Zero sequence
Flg.1248 Connection sequencenetworks of with faultedline switchedoff
(at)'
lEl=1'z
P
l v l= 1 ' o
P/ o:4
4. 63xI 0 4
5.4
Time when single/threepole switching occurs = 0.075 sec (during middle of At) Time when reclosing occurs = 0.325 (during middle of at) .,
Fig. 12.49 Reduced networkof Fig. 12.49givingtransferreactance
Power angle eguations PreJault p,= rE vr sin d= 7 . 2 x 1 S l I l X,, 0.8
d= l.) stn b
L
Table 12.9 swing curyecalculationthree poleswitching
P,u,o
6
sec
)
P " (pu)
P,, (pu)
.5.4P,,
ifi
elec deg elec deg
6 elec deg 41.8 41.8 41.8 43.0 46. 6 55. 6 70. 0 89. 8 I 15.0
0
o*
ouu, 0.05
Initial load= 1.0 pu Initial torqueangleis givenby
1 = 1.5 sin 5o
or Duringfault
6o= 47.8"
F'",,= I)uring
l.Zxl
lff
sin d = 0.827 sin
single pole switching
= Perrr
+#
sin d = 0.985 sin
0.075+ 0.10 0.15 0 . 20 o.25 0 . 30 0.325+ 0 . 3s 0.40 0. 45 0. 50 0 . s5 0 . 60 0.65
1. 5 0. 667 r . 0 0.0 0. 827 0. 667 0. 552 0. 448 0.224 0.827 0.682 0. 564 0. 436 0.0 0. 0 0.0 0. 0 0. 0 1. 5 1. 5 1.5 1. 5 1.5 r.5 0.726 0.0 0.0 0.0 0.0 0.0 0.85 0.078 0.827 r . 48 0.98 0. 146 1.0 1.0 1.0 1.0 1.0 0.15 0.922 r.827 2. 48 1.98 0. 254
1.2
N A
1.2 3.6 9. 0 r 4. 4 19. 8 2s. 2 30.6 31.4 36.4 46.3 59.7 10.4 71.8
5. 4 5. 4 5. 4 5. 4 5. 4 0. 8 5. 0 9. 9 13.4 10.7 t.4
0. 565 0. 052  0.55  0. 984 0 . 6 5 10.497
r45.6 177.O 2r3.4 259.7 3r9.4 389.8 461.6
The swing curve is plotted in Fig. 12.50 from which it is obvious that rhe
\4gqe!_lg',gf tyq]g1_Analysis 0.40 0.45 0.50 0.55 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00 1.10 1.15 1.20 r.25 1.30 1.35 r.40 1.45 1.50
r
PowerSystem Stability l.s 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 0.998 1.5 1.0 1.5 1.0 1.5 0 . 9 9 8 51 . 5 0.96 0.894 0.781 0.62 0.433 0.2s9 0.133 0.079 0.107 0.214 0.38 0.566 0.738 0.867 0.946 0.983 0.997 t.44 1.34 t.l1 0.932 0.6s 0.39 0.2 0.119 0.161 0.322 0.57 0.84 1.11 1.3 t.42 1.48 1.5 0.5 0.5 0.s 0.5  2.7  2.7  2.7 2.7 2.8 0.1  2.6
505;  86.6 89.4 89.5 86.9
I
I
300 !
t I
I
I 250 I
i
I
\
r pote ii switchoff i
I
L
6 q)
g)
o E o
L
2oo / / l iI
I
r. 0 5 r . 5
MACHINE UNSTABLE
q)
(.) 1 5 0
100
J)
to
 0.44  2.4 0.34 1.8  0.17  0.9 0.068 0.4' 0.35 r.9 0.61 3.3 0.8 4.3 0.881 4.8 0.839 4.5 0.678 3.7 0.43 2.3 0 .1 6 0.9  0.11  0.6  0.3  1.6  0.42  2.3  0.48  2.6  0.s  2.7
10.3 12.r 13.0 12.6 10.7  7.4  3.r 1.7 6.2 9.9 12.2 l3.l 12.5 10.9 8.6 6.0 3.3 \ '
73.7 63.4 51.3 38.3 25.7 15.0 7.6 4.5 6.2 r2.4 22.3 34.5 47.6 60.1 7r.O 79.6 95.6 gg.g
r
I
,r
l,r r Time (sec)
]...r
o'5
1.0
The swing curve is plotted in Fig. 12.51from which it follows that the sysrem is stable.
Single pole switch off
Fig. 12.50 swing curuefor threepole switching with reclosure Table 12.10 swing curvecalculationsingle pole switching
I SEC I',uu^
A
/'/'
I I
 ' /
(fault Reclosure cleared)
,s:itth
t,,,
(pu)
I
BO MACHINE STABLE 60
(pu)
l'r, (pu)
5'4P,, elec deg
Ab
elec deg
b elec deg
0
o,
f
1.5 0.661 1.0 0.827 0.667 0.s52 0.827 0.682 0.s64 0.98s 0.726 0.98s 0.784 0.985 0.848 0.98s 0.908 0.985 0.956 1.5 0.7Ls 0.77 0.834 0.893 0.940
oory
0.05 0.0750.10 0.15 o.20 o.25 0.30 4325+ 0.35
0.0 0.448 0.224 0.436
1.2 2.4
r.2 3.6 5.1 6.3 7.2 1.8 8.1 5.5
4t . 8 0 41.8 41.8 43.0 46.6 5r.7 58.0 65.2 73.0 81.1
(Contd....)
o) o o) o) o
L
C)
o o o)
ra
0.285 1.5 0.230 r.2 0.166 0.9 0.107 0.6 0.060 0.3 0.485  2.6
i
lr
05
Time (sec)
0.988 1.485
Fig. 12.51 Swingcurvefor singlepoleswitching with reclosure
506 
po@is Modern
I ina
PROB TEIlIS
t 2 . l A twopole, 50 Hz, 11 kv turboalternatorhas a rating of 100 Mw,
powel factor 0.85 lagging.The rotor has a momentof inertia of a 10,000 12.2 Two turboalternatorswith ratings given below are interconnectedvia a short transmission line. Machine 1: 4 poIe, 50 Hz, 60 MW, power factor 0.g0 lagging, moment of inertia 30,000 kgrn, Machine 2 pole, 50 Hz, 80 MW, power factor 0.85 lagging, moment of inertia 10,000 kg' Calculate the inertia constantof the single equivalentmachine on a base of 200 MVA. t 2 . 3 Power station t has four identical generatorsetseachrated g0 MVA and each having an inertia constant 7 MJA4VA; while power station 2 has three sets each rated 200 MVA, 3 MJA4VA. The stations are locatld close togetherto be regarded a single equivalentmachinefor stability as studies. Calculate the inertia constant of the equivalent machine on 100 MVA base. I2.8
from its prefault position, determine the maximum load that could be transferredwithout loss of stability. A synchronousgenerator is feeding 250 MW to a large 5O Hz network over a double circuit transmissionline. The maximum steadystateDower that can be transmitted over the line with both circuits in operation is 500 MW and is 350 MW with any one of the circuits. A solid threephasefault occurring at the networkend of one of the lines causesit to trip. Estimate the critical clearing angle in which the circuit breakersmust trip so that synchronismis not lost. What further information is neededto estimate the critical clearing time? 12.9 A synchronousgenerator represented a voltage source of 1.05 pu in by serieswith a transient reactanceof 70.15 pu and in inertia constant F/ = 4.0 sec, is connectedto an infinite inertia system through a transmission line. The line has a series reactance of70.30 pu, while the infinite inertia system is representedby a voltage source of 1.0 pu in series with a transient reactanceof 70.20 pu. The generatoris transmitting an acti're power of 1.0 pu when a threephasefault occurs at its terminals. If the fault is clearedin 100 millisec, determine if the system will remain stable by calculating the swing curve. 12.10 For Problem 12.9 find the critical clearing time from the swin! currrefor a sustained fault.
/..^^+^+r
\Lt,t:ltdlllr.
12.4 A 50 Hz transmission 500 km long with constants line given below ties
up two large power areas R = 0 .1 1 f)/k m C = 0.009 lFlkm
F i n d t h e ,s. f  ' : ,t,rJl v s l z f e c f e h i lrirtr r r r i mrirtr lrrr c. J
,,
L  1.45mH/km G = 0
i f ,l rl ,, t/t  _
 v
ll/
Rt
 _
t \n v l l , r
A
t,\/
V
What will the steadystate stability limit be if line capacitanceis also neglected?What will the steadystate stability limit be if line resistance i s a l s o n e g l o c tc d C o rn n rc l tt tl rc rcsul ts. '/ on t 2 . 5 A power deficient area receives 50 MW over a tie line from another area.The maximum steadystatecapacityof the tie line is 100 MW. Find the allowable sudden load that can be switched on without loss of stability. 1 2 . 6 A synchronous motor is drawing 30vo of the maximum steady state power from an infinite bus bar. If the load on motor is suddenly increasedby 100 per cent, would the synchronismbe lost? If not, what is the maximum excursionof torque angle about the new steady state r<ltorposition.
t 2 . 1 The transfer reactancesbetween a generator and an infinite bus bar
o p e ri l ti n g rf 2 0 0 k V trn d e r a r i ouscondi ti ons fhe i nterconnector i v on aro: Pretault During fault Postfhult S0 0 per phase m O per phase 2ffi {) per phase
A synchr t ) nous gener at or epr esent cd a volt ageof l. l5 pu in ser ies r by with a transientreactanceis c<lnnected a large power system with to volt t t gc 1. 0 pu t hr ot lgh l powcr r r clwor k.Thc cquivalcntt lar r sient transf'er reactanceX betweenvoltage sourcesis 70.50 pu. After the occurrence a threephase grouncl of to fault on one of the lines of the power network, two of the line circuit breakersA and B operate sequentially as follows with correspondingtransient transfer reactance given therein. (i) Short circuit occurs at 6 = 30", A opens instantaneouslyto make X = 3.0 pu. (ii) At 6 = 60o, A recloses, X = 6.0 pu. (iii) At 5=75o, A reopens. (iv) At d = 90o, B also opens to clear the fault making X = 0.60 pu Check if the systenrwill operate stably. 12.12 A 50 Hz synchronous generatorwith inertia constant H = 2.5 sec and a transientreactanceof 0,20 pu feeds 0.80 pu active power into an infinite bus (voltage I pu) at 0.8 lagging power lactor via a network with an equivalent reactance of 0.25 pu. A threephase fault is sustainedfor 150 millisec across generator terminals.Determine through swing curye calculation the torque angle6, 250 millisec, after fault initiation.
l 2.l l
PowerSystem Stabitity
12.13 A 50 Hz, 500 MVA,400 kV generator (with transformer) is connected to a 400 kV infinite bus bar through an interconnector. The generator has F1 2.5 MJA4VA, voltage behind transientreactanceof 450 kV and is loaded 460 MW. The transfer reactancesbetween generator and bus
5(D I
12. Kundur, P., Power SystemStability and Control, McGrawHill, New York, 1994. 13. Chakrabarti,A., D.P. Kothari and A.K. Mukhopadhyay, Performance Operation and Cqntrol of EHV Power TransmissionSystems,Wheeler Publishing, New Delhi, 1995. 14. Padiyar,K.R., Povter System
lcatlons, Hy Stability and Control, Znd
During fault Postfault
1 . 0p u 0.75pu
15. Sauer,P.W. and M.A. Pai, Power SystemDynamics and Stabiliry, PrenticeHall, New Jersey,1998.
Calculate the swing curve using intervals of 0.05 sec and assuming that the fault,is cleared at 0.15 sec. I2.I4 Plot swing curves and check system stability for the fault shown on the system of Example 12.10for fault clearing by simultaneousopening of breakers at the ends of the faulted line at three cycles and eight cycles after the fault occurs.Also plot the swing curye over a period of 0.6 sec if the fault is sustained. For the generator assumeH = 3.5 pu, G = 1 pu and carry out the computations in per unit. 12.15 Solve Example 12.10 for a LLG fault.
Papers
16. Cushing, E.W. et al., "Fast Valving as an Aid to Power System Transient Stability and Prompt Resynchronisationand Rapid Reload After Full L,oad t Rejection", IEEE Trans, L972, PAS 9I 1624. 17. Kimbark, E.W., "Improvement of Power System Stability", IEEE Trans., 1969, PAS88:773. 18. Dharma Rao, N. "RouthHurwitz Condition and Lyapunov Methods for the TransientStability Problem", Proc. IEE, 1969, 116: 533. 19. Shelton, M.L. et al., "BPA 1400 MW Braking Resistor",IEEE Trans., 1975,94: 602.
NCES REFERE
Books
1 Stevenson, W.D., Elements of Power SystemAnalysis, 4th edn., McGrawHill, New York, 1982. Elgerd, O.I., Electic Energy Systems Theory: An iniroduciion, McGrawHill, New York, 1982. 3. 4. 5. 6. 7. 2nd edn.,
20. Nanda, J., D.P. Kothari, P.R. Bijwe and D.L. Shenoy,"A New Approach for
Dynamic Equivalents Using Distribution Factors Based on a Moment Concept", Proc. IEEE Int. Conf. on Computers, Systemsand Signal Processi4g, Bangalore, \ Dec. 1012, 1984. 'Dynamic Modelling 2 r . Dillon, T.S., and Control of Large Scale System", Int. Journal of Electric Power and Energy Systems,Jan. 1982, 4: 29. 2 2 . Fatel,R., T.S. Bhatti and D.P. Kothari, "Improvement PowerSystemTransient of stability using Fast valving: A Review", Int. J. of Electric Power components and Systems,Vol. 29, Oct 2001, 927938. 23. Patel, R., T.s. Bhatti and D.P. Kothari, "MATLAB/simulink Based rransient Stability Analysis of a Multimachine Power System,IJEEE, Vol. 39, no. 4, Oct. 2002, pp 339355. Patel R., T.S. Bhatti and D.P. Kothari, "A Novel scheme of Fast valving Control", IEEE Power Engineering Review, Oct.2002, pp. 4446. 25. Patel,R., T.S. Bhani and D.P. Kothari, "Improvementof Power system Transient stability by coordinated operation of Fast valving and Braking Resistor", To appearin IEE proceeriingsGen., Trans and Distribution.
. A
Anderson, P.M. and A.A. Fund, Power SystemControl and Stability, The Iowa State University Press,Ames, Iowa, 1977. stagg, G.w. and A.H. oAbiad, computer Methods in Power system Analysis, Chaps9 and 10, McGrawHillBook Co., New York, 1968. Crary, S.8., Power System Stability, Vol. I (Steady State Stability), Vol. II (Transient Stability), Wiley, New York, 19451947. Kimbark, E.W., Power System Stability, Vols 1, 2 and 3, Wiley, New york, 1948,
L+.
Veuikorz,Y.A., TransientPhenomena Electrical Power System(translatedfrom in the Russian),Mir Publishers, Moscow, 1971. 8. Byerly, R.T. and E.w. Kimbark (Eds.), stability of l^arge Electric power Systems,IEEE Press,New York, 1974. 9. Neuenswander, J.R., Modern Power Systems,International Text Book Co., 1971. t0. Pai, M.A., Power SystemStability Annlysis by the Direct Method of Lyapunov., NorthHolland, System and Control Services, Vol. 3, 1981. I 1. Fouad, A.A and V. Vittal, Power System Transient Stability Analysis using the Transient Energy Function Method, PrenticeHall, New Jersy, 1992.
13.1
INTRODUCTION
In Chapter7, we have beenprimarily concerned with the economical operation of a power system' An equally important factor in the operation of a power system is the desire to maintain system security. System security involves practicessuitably designedto keep the system operating when componentsfail. Besideseconomizingon fuel cost and minimizrngemission of gases(co, cor, Nox, sor), the power systern should be operationally,.secure,,. operation_ An ally "secure" power system is one with low probability of, systern black out (collapse) equipment or damage. the pro."r, uf cascading If failurescontinses. the systernas a whole or its tnajor parts may completely collapse. This is normally referred to as system blackout. All these aspects require security c ons t r a i n ep o w e r s y s te m p ti mi z a ti on C O). d o (S Since security and economy normally have conflicting requirements, it is inappfopriate treat them separately. to The fina.laim of economy is the security lunction of the utility company.The energy management system (EMS) is to operate the system at minimum cost, with the guaranteed alleviation of emergency conditions.The emergency condition will dependon the severity of t iolat io n s f o p e ra ti n g i rn i ts(b ra n c h l ow sand bus o l f' vol tagel i mi ts).The most severeviolationsresult fiom contingencies. irnportant An part of securitystudy, therefbre,moves around the power system'sability to withstanrjthe effects of contingencies.A particular systemstateis said to be secureonly with reference to one or more specific contingency cases, and a given set of quantities monitoredfor violation. Most power systemsare operated in such a way that any singlecontingencywill not leaveother.opon"nts heavily overloaded,so that cascading failures are avoided.
Sll I Most of the security related functions deal with static "snapshots" of the power system.They have to be executedat intervals compatiblewith the rate of changeof systemstate.This quasistatic approach to a large extent,the is, only practical approachat present,since dynamic analysisand optimization are conslder4bly mole {!fficu!! 4nd cqmpurallo44lly 1aqtelime corrsulurg, System security can be said to comprise of three major functions that are carried out in an energycontrol centre: (i) systemmonitoring, (ii) contingency analysis,and (iii) comectiveaction analysis. System monitoring suppliesthe power systemoperators dispatcherswith or pertinentuptodateinformation on the conditionsof the power system on real time basisas load and generation change.Telemetrysystemsrneasure, monitor and transmit the data, voltages,currents,current flows and the statusof circuit breakersand switchesin every substation a transrnission in network. Further, other critical and important information such as frequency, generator outputs and transformertap positions can also be telemetered. Digital computers in a control centre then processthe telemetereddata and place them in a data base form and inform the operatorsin case of an overload or out of limit voltage. Important data are also displayed on large size monitors. Alarms or warnings may be given if required. Stateestimation (Chapter 14) is normally used in such systemsto combine telemetereddata to give the best estimate (in statisticalsense)of the curreltt systemcondition or "state". Such systemsotten work with supervi$orycontrol systemsto help operatorscontrol circuit breakersand operateswitches and taps remotely. These systemstogetherare called SCADA (supervisorycontrol and data acquisition)systelns. The second maior security function is contingency analysis. Modern operation computers havecontingency analysis programs storedin them.These loreseepossiblcsystetntroubles(outages)beforethey occur.They study outage events and alert the operators to any potential overloads or serious voltage vi ol ati tl ns. For exalnple, he sir nplest ir r m of cont ingency t f analysis can be put together with a standard LF program as studied in Chapter 6, along with proceduresto set up the load flow dafa for each outageto be studied by the LF plogram. This allows the system operatorsto locate def'ensiveoperating stateswhere no single contingencyevent will generateoverloadsand/or voltage violation:;.This analysis thus evolves operatingconstraintswhich may be cntpi oycdin t hc liD ( ccot r olnic dispat ch) and UC ( unitcor nnr it r nclrpr ogr ar r r . t) Thus contingencyanalysiscarricsout ornergcncy identil'ication ancl"what if'' simulations. The third major security function, corrective action analysis, permits the operatorto changethe operationof the power systemif a contingencyanalysis program predicts a serious problem in the event of the occurrenceof a certain outage.Thus this provides preventive and postcontingency control. A simple example of corrective action is the shifting of generationfrom one station to another.This may result in change in power flows and causing a change in loading on overloadedlines.
Power System Security
j5! 2 ,1
M o d e rnP o @i s
I srr
o o o o o o c o o o
Threse threeftrnctionstogetherconsist oi a very compiex set of toois that heip in the secureoperationol'a power system. T3.2 SYSTEM STATE CLASSIFICATION
= 3
pb. 6 q
o o oo
L
Dyliacco [13] and further clarified by Fink and Carlsen l23l in order to define relevant EMS (Energy ManagementSystem) functions. Stott et. al [15] have also presenteda more practical static security level diagram (see Fig. 13.1) by incorporating correctively secure(Level 2) andcorrectableemergency(Level4) security levels. In the Fig. 13.1, arrowed linep represent involuntary transitions between Levels 1 to 5 due to contingencies..The removal of violations from Level 4 normally requiresEMS directed "corrective rescheduling" or "remedial action" bringing the system to Level 3, from where it can return to either Level I or 2 by further EMS, directed "preventive rescheduling" depending upon the desired operational security objectives. Levels I and 2 representnormal power system operation.Level t has the ideal security but is too conservativeand costly. The power system survives any of the credible contingencieswithout relying on any postcontingency corrective action. Level2 is more economical,but dependson postcontingencycorrective rescheduling to alleviate violations without loss of load, within a specified period of time. Postcontingencyoperating limits might be different from their precontingencyvalues. 13.3 SECURITY ANALYSIS
o, c o (!6
L
EE
8E
 o =
q)
z
o. o o
BE
r: .9
qF
E o oo_ o o o. oo o o)
o E o
=E;.
o
gH€
' = O 
o_o x
oc.Y
^;E t bE . E
abg
E O
!ie =98
= O
o o
c
g, c
o (J
o o o
L
o
o
P c
9 c
()
f
d H
9 'oE .b
.= g, g)'E c c F O (g(J
d 8
o c c) o) c
x
.FOE
b o d
=E fi E8E
z E C g
g >
U) C) (E o
o
E
o o o
: 5' 58 d
dl )  . o (
. = 9 =
8bb
o
9 c
o
o o o_
System security can be broken down into two major functions that are carried out in an operationscontrol centre: (i) security assessment, and (ii) security control. The tormer gives the security level of the system operating state. The latter determines the appropriate security constrained scheduling required to optimally attain the target security level. The security functions in an EMS can be executedin 'real time' and 'study' modes. Real time application functions have a particular need for computing speed and reliability. 'fhe static securit.v level of a power systemis characterised the presence by or rrtherwise of emergency operating conditions (limit violations) in its actual (precontingency) or potential (postcontingency)operating states. System security assessment the process by which any such violations are detected. is System e.:^ sstlt€rrt involves two func tions: s€ (i) system monitoring and (ii) contingencyanalysis.Systemmonitoring provides the operator of the power system with pertinent uptodate information on the current condition:;clf the power system. In its simplest form, this just detects violations in the actual systemoperating state.Contingency analysisis much
los i
5t ) c o (
.=
at
6 H
F.O
HE ' =
e o y
=ee d 8
6L
9 c
g#EE
('5 o 3
o 0) o) E o) o o
r'/^ Y q)
L
.d, lr
o c) a o
o o
L
 E a d
gE
:( l:) 0
([ ()
8o s o
9= ':

, ^'ku t * ao
ocfr > o x C. Az J3(D
, ^9: * i
o)c, > o Ail J3
A
_9 !)
