0% found this document useful (0 votes)
309 viewsB R Gupta Ch-10 Power Flow
Uploaded by
Pe KaCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
309 viewsB R Gupta Ch-10 Power Flow
Uploaded by
Pe KaCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 64
{conductor adius rand inside
ers of dilelecttic. Find expres.
CHAPTER
| Ovcssmemes
@ oad flow study in a power system constitutes a study of paramount importance, The study
weals the electrical performance and power flows (real and reactive) for specified conditions
the line and transformer loads (as well as losses) throughout the system and voltages at
points in the system for evaluation and regulation ofthe performance of the power system
conditions known a priori. Further alternative plans for future expansion to meet new
ids can be analyzed and a complete information is made available through this study.
ybusis a node at which one or many lines, one or many loads and generators are connected.
lations can be made by using bus admittance matrix or bus impedance matrix.
is simpler to use the bus admittance matrix in these studies since it can be generated
tion,
power system consists of a large number of buses interconnected through transmission
3. Some of these buses have generators connected to them while loads are connected to the
ing buses. Some buses may have both generators and loads while some others may have
ic capacitors (or synchronous condensers) for reactive power compensation or voltage control.
x the simple 4-bus system of Fig. 10.1(a). S,), denotes the 3-phase complex generator
flowing into the ith bus and S,, denotes the 3-phase complex load power flowing out of
Sey" Poy* ies (10.19)
5,>P,+0,, (10.16)
237 238 Power System Analysis and Design
where P and Q denote the real and reactive power. At each bus the generator and load
ccan be combined so that the net 3-phase complex power flowing into ith bus can be written
S.* Sq Su Par Pa) + MQ 21) =P, +5Q, (0
® sy
elt
Sy Sia
Sot 1 Sea
1 a
7 2
Sor
Sa
@
Fig. 10.1 Four bus system (a) One line diagram
Figure 10.1 (b) shows the equivalent circuit of the network of Fig. 10.1 (a). S,, 5, S, Si
denote the net 3-phase complex power flowing into the buses and I, [yl J, denote the curr
flowing into the buses. Each transmission line is represented by a m circuit.
‘The equivalent circuit of Fig. 10.1 (b) has been redrawn in Fig. 10.1 (e) where al the sour’
have been shown connected toa common reference at ground potential and the shunt admittanes
at the buses have been lumped. Since the system has 4 buses, the number of nodes (m) is 8. It
general, denotes the admittance of ith node with respect to reference, y, denotes the admittance
between ith and pth nodes (y, =y,) and ¥, denotes the voltage of the ith node. We assume tht
there is no mutual coupling between the lines. Application of Kirehhoff’s current law to the fou
nodes of Fig. 10.1 (c) gives us the following equations:
Vet VV VK —-Vnt %— Mu
‘Yao VeVi. * Vz~ Vers
VV 0+ VV Wg + Vy~ Vn * Va Fe (0sy
Pat Ve Vt Ve Voy
Equation (10.3) can be re-arranged and written in matrix form as:
1] [Owtmatyt) Me => ma YK
4] | “a Om +92 +9) “> ° %
4 in ¥s Om tit atyy) eA
(a0 + Ya + Ye) ILM.
(10.4)
I “Na 0
Equation (10.4) can be written as:
st
Ho
Fig, 10.2 is the generator and load poy]
ig into ith bus can be written
OQ.) =P, +50,
Sis
Sos
x of Fig, 10.1 (a). S,, Sy Sy
1 yy Lente the
ya rcircuit.
10.1 (c) where al the sout
ential and the shunt admit
he number of nodes (m) is $f
rence, y,, denotes the admittany
of the ith node. We assume tk
rchhof?'s current law to the f
Wis*%~ hire
Was 4
Wat-Vo.
Way
mas:
ma
°
3 +934) “Yu
Owe t ne +950),
Load Flow Studies. 239
(10.5a)
©
Fig. 10.2 (b) Equivalent circuit (c) Modified equivalent circuit,
Ny Yot Yat WaMu
Yann tn, 240 Power System Analysis and Design
Ya =Yaut In Yay Ye
From network theory we know that each admittance ¥, is known as self admittance (or driving
point admittance) of ith node and equals the sum of the admittances connected to ith node. Each
off diagonal term y, is known as mutual admittance (or transfer admittance) between ith and
pth nodes and equals the negative of the sum of all admittances connected direetly between the
‘th and pth nodes.
