COIIII".tf!r llletl.otls in '·oll;cr Sf/sten •.
r·lnalysis
.'
Jr T ,4 e 'j 9L't
<,
~~/
(/ lZ ~L .. ,. _
.9.L~ c~7:! ,::dL •
McGrawHill Series in Electronic Systems
John G. Truxal and Ronald A. Rohrer Consulting Editors
Schuliz and Melsa Stagg and ElAbiad Timothy and Bona
State Functions and Linear Control Systems Computer I'rfethods in Po icer System Analysis State Space Analysis: An Introduction
CtulIllll,tc'r il'ctl,olls in 1#{Ht;C'r S,/stel", .ltn,(llgsis
Glenn u·. ,fit,agg l! ad, ,Engineering A na(\'sis and Computer Dirision ,\ tucrican Electric Poiccr Service Corpora/io'fl
Ahitned fl. ElJlbiud
Professor of Elcclrical EfI{/i"nccrin,q Purdue Unircrsil;y
INTER.VA 710.'J;,uoJ STu.nENT EDITIO.V
T')I.~lIfi A.I!~kla lid nil s . srl d.ar] ./,,/l.a.TI !1.1';~bl(rli f.A!1<i(m" Merico ,.,\'(,Uj 1)p1lhr "/'n_rUL_tnl'J S(UI /JrtnlfJ Sin,ytll)IJTt.· .Sllrlnf!Jj
Computer ltlethods in Power System Analysis
I.': TU!.vA TfOSAI. STULJfST ElJITlO.\'
Exclusive rights by McGrawHili Kogakusha, Ltd., for manufaclure and export. This book cannot be reexported from the country to which it is consigned by McGrawHill.
V
Copyright @ 19GB by McGraw· Hill, Inc. All Rights Resrrved. No part of this publication may be reproduced, stored in a ret r ieval system, or transmitted, in any form or by any means, electronic, mechanical, p horocopying , recording. or otherwise, without the prior written permission of the publisher.
Library o] CVII{/n'ss Cat ol oc Cord Number 67128(j:l
This book present s techniques that huvo been a pplied sur cessf u llv in solvint: pOI\'l'r syst crn pr ohlerns with a dicital computer. It ran thus serve 11.; a text ior ad "anced power svst crn courses to inform pruspr+t in' I"'" er cnuineers of met liod s currently emploYf>o in the electric utility inriutr y. Hccause of the inrr cairu; use of the computer as an indi~pensabl(' t",,1 ill power system enuincer ing, this book will also serve as a basic reference fur power system eneineers re"l'omihlr:
Ir.r the development of computer applications.
The material contained in the text has heen developed from not es for specia] twoweek courses offered since JD64 at Purdue Lniversity, The l·nivcr.,in· of ,risconsin and the {·niversit.y of Santa ('lam. These courses were altl'nJed hy representatives of universities. elect rip utilities, and equipment manuracturrr s.
Solution techniques arc presented for the three problems encountered mo.'t frtq·.:, n t ly in power system analysi, namely, short circuit, load flow. aud 1,,'\1('1' ytcm stability. In addition to an engineering description of these [If,,I,1.'I1;', :1,(' rna tho mnt ical t cchniquc that nrr required for a computer ~nlllti"n :Lri' rir,crilwd. Thus, relevant. mnt ei ial is included from rna t r ix all!:!'bra and n u rncr ic:lllll1aly~is. It is ass urned , however, that t he rradrr has a generallinderslandinl! of clernentarv power system analysis.
Chapter J presents, as a brief introdur t ion, the impact nf computer" on tower syt crn enuineer ina, the orient at ion of rnl!:inrrring problcru to comput crs. ami the advantages of digital computation. Chapter 2 COHr, thc ba ic pr inciplr, "f matrix ulcrbrn and provides ,uffil"ipnt b:l"k~rollnd in matrix t hcor v f"r ~ I", nrnni nd cr of t I,I' book. For fI'arl,'1" fami! inr 1\ it I, n,at r ix t<'f'itlli'lIl(,S, I hi, "I,a: 01 ('r serves 8.5 a. review and cst ablihes the not nt ion used throuuhout the t cxt . l ncidrnrr anri net ,I'ork matrices arl' introdurod in Chapter :~, which prrsr nt s t hr: 1 ",.j, n iq uos for dc.'crit.irlg t),P ~""n\f·t ri" "I r nrt urr of a nrt work anel ()\!t Ii 11<''' the' t r ansf or mnt ions required to drri ,'P nrt "', .rk mal rices. Titr Iorrnnt ion (If I hc~e rnat r iccs is t hc first ~t"l' in the annlvi. (,f I'''\\,('i' ".\·,trm jlr(l"lcm~, (,h:ll'tt'r 4 presrnts aknrithm!< whirh r a n 1,(, lI"ril in all a lt er nn t i vc mt hr.d fM thr Ior mnlion of certain network rnatriccs. Th"~l' ~lg"rithll\:i have proved to 1,(, dfL'l'iin' [lOr use in computer calculat ion. The ruct hods dcscr ihed in (,h3I't('r,;:3 and .j 31'(' developed for s inglephase representation of power systems. Chapter S extend,
these methods for threephase representation. The application of network matrices to short circuit calculations is presented in Chapter G. Several methods are included and a typical computer program i" dcrribcd to illu: .. t r.ur a practical application uf the techniques.
Chapter 7 contains a hrid int rndur t iou til t l.c .";OIUli()1I (,f linrar :IlId ""II· linear simultuncous algehrai« «quat ion. This IllalPl"l:d i, l,n'.,,;clIled ill a n.uurur that affords direct application to the solut iOB of the 11I3d flow problplIl. Til,' formulation and solution of the load A,)w pr oblcu: i~ IJI'('''l'nt('d in Ch:'ptf'l ~. This chapteralso describes the procedure" for handline volt ngcrout rol lcd Lues. transformers, and tie line control. The different mct l.od, :HI' compared Ir..m several points of view and a description i., giv('n of an a ct uul I'r"J!l'alll lH'd fur load flow calculations. In a manner similar to t hnt in Ch aptc r I, C],apter Sl introduces methods for the numcricul olut io» "f t l. r dirr .. r<'loIi,,1 "'1,,;oli(Jlh t!.:lt are required fur transient stabilit y t udivs. Chaptc'r IU ,'(J"l'r.' th,· ;"mlhul:oli"ll and solution techniques emplovcd in t ra nsient st udics and pr".,('nt$ proced ur .. , for the detailed representation of svnchronous and induct im rnachinc. {'v'i:n and governor systems, and the distance relays. Au actual t runirnt :,t:dlility computer program is described.
The first efforts in the development of this material W('Ie rnude ill the cu rlv 1950s at the American Electric Power Service Cor porat ion as a result of the interest in the application of computers to the planning anr! opera t io n of electric power systems. In J959, the authors had an o pport uuit v to work t ozet hrr ns members of the staff of the American Electric Power ::\el'\'ire Corporation and continued to work together on a parttime basis for several years. This made possible the f~urther development of basic computer met hod . .; established in previous years.
This research work was endorsed enthusiast ically hy the manauernent of the American Electric Power Company. The authors wish to express their a pnrec intion for this support.
It is. a pleasure also to acknowledge the conutnutions of those who have helped in the preparation of this book. The a u t hors would like I':lrt'il'lilariy to thank Jorge F. Dopazo, who studied the. text in derail and made ma nv ,uggc~tions; Marjorie Watson, for her contribution related to th€' rnat hemat ical t cchniques and for edit ing the manuscript; and G. Hobert Hail(!y, Dcnni \\'. John.,t,nl. Kasi Nagappan, Janice F. Hohenstein, and other mcrnhors of the Engincerirur Analysis and Computer Division. Tho authors would like :1],') to thank I'rof(',~(Jr~ Ar un G. !'l.:ulke and Daniel K. 11";1:111 hf TIlt' ('lIiq','sil_\' .. [ Wi'n,II,ill f",' their helpful comments in re\'icwill~ th .. trxt . Last, l.ut "(!r\ailll~' not 1":I.,t, sincere thanks to Constance Aquila for her ex .... .llcnt work iu thr tYI,illg and general preparation o~ the rnanucript.
Gl"lIn IP. Sruf!.f!.
Allnl(''/ 11. £1.AtJiocl
l n t r arl u c t i o n
I I i~t"rir.11 n(1I('
.2 Irn pnct of rorn put ors 2
.:;; nri,~n:ati"n of ('nl!in"Prin~ pr oblcrns to ccm puters 1
.\lntrix alg .. br a
£.1
2_'2 H:l!'ic concepts and dpEnit ions ;
2.8 iJ('tcrnlinnn:s II
:?.; '.bl"i~ operations 14
.: .. ', I.;IH'"r dClwn,J~flcc and rank of n matrix l!l
2 r: Li nca r cqun t ions :21
ch a p t .. r 3
Ln ci do nr e and rlpt.k'nrk macric es
s.t Intr(1duction:27
s.e Graphs 2.';
.).,) Inf'icl~n('c mn t rice :) I
8,.~ Prilllill\·C'TH·twllrk III
.1._:; FlIrlllariCiri (If nrt w nr k 11I:ltrjc('~ by singular transformations 42
'j)j FOfillali(dl ni nrt wor k rll:1trj(·f>~ by nonsincular t ransfor mat ious 4:<
.); L\:,mplp (~r fllrm:ltinn oi in('ld"flr'~ nnr:l network mA.trir:~R. ;,~)
c/II'I'I('r 4
Alunr; III TIl.' [nr [ar m a tinn of network matrices
.:.1 Lnt rod urt.iou ;~l
.;.2
.\!J.;orillllll i"r [nr rua t mn pr bus impcdanre matrix 79 :'>lodificalinll fir the btl" impedance matrix
for ha nccs in tl,c nrt wnrk W)
r:':lrllpk of ["II"":i",, a nd modification of bus impedance matrix 92
.~ .. j
, , ., .. "
chapter 5
chapter 6
chapter 7 7.1 7.2 7.8 7.4 7.5 7.6 7.7 7.8
chapter 8
4.6 Derivation of loop admittance matrix from bus impedance matrix lP4
4.6 Example of derivat ion of loop admittance matrix from bUB impedance matrix 110
Threephase networks
5.1 Introduction 117
6.2 Threephase network e lerncnts 117
5.3 Threephase balanced network clements J:.!U
5.4 Transformation matrices 1:l3
6.6 Thrccphasc unbalanced network element a I:!::'
5.C Incidence and network matrices for thrcephuse net wr.r k s 1:!(j
6.7 Algorithm for formation of threephase bus impcdu nce matrix 1"(;
6.8 Modification of the threephase hua impedu nrc m at r ix for changes in the netwurk I:ltj
5.!! Example of formation find rnodifira uon uf
thrcc·plllll'tC network IIJhlriC'I'1i 1 :r;
Short circuit studie~
0.1 In trod uct.ion IG7
6.2 Short circuit calculations using Z"t's 1Gb
6.3 Short circuit calculations for balanced
threephase network using Z sus 173
6.4 Example of short circuit calculations using Znv., Ib4
6.6 Short circuit calculat ions using Z LOOP 201
6.6 Example of short circuit calculations using ZLOOP 20G
6.7 Description of short circuit program 21.'>
Solution of si,nultaneoU3 algebraic equation. Introduction 223
Direct methods for solution of linear algebraic equations 224 Example of solution of linear equation, by direct methods :!:li Iterative methods for solution of linear algchrnic equut ions 240 Example of solution of linear equations by it ers tive methods 24:) Methods for solution of nonlinear algebraic equations 241\ Example of solution of nonlinear equations 250
Comparison of methods 253
Load!Ww 8tudie~
8.1 Introduction 257
8.B Power system equations 259
8.S Solution techniques 261
8.1; Acceleration of convergence 281
8.6 Examplcs of load flow calcula t ions 283
8.6 Voltage controlled buses 312
8.7 Representation of transformers 31 i
8.8 Tie line control 324
8.9 CompariBon of methods 325
8.10 Description of load flow program 332
cho p ter 9
Numerical aolution of differential equationa
9.1 Introduction 343
B.t Numerical methods for solution of differential equations 343
9.8 Solution of higherorder differential equations 353
9.1, Exnrnples of numerical solution of differential equations 353
9.5 Comparison of methods 362
chapter 10
Tran~ient stability Hudies
10.1 Introduction 36j
10.t Swing equation 366
10.S Machine equat ions 3G9
10.1, Power system equations 374
10.5 Solution techniques 3i8
10.6 Example of transient st abil it.y calculations 386
10.7 Exciter and governor control systrms :;!)j
10.8 Distance relays 4n3
10.9 Description of transient stability program 410
Lndez 4£1
W'
/~~
/ .~ ...  .. . I j/ ~:,, '
t • ~.,,___ ~. _ ..
~
c~ ... ' ... :i .. 't.e ........•. r. I.
I _' i c "<.,
Intr ..... ; .. ~.
1.1 Historical note
The grcat t cchnicnl :l(h'a!lc(,~ in the de"igll and production of commercial .uid scientific generalpurpose digital computers since the early 19,')Os have pl.iccd a powerful tool :t t t he disposnl (;[ the cngi nccring profession. This .rdv.mccmcut 11:\" m.idc eCI)Il()mic:!lly fea"iblc the utilization of digital r"m]l\ltrr.~ for rou t ino c.ilcul.u ions encountered in everyday engineering work. III :1c1clitioJ1. it li:!s prrl\'idet! the c a pabilit y for performing more :ldY:tnced l'Ii~IIl('crill~ :'llid "r:iClltliic comput n tions that were previously Inlp"s"iL](: ll('(';'ll."(: ,,1 !l'!'ir cllmpl!:;; (ll' t imcconsum.ng nature. All t lu«: t rcnrls 11:1\'(' ill(:n'''':l'd irnrncnrlv tlte i nt crcst in digital computers :Inrl han; ll!'c(''',itat!''] ;1 bct tcr 1IIltier;;t:lnding of the engineering and Illath('m;\tic;;] h:t,,(:s for pl'"hlclll ."(li\'iTlg.