> o
e ai
C)
,.sfe ;l
Modern power Svstem Analvsis
SLACKBUS
Power System Security
I srai _
f+o*ys
41.0236t5.3"
Only a small proportion of work on optimal power flow (OPF) has taken into account the security constraints.The most successfulapplications have beento the security constrained MW dispatch OPF subproblem. The contingencyconstrainedvoltageivar reschedulingproblem, as of the writing of this text, still remains to be solved to a satisf desree. The total number of contingency constraintsimposed on SCO is enormous. The SCO or contingency constrained OPF problem is solved with or without first optimizing with respectto the base case(precontingency)constraints.The general procedure adopted is as follows: (i) Contingency analysis is carried out and cases with violations or near violations are identified. (ii) The SCO problem is solved. (iii) The rescheduling in Step 1 might have created new violations, and therefore step 1 should be repeatedtill no violations exist. Hence, SCO represents potentially massiveadditional computing effort. a An excellent comprehensiveoverview of various available methods is presentedby Stott et. al [15]. There is still great potential for further improvement in power system security control. Better problem formulations, theory, computer solution methods and implementation techniquesare required. T3.4 CONTINGENCYANALYSIS
t +s*y ts
1.024215.0" 3
1 . O 6 t O "1
.>40.7 j1.2 _ 39.5_i3.0:_ +
,l
y8.6 ,f88.e
Ja.a/e.8 +
_ + 1 8 . 9 y 5 . 2 _ 1 6 . 9+ 1 3 . 2  <
i6.3j2.3
zt.si5.9 *js.a I z+.2 / 'L?:'q'''
t
* t az.s
*S4.9 +17.3 1.0474t2
.i;n;l+s i7.2 * 53.7 '
'"'' l+u
*oo*ito
5 i1.o17gt6.2" Y
+1t ) , zo o
!G) t40 +i30 Fig. 13.2
SLACKBUS
5 Base Case AC Lineflow for sample bus system
{+s*yts
1 . 0 1 0 7 1 5 . 9 " 3 38.4 i1.1
f+o+7s
4 1.00682 6.6" :
1.0610" 1
>48.6 + j5.2 46.7  j5.3< + 38.510.6
.8 +81 i 5.5
+ 3a?/'er
{  r . o 7 a. s
In the past many widespreadblackouts have occurred in interconnectedpower systems. Therefore, it is necessaryto ensure that power systems should be operatec! mosf economic:r!ly such that povrer is cle!i.rerecl reliably. Reliable operation implies that there is adequatepower generation and the same can be transmitted reliably to the loads.Most power systems designedwith enough are redundancyso that they can withstand all rnajor tailure events.Here we shall sttrdy thc possiblc consccprcnccs rcrncdialactions rcquircd by two nrain :rncl f ailur e ev e n ts ;l i n e .o rrta g e s c lg e n e ra t i ng tfai l ures. an rrni To explainthe problelnbriefly,we consider fivebussystem Ref'erence the of LI0J. The basecaseload llow resultslbr the exampleare given in t ig. 13.2and, s lt r r wir f ' lo w< tf2 4 .7 MW a n d 3 .6 MV AR on the l i ne f} om bus 2 to hus 3. L.ct t ls lls s t llll c a t a t p rc s c n l w e i l rc o n l y i n tcrestecl the MW l oadi ngof the l i ne. , th in Let us examinewhat will happenif the line from bus 2 to bus 4 were to open*. ' l' lr c r c s ulti n g i n c l ' l o w sa rtc v o l ta g c s l c show ni n l ri g. 13.3.tt nray be notcd l l u that the flow on the line 23 has increasedto 31.5 MW and that most of the other line flows are also changed. It may also be noted fhat bus voltage magnitudesalso get aff'ected, particularly at bus 4, the change is almost2To less from 1.0236to 1.0068pu. Supposethe line from bus 2 to bus 5 were to open. Figure 13.4shows the resultingflows and voltages. Now the inaximum change t:fgl.g.j 5 which is almost 107o less. ",jus xSimulation line outageis more complex than a generatoroutage, of since line outageresults a change system in in configurations.
* t eo.o
:63.1 + j1O.2 1.0468t2.
<  61.6 i8.9
i r.s ;t .r
5 1.O114t6.4"
rlr { 2s+ilo t40+i30
\ lJ./
loo *lt o
2 Fig. 13.3 PostoutageAC LoadFlow(Linebetween and4 is open)
SLACKBUS
 +80.e+13.21
0.se35/ 6.s" o1  1 ' 0 6 t 0 "1 l ' 5 4 . 4+ 7 i 1 i . 3 s 2 . 1  i 9 .    s 1 . 4+ i 7 . 3  s 1 . 1 _1__r
(c)
t a sr r t s
, I ao7s
tl;f _ q q . z _ j 1 z . s Itl
*  +s,s 1tz.t
':: )ll
+ 120 i10
':t't'
 52.s 13.0 i +
 5 0.8994t16.?
'foo*yro
2 AC Fig. 13.4 Postoutage Lcad Flow(Linebetween and5 is open)
;'518 t

5LAUK HUS
rr,rodern Po@is
e t 4 s+ l 1 5
l+o*1s
_ 1 . o o 6 1 ts . 7 " 3 1  1 . 0 6 t 0 "'l_ !.6 + ^ 1l I _j7.6 <_i __22.4 j2.6 _22 .3 +j0.7 <_ 4 1 . O 0 4 3 t  6 . 1 " _ 1t.s _ 4s.9 Li1I {nze+iz1.7l { t.s  1t.t I  V_21.5_j4.8 25.2j4.4 +
1  r z o .g t s . t (  '  _ 1 l
2..]u+ _+ _53.6+16.8 1.o24st3.7. i Y + l,zo 1lo
l l r
I
t
i
'''"
GiveAlarm signal
fzs7s.s
<52.5 16.5 5
o 9 9 5 6 / _  71 " .
{oo*iro
GiveAlarm signal
Fig. 13.5 Postoutage AC Load Flow(Generator outage,lostgeneration 2 is picked by generator up 1) Figure 13.5is an exampleof generator outageanclis selected explainthe to tact that generatoroutagescan also result in changes in line flows and bus voltages.In the exampleshowr in Fig. 13.5 all the generationlost from bus 2 is picked up on the generatorat bus 1. Had there been more thanZgenerarors in the sample system say at bus 3 also, it was possible the loss of leneration on bus 2 is made up by an increase in generation at buses 1 and 3. The differencesin Iine flows anclbus voltagesvrould show how the lost gener.ation is shared by the remaining units is quite significant. It is important to know which line or unit outageswill render line flows or voltagesto crossthe lirnjts. To find the eff'ects outages,contingencyanalysis of techniques are empioyeci. Contingency analysis models single failure events ( i' e' one l i n co u ta g c s iro l l c u n i t o u tu g cs) nrul ti pl c o or ccpri pnrcntri l urccvc^ts fi (failure of multiple unit or lines or their combination) one after another until all "credible outages"are considered.For each outage,all lines ancl voltagesin the netrvork are checkedagainsttheir respective limits. Figure 13.6 depictsa flow chart illustrating a simple method for carrying out a contingency analysis. One of the important problems is the selection of "all credible outages,,. Execution time to analyseseveralthousand outagesis typically I min basedon computerand analytical technologyas of 2000.An erpproxirnate rnodelsuclras DC load flow may be used to achievespeedy solution if voltage is aiso required, then full AC load flow analysishas to be carried out.
1;no*rifii.iuior".""i;{offi 'l ' ''l'''' l'r'.,o _ _ l
't A..
,rr\=_
violation? >*t_qJ
ves jciventarn
T
,*"
, Noi 'I
I
Ail
\
FIg. 13.6 A simpletechnique contingency for analysis
620j;:f
I
^
uodern PowerSystem Anatysis
Y .FAUIL'I(s
IJ.C
a
?
DIIIYDIIIVII
frErrarE?rtrElt
A .security analysisprogramis run in a load dispatchcentre very quickly to help the operators.This can be attempted by carrying out an approximate analysis and using a computer system having multiple processors or vector processors y anarysts. ne s I may uate an equrvalent should be used for neighboursconnectedthrough tielinos. We can eliminate all nonviolation casesand run complete exact program for "critical" casesonly. This can be achievedby using techniquessuch as "contingency selection"or "contingency screening", or "contingency ranking". Thus it will be easy to warn the operation staff in advance to enable them to take corrective action if one or more outageswiil result in seriousoverloadsor any violations. One of the simplest ways to present a quick calculation of possible overloads is to employ network (linear) sensitivity factors. Thesefactors give the approximate change in line flows for changes in generationin the system and can be calculated from the DC load flow. They are mainly of two types: 1. Generation shift factors 2. Line outage distribution factors Briefly we shall now describe the use of those factors without deriving them. Reference [7] gives their deri iation. The generation shift factorsl cr.,;are defined as:
limits and those violating their limit can be informed to the operator for control action. necessary generation shift sensitivity factors are linear estimatesof the change in The line flow with a change in power at a bus. Thus, the effects of simultaneous principle of superposition. Let us assumethat the loss of the ith generatoris to be made up by governor action on all generatorsof the interconnectedsystemand pick up in proportion to their maximum MW ratings. Thus, the proportion of generationpick up from unit k (k * i) would be
(13.7)
where = Pn,,,,,u,^maximum MW rating for rnth generator g*i= proportionality factor for pick up on kth unit when ith unit fails. Now, for checking the /th line flow, we may write
j, = ff * 0r; APo, \ri LPotl E,lau,
(13.8)
(13.4)
where, 4t = Change in MW power flow on hne I when a change generain tion, AP", takes place at the ith bus
that no unit will violate its maximdm limit. For In Eq. (13.8) it is assumed unit limit violation, algorithm can easily be modified. Similarly the line outage distribution factors can be used for checking if the line overloadswhen solllc of the lines are lost. The line outage distribution factor is defined as: d,,,= * where dt,i = line outagedistribution factor when monitoring /th line atter an outage of ith line. Aft = change in MW flow on /th iine'  precontingency line flow on ith line fi lf precontingency line flows on lines / and i, tlre power flow on line / with line i out can be found out employing "d" factors.
Here, it is assumedthat LPotis fully compensated an equal and opposite by change in generation at the slack (reference) bus, with all other generators remaining fixed at their original power generations. The factor al,then gives the sensitivityof the /th line flow to a changein generationat ith bus. Let us now study the outage of a large generating unit and assume that all the lost generation (Pod would be supplied by the slack bus generation. Then
(13.e)
Ji
APo,  P1i
(13.s)
and the new power flow on each line could be calculatedusing a precalculated set of " d' factors as given below. ft = f i * dti APc, where, ft for all lines V /
(13.6)
Here, fi
?,= ff *d,,,f,o
(13.10)
 power flow on /th line after the failure of ith generator
f i = power flow on /th line before the failure or precontingency power flow
^d foi =precontingency or preoutageflows on lines / and i respectively
fr = power flow on /th line with ith line out.
iii
l
a
Po Modern
j=l j=2 j=3 j=4 j=5 j=6
J"srr
j = 7
'd' factors ali the lines for Thus one can check quickiy by precaiculating overloading for the outage of a particular line. This can be repeated fbr the outage of each line one by one and overloads can be found out for corrective action. It may be noted that a line flow can be positive or negative. Hence we must check / agarnst  Jt ^u* as well as h **. Llne tlows can be louncl out usmg If telemetry systemsor with stateestimation techniques. the network undergoes any significant structural change, the sensitivity factors must be updated.
(line 12)(linel3)(Iine 23)(line24)(line25)(line34)(line45) (l i n e l 2) 1. 00010. 3331 0. 2685  0. 2094 0. 3735 0. 2091 0. 0
v = J (line L  J ) 3 \ r r \ r w 23) 0.4542 0.4545 0 
= = = =
4 5 6 7
0.4476 (hne 24) 0.3634 0.3636 0.4443 0.0 (line 25) 0.1819 0.1818 0.2222 0.2835 (line 34) 0.5451 0.5451 0.6662 0.7161 (line 45) 0.18i6 0.1818 0.22220.2835
0.3488 0.6226 0.3488 0.4418 0.6642 O.44r8 0.0 0'3321 1.0 0.5580 0.5580 0.0 1.0002 0.3321 0'0
Find the generationshift factors and the line outagedistribution factors for the fivebus sample network discussedearlier. Solution Table 13.1 gives the [x] matrix for the five bus sample system, together with the generation shift distribution factors and the line outage distribution factors are given in Tables I3.2 and 13.3 respectively. Table 13.1 X Matrixfor FivebusSampleSystem(Bus 1 as a reference)
0 0 0 0 0. 0.05057 0.03772 0.04029 0.4714 0.03772 0.08914 0.07886 0.05143 0.04029 0.07886 0.09514 0.05857 0.04714 0.05143 0.05857 0.13095
It has been found that if we calculate the line flows by the sensitivity close to the valuescalculatedby the methods,they come out to be reasonably full AC load flows. However, the calculations carried out by sensitivity methods are faster than those made by full AC load flow methodsand therefore are used for real time monitoring and control of power systems.However, where reactive power flows are mainly required, a full AC load flow method (NR/FDLF) is preferred for contingency analysis. The simplest AC security analysisproceduremerely needsto run an AC load flow analysis for each possible unit, line and transformer outage.One normally does ranking or shortlisting of most likely bad caseswhich are likely to result in an overload or voltage limit violation and other casesneed not be analysed. Any good P1(performanceindex can be selected)is usedfor rankirig. One such P/ is
(13.11)
For large n, PI will be a small numberif all line flows are within limit, and will be large if one or more lines are overloaded. ciur be done fbr P1. P1 tablecan be orderedfrom For rr = I exact calculations value to least. Suitable number of candidatesthen can be chosen for largest further analysis[7]. If voltagesare to be included, then the following PI can be employed.
Factorfor Fivebus System Table 13.2 Generation Shift Distribution Bus I
l=1(line12) /=2(line13) /=3(line23) I=4(line24) /=5(line25) /=6(line34) I=l (line45)
Bus 2  0.8428  o.t572 o.0714 0.0571 0.0286  0.0857  0.0285
(13.12)
Here, Alvil is the difference between the voltage magnitude as obtained at the end of the lPlQ FDLF algorithm Alvlu* it the value fixed by the utility. ue Largest vaiue oi Pi is piaceciat the top. The security arraiysistrray rrow startedfor the desired numbel of casesdown the ranking list. $ummary and Further Reading: Reference [25] has discussedthe concept for screening contingencies. Such techniquesfonn the foundationfor many realcontingency selection/screening time computer security analysis algorithms.
l#GrffiE
IE!F}ILfr
Reference [15] gives a broad overview of security assessment contain an and bibliography covering the literature on"security assessmenrup to ;;;;t]r", Reference [11] gives an excellent bibliography on voltage stability. This topic is discussedbriefly in the next section. 13.6 POWER SYSTEM VOITAGE STABILITY
stability, the dynamics mainly involves the loads and the means for voltage control. Ref [11] provides a comprehensivelist of books, reports, workshops and technical papers related to voltage stability and security. Definitions: [2]
Power transmissiol tuglgility has traditionally been Iimited by either rotor angle (synchronous) stability or by thermal loaoing capabilities. The blackout problem has been linked with transient stability. Luckily this ;roblem is now not that serious because of fast short circuii clearing; po*Lrru excitation systems, and other special stability controls. ElectriJ *rnpuni"s are now required to squeezethe maximum possible power through networks owing to various constraintsin the construction of generation "ii.G transmission and facilities. voltage (load) stability, however, is now a main issue in planning and operating electric power systemsand is a factor reading to limit pow; transfers. voltage stability is concerned with the ability of a power system to maintain acceptable voltages at all busesin the system under normal conditions and after being subjected to a disturbance.A power system is said to have entereda state of voltage instability when a disturbante resurts in a pro!."rriu" and uncontrollable decline in voltage Inadequate reactive.power support from generatorsand transmission lines leads to voltage.instability ot uoitug" collapse, which have resulted in several major system failures in the world. Th"y ", (i) south Florida, usA, system disturbanc e of 17 May r9g5, (transient, 4 sec) (ii) French systemdisturbancesof Decembe r 19, r97g and Januar 12, rgg7, y (longer term). (iii) swedish systemdisturbanceof December 27, rgg3 (longer term, 55 sec) (iv) Japanese(Tokyo) system disturbance of July 23, 1gg7irorg", term, z0 min) (v) NREB grid disturbancein India in 19g4 and r9g7. (vi) Belgium, Aug 4, 1992. (longer term, 4.5 min) (vii) Baltimore, washington DC, usA, 5th July 1990 (longer rerm, insecure for hours) Hence, a full understandingof voltage stability phenomena and designing mitigation schemesto prevent voltase instabilitv is nf o,raqr'or,,o+^,.+:r:a:^^
A power system at a given operating state is smalldisturbance voltage stable if, following any small disturbance, voltages near loads are identical or close to the predisturbancevalues. The concept of smalldisturbance voltage stability is related to steadystatestability (Chapter 12) and can be analysedusing smallsignal (linearised) model of the system. A power systemat a given operating stateand subject to a given disturbance ts voltage stable if voltages near loads approach postdisturbanceequilibrium valuei. The concept of voltage stability is related to the ffansient stability of a power system. The analysis of voltage stability normally requires simulation of the system modelled by nonlinear diffdrentialalgebraic equations. A power systemat a given operating stateand subject to a given disturbance undergoes voltage collapse if postdisturbanceequilibrium voltages are below acceptable limits. Voltage collapse may be total (blackout) or partial. The voltage instability and collapse may occur in a time frame of a second.In this case the term transient voltage stability is used. Sometimes it may take up to tens of minutes in which case the term longterm voltage stability is used. The term voltage security means the ability of a system, not only to operate stably, but also to remain stable following any reasonably crediblebontingency or adverse system change such as load increases[2]. Voltage stability involves dynamics, but load flow based static analysis methods are generally used for quick and approximate analysis. Figure 13.7 depicts how voltage stability can be classified into transient and longterm time frame l2l.
Transientvoltage stability Inductionmotor dynamics Generator/excitation dvnamics Longerterm voltage stability Increase in load/powertransfer
LTCtransf& Distvolt.Reg.
Load diversity /thermostat Excitationlimiting
Primemovercontrol
Mech. switched capacitors/reactors
Gasturbine startup
Under voltageload shedding
', ;,ilir; ;;,:;;
;agestability. lf phenomena.Becadseof this, voltage 'ent engineers. Voltage instabilitv and rterchangeably.by many ."r"ur"h..r. voltage instability or co[apse is a faster dynamicprocess. As opposedto angle
ffiri#il;";
Protective relaying includingoverload protection
100 Timeseconds Fiq. 13.7 Voltage stabilitVphenomenaand time responses
SPfti*f uooern po@is I Voltage stability problems normaliy occur in heavily stressed systems. voltage stability and rotor angle (or synchronous) stability are more or less interlinked' Rotor angle stability, as voltage stability is affected by .a3.wex reactive power control. voltage stability is concern"o *itt, load areas and load
rvr^rvlv vv'Yvr Pr'lurs tu a large system over long transmission lines. Voltage ':::i:y,"::,*:,','31^
wffi;Ewe
desirable.
;;;''',h.;A#";
Unity powerfactor (knee) Nosepoint
t
^,?::^s,112itiu Seneratorstability.In a large interconnectedsystem,voltage coilapse of a load area is possible without lois of synchroni* or any generators. The slower forms of voltage instability are often analysed as steadystate problems' 'snapshots' in time following an outage or during load buildup are simulated' In addition to postdisturbancl load floivs, two other load flow based methods are widely used: PV curves and ev curves. pV curves are used for especially for radial systems. / curves (Fig. 13.g), eV curves (Fig. 'ig. 13.8) and methods to quantify nose puted. Power flow analysis cletermines
voltage exist for each value of load. The uppe. on" indicates stable ;;i;;; whereas lower one is the unacceptable value (multiple load flow). At limiting of voltage stability i.e. at nose point single l!ug" ioad flow solution exists. Nearer the nose point, lesser is the staLility maigin, Effective counter Instability Measures to prevent or contain voltage
and ,oto."ungl"stabilityis basicauy
\
Loaw of V6 and Pr"r; same for const Z load for other types of load, V degradationis faster.
PplP6zy
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
as vanous parameters svstem and ""lT:ir'#n#'f
HI'I,:IJHI#
Fig. 13.8 PV curveswith different load powerfactors Only the operating points above the critical points represent satisfactory operating conditions. At the 'knee' of the VP curve, the voltage drops rapidly with an increasein load demand. Powerflow solution fails to convefgebeyond this limit indicating instability. Operation at or near thel stability limit is impractical and a satisfactory operating condition is ensured by permitting sufficient "power margin". 1400 1200 1000
P> P, > P,, iNMW
(i) Generator terminal voltage shourd be raised. (ii) Generator transformer tap value may be increased. (iii) Qinjection should be carried out at an appropriate location. (iv) Loadend oLTc (onload tap changer) should be suitably used. (v) For under voltage conditions, strategic load shedding should be resorted to' System reinforcement may be carried out by instpiling new transmission lines between generation and load centres. series and shunt compensationmay be carried out and svcs (static var compensation) may be installed. Generation rescheduling and s_tartingup gas ,rrui"* of ,.y i" .,*.i.,t out. Practical aspectsof Qflow problemsleading to voltage collapsein EHv lines: (i)
For lons line.s urifh 'r^^nrrnu^r L^^uus's, recelvlng_end or road voltages "!' s'vv'Lr.,'e' increase for light load conditions and decrease for heavy load conditions. (ii) For radial transmission lines, if any loss of a line takes place, reactance goes up, I2x loss increases resulting in increase in voltage drop. This
j+Allowable
range
ffuo. Boo
Systemcharacteristics
Capacitor characteristics
0 . 9 0 . 9 51 . 0 1 . 0 5 Vin pu
should be suitably compensated local by e injection.of course this involvescost.If thereis a shortage rocal q of ,ourrr, thenimport of e
Fig. 13.9 Systemand shunt capacitor steadystate QV characteristics, capacitorMVAr shown at rated voltage
M
Voltage Collapse
; (v) Postdisturbance MwA4vAR margins should be ffanslated to predisturbance operatinglimits that operators monitor. can
(vi) Training in voltage stability basis (a training simulator) for control centre and power plant operators should be i
Voltage collapse is the processby which the seguence events accompanying of voltage instability le_ads unacceptable voltag6 profile in a significant part to of the power system. It may be manifested in ieverat different"ways. Voltage collapse may be characterised as follows:
SUIvIMARY
Power system security (including voltage stability) is likey to challenge planneqs,analysts; researchersand operators for the foreseeablefuture. As load grows, and as new transmission lines and new generations would be increasingly difficult to build or add, more and more utilities will face the security challenge. Deregulation and socioeconomic ffends compounded by technological developmentshave increasedthe likelihood of voltage instability. Luckily many creative persons are working tirelessly to find new methods and innovative solutions to meet this challenge.