Eq, (10.5.a) can be written in compact form as:
og = Dado (10.36)
For an n bus system J, is an n x 1 vector with general entry J, Vis an n x 1 vector with
general entry ¥,
(fe) i8 ax 1 matrix with entries
17) 2 7%,= short circuit driving point admittance at ith node
¥,=[%,| 2 %_= short circuit transfer admittance between ith and pth nodes.
[y,.]is known as bus admittance matrix.
Equation (10.5b) can also be written as:
Zed Tou (106)
where = Z=¥,,
Z,,, is known as bus impedance matrix and is the inverse of ¥,, (see Appendix B for matrix
inversion). Bus impedance matrix is used for fault calculations using computer.
10.3 FORMATION OF Y,,, USING STEP BY STEP METHOD
‘The admittance matrix can be formed from the parameters of system components as discussed
above. A diagonal element Y, is the sum of all admittances connected to ith bus. An off diagonal
clement ¥, is negative of the total admittance directly connected between ith and pth buses. The
following'step by step procedure is very suitable.
‘We start with [Y,,] array initially set to zero, The dimensions of [¥,,_] is n x m where m isthe
number of buses (the total number of nodes m= n+ 1 including the reference node). Consider
‘an element having admittance y,, connected between buses i and p. Four entries in (Y,,] a€
affected: ¥,, ¥,, ¥, and ¥,. We modify these entries as under:
Va aw = Yaau*Yp (40.78)
Yo. naw = Yo, ctt Ye (10.78)
Ba n= I Yo (10.70)
Tew Fatt Yy (10.74)
We ada the elements one by one and modify the entries of [Y,.] as per Eq, (10.7). Ifan element
is connected from ith bus to reference, only entry ¥, is affected.
jpsaniple 10.4
Formulate [Y,,] for the n
Solution
Since the network has tw.
‘Asan alternative we can
‘Add —j 0.5 element
‘Add -j 0.3 element
Add-j 0.4 element
10.4 FORMATIC
FORMATIC
‘Another method for forn
we first briefly review th
10.4.1. Graph,
each element of a netv
isknown as a graph oft
asa single element. Ifea
graph. The directionis,
of current in the element
A tree of a graph is th
all the nodes without for
of the connected graph.
of buses in the network
than the number of buse Load Flow Studies 241
Formulate (Y,.] for the network shown in Fig. 10.2. The values shown are admittances.
Solution =a
ce the network has two buses, [Y,,] isa 2x 2-matrix, 8885
as self admittance (or drivig
: connected to ith node.
admittance) between ith
anected directly between
th alternative we can use the step by step approach (Eq. 10.7). We start with
00
wa=[0 ol
(ao. 0.5 element
| Vou is an n x 1 vector y a[s 108
i Mal" Lejos -j05
ce at ith node Aiid-j 0:3 element
ween ith and pth nodes. y ju [-5 408
Chel | 4705 -j08
{dd -j 0.4 element
—j09 +05
l-[ ns ios
_(see Appendix B for mat ical
ing computer.
10.4 FORMATION OF Y,,, AND Z,,, USING SINGULAR TRANS-
FORMATION
‘method for formation of Y,, is by singular transformation, To understand this method
first briefly review the graph theory.
4.1 Graph, Tree and Co-tree
ch clement of a network is represented just by a line, the resulting diagram of the network
known as a graph of the network. Each source and the shunt admittance across it are taken
2 single element. I'each element of the graph is assigned a direction, it is known as oriented
“gaph. Tae direction is, generally, so assigned as to coincide withthe assumed positive direction
of current in the element. Fig. 10.3 shows the oriented graph of the network of Fig. 10.1.
“ Attee of graph is that part of the graph which has sufficient number of branches to connect
‘the nodes without forming a closed path. The tree branches form a sub-set of the elements
‘the connected graph. The number of branches b required to form a tree equals the number
F buses in the network (the total number of nodes, including the reference node, is one more
‘than the number of buses). 242 Power System Analysis and Design
Those elements of the graph which are not included in
the tree are known as links (or link branches) and form a
sub-graph, not necessarily connected, known as co-tree.
‘The number of links equals the number of elements minus
the number of tree branches.
Figure 10.4 (a) and (b) show a tree and co-tree of the
graph of Fig. 10.3. Ifa link is added to the tree one loop
is formed in the graph, A graph has as many loops as the
‘number of links. Forac steady state condi
‘The above system has 9 branches. Therefore, it has sumbers.