The planninz, cJ('si}!Tl. and opcrat ion of power systems require cont inu ous ;\n<1 r"mprchell~i\'c analysi, to cvnluatc current system perftll'mallre .uid to :\"c(,rt:lill the cff cct i vcncss of alternn t.ivc plans for system expansion. Tltcsr "t\ldIC~ play an import ant role in providing a high t andard of piwcr svst crn I'ell:lbility and ensuring the maximum utilization of capital invcst mcnt.
The comput at ional task of determining power flows and voltage levels rcsul t i nu frol11 a single opcra tiug condition for even a small network is :111 but insurmountable if performed by manual methods. The need f"l' computational aid . .; in power syst crn engineering led .in 1929 to the design of a specialpurpose analog computer called all ac network analyzer. This device made possible the study of a greater variety of system operating conditions for both present and future system designs. It provided the abili ty to determine power flows and system voltages during normal and emergency condi tions and to study the transient behavior of the
2 Computer method" in power lIyHem analYllill
system resulting from fault conditions and switching operations. By the middle 1950550 network analyzers were in operation ill the Ullited States and Canada and were indispensable tools to planning, relaying, and operating engineers.
The earliest application of digital computers tv power system problems dates back to the late 19408: However, most of the early applications were limited in scope because of the small capacity of the punched card calculators generally in usc at that tim«. The a v.ul.rbi l i t v of In ruescale digital computers in the middle 1 \).,)Os provided eq uipmcn t uf su llicient capacity and speed to meet the rcquiremcnt s of major power system problems. In 19.')7 the American Electric Power Servicc Corporut.i on completed a largescale load flow program for the IB.\[ 704 which calculated the voltages and power flows for a specified i'()\\'cr ;;)"';\('111 not \\·mk.
The initial application of the load flow )!r()~rall1 tu t r.uimission planning studies proved so successful that all subsequent studies empl()yed the digital computer instead of the network :walyzer. The success of this program led to the development of progrums for short circuit and transient stability calculations. Today the computer is an indispcnsaule tool in all phases of power system planning, design, and opcru t i on.
1.2 Imp{lct of computers
The development of computer technology has provided the f ollowiug advantages to power system engineering:
l. More efficient and economic means of performing; routine engilleering calculations required in the planning, design, and opor.it ion of a power system
:2. A better utilization of engineering talent by relicving the engineer from tedious hand calculations and permitting him to spend more time on technical work
3. The ability to perform more effective engineering studies by applying calculating procedures to obtain a number of alternate solutions for a particular problem to provide a broad Lase for engineering decisions
4. The capability of performing studies which heretofore were not possible because of the volume of calculat ions involved
Two major factors which have contributed to the realization of these benefits are the declining cost of computing equipment and the development of efficient computational techniques. Now that a substantial reduction in computing cost has been effected, principal effort must be dirccted toward the orientation of engineering problems to computer solutions.
Chapter 1
Introduction
3
1.3 Oriefltation of engineering problems to computers The process of applying a computer to the solution of engineering problems involves a number of distinct steps. These steps are:
1. Problem definition Initially, the problem must be defined precisely and the objectives determined. This may be the most difficult step in the entire process, Cnnsldcrnt.ion must be given to the pertinent data availnblc for input, t ho scope of the problem and its limitations, the desired results, and their relative 'im portuncc in making an engineering decision. This phase requires the judgement of experienced and capable engineers.
2. Jf ath emniical formulation After the problem has been defined, it is necessary to develop a mathematical model to represent the physical system. This requires specifying the characteristics of individual system components as well as the relations which govern the interconnection of the clements. DifTerent mathematical models may be used to represent the same system and, for many problems, complementary (dual) formulations may be obtained. One formulation may result in a different number of equations than another as, for example, in the case of network problems which can be solved using either loop equations or node equations. The mathematical formulation of the problem, therefore, includes the design of a number of models and the selection of the best model to describe the physical system.
8. Selection of a solution techriique The formulation of most engineering problems invol ves mat hema tical expressions, such as sets of nonlinear equations, differential equations, and trigonometric functions, which cannot be evaluated directly by a digital computer. A computer is capable of performing only the four basic arithmetic operations of addition, subtraction, multiplication, and division. A solution for any prol.lern , thcrof'orr, must be obta incd hy numerical techniques which employ the four l.nsic arithmetic operations. It is important in this phasr: to s('l('('t n mot hrxl which is prncticnl for machine computation and, ill pn rt iculur. will prodllcc t hr: desirnd results in a reasonable amount of computer time. Since numerical approaches involve a number of assumpt ion», (':1.r('[111 considrr.rt ion must hf' given to the degree of accuracy ["C'III ired.
4. Program dcsiqn The sequence of logical steps by which a particular problem is to be solved. the allocation of memory, the access of data, and the assignment of input and output units are important aspects of computer program design. The objectives are primarily to develop a pro
" Compuur methods in power sy.um analysis
cedure which eliminates unnecessary repetitive calculations and remains within the capability of the computer. The program design is usually prepared in the form of a diagram called a flow chart.
5. Programming A digital computer has a series of instructions consisting of operation codes and addresses which it is able to interpret and execute. In addition to the arithmetic and input! output instructions, logical instructions are available which are used to direct the sequence of calculations. The translation of the precise detailed steps to be performed in the solution of a problem into an organized list of computer instructions is the .process of programming. A program can be developed by using computer instructions in actual or symbolic form, or it can be written in a generalized programming language, such as FORTIL\X.
6. Program verification There are many 'opportunities to introduce errors in the development of a complete computer program. Therefore, a systematic series of checks must be performed to ensure the correctness of problem formulation, method of solution, and operation of the program.
7. Application Engineering programs, in general, can be classified into two groups. The first consists of specialpurpose programs, which are developed in a relatively short period for the solution of simple engineering problems. Such a problem is usually well defined, and often the program completely serves its purpose after the first series of calculations has been completed. Some small programs are used on a continuing basis but are restricted in their use because of their specialpurpose nature.
The second group consists of generalpurpose programs that are designed for the anulysis of large engineering problems. T'h esc pr ogrnms are applied extensively ill the regular studies of one or mure engineering departments. Their use may have an effect on the approach to an engincering problem and the organization of IL study, Thus, it ill important that consideration be given to the manner in which a program is to be empioyed in an engineering activity. Some aspects which must be considered are means of collecting and preparing data, processing time, and presentation of results. Programs of this type are becoming an integral part of power system engineering.
The relative importance of each of these steps varies from problem to problem. Moreover, all steps are closely related und play Il.ll important role in the decisions that must be made. Of primary importance is the interrelation of the mathematical formulation of a problem and the selection of a solution technique. Frequently, it is difficult to evaluate the
Chapter I
Introduction
5
influence of lhese two step~ Oil each other without developing a complete program and performing nrt u.rl calculations to compare the alternatives.
The material covered ill this hook pertains to the first three steps, with particular clllplia"is 1>11 t l u in\nrciatil>!ls 1,[ qt:ps ~ and :). Simplified flow charts are used to illustrate the method presented.
Biblio gra ph v
Gross, Eric T. B.: :\et work Analyzrr TnstallatilJ1l~ in Canada n n d the United States, Pro: .. 1 III. Pou:cr COII/., vol, 21, pp. GG:iG(j9, l!J;j9.
St. Clair, H. P., :ilid G w Stagg: Digital Computer ~olutioll of Power System Problems. Proc .. 1 Ill. Poircr Conj., vol. 20, pp. fiJ4G2\l, l!J.")8.
St. Clair, II. 1'., G. \Y. St:,gg. and \Iaria Tvchrr nr: : Digit.al Computer Takc,.; O\,pr LI,:td FI')II' C'alcnlations, Electrical !VII/·lri. Sept. :~(), 1~F)7.
Stagg, G. \\'., .. \. I.'. C;:t\'f1cile. and .T. F. Hohcnstein: Digital Computer Calculates Trnuirnt Stability Problems, RII,I;lriwl Wor/ri, Sept. J, 1958.
Stagg, G. '\'., and E. L. \\"izemann: Computer Program for Load Flow Study Handles Tensystem Interconnection, Electrical World, Aug. 1, 1\)00.
Zuckernick, S . .T., and C. \\'. Stagg: Computer Solves Relay Problems, Electrical World, July 20, l\J;jD.
W'
/~~
/ .~ ...  .. . I j/ ~:,, '
t • ~.,,___ ~. _ ..
~~
cWl!!!:!5r 2
]"atri~~!!._,
2.1 'ntroduction
In recent years, the use of matrix algebra for the formulation and solution of complex engineering problems has become increasingly important with the advent of digital computers to perform the required calculations. The application of matrix notation provides a concise and simplified means of expressing many problems. The use of matrix operations presents a logical and ordered process which is readily adaptable for a computer solution of a large system of simultaneous equations.
2.2 Basic concepts and definitions
.Hatrix n o t.a t.io n.
Matrix notation is a shorthand means of writing systems of simultaneous equations in a concise form. A matrix is defined as a rectangular array of numbers, called elements, arranged in a systematic manner with m [()WS and n columns. These elements can be real or complex numbers. A doublesubscript notation a., is used to designate a matrix elemcn t. 'I'h« Ii rxt subscript i d('sil';l1atp>l t.he row in which Lhe clcrnen t IiI;>;, and t.h« second subscript j designates the column.
In the following system of equations,
allX1 + a12X2 + aux. = Yl a.1X, + a2ZX. + anxa = y. a31X' + a32XZ + a33Xa = YJ
X" Xz, and Xa are unknown variables; all, all, an, , an are the coefficients of these variables; y" Yt, and Y3 are known parameters. The
(2.2.1)
8 Computer m et.h.od« in power syotem arusl ysi s
coefficients form an array
[till till
a'l a22
a31 aJ2
(:!.:2.:.n
which is the coefficient motrix of the system of equations (:2.2.1).
Similarly, the variables and parameters can be written in matrix form:
and
(2.2.3)
The matrix (2.2.2) is designated by a capital letter A and the matrices (2.2.3) by X and Y, respectively. In matrix notation the equations (2.2.1) are written
AX = Y
A matrix with m rows and n columns is said to be of dimension m by n, or m X n. A matrix with a single row and more than one column em = 1 and n > 1) is called a row matrix or row vector. A matrix with a single column and more than one row is called a column matrix or column vector.
Types of matrices
Some matrices with special characteristics are significant in matrix operations. These are:
Square matrix When the number of rows equals the number of columns, that is, m = n, the matrix is called a square matrix and its order is equal to the number of rows (or columns). The elements in a square matrix a,j for which i = j are called diagonal elements. Those for which i ~ j are called offdiagonal elements. For elements a., to the right of the diagonal i is less than i. and for those to the left of the diagonal i is greater than j.
Upper triangular matrix If the. elements a., of a square matrix arc zero for i > i. then the matrix is an upper triangular matrix. For example:
Chapter 2
l'rfatriz algebra
9
Lower triangular matrix If the elements aij of a square matrix arc zero for i < i, then the matrix is a lower triangular matrix. For example:
Diagonal matrix If all offdiagonal clements of a square matrix are zero (a.;j = 0 for all i ,e ij, then the matrix is a diagonal matrix. For example:
[all
A = 0 o
o
Unit or identity matrix If all diagonal elements of a square matrix equal one and all other clements are zero (o;j = 1 for i = J and a., = 0 for i ,e j), the matrix is the unit or identity matrix, designated hy the let ter U, For example:
Null matrix If all elements of a matrix are zero, it is a null matrix.
Transpose of a matrix If the rows and columns of an m X n matrix are interchanged, the resultant n X m matrix is the tronspose and is designated by At. For the matrix
the transpose is
Symmetric matrix If the corresponding offdiagonal elements of a square matrix are equal (c., = aj;), the matrix is a symmetric matrix. For example:
5 2 6
~l
10 Computer rnethods in power system analy~is
The transpose of a symmetric matrix is identical to the matrix itself, that is, A' = A.
Skewsymmetric matrix 1f A = :1' for u square matrix, A. is a skcu:symmetric matrix. Thc corresponding oITdi:lg()!lal elements arc equal l.ut of opposite sign (c., =  aji) and the diagonal elements arc zero. Fur example:
A
[ ~
3
5 o u
3]
()
o
Orthogonal matrix If A'A = U = .'1;1' for a sqllarc ma t rix with r('al elements, then A is an orthogonal matrix.
Conjugate of a matrix If all the clements of a matrix arc replaced by their conjugates (replace the clement a + jo by a  jo), the resultant matrix is the conjugate and is designated by .4 *. For a matrix
[ J3
A = 4 + J2
the conjugate is
A." = [4  J3 5]
 J2 1  )1
If all the elements of A are real, then A. imaginary, then A. = A *.
A *. If all clements nrc pure
Hermitian matrix If A = (A *)' for a square complex mnt rix, :1 is a Hermitian matrix in which all diagonal clements arc reed. For example:
[ 4 2 5 j3]
A = 2 +]3
SkeuiHermiium matrix If.4 = (A *)' for a square complex matrix, A is a skeicHermitian. matrix ill which all diagonal elements arc either zeru or pure imaginary. For example:
A = [ 0 2 0 J3]
2 j3
Unitary matrix If (A *)'.1. = U = A(A *)' for a square complex matrix, A. is a unitary matrix. A unitary matrix with real elements is an orthogonal matrix.
Chapter 2
j~fatrj% algebra
11
Table 2.1 summarizes some types of special rna trices.
Con d itio n Typ( of mnlr ix
~.