(iii) The voltage collapse generallymanifests itself as a slow decay of voltage. It is the result of an accumulative process involving the actions and interactions of many devices, controlJ, and protective iyrt"*r. The time frame of collapse in such caseswould be of the order of several minutes. Voltage collapse is strongly influenced by system conditions and characteristics. (iv) Reactivecompensationcan be made most effective by the judicious choice of a mixture of shunt capacitors, static var system and possibly synetuonouscondensers. Methods of Improvlng Voltage Stabllity
NCES REFERE
Books L l.J. Narguthand D.P. Kothari, PttnterSystem Engincering,'fataMc0raw.Hill, New Dclhi, 1994. 2, C.W, Taylor, Power SystemVoltag,eStabiliry,McGrawHill, New york, 1994. 3. P. Kundur, Power SystemStability and Contol, Sections2.12, ll.2 and Chapter 14, McGrawHill, New York, 1994. 4, T,J.E,Miller, Editor, ReactivePower Control in ElectricSyslens,John Wiley and Sons, New York, 1982. 5. A. chakrabarti, D.P. Kothari and A.K. Mukhopadhyay, Perforurutnce, operation and Control of EHV Power Transmission Systems,Wheeler Publishing, New Delhi. 1995. 6. T.V. Cutsemand C. Vournas, VoltageStability of Electric Power Syslerns,Kluwer Academic Publishers,London, 1998. 7. A.J. Wood and W.F. Wollenberg, Power Generation,Operation, and Control, Znd Edn, John Wiley, New York, 1996. E. John J. Gratngerand W.D. Stevenson,Power SystemAnalysis, McGrawHill, New York, 1994. 9. G.L. Kusic, ComputerAidedPower SystemsAnalysis,PrenticeHall,New Jersey, 1986. 10. G.W. Stagg and A.H. ElAbiad, Computer Methods in Power System Analysis, McGrawHill, New York, 1968.
voltage stability can be improved by adopting the following rneans: (i) Enhancing the localisedreactivepower support (SVC) is more efl,ective and Cbanks are lnore economical. ratS devices or synchronous condenser may also be used, (ii) Compensating the line length reduces net reactance and power flow increases. (iii) Additional transmission linc muy be crectccl. also improves reliability. It (iv) Enhancing excitation of generator, system voltage improves and e is supplied to the system. (v) HVDC tie may be used between regional grids. (vi) By resorting to strategic load shedding, voltage goes up as the reactive burden is reduced. . Future
(i\
\./
Trends and Challenges
oifiurrrrr6 n'F IEA/'nr vr nv
i,
Ontirnql
,{^*,:^^^
\rlvv.ltvgD.
(ii) Betterand probabilistic loadmodelling. (iii) Developtechniques models studyof nonlinear and for dynamics largr of siz.e systems. example, For new methodsto obtainnetworkequivalentr suitable voltagestabilityanalysis. for
Papers
11. v. Ajjarapu and B. Lee, "Bibliography on voltage stability", IEEE Trans. on Power Systems,Vol. 13, No. 1, February 1998, pp lI5125, 12. L.D. Arya, "Security ConstrainedPower System Optimization", PhD thesis, IIT Delhi, 1990. 13. T.E. Dyliacco, "The Adaptive Reliability Control System", IEEE Trans. on pAS, Vol. PAS86, May 1967, pp 517531 (This is a key paper on system security and energy control system) 14. A.A. Fouad, "Dynamic Security AssessmentPractices in North America", IEEE Trans. on Power Systems, Vol. 3, No. 3, 1988, pp 13101321. 15. B. Stott, O. Alsac and A.J. Monticelli, "security Analysis and Optimization", proc IEEE, VoL 75, No. 12, Dec. 1987, pp 16231644. 16. Special issue of Proc. IEEE, February 2000. 17. P.R. Bijwe, D.P. Kothari and L.D. Arya, "Alleviation of Line Overloads and voltage violations by corrective Rescheduling", IEE proc. c, vol. 140, No. 4, July 1993, pp 249255. 18. P.R. Bijwe, D.P. Kothari and L.D. Arya, "Overload Ranking of Line Outageswith postourage generation rescheduling", Int. J. of Electric Machines and Power Systems, Yol. 22, No. 5, 1994, pp 557568. 19. L.D. Arya, D.P. Kothari et al, "Post Contingency Line Switching for Overload Alleviation or Rotation", Int J. of EMPg Vol 23. No. 3, 1995, pp 345352. 20. P.R. Bijwe, S.M. Kelapure, D.P. Kothari and K.K. Saxena,"Oscillatory Stability Lirnit Enhancement by Adaptive Control Rescheduliig, Int. J. of Electric Power and Energy Systems,Vol. 21, No. 7, 1999, pp 507514. 21. L.D. Arya, S.C. Chaube and D.P. Kothari, "Line switching for Alleviating Overloadsunder Line OutageCondition Taking Bus Voltage Limits into Account", Int. J. of EPES, Yol. 22, No. 3, 2000, pp 213ZZl. 22. P.R. Bijwe, D.P. Kothari and S. Kelapure, "An Effective Approach to Voltage 'Int. J. of EPES, Yol. 22, No 7, 2000, pp 4g34g6. Security and Enhancement", 23. L. Fink and K. carlsen, "operating under sttress and Strain", IEEE spectrum, March 1978, pp. 4850. 24. S.M. Kelapure, "Voltage Security Analysis and Enhancement", Ph.D. thesis, IIT Delhi,2000. 25. G.C. Ejebe, et. al, "Fast Contingency Screening and Evaluation for Voltage security Analysis", IEEE Trans. on Power systems,vol. 3, No. 4, Nov. l9gg, pp 15821590. 26. T. Van Cutsen, Voltage Instability: "Phenomena,Counter measures,and Analysis Methods", Proc. IEEE, Vol. 88, No. 2, Feb. 2000, pp 208227.
t4
T4.I
INTRODUCTION
State estimation plays a very important role in the monitoring and control of modern power systems. As in case of load flow analysis, the aim of state estimation is to obtain the best possible values of the bus voltage magnitudes and angles by processing the available network data. Two modifications are, however, introduced now in order to achieve a higher degreeof accuracy of the solution at the cost of some additional computations. First, it is recognised that the numerical values of the data to be processed for the state estimation are generally noisy due to the errors present. Second, it is noted that there are a larger number of variables in the system (e.9. P, Q line flows) which can be measured but arc not utilised in the load flow analysis. Thus, the process involves imperfect measurements that are redundant and the process of estimating the system statesis based on a statistical criterion that estimates the true value of the state variables to minimize or maximize the selected criterion. A well known and commonly used criterion is that of minimizing the sum of the squares of the differences between the estimated and "u1le" (i.e. measured) values of a function. Most state estimation programs in practical use are formulated as overdetermined systems of nonlinear equations and solved as weighted leastsquares(WLS) problems. State estimatorsmay be both static and dynamic. Both have been developed for power systems.This chapter will introduce the basic principles of a staticstate estimator. ln a power system, the state variables are the voltage magnrtudes and phase angles at the buses. The inputs to an estimator are imperfect (noisy) power The estimator is designedto give the "best estimate" of system measurements. the system voltage and phase angles keeping in mind that there are errors in the The output measuredquantities and that there may be redundant measurements. data are then used at the energy control centres for carrying out several
or,01li,r:, systeT^studics as cconornic such clisplrch (Chaprcr 7), ::if.jl:t a n a l y s t s ( L n a p t e r strLurly
lJ).
to An lntroduction State Estimationof Power Systems
J =7'V
/ r1. J ,(, . 5 ) ( +/ 4 \ \
r42 LEAST souARES ESTIMATION:THE BAsrc soluTroN I7l IeI
ns wilf be seen iater in Section r4.3, the probrem of power system state estirnation is a special case of the more general problem of estimation of a random vector x from the numerical values of anotirer related random vector y with relatively little statistical information being available for both x and y. In such cases,the method of leastsquarederror estimation may be utilised with good results and has accordingly been widely employed. Assume that x is a vector of n random variabres x1tx2, ..., x' that y is another vector of m (> n) random variables !1, J2, ..., J^ and both are related as !=Hx+r (r4.1) where H is a known matrix of dimensionmx n ancl is a zero mean rancklm r variable of the same dinrension y. The vector as x represents the variablesto be estimated,while the vector y representsthe variables whose numerical v aluesa re a v a i l a b l eE q u a ti o n l 4 .l ) suggests . ( thatthe measurement vectory i s linearly related to the unknown vector x and in addition is corrupted by the vector r (error vector). The problem is basicallyto obtain thc best possible value of the vector x from the given values of the vector y. Since the variable r is assumedto be zero m ean,o n e ma y ta k e th e e x p e c ta ti o n f E q. (14.1) o anclget the rel ati on f = Hx (r4.2) I , , = cxpcctcclvalue ol' x iurd y, respec:tively. This shows that the load flow methods of chapter 6 could be used to estimate the mean valuesof the bus voltages.Howeu"r, tn. woulcl like to estimatethe actual values of bus voltagesrather than their averages. One possibleway of obtainingthe bestpossibleestimate of the vectorx from y lies in the use of the method of leastsquareestimation (LSE). To developthis rneth<td, assllmethat i rcpresettts desirecl the estirnateof ,r so that y given by the equation whcre
A lrr
for From Eqs. (14.31) and (14.4),one gets the following expression the index: ( r4.6) J =yt! y'H*. i<'H'y+ *.'HtH*. For minimizing J = f$), we must satisfy the tollowing condition. gradoJ = 0
\r+.t)
It is easyto check (see,e.g. [1]) that Eq. (1a.7)leadsto the following result. HtH*,H'yO
(14.8)
'notmal equation' and may be solved explicitly This equation is called the for the LSE of the vector i as
*. = (H,Il)t Ht y
(14.e)
E;il;il] r .
i
ln orclerto illustrate the methodof LSE, let us considerthe simple problem of estimating two random variables x, and .rz by using the data for a thret: vector y. dimensional
Assume
l' 0.1 H=10 ll
Lrrl
I is u'u =12 I unaits inverse Ll 2) l
The matrix Ht F1 is then given bY
=f u3 :':1 (H'm' ''':' 2131
^
f rrt lat . /1
It is easy to form the vector Hty and combining this with the itrverse of (H'H), the following estimateof x is obtained.
\Ltr)/r\rrt)\/2Yz)  * = /3)yr*(2/3)yr1(r/3)v, l L(t
t1\/ r l
j' = Hf,
(r4.3)
Weighted
mI ..
representsthe estimateof the vector y. The effor ! of the estimation of y is then given bv
LSE
Eilv€il !rl D9. \!+,2) ls ulrerr lrrsrrsu r\, (ri) rrru \rr\.rrrr(uJ r1,61ir
lnq
grsllll:.iltr
(r4.4)
The estimate i is definecl irc thc IsE if it is to cornputccl rninirnizingthe by estimation index J given by
squaresestimate and is obtainedby minimising the index function that puts of equal weightage to the effors of estimation of all components the vector y. weightageson the ditt'erentcomponentsof put.dift'erent It is often desirableto may be more reliable and accuratethan the y since some of the measurements should be given more importance.T'o achievethis we define othersand these the estimation index as
rvtouern Power
I = l'wf
/ 1 4 1 n \ \ r.i. r vi,
An lntroductionState to Estimation Power of Systems l" ,3iiSl: T
where W is a real symmetric weighting matrix of dimension m x ru. This is often chosen as a diagonal matrix for simplicity. It is relatively straightforward to extend the method of LSE to the weighted form of "I and to derive the foilowing form of the normal equation. HWH?_ H'Wy _0 This leads to the desired weighted least squares estimate (WLSE) *. = (H,W H)t H,W y (14.11b) This pertains to minimization as the hessian 2HWH is a nonnegative definite. Some Properties: Rewriting Eq. (14.ilb) as G4.Ila)
[o.r w =t l
I
I
L
The matrir HLWH ts
I o.tj
l
l''' H,wH
o'tl
10.1 1.lJ
and the matrix HtW is obtained as
H '  w = f o 'o o t  l t
L0 1 0.1j
Theweighted leastsquares estimate the vector is thenobtained (from of x as E q .( l a . 1 1 b ) ) ^ *=[ L ellr) y, (lr/21) y,  (r}/Zt) y, (t0/2t) yr
where
i=ky k = (H'WH)' Ht W.
(I4.I2a).
(r4.rzb)
Here the matrix k dependson the value of H and the choice of w. using Eqs. (14.1) and (r4.lzb) it is easy to get the relation as follows.
(2ot2r) y,
(r/2\ y, )
I
r
Or and
=f::;,7, ,,,wH) + kr x
(14.r3) (14.r4)
X =r + kr E{i} =E{xl
If this result is compared with the result in Example I4.1, the effect of introducing the weighting on the estimate is apparent.Note that the choice of W in this case suggeststhe data for y2 is consideredmore valuable and this resultsin the components of x being more heavily dependenton )2. The matrix ft is in this case found to be (Eq. (1a.12b))
ln IJq. (14.14) it is assumedthat the error r is statistically independent of columns of H and the vector r has a zero mean. An estimate that satisfied Eq. (l 4)4) is calledan unbiased estimate. This implies that the estimation error ls zero on an averase. x =kr The covariance of the error of estimation is therefore given by P, = KRK (14.15a)
rc=lrr/21
L rlzt
ro/2r ro/zrf
20/21 r/21J
If the covariance of the measurementerror is assumedto be R = L the covariance the estimation error is obtained as (Ref. Eq. (14.15c)) of
(14.1sb)
) The choice of W above yields unacceptablylarge estimation error variances. Let us now choose the weighting matrix W = I. The matrix Kis then obtained
AS
P,= (1tr47)*:_ .::1  r34 L67
where R is the covarianee the error vector r. Note that the covariancep" o1 is a lneasureof the accuracy of the estirnaiion and a smaiier trace of this matrix indicatesa better estimate.Eq. (14.15b) suggests that the best possiblechoice of the weighting matrix is to set w  R1.The optimum value of the error covariance'matrix is then given by
P,  (HtRrH)l (14 15c)
*=l''3
1/3t/3f
L U3 2/3 1/3J
6 3]
The error covariance matrix is then given by
D / \ r  x = \ r / tn> ) [ L L_ E 6 J The error variances are now seen to be much smaller as is to be expected.
** t*1Et<ample14;2 Assumethat in the Example I4.1, we want to obtain the WLSE of the variable x by choosing the following weighting matnx
Nonlinear
Measurements
The case of special interest to the power system state estimation problem corresponds the nonlinear measurement to model.
536  I
I
I
Modern Power SystemAnalysis
An lrylfgduction to_State Estimation PowerSystems of
I S3T
I
(14.16) !=h(x)+ r where h(x) represents rz dimensional vector of nonlinear functions of the an variablex. It is assumed that the components the vector h(x) arecontinuous of in their arguments and therefore may be differentiated r.vith respect to the components x. The problem is to extendthe methodof least squares order in of to estimate the vector x from the data for the vector y with these two variables being relatedthrough Eq. (14.16). case, assumethat i To mimic our treatment of the linear measurement y represents desired estimate so that the estimateol the measLlrement could the be obtained using the relation
useful result in the sensethat it shows a mechanismfor improving on the initial estimafeby making use of the available nleasurements. Having obtainednew estimatei , the processof linearisationis repeatedas many times as desiredand this leads to the {qllq*tng 4erylle {qryq qf the qqb]ro4 of thq 4onlinear estimationproblem.
i ( t + 1 ) = i ( t )+ K ( t ) { y  h t t ( / ) l }
where the matrix K(D is deflned as
(r4.23) (r4.24)
K(t)  LH/ wHil' al w
I = h(i)
This yields the error of estimationof the vectol y
(1.4.17a) (r4.r7b)
j =y_ h(3)
the index of estimation In ordcr to obtain the WLSE of ,,r,we nrust cho<lse J as follows.
J=U
h(fi)l'WlJ h(i)l
(14.18)
The necessarycondition for the index ,I to have a minimum at .x,is given by Eq. (14.1e).
The index / representsthe iteration number and H lrepresentsthe value of the Jacobian evaluated at x = i (l). Usualiy the iterative process is terminated wheneverthe norm of the differenceof two successive values of the estimate i 1t + l)  .i (/) reachesa preselecred rhresholdlevel. A flowchart for implementing the iterative algorithm is shown in Fig. 14.7. A major sourceof computationin the algorithmlies in the need to updatethe Jacobianat every stageof iteration.As discussed earlier in Chapter6 (seeEqs. (6.86) and (6.87)) it is often possibleto reducethe computations holding rhe by value of H a constant,possibly after / exceeds2 or 3. T,ris is in general, permissible in view of the fact that the change in estimate tends to be rather small after a couple of iterations.
tyh(x)lH(i)=o
(r4.re)
where H (fr) is the Jacobianof h (x) evaluatedat i. In general this nonlinear equation can not be solved for the desired estimate, i. n way out of this technique. Let us assumethat an difficulty is to make use of the linearisation a priori estimate xu of the vector x is available (say from the load flow solution). Using Taylor seriesapproximation,we get
! = h ( x d + H u &  x 1 ) +r
(14.20)
where 1/o standsfor the Jacobian evaluated at x = x9 and the noise term r is now assurned include the effects of the higher order terms in the'Taylor to series.Equation (14.20)can be rewrittenas: (14.2I) A ) ' = y  h ( x o ) =H s A x + r where Ay is the p^rturbed measurementand Ax is the perturbed value of the vector r. An \\LSE of .r is then ea^silv obtainedasdiscussedearlier and this leads to the desired expressionfor the lineaized solurion of the nonlinear problem. estjmat.ion
t .i. = .rrr [H,,' I+'HQ] H()' ]l' {r  h t.t,,,)}
(+..227 i ,00"," ) eq.
c"l::'1: ) \,* ! _I\,,ll ,=n
./
:\
I
/
''!'
ls
.2a.,
r11.22)
No
from E q. (11.22r i s goi ngto be I r is no r l i k e l v th a t th e e s ti ma te.i o b tai ned (Dtirnal ydue rf rhe \ect()r x. Hrt,*'ever, F,q.(14.22) prot'ides us wirh a very
I t  r fI r t u c i r u s e s i l t . ' c . i r t t e n e r . r i . t l t r ' . r i , t ' r t t t ' t c ' s t i r l u t e . l ' , . rl ) J \  I l r r t h ' c l o s e t i r t h f
r Yes
\
Stcp
rig. r+.1
',:83E, I
I
todern powerSystem Analvsis
of to An tntroduction State Estimation Power Systems
 539 _r
Consider simple caseo1'ascalar variablex and assunlethat the relationship the
_xls glven Dy
It is thus apparentthat the problem of estimation of the power system state 'is a nonlinear problenn or and may be solved using either the batchpr:ocessing Section 3.3 of Reference 1]. Also, if the sequentialprocessingformula [see condition. the voltageangles to svstem is assumed have reacheda steadystate problem is then a static problem and the methodsof Sec. 14.2 may be used. if so desired. To develop explicit solutions, it is necessaryto start by noting the exact forms of the model equations for the components of the vector y (k). Let P, arfi Qi denote the active and reactive power injections of ith bus. These are related to the components of the state vector through the following equations.
Y = x 3+ r The JacobianH, is easily obtained in this caseand the iterative algorithm takes the explicit form
i (t + 1)= i (t) + t3i (Dlz {y  ti (Dl3}
where we have used W = 1. Let the correct value of x/ be eclualto 2 and assumethat due to the effect of r the measuredvalue of y is found to be 8.5. Also, assumethat the initial estimatex (0) is taken to be equal to 1. The table below gives the results of the first few iterations.
P,= D
N
I Yj   Vjl lYij I cos ( 6t + 6, *
N
4ti)
(r4.2s) (r4.26)
j r
, (t)
0 I 2 1.0 3.5 2.56 2.t6
e,= D
j:1
 % l I v j l l Y i j l s i n(  6 , + 5 t + 0 ; )
_)
a
It is apparent the algorithmwould yield the correctsolutionafter several that iterations. 14.3 STATIC STATE ESTIMATION OF POWER SYSTEMS
the uragnitudeand l/U representsthe angle of the where I IU I repre5ents admittance of the line connecting the lth and7th buses.The active and reactive componentsof the power flow from the ith to the7th bus, on the other hand are given by the tbllowing relations. 9 P,j=  vil I Vjl I Yij I cos (d,  6i * 0,)  l v i P l Y , , l c o s , ,
(r4.27)
Q , j = l % l I V j l l Y i j l s i n ( d ' 6 i * 0 t ) I viP I Y,,I sin Iij
lrollr2l
As noted earlier, for a system with N buses,the state vector x may be defined as t hc 2N   v c c l tl rtl l ' l l tc N   v o l l agcangl cs62,..., 6" and thc N vol tage magnitudes/1, v2, ..., v". The load flow data,dependingon type of bus, are generally comrpted by noise and the problem is that of processing an adequate set of available data in order to estimate the state vector. The readily available data may not provide enough redundancy (the large geographical area over which the system is spreadoften prohibits the telemeteringof all the available tnedsurements the central computing station).The redundancyfactor, defined to as the ratio m./n should have a value in the range 1.5 to 2.8 in order that the computedvalue of the statemay have the desiredaccuracy.It may be necessary to irlehrde data for the power flows in both the directions of some of the tie the lines in order to increase the redundancy factor. In fact, some 'psuedo measurements' which represent the computedvalues of such quantitiesas the active andreactiveinjections some remotebusesmay alsobe includerlin the at vector y (k).