A .4 :"ull
:I A' Svrnrnctric
A A' ~krw·.'innrnctric
!l .4* Heal
.4. .4. * Pure imaginary
A (A *), Hcrrn itinn
.1 (A *)' ~kc\\'·1 r r r iu it ian
.1 '"I = U Orthogonal
(.! ')' il = U Un it a rv 2.3 Determinants
lloji ni t ion. a n d pr o p er t.ie. of de t.cr rn i n a n t.s
The solution of two simultaneous equations
011.'[1 + alZTZ OZlX1 + anXz
(2,3,1)
can be obtained by eliminating the variables one at a time. Solving for 1', ill terms of II from tIle second equation and substituting this expression I or I2 in the rirst cqun t ion, the following is obtained:
!111.fl + 012 (h  Q2_l Xl)
(122 0"
Qll!1nTl + ol,k,  Q1Z021Il (a11a22  a12(]Zl)Il
02zkl
Qnkl  012k2
Theil, subst it n t ing .fl 11\ either of the equations (2,3.1), X2 is obtained:
The expression (1111a22  (12021) is the value of the determinant of the coefficient matrix .1, where IA I denotes the determinant
1:1! =
12 Computer method. in power .y.tem analy.i.
The solution of the equations (2.3.1) by means of determinants is
I kl all I
k% a22 i
Xl
I all aul
au aa
and
I all kll
a2l k2
X2
I all all I
an an A determinant is defined only for a square matrix and has a single value. A method for evaluating the determinant of an n X It matrix is given in Chap. 7.
Determinants have the following properties:
1. The value of a determinant is zero if
a. All elements of a row or column are zero
b. The corresponding elements of two rows (or columns) are equal c .• A row (or column) is a linear combination of one or more rows
(or columns)
2. If two rows (or columns) of a determinant are interchanged, the value of the determinant is changed in sign only
3. The value of a determinant is not changed if
a. All corresponding rows and columns are interchanged, i.e.,
IAI = IA'I
b. k times the elements of any row (or column) are added to the corresponding elements of another row (or column)
4. If all elements of a row (or column) are multiplied by a factor k, the value of the determinant is multiplied by k
5. The determinant of the product of matrices is equal to the product of the determinants of the matrices, i.e.,
lA B CI = IAIIBIICI
6. The determinant of the sum (or difference) of matrices is not equal to the sum (or difference) of the individual determinants, i.e.,
IA + B  CI ~ IAI + IBI  ICI
The application of these properties can reduce the work In evaluating determinants.
Chapter 2
Matrix algebra
13
lUinors and cofactor»
The determinant obtained by striking out the ith rnw and jth column I~ called the minor of the clement aij. Thus, for
all al2 au
IAI = a21 an au
a31 a32 a33
the minor of a21 is
I al2 al31
a32 a33 I The order of this minor is one less than that of the original determinant. By striking out any two rows and columns a minor of order two less than the original determinant is obtained, etc.
The cofactor of an element is
(l)i+j(minor of aij)
where the order of the minor of ai, is n  1. The cofactor of a21, designated by .421, is
alII = I an all I al2
The following relationships between a determinant and cof'actors exist:
I. The slim of t.he products of the clements in any row (or column) and their cofactors is eq ual to the determinant:
(2.3.2)
2. The sum of the products of the elements in any row (or column) and the cof'actors of the corresponding elements in another row (or column) is equal to zero:
(2.3.3)
Adjoint
If each element of a sq unrc matrix is replaced by its cofactor and then the matrix is transposed, the resulting matrix is an adjoint which is designated by A+
14 Computer method. in power system analysis
2.4 Matrix operations
Equality of matrices
If A and B are matrices with the same dimension and each element a,j of A is equal to the corresponding element b,j of B, the matrices are equal, i.e.,
A=B
Addition and subtraction of matrices
Matrices of the same dimension are conformable for addition and subtraction. The sum or difference of two m X n matrices, A and B, is a matrix C of the same dimension, i.e.,
A±B=C
where each element of C is
eij = 0.;, ± bij
For n conformable matrices the sum or difference is
A±B±C±D±'" ±N=R
where the elements of the resultant matrix Rare
roj = a;j ± bij ± C;j ± d., ± . . . ± nij
The commutative and asaocintivc lnws apply to uddit.ion of matrices as follows:
A+B=B+A
commutative law
i.e., the sum of the matrices is independent of the order of the addition.
A + B + C = A + (B + C) = (A + B) + C
associative law
i.e., the sum of the matrices is independent of the order in which the matrices are associated for addition.
Multiplication of a matrix by a scalar
When a matrix is multiplied by a scalar, the elements of the resultant matrix are equal to the product of the original elements and the scalar. For example:
kA = B
where bi; = ka;j for all i and j.
Chapter 2
Matrix algebra
15
The multiplication of a matrix by a scalar obeys the commutative law and the distributive law as follows:
kA = Ak k(.1 + B)
commutative law
U + k.B = (.4 + B)k
distributive law
lUultiplication of matrices ~Iultiplication of two rnut.rices
AB =C
is defined only if the number of columns of the first matrix A equals the number of rO\\'O' of 8. Thus, for the product of matrix A of dimension m X. q and matrix B I)f dimension q X n, the matrix C is of dimension m X n. Any clement c,; of C is the sum of the products of the corresponding elements of the zth row of A and the jth column of B, that is,
or
Q
c., = L a"bkj k 1
t= 1,2, ... ,m;)
1,2, ... ,n
For example:
raIl
:In = a21 (j"1
(l12] r bll
022 l b21
(j'.2
bl21 [allbll + alZbll
bZ2 = a21bll + anb21
a,lbll + a,2b21
allb12 + al2bn] a21bl2 + a22bn
a31bu + a32b22
In the product An, :l prcrnultiplics 8 or B postmultiplies A. The product 8:1 i~ riot drfin(~d sincr the number of columns of 8 is not equal to the num bo r of rows of :\. \\'hcn the products A 8 and 8.1 arc defined for a square matrix, it can be shown that, in general,
An ,e. BA
Therefore, the commutative law does not hold for matrix multiplication. If the matrices A, B, and C satisfy the dimension requirements fer multiplication and addition, the following properties hold:
A (B + C) A B + AC distributive law
A (BC) = (A B) C = A BC associative law
However,
AB = 0 does not necessarily imply that A = 0 or B = 0 CA = C B does not necessarily imply that A = B
16 Computer method" in power """tem anal),si.'!
If C = AB, then the transpose of C is equal to the product of the transposed matrices in reverse order, i.e.,
This is the reversal rule.
Inverse of a matrix
Division does not exist in matrix algebra except in the ease of the division of a matrix by a scalar. This operation is performed by dividing; each element of a matrix by the scalar. However, for a given set of equations,
aUXI + al2X2 + aux. = Yl aUXI + a22X2 + anx. = Y2 aux, + anX2 + anx. = !Ja
or, in matrix form,
(2.4.1)
AX = Y
(2.4.2)
it is desirable to express Xl, x«, and XJ as functions of VI, Vz, and s», that is,
X = BY
If there is a unique solution for the equations (24.1), then mn t rix B exists and is the inverse of A.
If the determinant of A is not zero, the equations call be solved for the x;'s as follows:
All A21 A3l
Xl = !AT u, + fAT Y2 + fAT !Ja
A 12 A22 A 32
!AT Yl + fAT Y2 + fAT Ya
where A i i, A 12, .. ,A 33 are the cofactors of all, a12, ..• , a3l and \.1.\
is the determinant of A. Thus
All A21 A31
IAI IAI IAI
B Au A22 A32 A+
IAI IAI IAI \AI
Au A23 A ..
!AT IAI IA\ Chapter 2
Matri:r algebra
17
whrrr: A + is the: adjoint of A. It should be noted thut the clements of the adjoint A + are the cofactors of the elements of A, but are placed in transposed posif.ion. The matrix B is tho inverse of A and is written AI.
:\Iultiplying A. by its inverse,
All A21 A31 0 0
au al2 a13 \A\ \A\ \AI
A 12 I1n 11 32 0 0 U
Q21 On a'!:\ \A\ 111\ IAI
A.13 An A33 0 0 1
aJ! a32 a33 \A\ \A\ jAi results in the unit matrix. This follows from the relationships (2.3.2) and (2.:3.:3). A diagonaJ term of lJ, such as Ul1, equals 1 since
All ,.112 A 13 \A\
all fAT + al2iAl + a13lA! = iA1 = 1
and an offdiagonal term, such as U12, equals zero since
..121 An A23 0
all !AT + al2 fAT + a1alA! = iA1 = 0
Thus
.4AI = AI.4 = [:
TO :01\,1' for X from the matrix COllation (2.4.2) both sides of the equation arc prcrnultiplicd by AI.
AX = Y AlAX = A_I}'
UX = kl}' X = AIY
The order of the matrices in the product must he maintained.
If the determinant of a matrix is zero, the inverse does not exist.
Such It mat r ix i;; c.illrd a sinqular matrix. If the (jr.tr:rminant. of a mat rix is not zero, the matrix is a non sirujulnr matrix and has a unique inverse.
The inverse of t hr prurillct of matrices can he obtained by the reversal rule, i.e.,
(:1 T?)I = N'.,j'
The trn nspose and inverse operations on a matrix can be interchanged, i.e.,
18 Co mpwt er methodA in power system analy .• i«
Partitioning of matrices
A large matrix can be subdivided into several submatricrs of smaller dimensions:
A
If the diagonal submat rices Al and .4" are square, the subdi\'i"ioll IS called principal partitioning.
Partitioning call be used to show the spcc ii:c <t r urt urc of ,I :ll1d tl) simplify matrix computation, Each suhm.u nx is CUI1SIU,'rl'U :l' :I!I d('ment in the partitioned matrix, Additiull or "UblLI1,ti()11 IS pnfulIll("d as Iollows:
Aa.
~1 __ ,_1_2 __
±
where the dimensions of corresponding submut riccs must I'e c()llfurmable, Multiplication is performed as follows:
where
CI AIBI + A2Ba
C2 AIB2 + A2B.
c, A3BI + A.B. C. = AaB2 + A.B,
The rule for partitioning two matrices whose product is to be f (lund is: the n columns of the premultiplier are grouped into k and n  k columns from left to right, and the n rows of the postmultiplier are grouped into k and n  k rows from top to bottom in order that the submatrices are conformable for multiplication.
Chapter 2
Ma tri» algebra
19
The transpose of 3. partitioned matrix is shown below.
A
'{l I A2
I A,
A, I
I
~I Al'
I
A I I A.'
i 2 I A'
The inverse of a pn r t it i onrd matrix is ob t a i n crl as fr.ll ows :
where
(AI  .1,,.1,1 .: 13)1 J3IA,.\,1
A,q,B1
(2.·1.:3.)
B, A,'  ..1.,I'\,B2
and Al and A, must he ~'1uarc matr ices.
2.5 l.inear dependence and rank of a matrix
Linear de p e n d e n c e
The columns of n n III X II mn t rix A can bc wr it tcn as n column vectors.
r ell: c,l . . . i c, I
:\150, the rows of matrix A ca n be written as 111 row vectors. frtllr21 ... 11'..,1
20 Computer methods in poteer sy s t em. analysi6
The column vectors are linearly independent if the equation
(2.')1)
is satisfied only for all p. = 0 (k = 1, 2, ... ,n). Similarly the rov vectors arc linearly independent if only zero values for the sc.clnrs qr (r = 1, 2, ... ,m) satisfy the equation
(252)
It is not possible to express one or more linearly independent column vectors (or row vectors) as a linear combination of others.
If some Pk y! 0 satisfies (2.5.1), the column vectors are linearly dependent. If some qr y! 0 satisfies (2.5.2), the row vectors are linearly dependent. That is, it is possible to express one or more column vectors (or row vectors) as a linear combination of others. If the column vectors (or row vectors) of a matrix A. are linearly dependent, then the deterrninant of A is zero.
Ra nk of a ma trix
The rank of an m X n matrix A is equal to the maximum number of linearly independent columns of 11 or the maximum number of linearly independent rows of I1. The former is called the COIUIIIIl rank and the latter the row rank. The column rank i" equal to the row r a n k. The rank of a matrix is equal to the order of the largest nonvanishiug determinant in A. For example, consider the matrix
2 4 8
The rows are linearly dependent since the equa t ion
qd124) +qzl2481 +q3138101 =0
is satisfied for
ql = 2
q2 = 1 q3 = 0
Similarly, the columns are linearly dependent since the equation
Chapter 2
,\latrix algebra
21
is satisfied for
PI 6
pz I p, I
However, Tin two colurn ns a rr: linrnrlv df!IH'Tld(,rlf :llld, t licreforc, t.hn rank of the matrix is 2.
A linear svst crn of m equations in n unknowns is written
OllXl + 012.(2 + + OI"X" = YI
!l21Xl + a22X2 + + a2""" = Y2
(2.0.1 )
known coefficients or parameter!' of the system unknown variables of the system
!), known constants of the system
The system of equations (2.6.1) in matrix form is
AX = Y
The augmented matrix of .1, designated hy A, is fonned by adjoining the column vector Y as the (n + I)st column to A.
I"" al1 al" . ;:"
A = a21 a22 a1"
Lam! a~2 G"," y ... If Yl, !J2,. .,!J", are all zero In (2.6.1), the linear equations are
homogeneous and
AX = 0
If one or more Y, arc nonzero. the linear equations are nonhomogeneous.
The necessary and sufficient condition for a system of linear equations to have a solution is that the rank of the coefficient matrix A be equal to the rank of the augmented matrix A. A unique solution exists when A is a square matrix and the rank of A is equal to the number of columns (variables). The unique solution is nontrivial for nonhomogeneous equations and trivial (i.e., zero) for homogeneous equations. If the rank of A is less than the number of equations, some of the equations are redundant and do not place any further constraint on the variables. If the rank of A is less than the number of variables of the system, there are all infinite number of nontrivial solutions.