(14.28)
Let us assume that fhe vector y has the general form J = lPt . . . PN Q t . . . Q r u Pr z . . . Pu )2 l, N Qn ... Q7,J 1, 7s, 
ll b i n v 1 l , . . . . ,I y N I l /
(r4.2e)
The Jacobian H will theh have the form Ht H3 Hs H_ H7 Hz H4 H6 Hs
(14.30)
/rur o
o IN
(N  l; submatrix where /" is the iclentity matrix of dimension N, H, is the N x power injections wrt 6's, H2 is the of the partial derivatives of the active
54q f
Modern Po
T
N x N submatrix of the partialderivatives of the active power injections wrt  7l' and so on. Jacobian H will also be a sparsematrix since I is a sparse matrix. Two special casesof interest are those correspondingto the use of only the active and reactive iniections and the use of onl flows in the vector y. In the first case, there are a total of 2N components of y compared to the 2N  1 components of the state x. There is thus almost no redundancy of measurements. However, this case is very close to the case of load flow analysis and therefore provides a good measure of the relative strengthsof the methods of load flow and state estimation.In the second case, it is possible to ensurea good enough redundancy if there are enough tie lines in the system. One can obtain two measurements using two separatemeters at the two ends of a single tieline. Since these two data should have equal magnitudes but opposite signs, this arrangement also provides with a ready check of meter malfunctioning. There are other advantages this arrangement of as will be discussed later. The Injections Only Algorithm
I 541' I (14.33) may be used to determinethe Jacobianat any specifiedvalue Equation of the systemstatevector. The injections only stateestimation algorithm is then obtained directly from the results of sec 14.2. Since the problem is nonlinear, it is convenientto employ the iterative algorithm given in Eq. (14.22). ine. the submatricesH., andH^ become null result that the linearised model equation rnay be approximatedas: with the of to An tntroduction State Estimation Power Systems
^
fH, ol
'
rol
(r4.34)
parrition,J";,1:' rrwe ,,o*^::::
= Av f A y , 1 '= l A r r l '= l j l^'r'r.j' Lor"l, L;
then, Eq. (14.34) may be rewritten in the decoupledform as the following two separateequationsfor the two partitionedcomponentsof the statevector (14.35) Alp= H1 Ax5+ ro Alq= H4 Axr , + r n
(r4.36)
In this case, the model equation has the form
!=hlxl+ r with the components the nonlinear of functiongivenby
(14.31)
Based on thesetwo equations,we obtain the following nearly decoupledstate estimation algorithms.
i u Q + r ) = x d 0 + t H i 0 w p H tU ) l  ' H , ( D { t o  h r l i U ) l }
.l= 0, 1,2, ...
\ (14.37)
(14.38)
h,[x)= D
N
l V i l l V j l l Y i j l c o s( 6 ,  6 i + 0 i ) ,i = I , . . . , N
i, (/ + 1)= i, Q) + tHl u) w, Hq (j)l' nl u) wo ur hq ti (j)ll
j = 0  2
p where the subscripts andq are usedto indicatethe partitionsof the weighting and the nonlinear function h (.) which correspond to the vectors yp matrix W *rd lerespectively. As mentioned earlier, if the covariances R* 11d Rn of the effors r, and r(t are assumedknown, one should select Wo = R o' and Wn =
Rq"
j:1
Qa32a)  D t v * _ i l l v j l l y N _ i , ; l s i( 6 ,  6 i +7 i i ) n i=N+1,N+2...,2N
j:1
(r4.32b)
The elements of the submatrices Hp H2, H, and Ho are then determined easily as follows. H t ( i , j ) = l V i l l V j l l Y i j I s i n ( r { . 6 t+ 0 t ) i = 1 , 2 , . . . ,N , j=I,2,"',N1 = l v i l l Yi j I c o s (d,  6i * 0;) i = 7,2, ..., N , H z Q ,i ) j = 1,2,"', N'
Ia3 \tt J)=
r f / r ! \


r t I I t r   r ,  yil I vjl I IUI

COS \Oi.Oj+
/
C
a

Aij) I =
n
\
.
n
I,
A
Z) ..., lY,
j=I,2,"',N1 j H q ( i , ) = l V i l l Y i j l s i n( {  6 i + 0 , ) i = I , 2 , . . . , N , j = I,2, "', N' (r4.33)
that Eqs. (14.37) and (14.38) are not truly decoupledbecausethe partitions of the nonlinear function dependon the estimate of the entire state vector. It may be possible to assume that vi U) = I for all i and 7 while' Eq. (14.37) is being used in order to estimatethe angle part of the statevector. similarly one may assume 6i U) = 0o for all i and 7 while using Eq. (14.38) in order to estimate the_yllltage part of the state vector. Such approximations allow the two equationsto be completely decoupledbut may not yield very good solutions.A betterway to decouplethe two equationswould be to use the load consnnt vaiues in Eq. (14.37) flow sglutions for x, and x6 as therr suppose<iiy and Eq. (14.38) respectively. There are several forms of fast decoupled (seee.g. [13], tl4l).A flow estimationalgorithmsbasedon suchconsiderations chart for one schemeof fast decoupled stateestimation in shown in Fig. I4.2.
'Note
i
542 
Modern Po*er System Analysis
of to An Introduction State Estimation Power Systems
 543 1_
F i g .1 4 . 3
*11
Application of the LSE then yields the following expressions for the estimatesof the perturbationsin the three state variablesaround their chosen initial values:
A i ,  AP,  AP,
I , o , r l l i 1 i * 1 t )  * ( j ) l l = , 'j
on,^t t\u
AT,= 0.78 AQz  0.26AQr
AV' = g39 AQ ' + 0'14 AQ r values of These equationsshould be used in order to translatethe measured int power inject ions o t heist inr at es in thc pcrturhat ions t he act iveanclr eact ive of of the perturbations the state variables. partitionsHranrl Htare to It is interesting note that for lhe sirnpleexample, are estimators the sameas thosegiven that the decoupiedstate null matricesso above. The Line Only Algorithm
ls
,2 \ an2
>
\\r'
\ (3'9)
Fig. 14.2
,
1"", I
j Example14.4
In order to illustrate an applicationof the injections only algorithm, consider the "imple 2bus system shown in Fig. 14.3. A s s u m i n go s s l e s lsi n e , 0 , j = 9 0 " . A l s o l e t Y r r = j z z = 2 a n d y n = y z t = l . The power relations in this casewould be Pt=   Yr  V2l lYrrl sin E Pz=lVtl lVzl I Y,rlsin 6, Q t = i r i l i i V r i '   y n l l y l   V r l c o s$ Q z = l Y z z l V r l '  l y t 2 l l y r   v r l c o s6 I f we c h o o s e e i n i ti a lv a l u e s Vf l = l V ; l = 1, 6o2= 0" , the correspondi ng th l power values are P f = Pf  0, Qf = Qi = 1. The value of the Jacobian matrix evaluated at the above nominal values of the variables turns out to be
I
This algorithm has been developed in order to avoid the need for solving a nonproblem.which as seenearlier,requiressome approximation linear estimation ln the line tlow only algorithm,the data tor the active and reactive tie or other. the in line flows are processed order to generate vector of the voltage difference acrossfte tielines.Let z denote this vector.A model equationfor this vector as may be expressed z= Bx + r
(r4.39)
incidencernatrix and r is the vector of the errors where B is the nodeelernent Since this is a linear equation.one may use the WLSE in the voltage elata. techniqueto generatethe estimate as
 *, = lB, wBl Bt wz
(r4.40)
where the weighting rnatrix may be set equai to the inverse of the covariance of rrif this is known. The main problem with Eq. (14.40) is that the vector z is not directly measurablebut needs to be generatedfrorn the tie line flow data.
r++
 I
power System Analysis Moctern
An lntroduction state Estimation powersystems to of Network Obsenrability [17]
V,tdenotesthe voltage acrossthe line connectingthe ith ancltheTth buses,the followine relation holds. Vi i = Z i i [@ i i  j Qi ;tV 7 V i Y ,i ] Here Z,.,stands the impedance the line. for of This shows that the vector z is related to the vectorsx and
ashion and one ffidy use the notation
 545 T
(14.41)
z = g (x, y) 04.42) In view of this nonlinear relation, (14.40) Ecl. maybe expressecl theform in i = LB,Wnyr B,W I (i, y) (14.43) This, being a nonlinear relationship, not be solvedexceptthrougha can numerical approach (iterative solution). The iterative form of Eq. (14.43)is
Considerthe staticWLSE formula lEq.(14.llb)l which servesas the srarting point fbr all the algorithrns.lnverse of information matrix Mn,, = Ht wH should exist otherwisethere is no stateestimate.This will happenif rank of f/ ls equal to n (no. ot state vartables).Since one can always choose a nonsingular l4l, so if lt1 has a rank n, the power network is said to be observable. Problem of lllconditioning
i (j + r) = lBt wBll H w s [i (i), y],  /= 0 , 1 , 2 ,. . .
(r4.44)
Note that the original problem of estimation of x from the data for z is a linear pr ob l e m s o th a t th e s o l u ti o ng i v e n by E q. (14.40) i s the opri nralsol rrti on. However, the data for z needto be generatedusing the nonlinear transforma_ tion in Eq' (14.42), which in turn has necessitatecl use of iterative Eq. the (14.44)' Compared to the injections only iterative algorithm, the prese1t algorithm has the advantage a constantgain matrix [8, W B]t gr I4z. of This resultin a considerable computational simplification. The concept decoupled of estimationis easily extencled the case of the lirre flows [15]. to T4.4 TRACKING STATE ESTIMATION SYSTEMS t16l OF POWER
Even if the given power system is an observable system in terms of the measurements selectedfor the state estimation purposes,there is no guarantee that the required inversion of the information matrix will exist. During multiplications of the matrices,there is some small but definite error introduced due to the finite word length and quantisation.Whether or not these errors create illconditioning of the information matrix may be determined from a knowledge of the condition number of the matrix. This number is defined as the ratio of the largestand the smallest eigenvaluesof the inforrnation rnatrix.The maftix M becomesmore and more illconditioned as its condition number increases magnitude.Somedetailedresultson power systemstateestimation in using Cholesky factorizationtechniquesmay be fbund in tl8l. Factorization helps to reduceillconditioning but may not reducethe computationalburden. A techrrique reduce computationalburden is describecl Ref. to in []91. 14.6 EXTERNAL SYSTEM EOUTVAIENCTNG [20]
Trackingthe stateestimation a given power systemis importantfor real tirne of monitoring of the system. is well known, the voltagesof all real systemvary As randomly with time and should therefore be considered to be stochastic processes' is thus necessary make use of the sequentialestimation It to techniquesof Ref. [1] in order to obtain the state estimate at any given time point. The power relationsin Eqs. (14.25) and, (74.26) are still valid bur must be rewritten after indicating that the voltage rnagnitudes and angles are new functions of the discrete time index ft. I4.5
Brrfh the
One of the widely practiced methods used for computationalsimplification is to divide the given system into three subsystems shown in Fig. 14.4. One of as these is referredto as the 'internal' subsystemand consistsof those buses in which we are really interested. The secondsubsystem consistsof those buses which are not of direct interest to us and is referred to as the 'external, subsystem'Finally, the buseswhich provide links betweenthese internal and external subsystems constitute the third subsystem referredto as the 'boundary' subsystem. For any given power network, the identification of the three subsystems may be done either in a natural or in an artiflcial way.
t,
/
SOME COMPUTATIONAL
qtafin qnd fhe trqo r 16 r r u v rl \z lrn o o r )frirmro qiln u l r v c r rf r 
,/
\
,
CONSIDERATIONS
^1^*i+L*^ ai^,r^61^r Ctr5\Jl tLllll.lJ PIEi)t'f lttrlt illl rL^ tll€ ^r: pICL:CUlllg
sections are computationally intensive, particularly for large power networks which may have nlore than200 importantbuses.It is, therefore,very important t o pa y a tte n ti o nfo s u c h c o m p u ta ti onalssuesas i l l condi ti oni ng, i computer storageand tinnerequirements. However, we need to first considerthe question of existenceof a solution of the state estimation problem.
\"
Internal system Boundary system
,/
Flg. 14. 4
fl6 l
I
Po Modern
An lntroductionto state Estimationof Power systems Bad Data Detection [231
t{7 I
To illustrate the simplification of the stateestimation algorithm, considerthe equationfor the injections only case.Since the system measurement linearised this equation may be written as is partitioned into three subsystems, Ayi
H,,
Hu, 0
H,,
Huu Hra
o
Ho, H""
A*, A*, A*,
Ayr,
,ol ,"]
(r4.4s)
ay"
It may be noted that the internal me ..rrerrrert vector Ayt is not completely of inclepenclent the external subsystem state Axu since Ay, dependson the boundary subsystetn state Axr, and Ax6 depends on Axr. Ayr= Ayt,i + Aynr,+ Ayu, + ru
(r4.46)
the where Ayo" represents injections intc the boundary buses from the external Ay66is the injection from the boundary buses znd Ay6, is the injection buses, that the term Ayu, (= Ht,, Axr), may be fiom the internal buses.It is assurned approximatedas n A*, where A It estimated from the relation
H = Ayt,nlAxl,
(14.47)
The conrponent Ayu" may be estimated if the terms Ayy, and Ayuo ne Ax1, and then subtractedfrom the measuredvalue computedas H,,, Ax,and H1,1, woulc! result in part of Eq. (14.45) to be rewritten as of Ayo. This
',;:;)l*,1 .[;] l^;,)=l',;;,
Ol. BAD DATA 127,221
( r4.48)
ivc fhc whcrc Ht,, = H t,t,t U pcrprcscrnts cffccf Jacohiltn if the hottnclitry the external subsysternon the that accounts for the eff'ects of subsystern Equation(14.48) has a lower dimensionthan the original boundarysubsystem. l s. u cc t nc a s u rc l n c rtt l L rl ti o n n d w o rrl rltl tcrcl orc i rtvol vcrcss ctl tttl l tr(.tl i ol lTl tc may be employed with the line or conceptof external systemequivalencing also. rnixedCatasituations 14.7 TREATMENT
A convenient tool fcr detecting the presenceof one or more bad data in the vector y at any given point of time is based on the'Chi Square Test'. To appreciate this, flrst nc:te that the trrethod tll least square ensures that the /ltgr haritrminimum rflLW Ta: 1r (t)lvalue when x  i. Since the variable r is random,the minimum value ,I.in is also a random quantity. Quite often, r may be assumed to be a Gaussian variable and then .I*1nwould follow a chi squaredistribution with L = m n is degrees of freedom. It turns out that the mean of ./,o1n equal to L and its variance is equal to 2L. This implies that if all the data processedfor state should be close to the estimatiorrare reliable, then the computed value of ../r,r,n averagevalue (=L). On the other hand, if one or ntore of the data for )' tre estimation are violated and of unreliable, then the assumptiorrs the least squares the computed value of J*1nwill deviate significantly from I. for It is thus possibleto developa reiiable schetrre the detection of bad data  h(r)], i being the estimate of [y  h(i)l'w ly in y by computingthe value obtainedon the basisof the concernedy. If the scalarso obtainedexceedssome threshold Ti  c L, c: being a suitable nuutber, we conclude that the vector y includes some bad data. (Note that the data for the.cornponenty;, i  1,2, ..., bacl nr will be consiclerecl if it deviatesfrom the tneanof r', by more than t 3e,. deviationof r;). Care must be exercisedwhile choosing where o, is the standard c.If it is closeto 1, the test maf producemany parameter the value of threshold 'talse alal1s' anclil it is ttrtr large, the test worrld fail to detect nlany bad data. To select an appropriatevalue of c, we may start by choosing the level d of the test by the relation' significance I'lJ (x) > cLlJ (.r) lbllows chi squaledistribution) = 6/ to We rnay select,fbr example,d = 0.05 which corresponds a 5Vofalse alarm of to [t is thcn possible find the valr.re c hy making use of the table sitrrltign. test X@).Once the valuc of c is determined,it is simple to cary out the cL. whether or not .l(x\ exceeds of Bad Data [231 Identification Once the presenceof bad data is detected,it is iraperative that these be before identified so that they could be removed from the vector of measurements it is processed.One way of doing this is to evaluatethe components of the that the residual !, = y,  h, (.x\, i = 1,2,..., m. If we assLlme measurement residualshave the Gaussiandistribution with zero mean and the varLance 4, then the magnitucieof the resiciuaiy, shouiciiie in the range 3o i ( )i ( 3or with 95Voconfidenceievel. Ihus, if any one of tlte cotnputedresidual turns out to be significantly larger in magnitude than three times its standarddevia,tion,then for corresponding data is taken to be a bad data. If this happenS more than one t r lur t t con)p ot r ct ol'_y, hcn t hc con'lponcnt vinglhc lar gcst csidualis assunt edo r( algorithmis rerun with be thc bad dataand is removedfrom y. The estimation the remaining data and the bad data detection and identification tests are
is The ability to detectand identify bad measurements extremely valuableto a load dispatch centre. One or more of the data may be affected by rnalfunctionsystemor both. or instruments the data transmission ing of either the measuring itscll rttay be rnay have becn wilcd incorrcctlyor tlto l.rattsduccr Transtlucers malfunctioning so that it simply no longer gives accuratereadings. If such faulty daia are included in the vector Ay, the estimationalgorithm of will yielcl unreliableestimates the state.It is therefbreimportant to develop techniquesfor detecting the presenceof faulty data in the measurementvector at any given point of time, to identify the fauity data and eliminate these from lbr the vector y befble it is processcd state estinration.[t is alstl ilttpoltant to rnodify the estirnationalgorithmsin a way that will permit tnore reliable state o i es t i ma ti o nn th e p !e s e n c e f b a d d ata.
,548 ,!
I
powerSystem Modern Analysis
performed again to find out if there are additional bad data in the measurement set. As we will see later bad rneasurement clataare detected,eliminated and replaced by pseudoor calculatedvalues. of Bad Data The procedures described so far in this section are quite tedious and time consuming and rnay not be utilized to remove all the bad data which may be presentin the vectory at a given point of time. It may often be desirableon the other hand to modify the estimation algorithms in a way that will minimise the influence of the bad data on the estimatesof the state vector. This would be possible if the estimation index J(x) is chosen to be a nonquadratic function. The reason that the LSE algorithm does not perform very well in the presence of bad data is the fact that becauseof the quadraticnature of J(x), the index assumesa large value for a data that is too far removed from its expectedvalue. To avoid this overemphasis the effoneousdata anclat the same time to retain on the analytical tractability of the quadratic: perlbrmanceindex, let us choose
549. I t The main advantage of the choice of the form (14.49) for the estimation in index is that it is still a quadratic the function g(.i) and so the LSE theory may be mimicked in order to derive the following iterative formuia for the state of to An tntroduction State Estimation Power Systems estimate.
Hr(D CTWCH Q)Tfl (D WCtfiDI (74.5ra) as are and its elements computed wherethe matrix C is diagonal
Ct = l, f ot l; lo, S ai = 0, f or lilo, > a'
(14.s b) 1
Comparing this solution with that given in Eq. (14.22), it is seen that the main effect of the particularchoice of the estimationindex in Eq. (14.49) is to ensurethat the data producingresidualsin excessof the thresholdlevel will not changethe estimate. This is achieved by the production of a null value for the matrix C for large values of the residual. I4.8 NETWORK OBSERVABILITY MEASUREMENTS AND PSEUDO
/(i) =s'(t) w sG)
(14.49a)
where 8(t) is a nonlinearfunction of the residual !. There may be several possiblechoicesfor this f'unction.A convenientform is the socalled'quadratic flat' form. In this case,the componentsof the function g (y) are defined by the following relation. g i (j ) = l i , = ai, fo r j ,l o , 1 a, ftx ' l ,/o ,> u,
(r4.49b)
where a, is a preselectedconstant threshold level. Obviously, the performance indcx ./("r)may bc cxpressed as
I(i) = Dy, c;)
; .  I I
m
(14.s0)
and each componenthas a quadratic nature for small values of the residual but has a constantmagnitudefor residual magnitudesin excessof the threshold. Figure 14.5 shows a typicai variation of J, (.r) for the quadratic and the nonquadratic choices.
A minimum amount of data is necessaryfor State Estimation (SE) to be effective. A more analytical way of determining whether a given data is enough for SE is called observability analysis. It forms an integral part olany real time state estimator.The ability to perform state estimation depends on whether are .sufficientmeasurements well distributedthroughoutthe system. When sufficient measurementsare available SE can obtain the state vector of the As whole system. In this casethe network is observable. explained earlier in Jacobiantlratrix is equal Sec. 14.5 this is true when the rank of measurentent to the number of unknown state variables. The rank of the measurement Jacobian matrix is dependent on the locations and types of available as rneasurements well as on the network topology. An auxiliary problem in state estimation is where to add additional data or to pseudomeasurements a power systemin orderto improvethe accuracyof the calculated state i.e. to improve observability. The additional measurements represent a cost for the physicat transducers,remote terminal or telemetry and software data processing in the central computer. Selection of sy.stem, filling of missing data,providing appropriateweightage pseudomeasurements, are the functions of the observability analysis algorithm.
UbsefvaDlllty Can De CIICCKCU UUtrrlB
I I  rI l:^ C^^t^i^+l^
lilulullzittLltrll.
Tf
Il 4IrJ Prv\rr
^.'
^i"^+
L^^^^.
'^.' r.rvtrrr.rrrlr .YvrJ
\
Fig. 14.5
small or zero during factoization, the gain matrix may be singular, and the systemmay not be observable. To finil the value ol an injection without nteasuringit, we tlrust know the currently being made.For example, we power system beyond the measurements MWs and MV ARs at generators through pormally know the generated telemetry channels (i.e. these measurementswould generally be known to the
An introduction state Estimation powqr systems to of
stateestimator). If thesechannelsare out, we can perhapscommunicate with the operatorsin the plant control room by telephoneand ask for these values and enter them manually. Similarly, if we require a loadbus MW and MVAR for a pseudo measurement' could use past recordsthat show the relationship we between an individual load and the total qyrtemload. We canestimate the tstal systetn load quite accuratelyby f)nding the total power being generated and estimating the line losses.Further, if we have just had a telernetry failure, we could use the most recentlyestimatedvaluesfiom the estimator(a.ssgming that i','is run periodic:ally) pseudorlreasureftlents. as Thus, if required,we can give . the state estimatorwith a reasonable value to use as a pseudomeasurement at any bus in the system. Pseudomeasurements increasethe data redundancyof SE. If this approach is adapted, care must be taken in assigning weights to various types of measurements. Techniques that can he usecl cleternrine rnctcror.pseu4g to the measurement locations obtaininga completeobservabilityof the sysremare fbr available in Ref. [251.A review of the principal observability analysis and m et erp l a c c rn c na l g ' ri th rn s s a v u i ra hrcn I(cr' . t i i [261. T4.9 APPLICATION ESTIMATION OF POWER SYSTEM STATE
I
I FSI t 
U ' E
t,
i,
L
bA l
!
o
9 F G .Yt
= t h
l m I
drx
F
]U
tr o t
o g a
CnF a trJ
I
I
o 1f o o o U)
I
Io
(! (E
I
o o o o
L
E
a
L n v ^ v
o (L I
I
c o o
o a U)
L
, Il '
In realtime environment state estimatorconsistsof different modules the such as network topologyprocessor, observabilityanalysis,stateestimationand bad data processing.The network topology processoris required for ail power systemanalysis.A conventional network topologyprogram usescircuit breaker statusinformation and network connectivity datato determinethe connectivity of the network. Figtrrc 1r4.6 it schcrnatic is cliagtarn showing the info.rnationflow between the variorrs functions to be performed in an operationscontrol centre compurer system. The system gets information from remote terminal unit (RTU) that cncodc lllcasLlrclllcnt tl'ullsduccr 0rrtputs opcnccl/closccl ancl staLus inlbmration into digital signals which are sent to the operation centre over communications circuits. Control centre can also transmit commands such as raiseflower to generators and open/closeto circuit breakers and switches. The analog measurementsof generator output would be rJirectlyused by the AGC program (Chapter 8). However, rest of the data will be processed by the stateestimator befbre being used fbr other functions such as OLF (Optimal Loacl Flow) etc. Before running the SE, we must know how the transmission lines are
rnnnenfprl fn l l ^^ L \ / fLrror u lrn\^/'c u ( r J r L rl i c r r r s r . u . ( r rl .,us c l i l . e . n e l w o r K r
a
E
. n F 
I I
t , 
q
o o I E o
L
A
'=.o
0(E
L
o
I
I I
I I
I
I
XH
z 8 F
;oii' >,
o a U) o
L
o
f a (!
,tl ncl l 6
i l_
o o
I
=
E
o) o
o)
i LLI > i o .!
I
LUI
q
sf E; ll
U)
F O o
6 E o
b 8g E E.E E E E
C O e C N'E
topology,
lhrs
changingand hcncel.hc r.clentetered currertt breaker'/switch stertus must be used to restructurethe electricalsystem model. This is called the network tr,tpop.tgy program or system status proce,\.tor or netwc;rk configurator.
kceps
6n
ilr l
I
Po Modern
An lntroductionto State Estimationof Power
The output of the stateestimator i.e. lvl, 6, P,j, Q,jtogether.with latestmodel form the basis for the economic dispatch (ED) or minimum emission dispatch (MED), contingency analysis program etc. Further Readin
The weighted leastsquares approach to problems of static state estimation in power systems was introduced by Schweppe 11969741. It was earlier originated in the aerospaceindustry. Since 1970s, state estimators have been installed on a regular basis in nern'energy (power system or load dispatch) control centres and have proved quite helpful. Reviews of the state of the art in stateestimation algorithms basedon this modelling approach were published by Bose and Clements[27] and Wu [28]. Reviewsof externalsystemmodelling A are available in 1291. generalisedstate estimator with integrated state,stutus and parameterestimation capabilities has recently been proposed by Alsac et al [30]. The new role of stateestimationand other advancedanalytical functions in competitive energy markets was discussedin Ref. [31].A comprehensive bibliography on SE from 196889 is available in Ref. [32].
F i g .P . 1 4 . 2
14.3Given a single line as shown in Fig. p 14.3, two measurementsare
available. using DC load flow, calculate the best estimate of the power flowing through the line.
4= 0 rad. ( ,r1 t tr
/1 
LEMS PROB
14.1 For Ex. 6.6 if the power injected at buses are given as Sr = 1.031 j0.791, Sz = 0.5 + 71.0and 'S3=  1.5  j0.15 pu. Consider Wt= Wz= Wz = l. Bus I is a referencebus. Using flat start, find the estimatesof lV,l and {. Tolerance= 0.0001. V\ lAns: Vl = l/0", V', = t.04223 10.4297", = 0.9982q l2.1864";
Meter
1
Mt2
y0.1pu (1ooMVAbase)
[l
Mzt
2
l Meter Reading
Fig. P. 14. 3 Full scale (MW) Meter Standard Deviation ( o ) in full scale
(Mw)
32 26
l1.356, Va  1.03831 Final values: Vt  1.04./.0", V2 = 1.080215 l 3.736"1. that the threemeters 14.2 For samplesystemshown in Fig. P. 14.2,assume have the followine characteristics.
Meter Mrz Mtt Mzz Full scale (MW) Accuracy (MW) o (pu)
Mrz Mzr
100 100
1 4
REFEREN CES
Books l. Mahalanabis, A.K., D.P. Kothari and S.L Ahson, ComputerAidetl pr;wer System Analysis and Control, Tata McGrawHill, New Delhi, 19gg. 2. Nagrath, I.J. and D.P. Kothari, Power SystemEngineering, Tata McGrawHill. Ncw Dclhi, I q94. 3. Mtrnticclli' A., State li.rtinrution in Eler:tric: Power Sysrems Gcnerali.recl A Apprctat:h, Kluwcr Academic Publishers, Boston, 1999. 4. Kusic, G'L., ComputerAided Power Systems Analysis,PrenticeHall, N.J. 19g6. 5. Wood, A'J. and B.F. Wollenberg, Power Generation,Operation and Control. Znd Ed., John Wiley, NY, 1996.
100 r00 100
 I lU{lDLll nmnnf t, t/l I lUl lL,)
+ 8 +4 + 0.8
0.02 0.01 0.002
d2 given the
Calculate the best estimate for the phase angles 4 *d
f,'11,...,:  1rl l\., W   l5
Meter
Meusured value (MW)
Mtz Mrz Mt,
70.0 4.0 30.5
'l
I
j
I
t
;554.;l
Analysls PowerSystem Modern

of to An Introduction State Estrmation Power Systems
_TI 555
6. Grainger, J.J. and W.D. Stevenson, Power System Analysis,McGrawHill, NY', 1994. 7. Deautsch,R., Estimation Theory, PrenticeHall Inc' NJ, 1965 g. Lawson, C.L. and R.J. Hanson, Solving Least SquaresProblens, PrenticeHall. inc., NJ., i974. g. Sorenson, H.W., ParameterEstimation,Mercel Dekker, NY, 1980'
26. Clements, K.A., "Observability Methods and optimal Meter Placement", Int. J. Elec. Power, Vol. 12, no. 2, April 1990,pp 8993. 27. BoSe,A. and Clements,K.A., "RealtimeModelling of Power Networks", IEEE Proc., Special Issue on Computersin Power Svstem Operations,Vol. 75, No. 12, Dee 1987;aP 76ff7=1ffi2: 28. Wu, F.F., "Power System State Estimation: A Survey", Int: J. Elec. Power and Energy Syst.,Vol. 12, Jan 1990, pp 8087. 29. Wu, F.F. and A. Monticelli, "A Critical Review on External Network Medelling for Online Security Analysis", Int. J. Elec. Power and Energy Syst.,Vol. 5, Oct 1983,pp 222235. 30. Alsac, O., et. aI., "Genetalized State Estimation", IEEE Trans. on Power Systems, Vol. 13, No. 3, Aug. 1998,pp 10691075. Management Dispatch and Congestion D. 31. Shirmohammadi, et. al., "Transmission Emerging Energy Market Structures",IEEE Trans. Power System.,Vol 13, in the No. 4, Nov 1998,pp 14661474. 32. Coufto, M.B. et. aI, "Bibliography on Power System State Estimation (1968Vol.7, No.3, Aug. 1990,pp 950961. 1989)",IEEE Trans.Power Sysr.,
Papers
F.C., J. Wildes, D. Rom, "Power SystemStatic StateEstimation,Parts 10. Schweppe, l, ll and lll", IEEE Trans', Vol. PAS89, 1970,pp 120135' ll. Larson, R.E., et. al., "State Estimation in Power Systems", Parts I and II, ibid, pp 345359. lZ. Schweppe,F.C. and E.J. Handschin,"static State Estimation in Electric Power System", Proc. of the IEEE, 62, 1975, pp 972982' and C. Rossicr,"A Fast DccouplcdStatic Statc H.P., J.C. Richarcl 13. Horisbcrgcr, for Electric Power Systems", IEEE Trans. Vol. PAS95, Jan/Feb1976, Estimator
pp 2 O8 2 1 5 .
1 4 . Monticelli, A. and A. Garcia, "Fast DecoupledEstimators", IEEE Trans' Power Sys/, 5, May 1990,pP 556564. 1 5 . Dopazo,J.F. et. al., "State Calculation of Power Systems from Line Flow PartsI and ll", IEEE Trans.,89, pp. 16981708,91, 1972,pp Measurcments, 145151. Dcbs, A.S. anclR.E. Larson, "A Dynamic Estimatorfor Tracking the Stateof a 16. Power System", IEEE Trans' 89, 1970,pp 16701678' A iZ. K1upholz, G.R. et. al.,"Power SystemObservability: PracticalAlgorithm Using Nctwork Topology", IEEE Trans. 99' 1980' pp 15341542' A lg. SirnoesCosta, and V.H. Quintana,"A Robust Numerical Techniquefor Power SystemStateEstimation", IEEE Trans. 100, 1981, pp 691698' Algorithm Row Processing "An Orthogonal A 19. SimgcsCosta, and V.H. Quintana, for Power System SequentialState Estimation", IEEE Trans., 100, 1981' pp 3'79r3799. 20. Debs, A.S., "Estimation of External Network Equivalentsfrom Internal System Data", IEEE Trans., 94, 1974, pp 12601268' A., A. Monticelli and P. Abreu, "Fast DecoupledStateEstimationand Bad 21. Garcia. Dara Processing",IEEE Trans. PAS98, Sept/Oct 1979, pp 16451652. E. 22. Handschin, et. al., "Bad Data Analysis for Power System State Estimation", IEEE Trans., PAS94, 1975, pp 329337.t2
4J.
I
i i
r.^r:6 r\V6lrrrt
rJ r . r Lt.
eL.
or @ e . , nl
"Flqd T)cre Detecfion
and ldentification".
Int. J. EleC. POWer,
Vol. 12, No. 2, April 1990, PP 94103' in "Bad Data Suppression Power SystemState 24. MeJjl, H.M. and F.C. Schweppe, Estimation",IEEE Trans. PAS90, 1971' pp 27182725' 25. Mafaakher, F., et. al, "Optimum Metering Design Using Fast Decoupted Estimator", IEEE Trans. PAS98, 7979, pp 6268.
in Compensation lglver Systems
 I
55?
in can be connecteci the system in two ways, in series and irr shunt at ihe line (or even in the midPoint)' ends Apart from the wellknown technologies of compensation, the latest technology of Flexible AC TransmissionSystem (FACTS) will be introduced towards the end of the chaPter. 15.2 LOADING CAPABILITY
There are three kinds of limitations for loading capability of transmission system: (i) Thermal (ii) Dielectric (iii) Stability Thermal capabitity of an overhead line is a function of the ambient temperature, wind conditions, conditions of the conductor, and ground clearance. There is a possibility of converting a singlecircuit to a doublecircuit line to the loading caPabilitY. increase Dieletric Limitations From insulationpoint of view, many lines are designed very conservatively. For a given nominal voltage rating it is often possible to normal operatingvoltagesby l07o (i.e. 400 kV  440 kV). One should, increase are however,ensurethat dynamic and transientovervoltages within limits. [See Chapter 13 of Ref. 71. There are certain stability issues that limit the tlansmission Stability Issues. capability. These include steadystatestability, transient stability, dynamic resonance. stability, frequency collapse, voltage collapse and subsynchronous topics.The FACTS on gooclbooks l, l, 2, 6,7 ,8) are available these Several technology can certainly be used to overcome any of the stability limits, in which casethe final limits would be thermal and dielectric' 15.3 LOAD COMPENSATION
15.1
INTRODUCTION
For reduction of cost and improved reliability, most of the world,s electric power systems continueto be intercor of diversity of loads,availability of sor to loads at minimum cost and pollr deregulatedelectric service environme to the competitive environmentof reli Nowadays,greaterdemands have beenplacedon the transmission network, and these demandswill continue to rise becauseof the increasingnumber of nonutility generators and greatercompetitionamong utilitics the'rselves.It is not easy to acquire new rights of way. Increased demands on transmission, absence of longterm planning, and the need to provide open access to generatingcompanies and customers have resultedin less securityand reduced quality of supply. compensationin power systemsis, therefore, essential alleviate some of to theseproblems'series/shunt compensation has beenin usefor past many years to achieve this objective. In a power system, given the insignificant erectricarstorage, the power generationand load mttst balance at all times. To some extent,the electrical system is self'regulating. generation If is less than roacl,voltageanclfrequency d'op, and thereby reducing the road. However, there is only a few percent margin for such selfregulation. ,,^r+^ : :_,lf systemcollapse' Alternatively, thereisinadequate if reactive power,the system rnay havc voltagecollapse. This chapter is devoteclto the stucly of various methods.f compensating power systems various and typesclfcompensating crevices, cailccr .,,,rrj",rru,urr, to alleviatethe problemsof power system outlined above.Thesecompensators
support, load rhen increase con"se;:;iilTi: ,:t#'ji;l:ffL; with
Load compensation is the managementof reactive pcwer to improve power quality i.e. V profile and pf. Here the reactive power flow is controlled by at installing shunt compensating devices (capacitors/reactors) the load end reactive power. and consurned generated bringing about proper balancebetween the power transfer capability of the system This is most effective in improving both economicallyand technicallyto ancl its voltagestability. It is desirable power factor. This is why someutilities impose operatethe systemnear uriity a penalty on low pf loads. Yet another way of irnproving the system is performanee to operate it under near balancedconditions so as to reduce the ho* of legative sequence currents thereby increasing the system's load capabilityand reducingpower loss' ( loading( ii) st eadyline A transm ission hast hr cccr it icalloadings i) nat ur al line and (iii) thermallimit loading.For a compensated the starestability limit is the lowest and before the thermal loading limit is reached, natural loading stability limit is arrived. steadystate
esJ
I
T5.4 LINE COMPENSATION
Ideal voltage profile for a transmission line is flat, which can only be achieved by loading the line with its surge impedance loading while this may not be achievable,the characteristicsof the line can be modi so that (i) Ferranti effect is minimized. (ii) Underexcitedoperationof synchronous generatorsis not required. (iii) The power transfer capability of the line is enhanced.Modifying the characteristicsof a line(s) is known as line compensation. Various compensatingdevicesare: o Capacitors . Capacitors and inductors . Active voltage source (synchronousgenerator) When a number of capacitors are connected in parallel to get the desired capacitance, is known as a bank of capacitors, it similarly a Uant< incluctors. of A bank of capacitors and/or inductors can be adjusted in stepsby switching (mechanical). . Capacitors and inductors as such are passive line compensators,while synchronous generator is an active compensator.When solidstatedevices are used for switching off capacitorsand inductors, this is regardeclas active compensation. Before proceedingto give a detailedaccount of line compensator, shall we briefly discussboth shunt and seriescompensation. Shunt compensation is more or less like load compensation with all the advantagesassociatedwith it and discussedin Section 15.3. It needs to be pointed out here that shunt capacitors/inductors can not be distributed uniformally along the line. These are normally connected at the end of the line and/or at midpoint of the line. Shunt capacitors raise the load pf which greatly increases the power transmitted over the line as it is not required to carry the reactivepower. There is a limit to which transmitted power can be increased by shunt compensation as it would require very iarge size capacitorbank, which would be impractical. For increasing power transmitted over the line other and better means can be adopted.For example,seriescompensation, higher transmission voltage,HVDC etc. When switched capacitors are employed for compensation,these shculd be disconnected irnmediately underlight loacl conclitions avoicl to er.cessive voltage rise and ferroresonancein presenceof transformers. The purpose of series compensationis to cancel part of the seriesinductive reactanceof the line using series capacitors.This helps in (i) increase of maximum power transfer (ii) reduction in power angle for a given amount of power transfer (iii) increasedloading. From practical point of view, it is
compensated,it will behave as a purely resistive element and would. cause series resonance even at fundamental frequency. The location of series capacitorsis decidedby economical factors and severity of fault currents.Series capacitor reducesline reactancethereby level of fault currents. on on vanous lssues ln in series and shunt compensatorsnow follows. 15.5 SERIES COMPENSATION
A capacitor in series rvith a line gives control over the effective reactance between line ends.This effective reactance eiven bv is Xr=XX, where Xl = line reactance Xc = capacrtor reactance It is easy to see that capacitorrcduccsthe cffectivc line rcactance*. This results in improvement in performanceof the system as below. (i) Voltage drop in the line reduces(getscompensated) minimization of i.e. endvoltagevariations. (ii) Preventsvoltage collapse. (iii) Steadystttte power transferincreases; is inversclyproportionalto Xl. it (iv) As a result of (ii) rransientstability limit increasbs. The benefitsof the seriescapacitorcompensator associated are with a problem. The capacitive reactanceXg fbrms a seriesresonantcircuit with the total series reactance X = Xt * X*.n * Xoun, The natural frequency of oscillation of this circuit is given by. ^ rfL  2rJrc
'2n
where l= system frequency
xReactive voltage drops of a series reactance added in a line is I2x It is positive if X is inductive and negative if X is capacitive. So a series capacitive reactance reduces the reactance voltage drop of the line, which is an alternative wav of saying that X't= \X,
I
ii ti ti
I
, ,
ModernPower System AnalYsis rL X =degree of compensation = 25 to J5Vo (recommended) ' technology, the capacitanceof the series capacitance bank can be controlled much more effectively; both stepwiseand smooth control. This is demonsffated by the schematicdiagramof Fig. 15.2 wherein the capacitoris shuntedby two nstors ln antl current in positive half cycle and the other in negative half cycle. In each half cycle when the thyristor is fired (at an adjustable angle), it conducts current for the rest of the half cycle till natural current zeto. During the offtime of the thyristor current is conductedby the capacitor and capacitor voltage is vr. During ontime of the thyristor capacitor is short circuited i.e. v, = 0 and current is conductedby the thyristor. The sameprocess is repeated in the other half cycle. This means that v, can be controlled for any given i, which is equivalent of reducing the capacitanceas C = vJi.By this scheme capacitancecan be controlled smoothly by adjusting the firing angle. l+{___a,.Currenuimitins
I
fc<f
which is subharmonicoscillation. Even though series compensationhas often been found to be costeffective compared to shunt compensation,but sustained oscillations below the fundamental system frequency can cause the phenomenon, referred to aS resonance(SSR) first observedin 1937, but got worldwide subsynchronous attention only in the 1970s, after two turbinegeneratorshaft failures occurred at the Majave Generating station in Southern Nevada. Theoretical studies line, pointed out that interaction between a series capacitorcompensated oscillating at subharmonic frequency, and torsional mechanical oscillation of turbinegenerator set can result in negative damping with consequentmutual is resonance often not a Subsynchronous reinforcementof the two oscillations. measurescan be and protective major problem, and low cost countermeasures applied. Some of the corrective measuresare: (i) Detecting the low levels of subharmonic currents on the line by use of sensitiverelays, which at a certain level of currents triggers the action to bypassthe series capacitors. (ii) Modulation of generator field current to provide increased positive damping at subharmonic frequency' under light load conditionsto for are Seriesincluctors neeclecl line compensation (Ferranti effect). counter the excessivevoltage rise As the line load and, in particular the reactive power flow over the line varies, thereis need to vary the compensationfor an acceptablevoltage profile. The mechanical switching arrangementfor adjusting the capacitance of the capacitor bank in series with the line is shown in Fig. 15.1. Capacitanceis with the capacitance of varieclby openingthe switchos individualcapacitances The C1, being startedby a bypass switch. This is a stepu'isearrangement. switch under any emergent whole bank can also be bypassedby the starting conclitions on the line. As the switches in series with capacitor are current However, breaker carrying suitablecircuit breaking arrangementare necessary. generallyavoided thesedays the,capacitoris switched capacitorsin seriesare either fixed or thvristor switched.
reactor ^  o ' l l  "r' " i
l1<r
. C r i i
l
t
] 7,c
Fig. 15. 2
Thyristors are now available to carry large current and to withstand (during offtime) large voltage encounteredin power systems.The latest device called a Gate Turn Off (GTO) thyristor has the capability that by suitable firing circuit, angle (time) at which it goes on and off can both be controlled. This meanswider range and finer control over capacitance. Similarly confrol is possibleover seriesreactorin the line. All controlled.".?nd uncontrolled (fixed) series compensators require a protective arrangement.Protection can be provided externally either by voltage arrester or other voltage limiting device or an approximate bypass switch arrangement.In no case the VI rating of the thyristors should be exceeded. Depending on (i) kind of solidstatedevice to be used (ii) capacitor and/or reactor compensation and (iii) switched (stepwise) or smooth (stepless) control, several compensatron schemeshave been devisedand are in use.Some of the more common compensationschemesare as under. (i) Thyristor Controlled Series Cappcitor (TCSC) (ii) Thryristor\Switched Series Capacitor (TSSC) (iii) Fllyristor controlled Reactor with Fixed capacitor (TCR + FC) (iv) GTO thyristor Coniiolled Series Capacitor (GCSC)
i'r ;
Modern Power System Analysis *562i I I (v) Thyristor Controlied reactor (TCR) Capacitor and/or reactor series compensatoract to modify line impedance. is approach to introducca controllablevoltagc sourccin series An altcrnativo with the line. This scheme is known as static synchronous series compensator (SSSC). SSSC has the capabitity to induce both capacitive and inductive wrdenrngme operatmg reglon o voltage ln senes wrtn [lne, It can be used for power flow control both increasing or decreasingreactive flow on the line. Further this schemegives better stability and is more effective oscillations. in damping out electromechanical Though various types of compensatorscan provide highly effective power flow control, their operating characteristics and compensating features are different. These differences are related to their inherent attributes of their control circuits; also they exhibit different loss characteristics. From the point of view of almost maintenancefree operation impedance modifying (capacitorsand/or reactors) schemesare superior. The specific kind of compensator to be employed is very much dependent on a particular application. 15.6 SHUNT COMPENSATORS
called static var switchesor systems.It means that terminology wise SVC = SVS and we will use theseinterchangeably. Basic SVC Configrurations (or Desigrns)
Thyristors in antiparallelcan be used to switch on a capacitor/reactor unit in stepwise control. When the circuitary is designeclro adjust the firing angle, capacitor/reactor acts as continuouslyvariablein the power circuit. unit Capacitor or capacitor and incluctor bank can be varied stepwise or continuouslyby thyristor control. Several important SVS configurationshave been devised and are applied in shunt line compensation. Some of the static compensatorsschemesare discussedin what follows. (i) Sutu.rutcdreuctlr
As alreadyexplainedin Sec. 15.4 and in Ch. 5 (Sec.5.10) shunt compensators are connected in shunt at various system nodes (major substations) and sometimesat midpoint of lines. Theseserveihe purposesof voltage control and load stabilization. As a result of installation of shunt compensators in the operateat nearunity pf and voltage emergencies system,the nearbygenerators in mostly clo not arise. The two kinds of compensators use are: (1) Static var compensators (SVC): These are banks of capacitors (somea t im e s i n c l u c to rs l s o fg r u s e u n d e rl i ght l oad condi ti ons) (ii) STATCOM: static synchronouscompensator motor running at noload (iii) Sync hronous condenser: It is a synchronous and having excitation adjustableover a wide range. It f'eedspositive VARs into the line under overexcited conditions and negative VARS when underexcited.(For details see Sec. 5.10.) It is to be pointed out here that SVC and STATCQM are stgtic var generatorswhich are thyristor controlled. In this section SVC will be detailed while STATCOM forrns a part of FACTS whose operafionis explainedin S e c .1 5 . l O . Statia VAR Compensator (SVC)
This is a multicore reactor with the phase windings so arrangedas to cancel the principal harmonics.It is consiclered a constantvoltageieactive source. as It is almost maintenance free but not very flexible with reipect to operating characteristics. (ii)'fhyristorcoilrroll,cd reuctor (T'CR) A thyristorcontrolledreactor (Fig. 15.3) compensator consistsof a combination of six ptrlseor twelve plusethyristorcontrollecl reactprs with a fixe<tshunt capacitor bank. The reactive power is changectby adjustingthe thyristor firing angle. TCRs are characterised by continuous control, no transients and gencration harmonics'k. controlsystenrconsists voltage(and current) of The of
  Line oi.rt
Power
I
transformer I
I
neaitr i
I Fixed caoacitor
i] l.: )tl I I
? {rf ' l
I n"r.to, f =..'x',rnc.I u v v v
I
I Neutral
Thesecomprisecapacitorbank fixed or switched(controlled)or fixed capacito bank and switched reactor bank in parallel. These compensatorsdraw reactiv (leadingor lagging) power* from the line therebyregulating voltage, improv
*A rcrrctlrrcc cor)ncctcd in shunt to tinc at voltage V draws reactive power Vzl] It is negative (leading) if reactance is capacitive and positive (lagging) if reactance is inductive.