22 Computer methods in power syatem analysia
Problems
2.1 Given:
A = [ ~ ~ ~l
7 2 3
and
4 2
~l
Determine:
a. C = A + B
b. What type of matrix C is
c. D = A.  B
d. What type of matrix D is 2.2 Given:
[3
A=
2
1 5
~J
and
Determine:
a. C = AB
b. D = BA
c. E = A'B'
d. What the relationship of matrix 1<) to matrix D is
2.3 Given:
[11 8 :1 [ _, n
A = ~ 1 and B = ~ 1
5 1
Determine:
a. C = B'AB b. What type of matrix A is
c. What type of matrix C is
d. D = CI
e. E = CCI
f. What type of matrix E is
2.4 Given:
o 1
and
Chapter 2 ;\1 a trix algebra 23
Determine:
a. B = AI
b. X from AX Y
2.5 Given:
~ l~ ;) 7 i:l
11 0 1
/ 0
1 7 0)
Determine:
(I. n = O.IA + O:,f[
i. \\"hat t ypc of III at rix A IS
c. That 11 is orth{)g()ll:ll
2.0 Given:
~ ~"l: .V;~ ' ~J~J
:1 .) )'T .) I 'J
 
1 v:i 1 . y':j
+j ~" :2  J '2
2 2
')(~I ('rIllIIW:
a. j) = (..I 0),..1
I). \\"hat type or mn t r: x :1 IS
2.7 Given:
~ [, ,) }2 , + J:l 1
A "OJ I '2 )
J
 ):1 2 +] :1
Determine:
a. B= (.'1 *)'
b. What type of matrix A IS
2.8 Given:
A = [~ 1 ] B = [j2 ~] and Y = UJ
:J J j.1
Determine:
(I. X from (.1 + jIJ)XO = y
b. What type of matrix A is
c. What type of matrix B is 24 Computer methods in power "yatem analY.'3
2.9 Given the partitioned matrix:
Determine B = AI using the formulas (2.4.3) for the inverse of a partitioned matrix.
2.10 Given:
1 2 0 3 4
2 6 0 0 1

0 0 14 5 1
3 0 5 10 2
4 1 1 2 12 1
Determine Az.
2.11 Given the partitioned matrices:
A
B=
1 2 0 0 0 0
1 1 0 0 0 0
 
0 0 1 3 2 0
0 0 2 3 4 0
0 0 1 5 2 0
 
0 0 0 0 0 s
1 3 2 ;; 6 1
4 7 1 3 2 4
 
3 4 2 6 5 1
4 6 3 1 2 5
2 6 7 3 8 i
 
7 2 3 1 4 ;) Determine C = AB.
Cbajner 2
Matr;" algebra
2.12 Given:
C ~ I: I: I
where
A = [~ ~] B = [~ 1 ]
4
and N is a null matrix.
Show that the inverse of C is
c' ~ I A' N :
N B1
2.13 Given: 1 6 5
Show t hn t A is a singular matrix and determine its rank. 2.11 Given:
i 6 0 3
I !
I
Al A'I I 2 4 2 2
A I 7 'J 3
,
A3 ..1'1 [ 6 4 2 2
Determine B = Al  112..1.1..13•
L.Lj Gin'll :
I 3 2 .; 1 B = [ 2 ~]
A = l2 [j and
1 8 12 1
C = [~ 0 ~]
1
Determine D = A  C'B'C. 25
26 Computer methods in po ioer syHem analysis
Ribliopraphy
Aitken. A. C "Determinants and Matrices." Iriterscience Publishers.
Inc .. Xew York. 1959 .
. Avros. Frank, Jr "Theory and Problems of Matrices." Schaum Publisbing Company .. \ew York, 1962.
ELLIE'. H '\Lltnx Algebra for Elec t ricul Engineers." AddionWelev Publihing C,nllpany. Inc .. Reading, '.[ass., 196:.i.
:'iil!:lllT. :' .\ll~Tell: "Tbe Elements of Det errniuan ts. \latriec,. uid Tensors fur Engiuecrs.' \[acDon:dd and Co., LTd .. London. l'6!1 Tropper..\ ,\1 "\Iatrix Theory for Electrical Engineers, AddisonWesley Publishing Company, Inc., Reading, :\lass., 19,12.
Incidencf!
3.1 Introduction
The formulat i..n of a sun able rnat hcmat.icul D'odejl" t h. first st cp in the analysis of all electrical network The mode! TLU,' de~('nh(' t h« C!i,tr,l(,teristics of individual llC't\I'(,rk ('omp(:llCT!t> '\0' well ,"', the relat ion that govern the interconnect iou of t hest' "iempnts,\ net II" irk rna t r;x eq llation provides a convenient mathematical nnde] i(Jf l' digital c.m.put er solution,
The elements of a network matrix depend '<;( t l« selp('T:(lll "f the independent variables. which caIl he either cu rrru r s or Vt)1t:lge" ('n!'respondingly, the elerncnt s of the network mllt!"" WIll he imped.u.ccs ur admittances,
The electrical char acteris ucs of t.he individual nct wor k (,lllllp"llenh can be presented conveniently in the form "f a primit iv. Ii('tw"rk matrix. This matrix, while adequatel:; describing the chnruct eristir of each ("'mponent, does not provide :111\' inf'ormatir.n p.r t au.ina to t hr lIP( wor k cnllnccti ons. It is IWr'e",:<an', therefore, t o transform tIl!' prim it i ve nct work matrix into a network rn a t nx that describes the pcrforruance of t h« nu erconnected network,
The form of the network m.u rix used ill thl' pcrf orm.mce equut ion depends on the frame of reference. namely, btl:' or loop, In the bus frame of reference the variables are the uo.ial voltages .uid )ludal rur rcntIn the loop frame of reference the variables are loop \'oltages .uid iuop currents.
The formation of the appropriate network matrix IS eUi irl1('gL,1 part of a digital computer program for the solution ,.)i PUm?!' system pf(,hlprn~,
28 Computer methods 'n. power system an.alysu
3.2 Graplas
T r. wdp,r t'l describe the geometrical structure of a network it IS sufficient u. replace the network components by single line seamem s irrespective of (,h characteristics of the components, These line segments are called elemen:» and their terminals are called nodes. A nodr md an Element are inricent if the node is a terminal of the element Node can be incident 1.(, ',nf' e>f more elements.
A ,;i'(jph shows the geometrical interconnection ,if tit" elements of a net ','()rk A subqra ph is any subset of elements of tho, grapl. A path is a subgraph of connected elements with no more than tw« clements connect ed t(i anyone node. A graph is connected if and only if there IS a path between every pair of nodes. If each element »f the connected graph is assigned a direction it is then oriented. A representation of a power system and the corresponding oriented graph are shown in Fig. 3.l.
+{~ t·· l_l=+@
(u ) ®
• '0 8)
CD ~ I
I
Ll
®
® 
( b) I
(c)
Fig. 3.1 Power system representations. (a) Single line diagram; (b) positit'e sequence riet worlc diagr a m; (c) oriented connected grnph.
Chapter 3
Incidence an.d n.cr.work matrices
29
  . ____
/' .
.>
 //
(!l~_:'
~~
~""
~
Branch
 Link
f":
Fig.3.2 Tree and co tree of the oriented connected ara ph .
A connected subgraph containing all nodes of a graph hut no closed path is called a tree. The elements .)f a tree are caller! branches and furm a subset of the elements of the connected graph. The number of branches b required to form a tree is
b = n  1
(3.2.1J
where n is the number of nodes in the graph.
Those elements of the connected graph that are not included in the tree are called links and form a subgraph, not necessarily connected. called the cotree. The cot ree is the complement of the tree. The number of links 1 of a connected graph with e elements is
l=eb
From equation (3.2.1) it follows that
1=en+1
(:~.2.21
A tree and the corresponding cotree of the graph given ill Fig. 3.1c are shown in Fig. 3.2.
If a link is added to the tree, the resulting graph contains one closed path, called a loop. The addition of each subsequent link forms one «r more additional loops. Loops which contain only one link are independent and are called basic loops. Consequently, the number of basic
.10 ComputeT rrulthoda in power system analY3u
®
~
6
./
_ ..  _
_/ .......
,/ "
,/ ."
,/ . "
F
G
3
®
Fig.3.3 Basic loops of the oriented connected graph.
loops is equal to the number of links given by equation (3.2.2). Orientation of a basic loop is chosen to be the same as that of its link. The basic loops of the graph given in Fig. 3.2 are shown in Fig. 3.3.
A cutset is a set of elements that, if removed, divides a connected graph into two connected subgraphs. A unique independent group of
A
 .... 
6
®
Fig. 3.4 Basic cutsets (jf the oriented connected graph.
Chapter 3
Incidence ond network nuJtrice3
31
cutsets may be chosen if each cutset contains only one branch, Independent cutsets are called basic cuisets. The number of basic cutsets is equal to the number of branches Orientation vf a basic cutset is chosen to be till' "arne as that of its branch The basic cutsets of the graph given in Fig ;3.2 are shown in Fig. :3.4
3.3 Incidence matrices
Elementnode incidence matrix A
The incidence of elements to Hodes in a connected graph is shown by the elementnode incidence matrix. The element:' .. f the matrix are as follows:
a" 1 if the ith element is incident to and orient ed away from the jtb node
a" 1 if the Ith element is incident t() and orieuted toward the )th !lode
a" = 0 if the rt h element is not incident to t he rb ",)de
The dimension of the matrix is c X II. where e is the number of elements and n is the number of nodes in the graph. The elementnode incidence matrix for the graph shown in Fig. 3.2 is
6
@ 'G; @ 0) G;
1 1
1 1
1 I I I I 1
I I I 1 I 1
I I 1 I 1 I
I 1 I 1 I I
1 1 2
3
A 4
5
7
32 Computer methods in power .y&tem analYlIis
Since
4
)' a., = 0
'' i == (1
1 = 1,2, ... ,e
the columns of A are linearly dependent. Hence. the rank of A < n.
ll u« incidence matrix A
Anv node of a connected graph can be selected as the reference node. Then. the variables of the other nodes, referred to as buses, can be measured with respect to the assigned reference. The matrix obtained [rom A by deleting the column corresponding to the reference node is the elcrnen tbus incidence matrix A, which will be called the bus incidence ma trix The dimension of this matrix is e X (n  1) and the rank is 71  1 = b, where b is the number of branches in the graph. Selecting node 0 as reference for the graph shown in Fig. 3.2,
.l
, bus
. CD @ (]) 8)
e
1 1
~~ ~~ ~~ ~
2 1
~~ ~~ ~~ ~~
3 1
~~ ~~ ~~ ~~
± 1 1
~~ ~~ ~~ ~~
;) 1 1
~~ ~~~ ~~
6 1 1
~~ ~~ ~ ~~
7 1 1 This matrix is rectangular and therefore singular.
I f the rows of A are arranged according to a particular tree, the matrix can be partitioned into submatrices Ab of dimension b X (n  1) and AI of dimension l X (n  1), where the rows of Ab correspond to branches and the rows of AI to links. The partitioned matrix for the graph shown III Fiz. 3.2 is
Chapter S
A
~us CD (2) ® 8)
e ,
1 1
   
2 1
   
3 1
   
4 1 ]
   
5 1 1
   
() 1 1
   '
7 1 1 Incidence and network matrices
Buses
A,
Ab is a nonsingular square matrix with rank (n  i)
Branchpath incidence matrix K
The incidence of branches to paths in a tree IS shown by ! he branchpath incidence matrix, where a path is oriented from; bus to the reference node. The elements of this matrix are:
kij = 1 if the ith branch is in the pat.h from the jth bus tp reference and IS oriented in the same direction
k., 1 if the ith branch is in the path from the jth bus to reference but
is oriented in the opposite direction
ki' = 0 if the ith branch is not in the path from the jr h bus to reference
With node 0 as reference the branchpath incidence matrix associated with the tree shown in Fig. 3.2 is
~ 0 ® ® ~
1
2
I I I '
1 !
1:
I 1 1 I :
1''
I 1 i 1 ,
 ;~:
. '
1 :
I i
K=;
3
4
This is a nonsingular square matrix with rank (n  li.
34 Computer methods in power system analysi&
The branchpath incidence matrix and the submatrix Ab relate the branches to paths and branches to buses, respectively. Since there is a onetoone correspondence between paths and buses,
AbK' = U Therefore.
(3.3.1)
13.3.21
Basic clltset incidence matrix B
The incidence of elements to basic cutsets of a connected graph is shown by the basic cutset incidence matrix B. The elements of this matrix are:
b., I if the ith element is incident to and oriented in the same direction as the jth basic cutset
b., } if the ith element is incident to and oriented in the opposite direction as the .ith basic cutset
b., = 0 if the ith element is not incident to the jth basic cutset
The basic cutset incidence matrix. of dimension e X b, for the graph shown in Fig. 3.4 is
~b e~
,,,,
R 4
2
1
}
1
Basic cutsets ABC D
3
5
() 1
7
The matrix B can be partitioned into submatrices Ub and B, where the rows of U; correspond to branches and the rows of BI to links. The partitioned matrix is
CluapUr J
Incidence and n.etwork matrices
35
Basic cutsets B C
D
'" b:
Basic cutset"
~_~i _
: l' I
i 'I I • 
:11:
2 i : 1, i
i~!i  I 3 iI
B=
4 ii' 1_
;~II_,
,5 I  1: 1
!I[' 
, ' ,
6 i 1 il'
i:I ,
7 i ,1; 1
L __ ! _
The identity matrix L'b shows the onetoone correspondence of the branches and basic cu tsets.
The submatrix B, can be nbtained from the bus incidence matrix A._ The incidence of links to ouses is shown by t he subrnatrix A I and the incidence of branches to bU5e'i~ shown by t hr submatrix A.,,_ Since there is a onetoone correspondence of the branches and baSIC cutsets, BlAb shows the incidence of link!' to buses. that IS,
BlAb = Al
Therefore,
B, = AIAbl
In addition, as shown in equation (3,3.2), Abl = K'
Therefore,
B, = A1K'
(3.3_3)
Augmented cutset incidence matrix B
Fictitious cutsets, called tie cutsets, can be introduced in order that the number of cutsets equals the number vf elements, Each tie cutset contains only one link of the connected graph. The tie cutsets for the graph given in Fig. 3.4 are shown in Fig. 3.5. An augmented cutset incidence matrix is formed by adjoining to the basic cutset incidence matrix additional columns corresponding to these tie cutsets, A tie
36 Computer method. in po_r .,..tem analYllu
/G
B
    .. ____.' 4' 5
F
~ ~  .... ~~
"
A
c
3
Basic cut  set
Tie cutset
Fig,3,') Basic and tie cu t=se ts of the oriented connected graph,
cutset is oriented in the same direction as the associated link. The augmented cutset incidence matrix for the graph shown in Fig, 3,5 is
, el Basic cutsets
e~ ABC D
;\1 I !.\ ~~
" 1\li~1
,,1,
I
3 i I ' , ' Ij;;
i I • I' ;
i: :
51i~II1
61~Il'I1 7 II~III
Tie cutsets E F G
1
Thi« is a square matrix of dimension e X e and is nonsingular, The matrix B can be partitioned as follows:
Ch4pUJr 3
Incidence and network matrices
37
B=
Basic cutset .. s ABC D
Tie cutseta E F G
o
Basic loop incidence ITUltrix (.