I
Fig. 15.3 Thyristor controlledeactor(TCR)with fixedcapacitor r
*Though ) connected TCR's are used here, it is better to use Aconnected TCR,s since it is better configuration.
Modern Power SystemAnalysis measuring devices, a controller for errorsignal conditioning, a Iinearizing circuit and one or more synchronising circuits. (iii) Thyristor switcherl capacitor (fSC) It consists of only a thyristorswitched capacitor bank which is split into a num De o unrts equal ratrngsto achrevea stepwisecon r o '0d6' Type of Var Generator
 ; fe" system (e.9. line faults, load rejection etc) TSC/TCR combinations are characterised continuouscontrdl, no transients, by low generations harmonof ics, low losses, redundancy,flexible control and operation. in Table 15.1. Table 15.1 Comparisonof Static Var Generators
TCR.FC (1)
TSC(TSR) (2)
Max. Comp. current is proportional to system voltage. Max. cap. var output decreaseswith the square of the voltage decrease. Low lossesat zero ouput. Losses increase steplike with cap. output
TCRTSC (3) Same as in (1) or (2)
/
Damping reactor

(TSC) Fig. 15.4 Thyristor switched capacitor
As suchthey are applied as a discretlyvariablereactivepower source,where this type of voltagesupportis deemedadequate. switchingtakesplace when All the voltage acrossthe thyristor valve is zero, thus providing almost transient tree switching.Disconnectionis eff'ected suppressing firing plus to the by the thyristors, which will block when the current reaches zero. TSCs are, charetcLorisccl stcp wisc control, no transients, by vcry low htlrnronics,low losses, redundancyand flexibility. (iv) CombinedTCR and TSC Compensator
VI and VQ Max comp. current characteristics is proportional to system voltage. Max cap. var output decreaseswith the squareof the voltage decrease. Loss Vs var High losses at zero output. output. Losses decreasesmoothly with cap. output, increirse with inductive output Hannorric generation I n t c r n a l l yh i g h (large pu TCR) Requires significant filtering l/2 cycle Poor (FC ('iluscs transientovervoltages in response to step disturbances)
A combinedTSC and TCR (Fig. 15.5) is the optimum solution in majority of cases. With this, continuousvariablereactivepower is obtainedtirroughoutthe c ot t t l; lc tcc o n l l o l rl n g c . F ru ' tl tc l rn o rci rl l control o1' botl r i nducti ve and r l capacitiveparts of the cornpensator obtained.This is a very advantageous is
{
Neutral
Max. theoret. delay Trrnsir:nl behaviour under systent voltagr: disturbances
hrtcrnallyvcry low Resonancemay necessirate tuning reactors I cycle
Low losses at zero output. Losses increase steplike with cup. output, smoothly with ind. output Internally low (small pu TCR) Filtering required I cvcle
Cun bc lrcutral. S a n r ca s i n ( 2 ) (Capacitorscan be switchedout to minimise trattsicntovcrvolt ages)
T5.7 COMPARISON BETWEEN STATCOM AND SVC
It
ll
Capacitor
a
mqw "'*J
he
nnterl
t hro tL rr q
ir r r
t h ov Lrr
nrn r r ll,a ll r vl , '
l;^^lrlrt/q.r
^*^,f:^\,Pgr4trlrE
rdlrBc
^r ul
r ule
r t t v l
Neutral
Fig. 15.5 A combined TCR/TSC compensator
characteristicand functional compensationcapability of the STATCOM and the sVC aie similar l2l. However, the basic operating principles of the STATCOM, which, with a converter based var generator,functions as a shuntconnectedsynchronousvoltage source,are basically different from those of the SVC, since SVC functions as a shuntconnected, controlledreactive admittance. This basic operational difference renders the STATCOM to have overall
.t lvlooernPower SystenrAnalysts t superior functional characteristics,better performance,and greater application flexibility as comparedto SVC. The ability of the STATCOM to maintain full capacitle output current at low system voltage also makes it rnore effective than the SVC in improving the transient (first swing) stability. Comparison between series and shunt compensation: tto
be effectively used for power flow conffol, load sharing among parallel corridors, voltage regulation, enhancement of transient stability anO mitlgation
* FACTStechnology havebeen proposed implemented. and FACTS devices ca1
(i) Series capacitorsare inherently self regulating and a control systemis not required. (ii) For the same performance, series capacitors are often less costly than SVCs and lossesare very low. (iii) For voltage stability, series capacitors lower the critical or coliapse voltage. (iv) Seriescapacitorspossess adequatetimcovelltrad capability. (v) Series capacitorsand switched series capacitors can be used to controi loading of paralledlines to minimise active and reactive losses. Disadvantages seriescompensation: of (i) Series capacitors are line connected and compensation is removed for outages and capacitorsin parallel lines may be overloaded. (ii) During tru.,vyioading, the voltage on one side of the seriescapacitormay be outof range. (iii) Shunt reactorsmay be neededfor light load compensation. (iv) Subsynchronous resonance may call for expensive countermeasures. Advantagesof SVC (i) SVCs control voltagedirectly. (ii) SVCs control temporaryovervoltages rapidly. Disadvantages SVC of (i) SVCs have limited ovcrload capability. (ii) SVCs are expensive. The best designperhaps a combinationof seriesand shunt compensation. is Because of higher initial and operating costs, synchronous condensers are normally not competitive with SVCs. Technically,synchronous condensers are better than SVCs in voltageweak networks. Following a drop in network voltage, the increasein condenser reactive power output is ilrslantaneous. Most synchrono.uscondenserapplications are now associatedwith HVDC installallons.
enablea line to carry power closer to its thermal rating. Mechanical switching has to be supplementedby rapid responsepower electronics.It may be noted that FACTS is an enabling technology, and not a oneonone substitute for mechanicalswitches. FACTS employ high speed thyristors for switching in or out transmission line cornponents such as capacitors, reactors or phase shifting transformer for some desirable performance of the systems. The FACTS technology is not a single highpower controller, but rather a collection of controllers,wtrictr can be applied individually or in coordination with others to conffol one or more of the systemparameters. Before proceeding to give an account of some of the important FACTS controllers the principle of operation of a switching converter will be explained, which forms the heart of these controllers. 15.9 PRINCIPLE AND OPERATION OF CON\ZERTERS
Controllable reactive power can be generated by dc to ac switching converters which are switched in synchronism with the line voltage with whicn tfr" reactive power is exchanged.A switching power converter consistsof an array of solidstate switches which connectthe input terminals to the output terminals. It has no internal storage and so the instantaneousinput and output power are equal. Further the input and output terminations are complementar, that is, if the input is terminated by a voltage source (charged capacitor or battery), output ii a cuffent source (which meansa voltage source having an inductive impedance) and vice versa. Thus, the converter can be voltage sourced (shunted by a capacitor or battery) or current sourced (shunted by an inductor). Single line diagram of the basic voltage sourced converter scheme for reactive power generation is drawn in Fig. 15.6. For reactive power flow bus voltage V and converter terminal voltage V, are in phase. Then on per phase basis i
r=vv4
x'i
'Fha ruv raa^+;"^ lvcvLtvly ^'.'^PrJw9r ^,^f^^J^^ E^Urr4rrBtr :^ l5
15.8 FLEXTBLE AC TRANSMTSSTON SYSTEMS (FACTS) The rapid development power electronics of technologyprovidesexciting opportunities develop to new powersystem equipment betterutilization for of
'
O=vI=
v(vvo)
X
563 l
I
Po Modern
System __ibus V 'i ' Coupling transformer
I
X
.J,
leakagereactance :l J,Transformer
capacitoris zero. Also at dc (zero fiequency) the capacitor does not supply any reactivepower. Thereibre, the capacitor voltage cloes not change and the capacitor establishesonly a voltage level for the converter. The switching causesthe converter to interconnectthe 3phaselines so that reactive current
can flow between thent.
Vd.
powergenerator Fig. 15.6 Staticreactive The switching circuit is capable of adjusting Vo, the output voltage of the converter. For Vo 1 V,1 lags V and Q drawn from the bus is inductive, while for Vo > V, I leads V and Q drawn from the bus is leading. Reactive power drawn can be easily and smoothly varied by adjusting Voby changing the ontime of the solidstate switches. It is to be noted that transformer leakage reactanceis.quite small (0.10.15 pu), which meansthat a amall differenceof voltage (VVo) causesthe required ,1and Q flow. Thus the converter acts like a static synchronous condenser(or var generator). twolevel, A typical convertercircuit is shown in Fig. 15.7.It is a 3phase sixpulse H.bridge with a diode in antiparallel to each of the six thyristors (Normally, GTO's are used). Timings of the triggering pulses ate in synchronism with the bus voltage waves.
Vo" Vot Vo,
The converter draws a small amount of real power to provide for the internal loss (in switching). If it is requiredto feed reai power to the bus, the capacitor is replace,l by a storage battery. For this the circuit switching has to be modified ro create a phase difference dbetween Vsand Vwith Vsleading V' The above explained converteris connectedin shunt with the line. On sirnilar with its terminals in serieswith the line. lines a converter can be constructed It has to carry the line current and provide a suitable magnitude (may also be phase)voltage in serieswith the line. In such a connectionit would act as an impedancemodifier of the line. 15.10 FACTS CONTROLLERS
[1' K
Va"
C
twolevel sixpulse bridge Fig. 15.7 Threephase,
The development of FACTS controllershas followed two different approaches. The first approach employs reactive impedancesor a tap changing ffansformer with thyristor switches as controlled elements, the second approach employs selfcommutatedstatic converters as controlled voltage sources. ln general,FACTS controllerscan be divided into tour categories. (iv) combined seriesshunt (i) series (ii) shunt (iii) combined seriesseries controllers. The generalsymbol for a FACTS controller is given in Fig. 15.8(a).which shows a thyristor arrow inside a box. The series controller of Fig. 15.8b could be a variable impedance, such as capacitor,reactor, etc. or a power electronics based variable source. All series controllers inject voltage in serieswith the line. If the voltage is in phasequadraturewith the line, the seriescontroller only or supplies consumesvariablereactivepower. Any other phaserelationshipwill involve real power also. Tlte shunt controllers of Fig. 15.8c may be variable impedance, variable source or a combination of these. All shunt controllers inject current into the controllers of Fig. system at the point of connection. Combined seriesseries seriescontrollers which are conffolled separate 15.8dcould be a combination of in a coordinatedmanner or it could be a unified controller. Combined seriesshunt controllers are either controlled in a coordinated manneras in Fig. 15.8e or a unified Power Flow Controller with series and shunt elementsas in Fig. 15.8f. For unified controller, there can be a real power exchangebetween the series and shunt controllers via the dc power link. magnet,or any Storagesource such as a capacitor,battery,superconducting added in parallel through an electronic interface other source of energy can be to replenish the converter's dc storage as shown dotted in Fig. 15.8 (b). A
ff.*J
power Modern System Analysis
in power Compensation Systems
Llne Line
controller with 'storage is much more effective for conffolling the system dynamics than the corresponding controller without storage.
Line
tffi*ffil T
(a)General symbolfor FACTScontroller (FC)
(a)
l
[,TTlFcl T
u
n
"
I
l
/
(b)
<oc
ac lines
Fig. 15.9 (a) STATCOMbasedon voltagesourced (b) currentsourced and converters. . A combination of STATCOM and any energy source to supply or absorb powei is called static synchronous generator(SSG). Energy source may be a battery, flywheel, superconducting magnet, large dc storage capacitor, another rectifi erlinverter etc. Statlc Synchronous Series Compensator (SSCC)
(c) Shuntcontroller
(d) Unifiedseriesseriescontroller
I it
FCH Coordinated
control
!.
dc power link
It is a seriesconnectedcontroller. Though it is like STATCOM, bu(its output voltage is in serieswith the line. It thus controls the voltage acrossthe line and hence its impedance. Interline Power FIow Controller (IPFC)
(e) Coordinated series and shuntcontroller
(f) Unifiedseriesshuntcontroller
Flg. 15.8 DifferentFACTScontrollers The group of FACTS controllers employing switching converterbased synchronous voltage sourcesinclude the STATic synchronous COMpensator (STATCOM), the static synchronous series compensator (SSCC), the unified power flow controller (UPFC) and the latest, the Interline Power Flow Controller (IPFC). STATCOM STATCOM is a static synchronousgeneratoroperatedas a shuntconnected static var compensator whose capacitive or inductive output current can be i;orrrroiieornoepenoentoi rhe ac system voltage. The STATCOM, like its conventional countetpart, the SVC, controls transmission voltage by reactive shunt compensation.It can be based on a voltagesourcedoi curtentsourced converter. Figure 15.9 shows a oneline diagram of srATCoM based on a voltagesourced converterand a current sourcedconverter. Normally a voltagesource converter is preferred for most converterbasedFACTS controllers. STATCOM can bc designedto be an active filter to absorb system harmonics.
^^^lll ! t a ,
This is a recently introduced controller 12,3]. It is a cornbinationof two or seriesconlpensators which are coupledvia a common more static synchronous dc link to facilitate bidirectional flow of real power between the ac terminals of the SSSCs, and are controlled to provide independent reactive series compensation for the control of real power flow in each line and maintain the desired distribution of reactive power flow among the lines. Thus it managesa comprehensiveoverall real and reactive power managementfor a multiline transmissionsystem. Unified Power FIow Controller (UPFC)
This controller is connectedas shown in Fig. 15.10. It is a combination of STATCOM and SSSC which are coupled via a iommon dc link to allow bidirectional flow of real power between the series output terminals of the SSSC and the shunt output terminals of the STATCOM. These are controlled fo provide concurrent real and reactive series line compensation without an external energy source.The.UPFC, by meansof angularly unconsffainedseries voltage injection, is able to control, concurrently/simultaneously selectively, or the transmission line voltage, impedance, and angle or, alternatively, the real
rr$il#4
I
Modern
l
l"r'.

r'
i:
and reactive line flows. The UPFC may also provide independentlycontrollabte shuntreactive compensation.
stability and cnnffol line flows. Voltage source converter based (selfcommutated) HVDC system may have the same features as those of STATCOM or UPFC. This system also regulatesvoltage and provides system A comparative perfonnance of major FACTS controllers in ac system is given in Table 15.2 U4l.
STATCOM
Table 15.2 A comparativeperformance major FACTS controller of Type of FACTS Controller SVC/STATCOM TCSC SSSC TCPAR UPFC
Load flow control
X XX XXX XXX XXX
Fig. 1S.10 UnifiedpowerFlowControiler UpFC Th yri s t or co n tro I I e d p h a s e sh ifti n g Tra n sfo rm er (TCp s r) This controller is also called Thyristorcontrolled Phase Angle Regulator (TCPAR). A phaseshifting transformer controlled by thyristor switches give to a rapidly variable phase angle. Th yristorCon troll e d VoI tdgre R egu trator flCVR)
V control
XXX
Transient stability
Oscillation Damping
XX XX XX XX XXX
x x
XX XXX
xxx xxx
XXX X XXX
* a;vslrong influence; xxaverage influence; xsmail infruence
A thyristor controlled transformer which can provide variable inphase voltage with continuouscontrol. Interphase Power Controller (IpC)
sutyffvtARy
Since the 1970s,energy cost, environmentalrestrictions, rightof*ay difficulties, along with other legislative sociai and cost problems huu. postponed the construction of both new generation and transmission systemsin India as well as most of other countries. Recently, becauseof adoption of power reforms or restructuring or deregulation, competitive electric energy markets are being developedby mandatingopen accesstransmission services. In the late 1980s,the vision of FACTS was formulated. In this various power electronics basedcontrollers (compensators) regulate power flow and transmission voltage and through fast control aciion, mitigate dynamic disturbances. Due to FACTS, transmissionline capacity was enhanced.Two types of FACTS controllers were developed. One employed conventional thyristorswitched capacitorsand react<lrs, and quadraturetapchangingtransformerssuch as SVC and TCSC. The second category was of selfcommutatedswitching converters as synchronousvoltage sources,e.g. STATCOM, SSSC, UPFC and IpFC. The two groups of FACTS controllers have quite different operating and performance characteristics.The second group usesselfcommutated dc to ac converter. The converter, supported by a de power supply or energy storage de.,,ice can also exchangereal power with the ac systembesidesconmollingreactive power independently. The increasing use of FACTS controllers in future is guaranteed.What benefits are required for a given system would be a principal justification for the choice of a FACTS controller. Its final form and operation will, ofcourse, depend not only on the successful development of the necessarycontrol and
A seriesconnected controller of active and reactive power consisting,in each phase, of inductive and capacitive branches subjected to separately phaseshifted voltages. The active and reactive power can be set inpedendently by adjusting the phase shifts and/or the branch impedances,using mechanical or electronic switches. Thyristor Controlled Braking Resistor ICBR)
It is a shuntconnectedthyristorswitched resistor, which is controlled to aid stabilization of a power system or to minimise power acceleration of a generatingunit during a disturbance.
Thyristorcontrolled
Voltage Limiter
FCVL)
A thyristorswitched metaloxidevaristor (Mov) used to limit the voltage across terminals its durins transient conditions 'TIVDC It may be noted that normally HVDC and FACTS are compleme ntary technologies. role of HVDC, for economic The reasons, to interconnect is ac system?,wherereliableac interconnection a would be too expensive. HVDC
r:5?4i I
I
Modern power System Analysis
communieationteehnologiesand protocols, but also on the final structure of ihre evolving newly restructuredpower systems.
REFERE NCES
Books
I ' Chakrabarti,A', D.P. Kothari and A.K. Mukhopadhyay, Performance Operation and Control of EHV Power TransmissionSystems. Wheeler, New Delhi, 1995. 2. Hingorani, N.G. and Laszlo Gyugyi, (Jnderstanding FACTS. IEEE press, New York, 2000. 3. Song, Y.H' and A.T. Johns,Flexible AC TransmissionSystems,IEE, Lbndon, 1999. 4. Miller, T'J.E., ReactivePower Control in Electric Systems, John Wiley and Sons, NY, lgg2. 5. Nagrath' I.J. and D.P. Kothari, Electric Machines, 2nd edn, Tata McGrawHill. New Delhi, 1997. 6. Taylor, c.w., power system vortage stability, McGrawHill, singapore, 1994. 7. Nagrath,I.J and D.P. Kothari, power S),stem Engineering, TataMcGraw_Hill, New Delhi, 1994. 8. Indulkar, C.S. and D.P. Kothari, Power SystemTransientsA StatisticalApproach, PrenticeHallof India, New Delhi, 1996. 9. Mathur, R.M' and R.K. Verma, ThyristorBasedFACTI; Controllers for Electrical TransmissionSystems. John Wiley, New york, 2002.
16.1
INTRODUCTION
Load forecasting plays an important role in power system planning, operation and control. Forecasting means estimating active load at various load buses ahead of actual load occurrence.Planning and operational applicationsof load forecasting requires a certain 'lead time' also called forecasting iritervals. Nature of forecasts,lead times and applicationsare summarisedin Table 16.1" Table 16.1
Nature of forecast Very short term Lead time A few seconds to several minutes Half an hour to a ferv hours A few days to a t'ew' rveeks A t'ew months to a few years Application Generation, distribution schedules, contingency analysis for system secunfy Allocation of spinning reserve; operarionilI planning ild unit commitment; maintenance scheduling Planning for seasonal peakwrnter. summer Planning generation growth
Papers
Edris A, "FACTS Technologv Development: An Update".\EEE Pott'er Engirteering Rerieu'. \'ol. 20. lVlarch 2@1i. pp f9. 1l  Iliceto F and E. Cinieri, "Comparative Analysis of Series and Shunt Compensation I0Schemes for Ac Transmission svstems". IEEE Trans. pAs 96 6). l810. 12. Kimbark. rg77. pp lglg
"Hou, E.W.. to Impror,e S.r'srem Stabiliri, u,ithout Risking Suhsrtchnrnous R('s(')nitncc"'. IEEE lr.rrs. P.tJ. 96 ti,I. Sept/Oc.t 1977. pp 160.S19.