The incidence of elements to basic loops of a connected graph is shown by the basic loop incidence matrix C. The elements of this matrix are:
c., 1 if the ith element is incident to and oriented in the same direction as the jth basic loop
c.j = 1 if the ith element is incident to and oriented in the opposite direction as the jth basic loop
c., = 0 if the ith element is not incident to the Jth basic loop
38 Computer methodlr in po_r .,.8tem GnGly.u
The basic loop incidence matrix, of dimension e X I, for the graph shown in Fig. 3.3 is
~ [I Basic loops
e",iE F G
_~'l_c._,
1 1 1. __ 1 I
:, I
:: 1 i 1 1
:11
3 ! 1 : 1
, ,
11
. !
4 ; 1 I
I '
,1
5! l'
11,
6' I 1 I
7 1111
c=
The matrix C can be partitioned into submatrices Cv and UI where the rows of Cb correspond to branches and the rows of UI to links. The partitioned matrix is
c
"" I i Basic loops
e ""; E F G
11 ~
211 ~11
:1' =:\=! 1 I ·1
5 I 1.
I i
6 iIl~
1_1 __
7 l I 1
~! ] 1,
o
~ !
~
Basic loops
L __
The identity matrix UI shows the onetoone correspondence of links to basic loops.
Augmented loop incidence matrix e
The number of basic loops in a connected graph is equal to the number of lin ks III order to have a total Dumber Qf loops eQ1,I~I to the number of
Chapurl
Incidence and netUJOrk matrices
E
A
.~ Basic loop Open loop
Fig. 3.6 Basic and open loops of the oriented connected graph.
elements, let (f  1.1 loops, corresponding i<, t he b branches. he designated as open loops. An open loop, then, is defined as a path between adjacent nodes connected by a branch. The open inops for the graph given in Fig. 3.3 are shown in Fig. 3.6. The orientation of an open loop is the same as that for the associated branch,
The augmented loop incidence matrix is formed by adjoinmg to the basic loop incidence matrix the columns showing the incidence of elements to open loops. This matrix, for the graph shown in Fig. 3.6, is
~ e i Open loops Basic loops
e~i ABC D E F G
j, .;
1 ill
, ,
:11:
: I !
2! I 1 I ,
i 1 ! !
II!I
3: ill' 1 :
I! i
111
(J = 4 I ' I i
~I~:~I
5~1~1~!
:1~1~I~i
1 :
1
1 1 I
; i
:!
~ .. ,~!
1
I
.~,
This is a square matrix, of dimension e X f. and is nonsingular.
40 Computer method. in power system analysis
The matrix e can be partitioned as follows:
" A; Open loops
t~~I_A __ B __ C __ D
1 1 1 ,Ii i
: I' I
     11
1 i 1 1: 1 I
1Ii1: =~I=~\
 1 , :11
I i 1 1 \
1:11 11
, I
1
I I 1 1
;  , I
,,,1
2
:l
c 4
S
6
7 Basic loops E F G
:
1 I
~I
.~
Open loops
Basic loops
o
3.4 Primitive network
:\ etwork components represented both in impedance form and in admittance form are shown in Fig. 3.7. The performance of the components can be expressed using either form. The variables and parameters are:
I'pq is the voltage across the element pq
Cpq is the source voltage in series with element pq Ipq is the current through element pq
Chapter 3
Incidence and network matrices
41
JJXi is the source current in parallel with element pq zpq is the selfimpedance of element pq
ypq is the selfadmittance of element pq
Each element has two variables. 'po and lpq. [D steady Uite these variables and the parameters of the elements Zp, and .'Ipq are ral numbers for direct current circuits and complex numbers fur alternating current circuits
The performance equation of an element in impedance f(.rm is
or in admittance form is
134.2)
The parallel source current in admittance form is related to thc seriosource voltage in impedance form by
A set of unconnected elements I, derined as " prinur ivo network.
The performance equations of a primit rve network can be derived frurn (3.4.1) or (3.4.2) by expressing the variables as vectors and the parame
Ep 8 Eq
• Zpq •
'/l) .'q'
epq ipq (a)
~~GD~
Ep ... _L~ YP'l I __ ~ • Eq
® Ipq '' ®
ipq+ jpq
(b)
Fig. 3.7 Re presercta tio n s of a netu:ork component. (a.; Impedance form; (b) admittance form.
42 Computer methods in power "y.tem OJ&4ly.ia
t"Tf. as matrices. The performance equation \D impedance form is
.... e = [zJi
r '!l admittance form is
_J,_ l = fy]v
\ dUlgonal element of the matrix [z] or [y] of the primitive network 18 tJw >,,'r,mpedance Zpq,pq or selfadmittance ypqrq An offdiagonal element is mutual impedance Zpq,," or the mutual admittance Ypq,., between 'hc " lq'lf'llt~ pq and r8, The primitive admittance matrix [y] can be it!l:lI11ed by inverting the primitive impedance matrrx [e]. The matrices :lflcl [y] are diagonal matrices if there is no mutual coupling between
elernen ts. In this case the selfimpedances are equal to the reciprocals uf the corresponding selfadmittances .
. 1.5 Formation of network matrices by singular transformations
\'ettcork perfor mance equations
\ network is made up of an interconnected set of elements, In the bus frame of reference, the performance of an interconnected network is described by n  1 independent nodal equations, where n is the number of nodes, In matrix notation, the performance equation in impedance form is
or III admittance form is 1 en',' = Y BUSESUS
where EBl,s
vector of bus voltages measured with respect to the reference bus
1 B['S vector of impressed bus currents
Zm's ='bus impedance matrix whose elements are open circuit driving point and transfer impedances
Y BG'S = bus admittance matrix whose elements are short circuit driving point and transfer admittances
In the branch frame of reference the performance of the interconnected network is described by b independent branch equations where b is the number of branches, In matrix notation. the performance equa
lion in impedance form is FoR = zBRIBR
Chapter j
Incidence (lnd network matrices
43
or in admittance form is
where EBR
vector of voltages across the branches vector of currents through the branches
branch impedance matrix whose eleYnf'ntf, are open circuit driving point and transfer lm;_;,,,b:!,,,·,; or '(;r branches of the netw ork
branch admittance rnatrix whose eien.ents life short circuit driving point and transfer admittur.ce ,,1 the branches of the network
In the loop frame of reference, the performance 'if au interconnected network is described by I independent loop equatious where I is the number of links or basic loops. The performance 2(1'liilll,n in impedance
form is
E LOOP = Z LOopI LOOP
or in admittance form is I LOOP = Y LoopE LOOP
where E LOOP = vector of basic loop voltages lLOoP = vector of basic loop currents ZLOOP = loop impedance rnat nx
Y LOOP = loop admittance matrix
Bus admittance and bus impedance matrices
The bus admittance matrix Y BUS can be obtained by using the bus incidence matrix A to relate the variables and parameters of the primitive network to bus quantities of the interconnected network. The performance equation of the primitive network
i + J = [y]i.i
is premultiplied by A', the transpose of the bus incidence matrix, to obtain
A'i + AtJ = A'[Y]i.i
(3.5.1)
Since the matrix A shows the incidence of elements to buses, A. 'i is a vector in which each element is the algebraic sum of the currents through the network elements terminating at a bus. In accordance with Kirchhoff's current law, the algebraic sum of the currents at a bus is zero. Then
A'i = 0
(3.5.2)
44 Computer methods in power .,..tem Gn4ly."
Similarly, At] gives the algebraic sum of the source currents at each bus and equals the vector of impressed bus currents. Therefore
I B,'S = AtJ
(3.5.3)
Subs tituting from equations (3.5.2) and (3 .. S.3) into (3.5.1) yields
I EWS = A'[yjv
(3.5.4)
P(J'.H~r into the network is U;us)'EilCS aud the sum of the powers in the prirnit i ve network is (J*)IV. The power in the primitive and interconnected networks must be equal, that is, the transformation of variables must be powerinvariant. Hence
(3.5.5)
Taking the conjugate transpose of equation (3.5.3) U;lJs)t = (J*)tA *
Since A is a real matrix
A* = A
and
(l~u.',)' = (J*)'A
Substituting from equation (3.5.6) into (3.5.5) (J*)tAEnus = err»
Since this equation is valid for all values of J, it follows that AEBc'S = fj
(3.5.6)
(3.5.7)
Substituting from equation (3.5.7) into (3.5.4), I He'S = At[yJ.4EIliIS
Since the performance equation of the network is I HU S = Y HU S E BC' S
(3.5.8)
(3.5.9)
it follows from equations (3.5.8) and (3.5.9) that Y BCS = At[yJA
The bus incidence matrix A is singular and therefore A.t[y]A is a singular transformation of [y].
The bus impedance matrix can be obtained from
Z eu s = Ynbs = (A t[y]A )1
ChGpter J
Incidence and neturork matrice&
45
Branch admittance and branch i nipedaru:e matrices
The branch admittance matrix Y BR can be obtained by using the basic cutset incidence matrix B to relate the variables and parameters of the primitive network to branch quantities of the interconnected network. The performance equation of the primitive network 111 admittance form is premultiplied by 8< to obtain
(3 . .5.10)
Since the matrix B shows the incidence (,f Iemrnts to basic cutsets, B'i is a vector in which each element is the algebraic sum of the currents through the elements incident to a basic cutset.
The elements of a basic cutset if removed divide the network into two connected subnetworks. Therefore, an element of the vector B'I is the algebraic sum of the current entering a subnetwork and by Kirchhoff's current law is zero. Therefore
B'i = 0
(3 . .5.11)
Similarly, B'J is a vector in which each element is t he algebraic sum of the source currents of the elements incident to the basic cutset and is the total source current in parallel with a brand). Therefore
(3.5.12)
Substituting from equations (3.5.11) and (3 . .5.12) into (;3 . .5.10) yields
(3 . .5.13)
Power into the network is (J~R)«E RR) and since power is invariant (l~R)IEBR = (J*)'fJ
Obtaining (l~R)' from equation (3 . .5.12), then (J*)<B*EBR = irv»
Since B is a real matrix
B* = B
and
Since this equation is valid for all values of J, it follows that fJ = BEBR
Substituting from equation (3 . .5.14) into (3 . .5.13) yields IBR = B<[yJBEBR
(3.5.14)
(3 . .5.15)
The relation between the branch currents and the branch voltages is
(3 . .5.161
46 Computer meth0d3 in power sy.tem analY8i8
It follows from equations (3.5.15) and (3.5.Hil that.
YER = B'[yJB
The basic cutset matrix B is a singular matrix and therefore B'[yJB is a singular transformation of [y].
The branch impedance matrix can be btained from
ZHR = YB~ = (B'[yJB)l
Loop impedance and loop admittance matrices
The loop impedance matrix Zwop can be obtained by using the basic loop incidence matrix C to relate the variables and parameters of the primitive network to loop quantities of the interconnected network. The performance equation of the primitive network
v + e = [zJi:
is premultiplied by C' to obtain
(3 .. 5.17)
Since the matrix C shows the incidence of elements to basic loops, C'i] gives the algebraic sum of the voltages around each basic loop. In accordance with Kirchhoff's voltage law, the algebraic sum of the voltages around a loop is zero. Hence
c» = 0
(3.5.18)
Similarly e'e gives the algebraic sum of the source voltages around each basic loop. Therefore
(3.5.19)
Since power is invariant (l~oop)IELOOP = (i*)'e
Substi tuting for BLOOP from equation (3.5.19), then Uroop)'C'e = (i:*)'e
Since this equation is valid for all values of e, it follows that (1*)' = Cl1oop)'C'
Hence,
i = C*lLOoP
Chapter 3
Incidence and netUlOrk matricea
47
Since C is a real matrix,
C" = C
and
i = CILQOP (3.5.20)
Substituting from equations (;L;i~) .j:, [q).ll:d (:3 .~i.20) int.o (3 . .').17) yields
c'[z]CILOop
(3.5.21)
The performance equation of the network if: the lccp frune of reference is
(3 .. 5.22)
and it follows from equations (3.521; and (3.522) that ZLOOP = C'[z]C
Since C is a singular matrix, Ct[zJC 18 a singular transformation of [zJ.
Tabk 3.1 Formation of network matrices by singular transformations
Network matrices
#J Computer method. in power aylltem analym
Table 3.2 Current arad voltage relatW~ betsoeen. primitive arad interconnected networks
Frame of reference
ISR = B'j
I
1 
Loop
Bus
Branch
t t
= cILoop
; ~I~
;;:; Ewop = c« 'I iJ = AE8us V = BEBs
~
The loop admittance matrix can be obtained from Y LOOP = Z"ibop = (C'[Z]C)l
The singular transformations for obtaining network matrices are summarized in Table 3.1. The current and voltage relations between the primitive and interconnected networks are summarized in Table 3.2.
3.6 Formation of network matrices by nonsingular transformations
Branch admittance and branch impedance matrices
The branch admittance matrix Y BR can be obtained also by using the augmented cutset incidence matrix B to relate the variables and parameters of the primitive network to those of an augmented interconnected network. The augmented network is obtained by connecting a fictitious branch in series with each link of the original network. In order to preserve the performance of the interconnected network the admittance of each fictitious branch is set to zero and its current source is set equal to the current through the associated link, as shown in Fig. 3.8a. The voltage across a fictitious branch is zero. Then a tie cutset can be interpreted as a cutset containing a link and a fictitious branch, as shown in Fig. 3.8b.
The performance equation of the augmented network in the branch frame of reference is
Chapter J
Ineidenc« and _tUJOrk matrices
49
Fictitious
node
Fictitiou» hrafl.Ch
.i
(a I
TiP cutset G
Fictitious . uxie
(I)
Fictitious branch
e__. C:~ 3
~I
Fig. 3.8 Repr es en tation. of an augmented n et u or k , a Fictitious branch in series ",ith a link; (b1 interpretation of a tie curset.
The matrix Y BR will be obtained directly from the admittance matrix 9' BR of the augmented network.
The performance equation for the primitive network
t + J = [y]v
is premultiplied by B' to obtain 13tl + Blf = Bl[y]ii
(3.6.1 )
50 Computer methods in power BYatem analyai&
Equation (3.6.1) can be written in the partitioned matrix form:
'~ i: ! BI' I ,,
jb !
I
:!
, jl i
r
_ _
y
(Hi2)
, ,

where the primitive current vectors 'I and j are partitioned into the current vectors Ib and Jb; which are associated with branches of the network. and t he current vectors 'il and li. which are associated with links. The left ,irjp of equation (3.6.2) is
!+
11
jl
where
and
Jb + Bz'J1
However
B'; = 0
and
Then the left side of equation (3.6.2) is
o
~I ~i , I s» +'
I
E
Iii + jl I I '
21
, JI L____c
Si nee each element of tl is equal to a current source of a fictitious branch, 11 + JI is a vector in which each element is equal to the algebraic sum of the source currents of a fictitious branch and its associated link. Therefore,
Chapter 3
Incidence and rw.twork matrices
51
and equation (36.1) becomes rBR = SI[y]i
(3.6.::»
Since the voltages across the ri('t.it!()!l~ branches are zero. the voltage vector of the augmented network lS
EiJ/1 ' t; Ilk = _. 
o
The voltages across the elements of t.hp origin"_] network from ('quatlo!! (;Ld41 art'
f = BEBR
However BEBR = StBR
then
Substituting from equation (3.h.4; mto equation (,:).0.0) ! BR = BI[y]Bf:BI?
Since the performance equation of the augmented network is fBR = rBRtBR
(3.G.G)
it follows from equations (3.6.5) and (:Hi.6) that the admittance matrix of the augmented network is
rER = Bt[y]B
Equation (3.6.7) can be written in the partitioned form
(3.G.7!
,.
Yl y~ ,
Y3 Y,
where [Yb6]
[Ybll _;...,
Ub BI' Yb/; U, i 0
.'Ib;
. _._ (;31)8)
0 C IN !:lit B, C, = primitive admittance matrix of branches
= [yu,l' = primitive admittance matrix whose elements are the mutual admittances between branches and links
[YII] = primitive admittance matrix of links
It follows from equation (3,6.8) that
YI = [Ybbl + B/[Ylbl + [YbdBi + B1'[YlllBI
(3.6,9)
52 Computer methods in power system analysu
) ; t: = :'MJ + BI'[Y'bl + iYbdB, + Ri[yulB, h"rr, equations (3.6.9) and (:Hl.lO). therefore. }'J;/i = Y1
\3.6.10)
The branch impedance matrix can be obtained from ZHIc = Y1l
Loop impedance and loop admittance matrices
The »OP impedance matrix ZLOOP can be obtained also by using the augmerited loop incidence matrix C to relate the variables and parameters of t h« 'Primitive network to those of an augmented interconnected network. The augmented network is obtained by connecting a fictitious link in parallel with each branch of the original network. In order to preserve t he performance of the interconnected network the impedance of each fictitious link is set to zero and its voltage source is set equal and opposite tl) 1 he voltage across the associated branch, as shown in Fig. 3.9a. The current through a fictitious link is zero. Thcn an open loop can be iu tcrprcted as a loop containing a branch and a fictitious link as shown in Fig. ::I.9b.
The performance equation of the augmented network in the loop fl':llIW of reference is
The matrix ZWOl' will be obtained directly from the impedance matrix ;? t.o.,» of the augmented network.
The performance equation for the primitive network
I' _;_ ,~ = [z]!
i prrmultiplicd by C' to obtain
C'l + ('c = C'[zji
(3.6.11)
Ch..pter 3
Incidence and nettDOrk matrices
S3
\
/:..,
"\ J
.,~.
(a)
,
"
':;_"'r_"~
Fictuious lInk   (li)
___ _
0' c.  ~.   411__:. (I)
\
\ \
\
\
,
"
, Open
", loop A ,
<,
f., <, <,
Fictitious "
link '_
(b) Co',
Fig. 3.9 Representation of an aug m en ted network. 'a) Fictitious link in parallel with a branch: ,b, interpretation of an open loop.
Equation (:3.6.11) can be written in the partitioned form
,' Cb' U1 i
z
'
(3.6.12)
54 Computer method» in powe,. ay .. tem onaly ...
where the primitive voltage vectors ii and e lire partitioned into the voltage v .. ctors ii&. and eb, which are associated with the branches of the network. and the voltage vectors iir and ih, which are as~;;c;ated with the links. 1')11' left side of equation (36.12) is
 ... 'I ,. ... ,
! I
'b I : eb I
. + 1:
ii"; l 1"1 i Cb'eb + CI
;
w hrr«
cv. + VI C'ii and Cb'CI + CI c»
H()\\t'ver
("f' = 0 and G'e = Bwop The left side of equation (3.6.12) is then
,~
I'b eb I Vb + eb
+ I
0 ELOOP 1 Ewop , Since each element of ~ is equal to a voltage source of a fictitious link. !'b + eb IS a vector in which each element is equal to the algebraic sum of the source voltages in an open loop. Therefore,
i~
[LOOP = 1
(3.6.13)
I ELOOP
and from equations (3.6.11) and (3.6.13) fwop = C'[z]I
(3.6.14)
Since the currents in the open loops are zero, the current vector of the augmented network is
=8
I ILoop I
The currents through the elements of the original network from equation (3.5.20) are
I = clLOop
Chapter 3
Incidence and network lfUJtrice~
55
However.
(3.6.15)
·;3 b lhl
~IIIV.~ thp pvrformaner eq uat iou of t he augmented r,p,. \\ )fK . ~ FU"0!" ,= ZU)(jP1LUOP
:11;.17
It follows from equations (3.6.16) and (3.6.17) that t h Impeclan('(' ruat nx of the augmented network is
i:UilSi
Equation (3.6.18) ran be written ltl the part.itioncd '"m:

i
r t L'b
.;.l.IU~
o
''
where [Zbb] primitive impedance matrix of branches
[zbd [zld' = primitive impedance matrix whos« elemen t,lrt the mutual impedances between branches and iinks
[zu] primitive impedance matrix of links
It follows from equation (3.6.19) that
Z. = Cb'lz6bjC'b + [Zlb]Cb + Cb'[ZblJ + [ZllJ (3.0.20
Since
ZLOOP = C'[z]C
or
,
ZLOOP =
then
ZWOP = Cb'[Zbh]Cb + [Zlb]Cb + Cb'[Zbd + [zu] From equations (3.6.20) and (3.6.21), therefore, ZLOOP = Z.
(3.6.21)
.'6 Computer methods in power system arwlysis
The loop admittance rnatrrx can he "ht1dnf'ri from
f)eriwlion of loop admittance matrix .from (lll!!trlenled n et uor k: admittance matrix
['J)I" ""p admittance matrix }'i ., IOP call i,p )htamed from the augmented
idll\;ll:lllt'l'm:ltnx fRR Frorn equ.n .ou ,:).(..7) and (::l.G.18),
(3.6.22 )
In p:lrtitioned form.
I
Co Cb Ub i Bt' I
_._
0 r, 0 I [.'/ Th« currents through the elements of the pnrnrtive network from equat ion ,:l.,;:!O) are
I = ('1 LOOP
Premult iplying by B',
IN = B'clLOop
(3.6,24)
However from equation (3.5.11) the left side of equation (3.6.24) is zero. Therefore. equation (3.6.24) can be written
(Ct, + Bi)l LOOP = 0
It follows that
Co = Bz'
Substituting from equation (3.6.25) into equation (3.6.23), (;11< = L
(3.6.2.5 )
(3.6.26)
III fl. similar manner it can be shown that
(3.6.27)
Substituting from equation (3.6.26) into (3.6.22), ZLOopf'BR = (;'[2][Y].8
Since
[z][y] = t:
Chapter 3
Incidence and network matrices
57
then
ZWOpfllR = ('IB
Therefore. from equation i;U;'27),
(~.6.28)
x. Z }, Y, \)
  . 
X:: l, Ye< r. n (. It follows that
ZI}"l + Z,Y, ZlY, + Z,Y.
[", o
(3.6.29)
Z,Yl + Z,Y, = 0 Z,Y~ T Z,Y, = l.·1
(:3.6.30) (:3.6.,'31 )
Solving for Z, from equation :.3.0.:'\0), Z3 = Z'Y3Yj1
and substituting into equation (3.6.31), Z,Y,Yl1Y, + Z,Y. = ["I
or
Since
Z,Y LOOP = ['I it follows that
Derivation of branch impedance matrix from augmented impedance matrix
The branch impedance matrix ZBR can be obtained from the augmented impedance matrix ZLOOP' Combining equations (;3.6.29) and (3.6.30) yields
(Z,  Z,Z,IZ3) Yj = Us
58 Computer metluHh in posoer .Ylftem analysilf
" .
•. lI1ce
ZBRY1 = U, it follows that
Irerirat io n of branch adntittance and irn pedance matrices from bus admittance and irnpe£iance mat.rices
l~lng the branchpath incidence matrix K the branch admittance matrix r u« can be obtained from Y BlS. From equation (3.3.1),
A.oK' = U,
and from equation (3.3.3),
Post multiplying A by K',
(3.6:3:2)
Su bst i tuting from equations (3.3.1) and (3.3.3) into (3.6.32),
AK'
B
Transposing,
f:A' = B'
l'"qmultiplying by [YlAK' yields
or
K(A'[vlA)K' = E'[ylB
(3.6.33)
From the singular t ransforrnat ions,
and
YER = E'[ylB
Chapter 3
Incidence and netUlOrk matrices
59
Hence equation (3.6.33) becomes
Y BR = KY BusK'
(3.6.34)
The branch impedance matrix is ZRR = Ylik = (K') TR~sKl
From equation (~);3.2),
(3.1).~~ii)
Substituting from equation (:3.(.UG) int« euuation ·3.6.:~;)'.
Derit'ation of bus admittance and impedance matrices from branch admittance and impedance matrices
Equation (3.G.:H) is prernultiplied hy K . and post multiplied by (Kt)l to obtain
(3.6.37)
Substituting from equation (:3.6.36) into equation 13.6.:37),
YBl"S = .·h'YBR.·h
Since
then
or
The nonsingular transformations for obtaining network matrices are summarized in Table 3.3.
3.7 Example of formation of incidence and network matrices
The method of forming the incidence and network matrices will be illustrated for the network shown in Fig. :3.10. The incidence matrices for a given network are not unique and depend on the orientation of the graph and the selection of branches, basic cutsets, and basie loops. However, the network matrices are unique.
i
I i
I
I I
..c! i I
.., '" I '"
~ .. .. I
e ! :...
... N I
~ t I • !
I I
ll~j i l·~
i ~ ~:
'" '" I :.;:. Q;_'
'" ., t> ..
.. ~ : ., :...
N  s, ~ I I
~ ~ ~ ~ i
I i i
I I
I i
" ., i
., t> '"
;:! .,
~ '" :... I
N
I
1
, I
c, e,
~ c c
'" c c
'" '" '"
...;: N s,
~ " ~
, Q I' '"
I o '"
tti .., ~ :"'1
, N ::
N II 11
I ~ I ~
N ~ :: ~
"t;
~ ,
~ ,
" :...
1:
"" II II
;:!
~ ~ "
Q .,
'" <:>..
~
'u 'CQ
~ ~
'G I '::Q
~
§ ~ .!:;
,_
Q"
uo.w.<aau[
r.JUlJpadul[ ~JUfl1ltlUpV I Chapter ,1
Incidence and network matrice$
61
I ~_j
(1)
(0'
 ,
Fill .. l,/(} Sample n e t : work.
Problem
'I Form the incidence matnces A. A, K, B. e. C. and r: for the network shown in Fig. 3.10.
b. Form the network rnatnces YBIS• YSR) and Zwop by singular t.ransf orma ti ons.
, Form the network matrices Z LUOP. Z BR. and Z 8';' by nonsingular
transformations.
Solution
The impedance data for the sample network is given JD Table 3.!.
Table 3,4 I rnpedance s for sample network
Self Mutual
Element BWl code l mpedance Bus code Impedance
number pq Zpq,pq r~s ZpIJ,f't
12(1) 0.6
2 13 0.5 12:1) 0
3 34 o 5
4 12(2) 0.4 12(1 ) o 2
5 24 0.2 The network contains four nodes and five elements, that is, n 4, and e = 5. The humber of branches is
b=nl=3
and the number of basic loops is
l=en+l=2
62 Computer methods in power ayatem analysis
3 0 C!'
i
I.
\
"5 i
\
\
._4 CD I I
L • J
4
Fi e . 3.11 Tree ana ("0 tree oj the oriented co n n.ec t e d gro ph. oj sample network.
 Branch
Link
11. The branches and links of the oriented connected graph of the nrt work are "hUWCl ill Fig. ;1.11. The elementnode incidence matrix is
f?""
'._/
1
1
,;
1 I 1 :
__ 1 :
, 1 ,
,1 1 i ! !