Short term
13. CIGRE/ 'wG 3801. .srrzric l/ar Conrp<,rrdrorr. CIGRE/. prrris.i 986 l{. Ptrvlr. f).. "Llsc'ofHDVC and F.ACTS".IEEE procc'edings. .\8. 2, Feb. 2000, vol. pp 235245. 15. Kinrbark,E.W. "A New Look At Shunt Compensation.,IEEE Trans. Vol PAS102. No. l. Jan1983.nn 2122!8
Medium term Long term
A good t'orecastreflecting current and future r.rends,tempered wi.Ji good judgement, is the key to all planning, indeed. financial success. to The accuracy of a forecast is crucial to any electric utility, since it determinesthe timing and characteristicsof major system additions. A forecast that is too low can result in low revenuefrom salesto neighbouringutilities or even in load curtailment. Forecasts that are too high can result in severe financial problems due to excessive investment in a plant that is not fully utilized or operatedat low
LoadForecasting Technique
can capacity factors. No forecastobtainedfrom analytical proceclures be strictly judgement of the forecaster, which plays a crucial role in r"ii.d upon the ariving at an acceptable forecast. Choosing a forecasting technique for use in establishing future load ing on nature o requirements is a nontrivial task in itself. may be superior to another. variations, one particular methcd The two approaches to load forecasting namely total load approach and component approach have their own merits and demerits.Total load approach has the merit that it is much smoother and indicative of overall growth trends and easy to apply. On the other hand, the merit of the component approach is that abnormal conditions in growth trends of a certain component can be detected,thus preventing misleadingforecastconclusions.There is a continuing need, however, to improve the methodology for forecasting power demand more accurately. The aim of the present chapter is to give brief expositions of some of the techniques that have been developed'in order to deal with the various load forecasting problems. All of these are based on the assumptionthat the actual load supplied by a given systemmatchesthe demandsat all points of time (i.e., there has not been any outagesor any deliberate sheddingof load). It is then possible to make a statisticalanalysisof previous load data in order to set up a suitable model of the demandpattern.Once this has been done, it is generally possible to utilize the identified load model for making a prediction of the estirnated demand for the selectedlead time. A major part of the forecasting task is thus concerned with that of identifying the best possible model for the the past load behaviour.This is best achievedby decomposing load demand at given point of tirne into a number of distinct components.The load is any dependent on the industrial, commercial and agricultural activities as well as the The weather sensitivecomponentdepends weathercondition of the system/area. cloudiness, wind velocity, visibility and precipitation. Recall on temperarure, the brief discussions in Ch. 1 regarding the nature of the daily load curve which has been shown to have a constantpari correspondingto the baseload and other variable parts. For the sake of load forecasting, a simple decomposition may serve as a cdnvenient starting point. Let y(k) representthe total load demand (either for the whole or a part of the system) at the discretetime k = l, 2,3, ....It is generallypossibleto decomposey(k) into two parts of the form T6.2 FORECASTING METHODOLOGY
Llif,iffi
I
Forecasting techniques may be divided into three broad classes.Techniques may be based on extrapolation or on correlation or on a combination of both.
rrnlnlstlc,
Extrapolation Extrapolation techniquesinvolve fitting trend curves to basic historical data adjusted to reflect the growth trend itself. With a trend curve the forecast is obtained by evaluating the trend curve function at the desired future point. Although a very simple procedure, it produces reasonable results in some instances.Such a techniqueis called a deterministic extrapolation since random errors in the data or in analytical model are not accounted for. Standard analytical functions used in trend curve fitting are [3]. (i) Straight line ! = a+ bx (ii) Parabola' !=a+bx+c*2 (iii) Scurve !=a+bx+ci+dx3 (iv) Exponential !=ce& (v) Gempertz ! = Inr 7a + ced''1 ' The most corlmon curvefitting technique for fitting coefficients and exponents(a4) of a function in a given forecastis the method of least squares. If the uncertainty of extrapolated results is to be quantified using statistical entities such as mean and variance, the basic technique becomes probabilistic extrapolation. With regression analysis the best estimate of the model describing the trend can be obtained and used to forecast the trend. Correlation Correlation techniques of forecasting relate system loads to various demographic and economic factors. This approach is advantageousin forcing the forecaster to understand clearly the interrelationship between load growth patterns and other measurable factors. The disadvantageis the need to forecast demographic and economic factors, which can be more difficult than forecasting system load. Typically, such factors as population, employment, building permits, business,weather data and the like are used in correlation techniques. No one forecasting method is effective in all situations. Forecasting techniques must be used as tools to aid the planner; good judgement a:rd experience can never be completely replaced. 16.3 ESTIMATION OF AVERAGE AND TREND TERMS
+ y(k)= ya(k) y"(k)
(16.1)
where the subscript d indicates the deterministic part and the subscript s indicates the stochasticpart of the demand. If k is consideredto be the present time, then y(k + j), j > 0 would represbnta future load demand with the index problem 7 being the lead time. For a chosenvalue of the indexT, the forecasting of estimating the value of y(k +/) by processing is then the sameas the problem adequatedata fbr the past load dernand.
The simplest possible form of the deterministic part of y(k) is given by ya &) = !a + bk + e(k)
(r6.2)
Modern PowerSystem Analysis where larepresents the averageor the mean value of yd(k), bk representsthe 'trend' term that grows linearly with k and e(k) representsthe error of modelling the complete load using the average and the trend terms only. The questionis one of estimating the valuesof the two unknown model parameters la aldb !o ensurea good model. As seenin Ch, 14, when little orlo st1listical information is availableregardingthe error term, the methodof LSE is helpful. If this method is to be used forestimating yo and b,the estimationindex "/is defined using the relation
Technique LoadForecasting
kt5f,*A I
it
I
I
l it
J  E{ez(D}
(16.3)
where E(.) represents the expectation operation. Substituting for e(A) from Eq. (16.2) and making use of the first order necessaryconditions for the index J to have its minimum value with respect to ya md b, it is found that the following conditions must be satisfiedl2).
(16'6a) and ln order to illustrate the nature of results obtainablefrom Eqs' of Fig. 16'l which give the (16.6b), consider the clatashown in the graphs jndust+ial values of the agricirttural and the iopurerion in millionT. Th= cash energy consumpdon ooiput, in millions r.rfrupees and the amount of electrical from (toaA demand) in MWs 1n Punjab over a period of seven years starting g5 data have been generatedfrom the graphs by sampling the 1968. A total of (16'6a) and graphs at intervals of 30 days. These have been substitutedin Eqs' and the trend coefficients of the four in if O.OUI order to compute the avetage variables.The results are given in Table 16'2'
Table 16.2 Variable Population Industrial output Agricultural output Load demand Average
T I l
E { y a  y a & )+ b k I = 0 + E {bkz  ya(k)k tdkl  0
(16.4a) (16.4b)
Trend Cofficient
Since the expectation operationdoes not affect the constant quantities, it is easy to solve these two equationsin order to get the desired relations.
ta= E{yd&)l b{E(k)} b  lE{ya&)kl yo E{kllt4{k2l
(16.5a) (16.5b)
13 million Rs 397 million Rs 420.9million 855.8MW
o.2
0.54 0. 78 r.34
If y(k) is assumedto be stationary(statisticsare not time dependent)one may involve the ergodic hypothesis and replace the expectatiori operation by the time for averaging fonnula. Thus, if a total of N data are assumedto be avai.labLe determining the time averages,the two relations may be equivalently expressed as follows.
(16.6a)
peprJlation millions in of in lndustry millions ruPees of millions rupees Agriculture in Loaddemand MW ,
(16.6b)
t
/
,
/
These two relations may be fruitfully employed in order to estimate the average and the trend coefficient for any given load data. Note that Eqs. (16.6a) and (16.6b) are not very accuratein case the load data behavesas a nonstationaryprocesssince the ergodic hypothesis does not It holcl for suchcases. may still be possibleto assumethat the data over a finite window is stationary and the entire set of data may then be consideredas the juxtaposition of a number of stationaryblocks, each having slightly different over the different statistics.Equations(16.6a)and (16.6b)may then be repeatecl the trend coefficient for each blocks in order to compute the average and window of data.
i16
I t
the data
ffi4
I
ModernpowerSystem Analysis
t I
I
600
{ I
ttr
400
"ertthe load model maY be assumedto be
in
I ,II
I
3
;:
y(k)fu3,+e(k)
i:l
(16.8a)
72
k (hours)_600
120
t I
i.
400
J
3 E 2oo o
0 1 192 216 240 264 k (hours)288 312 336
past load d where the coefficient b,needto be estimated from the index /<and would model above is obviously a nonlinear function of the time approach to nonlinear need L coefficients to be estimated. A much simpler form modelling of the load is to introduce an exponential ( 16. 8b) y( k) = c exp [ bk] + e( k) reducing the number of which involves only two unknown coefficients. Besides model has the additional advantageof being readily unknowns, th" exponential is to take the natural log of transformed into u linru, form. All that is required easily extended to estimate the given data. In either case,the method of LSE is the model parametersfrom the given historical data.
6,4ESTIMATIoNoFPERIoDIGcoMPoNENTS
periodic components in The deterministic part of the load may contain some the tn" uurrage and the porynomial rerms. consider for example ;ilri'.o,',; power supplied the active curve shown in Fig. tZ.Z wtrich givei the variation of over a period of t*o weeks. It is observed that tlrg daily load by a power utility some random fluctuations' variations are repetitive from day to day except for from those of the It is also seen that the curve for sundays differ significantly out that the curve ior ihe week days in view that Sundaysare holidays. I,,]rTt of one Sunday till the mid entire weekly period starting fiom, say, the mid night periodic waveform with night of the next Sunday behaves as a clistinctty superposedrandom variations' at an hourly interval, If it is assumed that the load data are being sampled so that the load pattern may then there are atotal of 168 load data in one period fundamental frequency ul be expressed in terms of a Fourier series with the load y(k) is then given by i.f"g'.q" ut to Z7rl\68rads. Arsuitable model for the y(k) = y + t
i=l
Fig. 16.2 Hourlyload behaviour Delhiover two consecutive of weeks Caution The 85 data, used in Example 16.r, are generally not adequate for making statistical caiculations so that the vaiues given ubour may not be entirely adequate. In addition, the statistical characteristics of the set of variables concernedmay have changed (i.e., the data may in fact be nonstationary) and this alsomav inffoduce some error in the results.Finally, the graphs in Fig. 16.1 are actually based on half yearly data obtained from the planning commission document and an interpolation processhas been employed in order to generate the monthly data. This may add some unspecified errors to the data which will also affect the accuracy of the estimates. Prediction of ya &+j)
Once the model for the deterministic component of the load has been determined, it is simple to make the prediction of its future value. For the simple model in Eq. (762), the desired prediction is computed using the relation
/ 4
[a, sin iuk + b, cos iuk] + e(k)
(15'9)
v'(k + i) = T, + h(k + i\
\ r / J A " U ' J l
(i6.7)
More General Forms of Models Before leaving this section, it may be pointed out that the load model may be generalizedby including second and higher order terms on the right hand side
present and a; and b; are the where L represents the total number of harmonics and the cosinusoidalcomponents' only amplitudes of respectively the sinusoidal to rlnrrrinqntt',ormonicsneed be included in the model' to make a prediction Once the harmonic load model is identified, it is simple the relation of the future load ya(k + 7) using
(ltJllllllgrr! lrq^r^\
9 a & + / ) = h ' ( k+ l ) i ( k )
(16.10)
'eifrz#d !6.5
ModernPo@is
LoadForecasting Technique
h v 2
Hffi T
ESTIMATION OF YS(^ft): TiME SERIESAPPROACH
= 15(ft) t
i:l
a, y, (k  r) +D b, w(k ) + w(k) (16.14)
j:l
Autoregressive Models The sequence y,(ft)is saidto satisfyan AR modelof ordern i.e.it is [AR(n)], if it can be expressed as
n
Estimation of two structural parameters n and rn as well as model parameters ap bi and the variance d of the noise term w(k) is required. Moie complex can be represented.The identification problem is solved offline. The acceptable load model is then utilized online for obtaining online load forecasts. ARMA model can easily be modified to incorporate the temperature, rainfall, wind velocity and humidity data [2]. In some cases,it is desirable to show the dependenceof the load demand on the weather variables in an explicit manner. The time series models are easily generalized in order to reflect the dependenceof the load demand on one or more of the weather variables. 16.6 ESTIMATION OF STOCHASTIC COMPONENT: KALMAN FILTERING APPROACH
y,(k)=
Do,r,(k
i_l
 i) + w(k)
(16.11)
lie inside the unit circle in the eplane. The problem in estimating the value of n is referred qc tha ntnhtaw )^s rn
The time series approachhas been widely employed in dealing with the load forecasting problem in view of the relative simplicity of the model forms. However, this method tends to ignore the statisticalinformation about the load data which may often be available and may lead to improved load forecasts if utilized properly. In ARMA model, the model identification problem is not that simple. These difficulties may be avoided in some situations if the Kalman ' , filtering techniques are utilized. Application to Shortterm Forecasting
An application of the Kalman filtering algorithm to the load forecasting problem has been first suggested Toyada et. al. [11] for the very shortterm by and shortterm situations.For the latter case,for example,it is possible to make use of intuitive reasoningsto suggestthat an acceptablemodel for load demand would have the form
v,(k) = y'(k) + v(li)
(16.15)
i,&) = D a,y, (k i)
i:l
(16.r2)
where y,(ft) is the observedvalue of the stochasticload at time ft, y,(k) is the true value of this load and u(fr) is the error in the observedload. In addition. the dynamics of the true load may be expressedas
The variance d of w(k) is then estimated using the relation
N n2 (1lnn F Zrrt \ r'l' ) \tr ) /,rc k:l
y,(k +I) = y{k) + z(k) + u1&)
(16.16)
( r 6 .l 3 )
Models
where a(k) represents the increment of the load demand at time k and u1@) represents a'disturbanceterm which accountsfor the stochasticperturbationsin y,(k).The incrementai ioaciitseif is assumerito remain constanton an average at every time point and is modelled by the equation
AutoRegressive
MovingAverage
In some cases,the AR model may not be adequate to representthe observed Ioad behaviour unless the order n of the model is made very hrgfi. In such a case ARMA (n, m) model is used.
z(k +l) = z(k)+ ur(k) wherethe term uz&) represents stochasticdisturbance a term.
(16.r7)
a.,.".lm
In order to make use of the Kalman filtering techniques, the noise terms u(k), uz&) and u(k) are assumedto be zero mean independent white Gaussian sequences. Also, the model equations are rewritten in the form
to use the solution of Eq. (16.18a)for the vector x (k+ d)to get the result
x (ft+l) = Fx(k) + Gu(k)
f;g+ d)= pdi(tcttc)
of thg load v(k + d\. it is necessarv fhaf the nnisc
(16.18a)
(16.199)
cfeficrinc qnr{ o^o nr}ra
k v(k where the vectorsx(ft) and u(ft) are definedas x(k) = ly,(k) = (Df and u(k) = fur(k) uz&))r The matrices G andh' arethen obtained 4 from Eqs. (16.15)(16.17) easily and have the followingvalues.
In order to be able to make use of this algorithm for generating the forecast information be available. The value of R(ft) may often be estimated from a knowledge of the accuracy of the meters employed. However, it is very unlikely that the value of the covariance Q(k) will be known to start with and will therefore have to be obtained by some means. An adaptive version of the Kalman filtering algorithm may be utilized in order to estimate the noise statistics alongwith the state vector x(k) tZ). Now let it be assumedthat both R(k) and Q(k) are known quantities. Let it also be assumed that the initial estimate i (0/0) and the covarianceP"(0/0) are known. Based on thesea priori information, it is possible to utilize Eq. (16.19a)(16.19e) recursively to processthe data for yr(l), yr(z), ..., yr(k) to generatethe filtered estimate ,(kl k). Once this is available, Eq. (16.19g) may be utilized to ger,erate the desired load forecast.
o1, ltl c [1 1l,=[1 1 l ' r=L o l "=L o u ' L o
Based on model (16.18), it is possible to make use of the Kalman filtering algorithm to obtain the minimum varianceestimateof the vector x (k) basedon the data y,(k): {y,(1), ),,(2) ... y,(k)}. This algorithm consisrsof the following equations.
i (k/k) = i (ktt< 1) + K" (k) ty"(ft) h'ft(k/kI)l i(ktk 1)= F i((kL)/(kr))
K,(k) = P,(k/kl) hlh' P,(klkI)h + R(k)lr P(k/k)= V  K,(k) h') P,(k/kl) P,(k/k  l)= FP,(kl/kl)F'+ where, Q(k) = covarianceof u(k) R(ft) = covarianceof y(ft) ft(klk) = filtered estimateof x(ft) ft (k/kL) = single steppretliction of x(ft) K.r(k) = filter gain vector of same dimension as a(ft) Pr(k/k) = filtering error covariance Pr(k/kl) = predictionerror covariance F r onr E q. (1 6 .1 8 b )o b ta i nth e p re d i c ti o ni ((k+ 1)l k) From this the one step aheadload fbrecast is obtained as GQGDG',
(16.19a) (16.19b) (16.19c) (16.19d) (16.19e)
To illustrate the nature of the results obtainable through the allorithm just discussed, the data for the short term load behaviour for Delhi have been processed. total of 1030 data collected at the interval of 15 minuteshave bee.n A processed.It has been assumedthat, in view of the short time interval over which the total data set lies, the deterministic part of the load may be assumed to be a constant mean term. Using the sample average Formula (16.7) (with b = 0), we get y = 220 Mw. The data for yr(k) have then been generated by subtracting the mean value from the measured load data. To process these stochastic data, the following a priori information have been used:
R(k)= 3.74,
= i (oro)
[;]
(16.1e0 Y " ( f t +I ) = h ' i ( ( k + l ) l k ) It may be noted that filtering implies removal of disturbance or stochastic term with zero mean. It is also possible to obtain a multistep aheadprediction of the load from the multistep ahead prediction of the vector x (k). For example, if the prediction
The results of application of the prodietion Algonthnn (16.19) are shown in Fig. 16.3.It is noted that the elror of 15 minutes aheadload predictionis around 8 MW which is about 3Voof the averageload and less than 2Voof the daily maximum load.
I,r6rsHs
1. Actualload 2. 15 mts. prediction error
ECONOMETRTC MODELS If theloadforecasts requiredfor planning are purposes, is necessary select it to
3
200
E c O I l I
the lead time to lie in thrro aJeulyearsIn sucLcasesthe load demand should be decomposedin a manner that reflects the dependence of the load on the various segmentsof the economy of the concernedregion. For example, the total load demand y(ft) may be decomposedas
M
o
E
O
E
I
I 180 1 6 =
l
y(k)=D",y,(k)+ e(k)
i:l
(16.2Oa)
o
(D
12.=1
9 l
I
I
o l
'1C
L
4
i3 14 15 F i g . 1 6.3 Comment 16 17 18
E
c o
l l
where aiare the regressioncoefficients, y,(k) arethe choseneconomic variables and e(k) represents the error of modelling. A relatively simple procedure is to rewrite the model equation in the tamiliar vector notation y(k)=h'(k)x+ where e(k)
(16.20b)
n f L
E
h'(k)= [yr(ft)yz(k)... yu&)] and x = far a2... ayl.
The regression coefficients may then be estimated using the convenient least squares algorithm. The load forecasts are then possible through the simple relation
i(k + 1) = i'(k) it(k +Ilk)
Application of Kalman filtering and prediction techniques is often hampered by the nonavailability of the required state variable model of the concernedload data' For the few casesdiscussedin this section, a part of the model has been obtainable from physical considerations.The part that has not been available include the state and output noise variances and the data for the initiai state estimate and the conesponding covariance. In a general load forecasting situation, none of the model parameters may be available to start with and it would be necessaryto make use of system identification techniquesin order to obtain the required state model. It has been shown that the GaussMarkov model described by Eq. (16.18) in overparameterised from the model identification point of view in the scnse that the data for y,(k) dr: not permit the estimation of ali the parametersof this model. It has been shown in Ref. [12] that a suitable model that is identifiable andis equivalent to the GaussMarkov model for state estimation purposesis the innovation model useclfor estimation of the stochastic component. on'line Techniques f,or Nonstationanl Load predietion
where i (k) is the estimate of the coefficient vector based on the data available till the ftth sampling point ana fr6 + Uk) is the onestepahead prediction of the regressionvector h(k). 16.8 REACTTVE LOAD FORECAST
\
(16.21)
Most practical load data behaveas nonstationaryand it is therefore imponant to consider the questionof adaptingthe techniquescliscussecl f,arto the nonso stationarysituation. Ref. [2] has discussed threemodels for this purposeviz. the (i) ARIMA Models, (ii) Time varying model and (iii) Nondynamic models.
Reactive loads are not easy to forecast as compared to active loads, since reactive loads are made up of not only reactive components of loads, but also of transmissionand distribution networks and compensatingVAR devicessuch as SVC, FACTs etc. Therefore, past data may not yield the correct forecast as reactive load varies with variations in network configuration during varying operating conditions. Use of active load forecast with power factor prediction may result in somewhat satisfactory results. Of course, here also, only very recent past data (few minutes/hours) may be used, thus assuming steadystate network configuration. Forecasted reactive loads are adapted with current reactive requirements of network including var compensation devices. Such rorecasts are neeoed ior secunty anaiysls, voitageireactivepower scheduling etc. If control action is insufficient, sffuctural modifications have to be carried out, i.e., new generating units, new lines or new var compensatingdevices normally have to be installed.
f  a
mitul
Modrrn Po*"r Syrt"r An"turi.
li':l;{lrJ.
sutut ARy ttil
Load forecastingis the basic step in power systemplanning. A reasonablyselfcontainedaccount of the various techniquesfor the load prediction of a modern prediction problems. Applications of time series,GaussMarkovand innovation models in setting up a suitable dynamic model for the stochasticpart of the load datahave been discussed. The time seriesmodel identification problem has also been dealt with through the least squaresestimation techniques developed in Chapter 14. In an interconnectedpower system, load forecasts are usually needed at all the important load buses. A great deal of attention has in recent years been given to the question of setting up the demand models for the individual appliances and their impact on the aggregated demand. It may often be necessary make use of nonlinearforms of load models and the questionof to identification of the nonlinear models of different forms is an important issue. Finally, a point may be made that no particular method or approach will work for all utilities. All methods are strung on a common thread, and that is the judgement of the forecaster. In no way the material presented here is exhaustive. The intent has been to introduce some ideas currentlv used in forecastingsystem load requirements. Future Trends
** and integration with conventional timeseries methodsin order to provide a more precise forecasting.
REVIEW QUESTIONS
L6.7 Which method of load forecasting would you suggestfor long term and why? 16.2 Which method of load forecasting would you suggestfor very short term and why? 16.3 What purposedoes medium term forecastingserve? 16.4 How is the forecaster'sknowledge and intuition consideredsuperior to any load forecasting method? Should a forecaster intervene to modify a forecast,when, why and how? 16.5 Why and what are the nonstationarycomponentsof load changesduring very short, short, medium and long terms?