'_' __ i_:i
ii' •
! i : i 1 I
5
Selecting node 1 as the reference, the bus incidence matrix is
~I @ ® 0
e 1 l_l!_I_1
2 I 1 I . I
'',
A 3 1 1 I 1 I
'_!_' ,
4 1 I I
!II
5 1 . \ 1 i
Chapter 3
The branchpath incidence matrix is
'~h: i 2: ;~\ "~
b J
 'I __ i r p
1 I
K :2 ! I I 1
3 J 1 Incidence .. nd nett.oork m«trices
63
The basic ar.ci tie cutsets of the orieutrd connected graph of the network are shown III Fig. 3.12 The basic cutset incidence matrix is
B
~i ABC
T 1  .] I
2 1= _, ~=I=I
3 ' iIi iI
, I
4 1 I
1
.') 1 1 I 1 I
..
B ;f
.0 3
\
\ lS ~
\
\ 
\
CD ®
I I
L "" __ .. . _J
/ 4
A\.,..
~: Basic cut set
t· Tie cui set
. 4 EI
Fig. 3.12 Basic and tie cutsets of the oriented connected graph of sample network.
64 Computer methods in power &'ystem anal.r&i&
. '..._
A
.... 3 ..... 0 \
\
\ 5).
\
\
\
E
._~~~.
D
,
~~~
4
Basic loop Open loop
The augmented cutset incidence matrix is
e'
B
D
E
c
A
J , I,
:11
~ I 1 1 I I
1·,1
3: iIi
1111
1 l 1 II!
::1,1 I 1 IIi 1 I 1 I
Fig. 3.13 Basic and open loops of the oriented connected graph oj sample network.
CluJpter 3
Incidence and network matrices
65
The basic and open loops of the oriented connected graph are shown in Fig. 3.1:1. The basic loop incidence matrix is
'''"', IDE
e'': . ..
1
1
c
1
4
5 !
The augmented loop incidence matrix is
""e 4. BCD F
~~,.
1 1 ;
__ ._ • __ 1 •
, ,
1 :
2
c 3
4
5
66 Computer methods in power sy s te m: analyai3
b. The primitive impedance matrix of the sample network from Table 3.4 is
[z) =
""e
e ',,
I 10.G~IO.l~Io.2~i~!
2 ~~I·IJI 3 I~iI
4 ~I~I
5 III!~~I
2
3
4
By inversion, the primitive admittance matrix is
""e 1 2 3 4 5
e "" . ~~~_.
2.n8:~ I 0.417
1.042
2
I I  
I 0.417 2.083
I 0.208
2.000
I 1.042 0.208 3.021
I 5.000 [y] = 3
4
5
~
en
C C'C
0
'';:;
c::
E:
S
~
5 ~l
..::
'
~
t;,JC
'y:
c:
>,
.c
c
CJ l::
c::
"
~ :;<
x
'i:
..., C'C
c::
E
<:) N
u
C
c::
_, ~
'
E ::;

.. 
"
~
.s '"
"" :;,
.c: '"
r ;:.... G e e
I I ,.... I I ,....
I I
I ,.... I ..... I I
I
..... I I I .< I .....
I I I I 0
0
0
I
I L()
C'l 00 .....
0 ~ C'1
0
..... 0 M
I
I 0
I 0
c:
I C'l
" C'0 00 I
I ..... 00 0
I ...,. 0 C'l
0 N 0
I I
M I r N I I
00 < ...,. I
I 0 ...,. c I
N 0 .....
I I I I ...... I I .....
I I I
..... I I
I
I ..... I ....
I
I ...... I
I
..... I I
I e 0) G
e
0 I 0 0
0 0 0
0 0 0
lQ N r.
I I
Cl M 0
0 00 0
N 0 c:
0 t< C'l
I I
0 0 0
C'l 0 0
c: C'l c:
C() 0 L()
I I I 0)
e e G
G e e
L() I ~
I
10
100 o Cl
o
I
8
o
L()
I
o 0
o 0
o 0
N N
I
I I I I :'1 I
I 10
i
:/ I I ~ i I
I
101 I
I I ~ I I
! I
I I I
_ I L') I I I I
ole I I I Ie 1 I
' 6 0 I
I C'1 I
10
"~Ll
e I  I I
CQ ~. ,I i i
C'< II i I
 ~J ~ I
'"ill II I 1
~ 12_1_ I I ' I I
I  C'l M ..,;< '<?
riI
I ,~ II C'l I
i 0 I
"1' C'~ I C'1 II
o 0
I~I
L!';
I 0
I
' v
0 0
I I I
..,;< >r.
I 0 0
I 'f;
C
.....
cj
E
....
.2
Ul
;;
:::
...
cj
§;
. iii
c
>.
.c:
?:!
'c;
.z:
c
'"
';::
~
~
<:)
''
~
'
E
c:::
U ~
~
c:
.~
<:) 0:
..<:: .;:
r C..)
C..)\ I  I I
I
~I I I I
I I ~
~I I 
 ' I
" 1
N ~ ""I' I~

, I I I 0
I 0
I I I ~
I~
I ,
':',1 I 00 I 
0 "1
0 N 0
 I 0 ~
I I
I ~
0
0
0
COl
I I~ I
1 <:";
 r/) 1
_,. ~
0 N ,0 I I
I I I
 0 0 I 0
0 0 0
0 0 I 0
.<) l~ I
I I
~
:::> I C':: I 0
0 00 0
COl 0 0
t:") l t:"::
I I I
0 I ::;:, I 0
C'l 0 0
0 C'l 0 I
00 I:': I 1:"::
I I I
I .
"
1 _I
,~ \~~I
1
~  I I
~ I I
COl I
1
o
0
I 0
I ':'1
<:0 I M
0 00
:1 0
0 N
I I
I
.... ' l I 1
6 ~
~
 0
I I I 1 I I
 Ii Ii I I
I
il I 1 I
I II I
I
1 \ \ I
I I \ I
I I I I I C'I
I I I .
~\ \ I ~ \
I I ~ I I
 I'~ i I I
I
~ \ ~.\ I N I
. I I~ ; I ~ ;: I C'; I 
I I I 1 .....
"": ..... I I  I C';
I
I I 1 I I I
I I;': I I It:
I I
1
 .r; I I ...... 1 ;
I I I

..... 1 I ...,.. I I~
 1
I I
I i I ...... 1 II I
I
I I I I I
 I I I I
I I I I
I I ...... I I
I I
I I I I
I
I ...... I I
I I
I
I I I I
..... I
I i 1 I I ~
I I
1 1 1
~ 1 1 I~ I t::
I I 1
I )'0: I I~
I I I
...... I .~ I I( I. 7
I " I I I
I I I
.~ I ~ I 1"( I .~
1 "
I I 
,
en
eo:> <.::>
0 ......
I
<.::> c:
0 0
I ::l.. C
C N
rn
;<
.;::
~
<::
6
C) M
<.J
.::: ~
<::
"0 I .
C) ~
P
.e N
,...
'0
.::: N
o:l
...
.D
C) 0:
..<:: 0;
~ N r:
I
I '"
0 I 0
I I I
I
..,. ':":
0 0
I
I:": C
c'; co
c.;:;
...
0 0
co L")
C".: e
00 M
0
l
~ I ..". L")
C I 0 0
I I
..". I .... I
? I 0 I
1 1
I 1 L")
0
.... L")
0 0
c.;:; I ......
0 I 0 .... I I~ I o:
0 I ':: I ':'1
.< i .< cv:
0 0 I 0
I I
1
G I ..,. I Ir::
,:" I r I l!":
I .< ~ 
I c I c I c
I I I
I
: I
I .< I 
r~ ':'1 C
C'l ...... ...... o 0
...,. l~ I ':'1
C ,:"; I r
 .< ......
0 0 0
I I
I I
c: I G L")
01 L') L":
0 ..... 
0 C 0
I I I
CO '0 ..,.
01 01 0
~ C .....
0 0 0
I I I I
I L")
I 0
.... L")
0 0
0 ....
0 0 'I;
 II I I
e )i I I
. I ' 'I
(:v 1" I
L I ~ I
c ! ,~ 'Y)
L": ~I
 i  i:
0 :::> c
I I
c I t I 1.,":"
':'1 I.,":"
~ ! .. I 
C I :::> c
I I
I
I
 I ~ I 0
r
':'1 I  I
c :::> c
i c
o
o cv: e ':"1
:::>
00

C
:1: I 0
e ("I ~..: ~
:::> I C I 0
;1;' i~
G ~ I: I~
(:j (0 0)
II I
I
e I' I
I I~ I t:
I.""::
 1  
c c I 0
I I
I I
I
C 00
,:"
 ._ 
0 0 0
I I I
I
 I .~ I s
r ':'1
':'1 I  ~i
0 I 0 I 0
I I I 1
I I (0 e ':)
i4 Computer methods in power aystem analysis
Problem«
a.r
Select for the sample network than that, used in the example. form;
a. The incidence matrices j. A. K, B. 8, C, and t and verify the
shown in Fig. 3.10 a different tree Retain node 1 as the reference and
following relations:
I. AbK' = e
ll. B, cAlK'
Ill. Co =.  Hi' IV. rB' = L'
b. The network rnat ricrs fa,·." YhH• and Zwop by singular transformations
c. The network matrices ZwoP. ZHR. and ZBUS by nonsingular transformations
B (N)
0 t (S\
~A c
~o
Fig. 3.14 Sample power syst.ern f or Prob . 3.2.
3.2 The positive and zero sequence impedance data for the sample power system shown in Fig. 3.14 is given in Table 3.5. For this system:
a. Draw the positive sequence diagram and an oriented connected graph.
Tubl e .1.5 Positive and sero sequence impedance do ta of sample poteer s y st.e m f or Prob . 3.2
Positive sequence Zero sfquenct ]futual
Element impedance Impedance Element impedance
Generator A o 0 + jO ? 0.0 + jO
_J
Generator B 0 0 + jl). 25 o 0 + iO
Generator C 0.0 + j025 (J.O + jO.l
Line Al3 o 03 + jO I" o OS + jO 45
.)
Line B·C(N) 0.05 + j02:! 0 13 ~ jO 75 Line BC(S) 0.08 + j048
Line BC(S) o 05 + jO 22 0 13 + jO 75
Line CD 0.02 + jO II 0 07 + jO 37 Chapter J
Incidence and network matrices
75
b. Selecting ground a.') reference, form the incidence matrices A. K, B, B, C, and (; and verify ';he relations:
1. AbK' = {'
n. B,'cc 4,K'
Ill. Cb C~  Bit
iv. tEP I
c. Neglectiug r(,~l.srall('r. form hr P(J';;UV( sequence network matrices ; j,; s. /Ii' S, YHH, ZHF c . L·"u'· ann YUlO? by singular trunsfnrrr.at 'un;;
d. Xeglecting ,:2:'lstl;llce. f orrn th~ zero> SPTif;nrf:> network matrices YBUS, Zp:'s, Ylil,. ZIIR. Zu"," tnd ·,', .•. I/L' by singular transformations.
e. Repeat c and d using nonsingular transformations.
f. Repeat c including resistance.
""T"""'!"'"T'A TTD
_j : L. __j i
Fit'. .1.15 Sample network for Prob, 3.3.
Table 3.6 Positive sequence reacWnces of sample network for Prob . 3.3
Element
Positive sequence reactance
GA. GB AB BC AD CF DF
0.04 0.05 0.04 o.os n 02
o O~
o 10
3.3 The positive sequence reactances for the network shown in Fig. 3.15 are given in Table 3.G. Designate elements A. B and DF as links and node G as the reference bus. Form:
a. The incidence matrices .4, A, K. B, 8, C, and C
b. The network matrices Y BUS, Y BR. and Z LOOP by singular transformations
76 Computer methods in power system analysis
" The network matrices Y 8{.·S, Znus, ZBR, Zwop, and Ywop by nonsingular transformations
:~ . .t Prove that when there is no mutual coupling the diagonal and offdiagonal elements of t he bus admit tance matrix Y'<I., ran b(~ cornputed from
r =)' u, ~ ...
where .II., is the sum of the admittances of all lines connect ing !>11"(,~ 1.' and j.
:L) C~ing the bus impedance matrix Zm.·s computed In Prob. 3.2 and t he internal generator voltages given in Table 3.7:
a Compute the positive and zero sequence bus voltages of the network.
o. Compute the positive and zero sequence currents flowing in the line BC(N).
To.ble 3.7 Internal eeruru tor iol t a ee s for I'rob. 3.5
1 nternal per unit voltaqes
Generator Positive sequence Zero sequence
A. .0/00 0
13 1. 1/100 0
C 1 O/W 0 1/0° 3.6 Using the relations between interconnected and primitive network variables prove the following:
a ... hKt = U
v. B, = AIKt
Bibliography
Byerly, R. T., R. W. Long, and C. W. King: Logic for Applying Topological :\Iethods to Electric :\ etworks, Trans. AI EE, vol. 77, pt. I, pp. 657667, 1958.
Edelmann, H.: :\ umerical and Algebraic Generation of Meshimpedance Xl atrices by Settheoretical Intersection on a Digital Computer,
Chapter 3
Incidence and netl(]()rk matrices
77
Trans. I E'EE on Power Apparatus and Systems, vol. 83. pp a97402. 1964.
GuillemlP, Ernst A.: "Introductory Circuit Theory." John Wj!(,.\! & Sons, Inc, New York, 1953.
Hale, H. W.: A Logic for Ident if'ying the Trees of a Graph. Tvui« .4.J EE, vol. 80, pt. III, pp. 19.'i19R, 1961.
Happ, H. fL: Orthogonal .:\et works, Trans. I EHE 07; h,j;, ,.4.pparatus and Systems, vol. 85, pp. 281 ~:294, 1966 .
. .. Special Cases of Orthogonal ?\etworks·Trel '!;"; '. E,K, 'I'rani IEKE Or! Power Apparatu., and Systems, vol. S;j, :i}:' S8li~91. [g(lt, Kron. Gabriel "Tensor Analysis 'if "\"etworks," John W::f;,' tV :;;ons. Inc.. "\"(·w York, 1939.
..: "Diakoptics," .\Iacdonald and Co., Ltd., Londcn [9(;;1
Le Corbeiller, P.: ".\Iatrix Analysis of Electric X etworks, . ,J one Wiley & Suns. Inc., New York, 19:10.
Sato, X.: Digital Calculation of X etwork Inverse and ,\le~l} Transformation .\Iatrices, Trans. A.IEE, vol. 79, pt. III, pp. 71\'725, 19hO.
Seshu, S., and .\1. B. Reed: "Linear Graphs and Electr.cal ~etlV()rks," Addison Wesley Publishing Company, Inc., Reading. \1,,>,,;. HWl Tinney, W. F., and C. :\1. .\1clntyre: A Digital Xlethod 'ni ()bLtinmg a Loop Connection Xl atrix, Trans. AlEE, vol. 79 pi LlI. pp. 740 746, 1960.
W'
/~~
/ .~ ...  .. . I j/ ~:,, '
t • ~.,,___ ~. _ ..
4.1 'ntroduction
The methods presented in Sees. :{.fi and 3.6 reo u.re transformation and inversion of matrices to obtain network matrices. An alternative method based on an algorithm can be used to form the bus impedance matrix directly from system parameters and coded bus numbers. The underlying principle of the algorithm is the formation of the bus impedance matrix in steps, simulating the construction of the network by adding one element at a time (Brown, Person, Kirchmaver, and Stagg, 1960)t. A matrix is formed for the partial network represented after each element is connected to the network.
In addition, an algorithm is presented for deriving the loop admittance matrix from a given bus impedance matrix.
4.2 Algorithm for formation of bus impedance matrix
Performance equation of a partial network
Assume that the bus impedance matrix ZEUS is known for a partial network of m buses and a reference node O. The performance equation of this network, shown in Fig. 4.1, is
EBU~ = ZBusiBUB
t Names in parentheses refer to the Bibliography at the end of each chapter.
Partial netuiork
Partial network
(al
Partial network
(b)
Reference
Element pq
Reference
Fig. 4.1 Representation of a partial network.
Fig. 4.2 Representations of a partial network with on added element. (a) Addition of a branch; (b) addition of a link.
Chapter 4
A.lgorithrru for formation of network matrice8
81
where EBUS = an m X 1 vector of bus voltages measured with respect to the reference node
1 BUS = an m X 1 vector of impressed hUi CUlTf"iJ ts
When an element P'1 is added to the partial network it may be a branch or a link 8.8 shown .n Fig, 4,2,
If pq is a branch, a new bus q is added te, the parual network and the resultant bus impedance matrix is of dimension 1,1/' , 1). X (Ill + I), The new voltage and current vectors are ,:,f C!ITIC!':3IUc ! 1'1 + 1) X l. To determine the new bus impedance matrix req;l,rb .J"'Y "ht, calculation of the elements III the new row and column,
If pq is a link. no new bus is added k the pqr\I:U network. In this case, the dimensions of the matrices in the pf'r!(Jrm3,!JI:," equation are unchanged, but all the elements of the bus impedance matrix must be recalculated to include the effect of the added link
Addition of a branch
The performance equation for the partial network with an added branch pq IS
:" ,I 'I
E,. ,
P n: 'I
, I
ZIl Z12 ZIp , : Zlm i z.;
: : , I
,,: l!;~r~'
;~i Z22 :_'_' _' !~:~~~l Z2m_I_~_'_"_
' ...... !".: ......
i,I1I'i 
1 i 'I '
pi Z1'\ ' Z1'2 . , . , . Zpp : ' .. ,Z".. Z'''l
., . , j. . . ;:, . '. . ·1. , .1, . ,,~~
• , !
'. 1: ",,
m. Z Iz ... Z i .. ·,Z iz
",1 I m2. mp 1 ,111m ~..,q
·i,j!i
s, Zql i Z.2 II ' i z.; I • ,. Z'1'" Zoo
I , i i
. 12 :
. I
I
I
Ell
I
~I
,I E1' 1 I
(4.2.1)
, ,
L; '
',
'E
. q
, !__j
It is assumed that the network consists of bilateral passive elements. Hence Zqi = Ziq where i = 1, 2, . , . ,m and refers to the buses of the partial network, not including the new bus q. The added branch pq is assumed to be mutually coupled with one or more elements of the partial network,
The elements Zqi can be determined by injecting a current at the ith bus and calculating the voltage at the qth bus with respect to the reference
Partial network
Element PQ
1110 pe; unit C:)
I
@)
Reference
Fig. 4.3 Injected current and bus voltages for calculation of Zq"
node as shown in Fig. 4.3. Sinre all other hus currents equal zero, it follows from equation (4.2.1) that
s, = Zhli £2 = Z2Ji
Em = Z",Ji E. = ZqJi
(4.2.2)
Letting I, = 1 per unit m equations (4.2.2), Zqi can be obtained directly by calculating s;
The bus voltages associated with the added element and the voltage across the element are related by
(4.2.3)
The currents in the elements of the network in Fig. 4.3 are expressed in terms of the primitive admittances and the voltages across the elements by
. ,
I I
i Ypq.pq i Yp •• pa Vpo:
11:
(4.2.4)
Ypa.pq i Ypa .o a
Chapter 4
.4.lgorithm..Jor fonn4tion oj network m4trices
83
In equation (4.2.4) pq is a fixed subscript and refers to the added element and fXT is a variable subscript and refers t,G all other elements. Then,
are, respectively, current through and voltage across the added element
are the current and voltage vectors of the elements of the partial network
is the selfadmittance of the added element
is the vector of mutual adrmt tances between the added element pq and the elements p=o of the partial network is the transpose of the vector y"",.,
is the primitive admittance matrix of the partial network
!i",N iY •• ,~l
The current in the added branch, shown in Fig. 4.3, is
(4.2.5)
However Vpq is not equal to zero since the added branch IS mutually coupled to one or more of the elements of the partial network. Xloreover.
v •• =EpE.
(4.2.G)
where E. and E. are the voltages at the buses in the partial network. From equations (4.2.4) and (4.2.5).
and therefore,
fjpq,P'Vp• Vpq =  
Ypq,pq
Substituting for v •• from equation (4.2.6), fjpq, •• (.E.  E.)
VpQ =  
(4.2.7)
Substituting for Vpq in equation (4.2.3') from ',4.2.7),
 (it it)
E = E + Ypq,.. •  •
q p
Ypq,pq
Finally, substituting for Eq, Ep, E" and E. frum equation (4.2.2) with I, = 1,
i = 1,2, ... , m I,_!q
(42.8)
84 Computer methoda in power sydem anal'ys~
The element Z •• can be calculated by injecting a current at the qth bus and calculating the voltage at that bus. Since all other bus currents equal zero, it follows from equation (4.2.1) that
£1 = Zlqlq
E2 = ZzqIq
(4.2.9)
Em = Z",.I. E, = Zo.I.
Letting Iq = 1 per unit in equations (4.2.9), ZqO can be obtained directly by calculating Eo.
The voltages at buses p and q are related by equation (4.2.3), and the current through the added element is
Z:po = 10 = 1
(4.2.10)
The voltages across the elements of the partial network are given by equation (4.2.6) and the currents through these elements by (4.2.4). From equations (4.2.4) and (4.2.10),
and therefore,
1 + ypq ••• v •• Vpq = 
Substituting for v •• from equation (4.2.6), 1 + Ypo.p.CE.  E.)
(4.2.11)
v =pq
Substituting for Vp• in equation (4.2.3) from (4.2.11), EQ = E; + 1 + YPQ ••• ct.  E.)
Ypq.pq
Finally, substituting for Eq, Ep, s; and E. from equation (4.2.9) with I , = 1,
Z = Z + 1 + Ypq .•• (Zpq  Z •• )
qQ pq .
Ypq.pq
(4.2.12)
If there is no mutual coupling between the added branch and other elements of the partial network, then the elements of Ypq.,.. are zero and
1
ZPQ,pq = y P •• P.
ZII' == Zpi
It follows from equation (4.2.8) that
Chapter"
Algorithrrufor formation of r&etlClOrk matrice6
85
i = 1,2, ... ,m
i ,.0 q
and from equation (4.2.12) that
Furthermore. if there is no mutual coupling and p ,s the reference node,
.. m
Zpo = 0
and
Zq, = 0 Also Z"" = 0
and therefore,
2 = 1,2, ... ,m
i ~ q
Addition of a link
If the added element pq is a link, the procedure for recalculating the elements of the bus impedance matrix is to connect in series with the added element a voltage source Cc as shown in Fig. 4.4. This creates a
Partial network
Fig. 4.4 Injected current voltage source in series with added link and bus voltages for calculation of Z".
®
Reference
fictitious node I which will be eliminated later, The voltage source e, is selected such that the current through the added link is zero,
The performance equation for the partial network with the added element pl and the series voltage source e, is
~
! E2 I
Since
1
p
m
ill , I i 1,1
111,_ Zzl~.I~~i~~:i~!··~,_~l~ Zll !_~~_l_,
I 21! z.; i .•. [ Z2P 'I  .. ! ZlR I 12 ,
1111 I 11
1'·\'" ... ! ...... i ... ""1'1".1.
[11 I!
PI~I~''· Zpp _'_'_'1 z_ ~I i, i
'I .. ·! .. ·I [ .
m I z., z., z., .. ·1 z •• ~ I I. I
I 1 ,I i
l, Za 1 Zu ,.. Zip [ ... I Z,... Zl! 1 II:
i...,_______J
(4.2.13)
el = EI  Eq
the element Zz, can be determined by injecting a current at the ith bus and calculating the voltage at the lth node with respect to bus q. Since all other bus currents equal zero, it follows from equation (4.2.13) that
E. = Z.J. e, = Z,Ji
k = 1,2, ... , m
(4.2.14)
Letting I, = 1 per unit in equations (4.2,14), Zli can be obtained directly by calculating el.
The series voltage source is
(4.2.15)
Since the current through the added link is
the element pl can be treated as a branch. The current in this element in terms of primitive admittances and the voltages across the elements is
Chapter J
AllJorithnufor for m.4 tion of n.etlllO,.k r7U1trice5
87
where
Therefore
i'pi =
Slllce
9 p!,r;c
fI pq,pa
and
then
: pi
11.2.16)
Substituting in order from equations i4.2.lfil. !!.2.o), and (1.2.14) with l , = 1 into equation (1.2.1.")} yields
1. '2.
(4.2.17)
The element Zli can be calculated by injecting it current at the lth bus with bus q as reference and calculatmg the \"(ltagl: at the lth bus with respect to bus q. Since all other bus currents equal zero, it follows from equation (4.2.13) that
E. = ZuII el = ZllII
k = 1, 2,
,m
0.2.18)
Letting II = 1 per unit in equation i4.2.18), Z/i caIl be obtained directly by calculating ei.
The current in the element pi is
lpl = II = 1
This current in terms of primitive admittances and the voltages across the elements is
Again, since
and
Ypq.P,
then
1 + Ypq.P'vp• Ypo.pq
(1.2.19)
88 Computer method» in power 'Y6tern arualYN
Substituting in order from equations (4.2.1.9). (4.2.6), and (4.2.18) with I, = 1 into (4.2.15) yields
Z Z 1 + ijpq,P'(Z pI  Z.r)
Z" = 1'/  ql + '
Ypq,pq
:4.2,20)
If there i~ 110 mutual coupling between the added element and other Plf'Ill(,'lt~ of t h« partial network, the elements of fjpq,p. are zero 8.Ed
Zp,;.pq = Ypq,Pli
11 l,dlo\\s from equation (4.2.17) that
i = 1, 2, .... m i ~ I
an d from equation (4,2.20),
Furthermore, if there is no mutual coupling and p is the reference node,
Z; = 0
i = 1,2, ... ,m
i r' l
and
1 = 1,2, . , . ,m
i ~ I
Also
Zp' = 0
and therefore.
ZII = Zqr + Zpq.pq
The elements in the Ith row and column of the bus impedance matrix for the augmented partial network are found from equations (4.2.17) and fl2.20), It remains to calculate the required bus impedance matrix t o include the effect of the added link. This can be accomplished by modifying the elements Z,;, where i, j = 1,2,. ., In, and eliminating t he it h row and column corresponding to the fictitious node.
The fictitious node I is eliminated by short circuiting the series voltage source f,. From equation (4.2.13),
(4.2.21)
and
('I = z.i «, + ZIIII = 0
(4.2.22)
Chapter 4
Algorithm&foT formation of network matrices
89
where i, j = 1,2, . .'7'. Solving for 1, from equation l4.2.22) and
substituting into (4.2.2n
EEr:s = (ZBU,'  ~i~lj') 11;';'>
.uti
which is t h« performance :cC]U:Hinti of t h.· parj'j;{l net work includi ng the link p.'(j. It follows r.ha: ;'nc rquired bus irandance matrix IS
2,/21,
Z BC· S (mod Hied i
where any element of Z 11; S :lrc.d·fi«il is
A summary of the eo uat.ions for the formation of the bus Impedance matrix is given in Table ·U.
4.3 Modification of the bus impedance matrix for changes in the network
The bus impedance matrix Z SIS can be modified to reflect changes in the network, These changes may be addition of elements, removal of elements, or changes in the impedances of elements.
The method described in Sec. 4.2 based on the algorithm for forming a bus impedance matrix can be applied if elements are added to the network. Then Z B( S is considered the matrix of the partial net wor kat t ha t stage and the new elements are added one at " time to produce the new bus impedance matrix Z~us
The procedure to remove elements or tn change the impedances of elements is the same. If an element is removed which is not mutually coupled to any other element, the modified bus impedance matrix can be obtained by adding, in parallel with the element. a link whose impedance is equal to the negative of the impedance of the element to be removed. If the impedance of an uncoupled element is changed, the modified bus impedance matrix can be obtained by adding a link in parallel with the element such that the equivalent impedance of the two elements is the desired value.
When mutually coupled elements are removed or their impedances changed, the modified bus impedance matrix can not be obtained by adding a link and using the procedure described in Sec. 4.2. However, an equation can be derived for modifying the elements of ZHl.·S by introducing appropriate changes in the bus currents of the original net