NCES REFTRT
Books
1. Nagrath I.J. and D.P. Kothari, Power System Engineering,Tata McGrawHill, New Delhi, 1994. 2. Mahalanabis,A.K., D.P. Kothari and S.I. Ahson, Computer Aided Power System Analysis and Control, Tata McGrawHill, New Delhi, 1988. 3. Sullivan, R.L., Power SystemPlanning, McGrawHill Book Co., New York, 1977. 4. Pabla,A.5., Electricol Power Systems Planning,Macmillan lndia Ltd., New Delhi, 1998. 5. Pabla,4.5., Electric Power Distribution,4th Edn, Tata McGrawHill, New Delhi. 1997. 6. Wang, X and J.R. McDonald (Eds), Modern Power SystemPlanning, McGrawHill, Singapore,1994.
Forecastingelectricity loads had reacheda comfortable stateof performance in the yearspreceding the recent waves of industry restructuring. As discussedin this chapter adaptive timeseries techniques based on ARIMA, Kalman Filtering, or spectral methods are sufficiently accurate in the short term for operational purposes, achieving errors of l2%o. However, the arrival of competitive markets has been associated with the expectation of greater consumerparticipation. Overall we can identify the following trends. (i) Forecast errors have significant implications for profits, market shares, and ultimately shareholdervalue. (ii) Day ahead, weatherbased,forecasting is becoming the rnost crucial activity in a deregulated market. (iii) Information is'becoming commercially sensitive and increasingly trade secret. (iv) Distributed, embeddedand dispersedgeneration rhay increase. A recentpaper [7] takesa selectivelook at some of the forecastingissueswhich are now associatecl with decisionmaking a competitivemarket.Forecasting in loads and prices in the wholesale markets are mutually intertwined activities. Models based on simulated artificial agents may eventually become as important on supply side as artificial neural networks have already become
Papers
7. Bunn, D.W., "Forecasting Loads and Prieesin CompetitivePowerMarkets"Pracof the IEEE, Vol. 88, No. 2, Feb 2000, pp 163169. 8. Dash,P.K. et. al., "Fuz,zy Neural Network and Fuzzy Expert System for Load Forecasting", Proc. IEE, Yol. 143, No. l, 1996, 106114. 9. Ramanathan, ef. a/., "Shortrun Forecasts Electricity Loads and Peaks". lnt R. of J, Forecasring, Vol. 13, 1997,pp 161174.
@
Mod"rnpo*", syrt"r An"ly.i.
10. Mohammed,A. et. al., "Shortterm Load Demand Modelling and ForecastingA Review," IEEE Trans. SMC, Vol. SMC12, No. 3, lggl, pp 370_3g2. ll. Tyoda, J. et. al., "An Application of State Estimation to Shortterm Load Forecasting",IEEE Trans., Vol. pAS89, 1970, pp l67g16gg. 12. Mehra, R.K., "Online Identification of to Kalmann Filtering", IEEE Trans. Vol. AC16, lg7l, pp lZ_21.
77
T7.T
INTRODUCTION
Voltage control and stability problems are very much familiar to the electric utility industry but are now receiving special attention by every power system analyst and researcher.With growing size alongwith economic and environmental pressures, the possible threat of voltage instability is becoming increasingly pronounced in power system networks. In recent years, voltagi instability has been responsible for several major network collap$s in New York, France, Florida, Belgium, Sweden and Japan [4, 5]. Researchworkers, R and D organizations and utilities throughout the world, are busy in understanding, analyzing and developing newer and newer strategiesto cope up with the menace of voltage instability/collapse. Voltage stability* covers a wide range of phenomena. Because of this, voltage stability meansdifferent things to different engineers.Voltage stability is sometimes also called load stability. The terms voltage instability and voltage collapse are often used interchangeably.The voltage instability is a dynamic process wherein contrast to rotor angle (synchronous) stability, voltage dynamics mainly involves loads and the means for voltage control. Voltage collapse is also defined as a processby which voltage instability leads ro very low voltage profile in a significant part of the system. Voltage instability limit is not directly correlated to the network maximum power transfer limit. A CIGRE Task Force [25] has proposedthe following definitions for voltage stability. Smalldisturbance voltage stability
A power system at a given operating state is smalldisturbancevoltage stable if, following any small disturbance,voltagesnear loads do not changeor remain *The problemof voltage stability hasalreadybeenbriefly rackled in Ch. 13. Here it is again discussed greaterdetailsby devotinga full chapter. in
lEq!tl#* l,rc'tr
close to the predisturbancevalues. The concept of smalldisturbancevoltage stability is related to steadystatestabiiity and can be analysed using smallsignal (linearised) model of the system. Voltage Stability
voltagestabilityare related to considered. Many of the indicesusedto assess NR load flow study. Detailsof static and dynamicvoltage stability will be 17.5. furtherin Section considered Some Counter Measures
lty are: n counter measuresto avord voltage r (i) generator terminal voltage increase (only limited control possible) (ii) increase of generator transformer tap (iii) reactive power injection at appropriate locations (iv) loadend OLTC blocking (v) strategic load shedding (on occurrenceof undervoltage) Counter measures to prevent voltage collapse will be taken up in Section 17.6. T7,3 REACTIVE POWER FLOW AND VOLTAGE COLLAPSE
to a certain disturbance,the voltages near loads approachthe postdisturbance equilibrium values. The concept of voltage stability is related to transient stability of a power system. The analysis of voltage stability normally requires simulation of the system modelled by nonlinear differentialalgebraic equations. Voltage Collapse
Following voltage instability, a power system undergoesvoltage collapse if the postdisturbanceequilibrium voltages near loads are below acceptablelimits. Voltage collapse may be total (blackout) or partial. Voltage security is the ability of a system, not only to operate stably, but also to remain stable following credible contingencies or load increases. Although voltage stability involves dynamics, power flow based static analysis methods often serve the purpose of quick and approximate analysis. T7.2 COMPARISON OF ANGTE AND VOLTAGE STABILITY
The problern of rotor angle (synchronous)stability (coveredin Ch. 12) is well understood and documented t3l. However, with power system becoming overstressedon accourrt of economic and resource constraint on addition of generation, transfofiners, transmissionlines and allied equipment, the voltage instability has become a serious problem. Therefore, voltage stability studies have attracted the attention of researchersand planners worldwide and is an active area of research. Real power is related to rotor angle instability. Similarly reactive power is central to voltage instability analyses.Deficit or excessreactive power leads to voltage instability either locally or globally and any increasein loadings may lead to voltage collapse. Voltage Stability Studies
Certain situations in power system cause problems in reactive power flow which lead to system voltage collapse.Some of the situations that can occur are listed and explained below. (1) Long Transmission Lines.' In power systems, long lines with voltage uncontrolled buses at the receiving ends create major voltage problems during light load or heavy load conditions. (ii) Radial TransmissionLines: In a power system,most of the parallel EHV networks are composedof radial transmissionlines. Any loss of an EHV line in the network causes an enhancementin system reactance. Under certain conditions the increasein reactive power delivered by the line(s) to the load for a given drop in voltage, is less than the increasein reactive power required by the load for the same voltage drop. In such a case a small increasein load causesthe system to reach a voltage unstable state. (iii) .Sftortageof Local Reactive Power: There may occur a disorganised combination of outage and maintenance schedulethat may cause localised reactive power shortageleading to voltage control problems.Any attempt to import reactive power through long EHV lines will not be successful. Under this condition, the bulk system can suffer a considerable voltage drop. I7.4 MATHEMATICAL
crrrr,lltE l.? Trnl't rit .l l.l.EDllrr I I
The voltage stability can be studied either on static (slow time frame) or dynamic (over long time) considerations.Depending on the nature of disturbance and system/subsystem dynamics voltage stability may be regardeda slow or fast phenomenon. Static Voltage Analysis
FORMULATION
OF VOLTAGE
littD/^t T Erilt r.ClLr.Cr!.GrlVl
Load flow analysis reveals as to how system equilibrium values (such as voltage and power flow) vary as various system parametersand controls are changed.Power flow is a static analysistool wherein dynamicsis not explicitly
The slower forms of voltage instability are normally analysed as steady state problems using power flow sirnulation as the primary study method. "Snapshots" in time following an outage or during load build up are simulated. Besides these postdisturbance power flows, two other power flow based
I 'sls{
analysis.
stability. conventional load flow programs can be used for approxi lmate PV. curves are useful for conceptual analysis of voltage stability and for The model that will be employed here to judge voltage stability is based on a single line performance. The voltage performun"" of this simile system is qualitatively similar to that of a practical system with many voltage sources, Ioads and the network of transmission lines. Considerthe radial two bus system of Fig. 17.1. This is the same diagrarn as that of Fig. 5.26 except that symbols are simplified. Here Eis 75 and yis vn and E and v are magnitudes with E leading v by d, Line angle"p: tunli XlR and lzl = X.
methods oftenused;pv curvesandve curves.(seealsosec. 13.6) are These two methods give steadystate loadabilitylimits which are relatedto vohags
V
'<Locus of V66 and Pr",
Noseof the curve
0.9 pf lead
0.8 pf lag
Fig.17.2 PV curvesfor variouspowerfactors As in the case of single line systerrr,,r, a general power system, voltage instability occurs above certain bus loading and certain Q injections. This condition is indicated by the singularity of the Jacobianof Load Flow equations and level of voltage instability is assessed the minimum singular value. by Certain results that are of significancefor voltage stability are as under. o Voltage stability limit is reached when
(r7.2)
F i g . 1 7 .1 In terms of P and e,the systemloaclenclvoltage can be expressed as [l]. where S = complex power at load bus Yrt= load bus admittance ', V = load bus voltage
L
1: i r ,  z z x  z :  _'il r z o x  E ) '  4 x 2 G 2 + o \ l , , _ < \
z
e7.t)
It is seenfrom Eq. (17.1) that Vis a doublevalued function (i.e. ithas two solutions) of P for a particular pf which determines in terms of p. The pV e curves for various values of pf are plotted in Fig. 17.2. For each value of pf, the higher voltage solution indicates stable voltage case, while the lower voltage lies in the unstable voltage operation zone. fhe changeover occurs at v".t (critical) and Pro*. The locus of v.r,p^u* points for variJus pfs is drawn in dotted line in the figure. Any attempt to iniiease the load abov" causes a reversal of voltage and load. Reducing voltage causes an "* increasing current to be drawn by the load. In turn the larger reactive line drop causesthe voltage to dip further. This being unstableoperation causesthe system to suffer voltage collapse. This is also brought out by the fact that in upper part of the curve ff. dP in lra l^.',o ^^* ..^.,+^Ll ,\ 0 and ^rr flrrv rvwwr panL /\u'Dr.rurc^ pdtL) > U (feduclng lOad means ,,, dV gV reducing voltage and viceversa).It may be noted here that the type of load tir" assumedin Fig' 17.2 is constantimpedance. In practical syste. type of loads are mixed or predomirrantlyconstantpower type such that system voltage degraclationis inore and voltagelnstability occurs much prior to the theoretical power limit.
Nearer the magnitude in Eq. (17.2) to unity, lesserthe stability margin. 'a o The loading limit of transmissionline can.be determinedfrom
lsl = v"3lx";
(r7.3)
X".i is the critical system reactancebeyond which voltage stability is lost. It can be expressed as (tan Q+sec Q) X"n= * 2P
Fz
(r7.4)
We have so far consideredhow the PV characteristicswith constant load power factor affect the voltage stability of a system. A more meaningful charrcteristic for certain aspectsof voltage stability is the QV characteristic, which brings out the sensitivity and variation of bus voltage with respect to reactive power injections (+ve or ve). Consideronce again the simple radial systemof Fig. 17.1.For p flow it is sufficiently accurate to assume X > R i.e. 0 = 90".It then follows that
o = EY.o,a 11
X X
(17.s)
or V"  EV cos 6+ QX = 0 Taking derivative wrt V gives d Q = E c o s6  2 V d V X
a
(r7.6)
(17.7)
Using ihe decoupiing principle 1.". dP = 0, we set dV
The QV characteristic on normalized basrs(etf**. VIE) for various values of P/P^ are plotted in Fig. 17.3.The system is voltage stable in the region where dQldv is positive, while the voltage stability limit is reached at d,eldV = 0 which may also be termed as the critical operating point.
+=."'r[#.+]
or or Isc=cosoO+11 dl
LdV X J X J LdV Voltagestability is achieved when
o
D , max
= Ers6 E cos5l !9 + 2Y]
E cos , (#
1.0 0.75 Unstable. . operation P I P ^ " , =9 . 5 Pr", is the maxlmum powertransferat upf
.
+)
> Ers,
(short circuitMVA of powersource)
(17.10)
cntenon dV voltage instability occurs when the system Z is such that
,..\ dz (lU 
0
0.2
0.6
0.8
1.0
VIE
Fig 17.3 QV characteristics the systemof Fig. 17.1 tor for variousvaluesof plp^^r. The limiting value of the reactive power transfer at the limiting stage of voltage stability is given by
(r7.1r) dZ dV Application of this criterion gives value of z";. \ (iii) Ratio of source to load reactance is very important dnd for voltage stability
Xsource a o2 X load
v M d o o o r  = u =
(17.r2)
a indicates the offnominal tap ratio of the OLTC transformer at the load end. T7.5 VOLTAGE STABILITY ANALYSIS
= Q,n^ t"o'26
(17.8)
The inferencesdrawn from the simple radial system qualitatively apply to a practicalsize system.Other factors that contribute to system voltage collapse are: strengthof transmissionsystem,power transfer levels, load chaiacteristics, generatorreactive power limits and characteristicsof reactive power compensating devices. Other (i) 34 Criteria of Voltage Stability
..it.rion: (E'=generator voltage; V=load voltage). Using this crite
rion, the voltage stability limit is reachedwhen
The voltage stability analysis for a given system state involves examining following two aspects. (i) Proximity to voltage instabitity: Distance to instability may be measured in terms of physical quantities,such as load level, rea power flow through a critical interface, and reactive power reserye.posiible contingencies such as a line outage,loss of a generating unit or.a reactiu. po*"rio*.. must be given due consideration. (ii) iuiechanismof voitage instabiiity: How and why does voltage instabitity take place? What are the main factors leading to instability?'ulhat are the voltageweak areas? What are the most effective ways to improv" stability? "of,ug,
cos{#.#}+sindffi}=o d
(r7.e)
.ffiiffi
.
Analysis System Power Modern
Static AnalYsis of systemconditionsat variousdme captures The staticapproach .snapshots
frames along the timedomain trajectory. At each of these time frames' X in
of system The static analysis techniques permit examination of a wide range the nature of the problem and give the main conditions and can descriUe of specific contributing factors. Dynamic analysis is useful for detailed study coordination of protection and controls, and testing voltage coliapse situations, and how the of remedial rneasures.Dynamic simulations further tell us whether point wr steadystateequilibrium Modelling Loads Detailed Load modelling is very critical in voltage stability analysis. representationin a voltageweakarea may be required' subiransmissionsystem and This may includeiransformer ULTC action, reactive power compensation' voltage regulators of loads' to It is essential. consider the voltage and frequency dependence Induction motors should also be modelled' Generators and their excitation controls Requirements of various Power system components
to redPce purely equations system the Thus, overau frame. ii h?lrJdil. allowing'h.t..Ytt ,t'1tit Tlv:i: ::*:1,::t"t' vlp qnrt equations algebraic :f vo by computin andvQ "'?'J"6;ffiffi;;;g."riuulirv s is determrnedusing static analysis havb been curves at selected load buses.Special techniques VQ sensitivity such as eigenvalue (or reported in literature. Methods based on methods give stabilityrelated modal) analysis have been devised. These and also identify areasof potential information from a systemwide perspective problemst13151. Proximity to InstabilitY
.vP
AVR, load It is necessary to consider the droop chatacteristics of the AGC, protection and controls should compensation,SVSs (static var system), also be modelled appropriatelyl4l. Dynamic 'lhe AnalYsis
by increasing proximity to smalldisturbance voltage instability is determined unstable or the load flow fails in steps until,the system becomes il;^";ruiion determining the point Refs. t16181 discussspecial techniquesfor to converge. instability' of voltage collapse and proximity to volage The Continuation Powerflow Analysis
analysis is rc generalstructu of the systcm moclel for voltage stability overall system equations may be similar to that for transient stability analysis. as cxprefihcd
*=f(X,n
and a set of algebraicequations I (X, V) = YxV (Xo' Ve)' with a set of known initial conditions X = system state vector where = but voltage vector Y 1= current injeition vector Ilv = rletwork node admittance matrix' (17.13)and (17'14)can be s Equations oi tt e numerical integration methods in clescribed Ch' 6' T analysisnrethocls the special models repl ,ninutes. As bee ieading to voltage collapse have is considerably differential equations models.Stiffnessis also called synchrc
(r7.r3) (r7.r4)
at the voltage stability limit. A! a result' The Jacobian matril becomes singular have convergenceproblems at operating conventional toadflow algorith*, rnuy The continuation powerflow analysis conditions near the stability limit. the loadflow equationsso that they overcomesthis problem by reformulating the possible loading conditions. This allows remain wellconditioned at all of the PV forioth upper *d lo*"t portions solution of loadflow problem *fi"t:lltinuationmethod and flexible and of powerflow analysis is ,obrrrt difficulties' However' with convergence is to suming. Hence the better approach (NR/FDLF) and continuaflow irethod :ase, LF is solved using a conventional increasingload levels ns for successively is Hereafter, the continuation method method is ns. Normally, the continuation point' I exactly at and past the critical Voltage StabilitY with HVDC Links
for extremely long distance (HVDC).litt tl:1r:t.:d High voltage direct current* ffansmissiclnanclftrrasynchronousinterconnections.AnHVDClinkcanbe
maY refer to [3]' *For dctailcd accountof HVDC, the reader
f00 
powerSystem Modern Anatysis
'
VoltaseStability
I
__{tr*g
either a backtoback rectifier/inverter link or can include long distance dc transmission. Multiterfninal HVDC links are also feasible. The technology has come to such a level that HVDC terminals can be connected even at voltageweak points in power systems. HVDC links mav present unfavourable "load" characterisficsfo fhe nnrr/rr converter consumesreactive power equal to 5060vo of the dc power. HVDCrelated voltage control (voltage stability and fundamental frequency temporary over voltages)may be studied using a transient stability program. Transient stability is often interrelated with voltage stability. Ref. t2i .onJia.r, this problem in greaterdetail. 17.6 PREVENTION OF VOLTAGE COLLAPSE
monitoringand analysisto identify potentialvoltage stabilityproblems are remedialmeasures extremelyhelpful. and appropriate T7.7 STATEOFTHE.ART, FUTURE TRENDS AND CHALLENGES
due The present day transmission networks are getting more and more stressed to economic and environmental constraints.The trend is to operatethe existing networks optimally close to their loadability limit. This consequentlymeansthat the system operation is also near voltage stability limit (nose point) and there is increasedpossibility of voltage instability and even collapse. Offline and online techniquesof determining state of voltage stability and when it entersthe unstable state,provide the tools for system planning and real time control. Energy managementsystem (EMS) provide a variety of measured and computer processed data. This is helpful to system operators in taking critical decisions inter alia reactive power management and control. In this regard autornationand specializedsoftware relieve the operator of good part of the burden of system management but it does add to the complexity of the systemoperation. Voltage stability analysis and techniques have been pushed forward by several researchers and several of these are in commercial use as outlined in researcheffort is being devoted this chapter.As it is still hot topic, considerable , to it. Pw et al. l8l considered an exponential type voltage dependentload model and a new index called condition number for staticvoltage stability prediction. Eigenvalue analyseshas been used to find critical group of busesresponsible for voltage collapse. Some researchers[26] have also investigated aspects of bifurcations (local, Hopf, global) and chaos and their implications on power systemvoltage stability. FACTS devicescan be effectively used for controlling the occurrence of dynamic bifurcations and chaos by proper choice of error signal and controllergains. Tokyo Electric Power Co. has developed a pPbased controller for coordinated control of capacitor bank switching and network transformer tap ctranging.HVDC power control is used to improve stability. More systematic approach is still required for optimal siting and sizing of FACTS devices.The availability of FACTS controllers allow operationclose to the thermal limit of the lines without jeopardizing security. The reactivepower compensationclose to the load centresand at the critical busesis essentialfor overcoming voltage instability. Better and probabilistic load modelling [11] should be tried. It will be worthwhile developingtechniquesand models for This may requireexploring study of nonlineardynamics of large size systems. new methods to obtain network equivalents suitable for the voltage stability analysis. AI is another approach to centralized reactive power and voltage control. An expert system [9] could assistoperatorsin applying Cbanksso that
(i) Application of reactive powercompensating devices. Adequate stability margins should be ensured by proper selection of compensation schemes terms of their size, ratingrund locations. in (ii) control of network voltage and generator reactive output Several utilities in rhe world such as EDF (France), ENEL (Italy) are developing specialschemesfor control of network voltages and reactive power. (iii) Coordination of protections/controls Adequatecoordinationshould be ensuredbetweenequipment protections/ controls basedon dynamic simulation studies. Tripping of equipment to avoid an overloadedcondition should be the last alternative. Controlled system s€paration and adaptive or irrtelligent control could also be used. (iv) Control of transfurmer tap chan.gers T'apchangerscan be controlled, either locally or centrally, so as to reduce the risk of voltage collapse. MicroprocessorbasedOLTC controls offer almost unlimitedflexibility for implementingULTC control strategies so as to take advantage the load characteristics. of (v) Under voltage load shedding For unplanned or extreme situations, it may be necessary to use undervoltageloadsheddingschemes.This is similar to under ir"q1r"rr.y load shedding,which is a common practiceto deal with extreme situationsresulting from generationdeficiency. Strategic load sheddingprovides cheapestway of preventing widespread
VOltaSg C O___r_ _ llanse Lnarl sherldino cnl,aoo uvrrvrrrvJ olrn,,r,r orr\_rrll\.t L ^ ug r^^: ^^) ugsrBlrtru
differentiate berween faults, transient voltage dips unJ lo* conditions leading to voltage collapse. (vi) Operators' role
^^ su
as
t()
voltage
Operators must be able to recognise voltage stabiiityrelated symptoms and take requiredremedial actions to prevenr voltage collapse.Online
a00.l,*l
t
ModernPo
I
. .
generatorsoperate near upf. The design of suitable protective measuresin the event of voltage instability is necessary. So far, computed PV curves are the most widely used method of estimating voltage security, providing MW margin type indices. PostdisturbanceMW or MVAr margins should be translated to predisturbance operating limits that operators can monitor. Both control centre and power.plant operators should be trained in the basics of voltage stability. For operator training simulator [10] a realtime dynamic model of the power system that interfaceswith EMS controls suc