18 variables (9 branch voltages and 9 branch currents). ‘The performance equati
However, it is easy to see that all these 18 variables are Cc
not independent. The number of independent variables is Fig. 10 ceed staph of te
found from the concept of tee, The number of tree branches ialoodt ofa. 104
gives the number of independent voltages. For any system the number of tree branches equols
the number of buses n. The number of links gives the number of independent current variables,
Fig. 10.5 Represent:
Yy
and for admittance represe
1,
The impedance and adt
Fig. 10.4 (2) Tice (b) Co-tee of the graph of Fig. 10.3, ‘Asset of unconnected el
Admittance (or impedance
10.4.2 Primitive Network inimpedance form is
Avnetwork consists of many branches. Each branch consists of passive and/or active elements, 1
Fig. 10.5 (a) and (b) show a network branch, containing both active and passive elements in
impedance and admittance representation. Let Load Flow Studies 243
2, = self impedance of element —p
y= Self admittance of element ip
¥,, = voltage across element i—p
E,, = voltage source in series with element i~p
1, = current through element i—p
J,, = current source in parallel with element i—p
10.8. Oriented ‘
ielwork of fg 10 =
amber of tree branches
ndependent current vai Wye i= ¥o
@
| g
v \,
ote HH
lot te
Vp=Vi- Vo
©
Fig. 10.5 Representation of network element (a) Impedance form, (b) Admittance form
(10.8a)
(20.86)
(10.9)
(10.96)
" Aset of unconnected elements is known as primitive network. The performance equations in
‘Admittance (or impedance) form can be written for all the branches, The set of these equations
‘inimpedance form is,
ssive and/or active elements. V+E=a (10.104) 244 Power system Analysis and Design
and in admittance form itt
Lif
ip = Ott
‘Thus bus incidence m:
is the number of eleme
If the rows of the bus in.
can be partitioned into tw
bis the number of tree br
10.4.4 Formulati
Substituting Eq. (10.12)
esa (10.105)
where V and £ are branch voltage and source voltage matrices, ! and J are branch current anj
source current matrices, z is primitive impedance matrix (1e., a matrix whose elements i
branch self impedances) and y is primitive admittance matrix (ie., matrix whose elements ay
branch self admittances), Since no mutual couplings have been assumed between lines, 2 an
y are diagonal matrices,
10.4.3 Bus Incidence Matrix
For the graph of Fig. 10.3 the 9-branch voltages (Vy, Vig .~. Vi) can be expressed in terms
4-bus voltages (V,... V,) as under:
Premultiplying Eq, (10.1
AL
Since the matrix A shi
(10.1 element is the algebraic +
per Kirchhoff's current k
Similarly, A‘J gives ¢
this as ,, in writing Eq.
Equation (10.11) can be written in matrix form as: Substituting Eq. 10.16
V=AV (10.12)
where 4, known as bus incidence matrix, is ‘Comparing Eq. (10.17
us 1 2 3 4 Bus [~ Buses
> Eq, (10.18) suggests t
ee transformation of y. The
1 1 0 0 0 Standard matrix multipli
using Eg, (10,18). As me
2 o 1 0 0 Ay
orem Ep
3 oo 1 0 branches
a- Formulate for the ne
‘ a °F - Solution
5 a ee
6 0 o a 1 a
Link
7 ns er) branches a0
8 1 0 4 0
9 1 1 0 06
The bus incidence matrix depicts the incidence of elements (branches) to the buses in the
graph of a network. The element a, of the bus incidence matrix is written according to the rules
given as under: toad Flow Studies 245
= 1 if the th element is incident to and oriented away from pth bus.
f the ith element is incident to and oriented towards pth bus.
a, = 0 ifthe ith element is not incident to the pth bus.
‘Thus bus incidence matrix is rectangular and, therefore, singular. Its dimension is e xn where
‘eis the number of elements (branches) and m is the number of buses (excluding the reference).
{fhe rows of the bus incidence matrix are arranged according to a particular tree, the matrix
can be partitioned into two sub-matrices 4, of dimension 6 xn and A, of dimension 1x n where
the number of tree branches and / is the number of link branches.
(10.1
(and J are branch curren
a matrix whose elements
e., matrix whose elements
assumed between lines, 2a
) can be expressed in terngl 0.4.4 Formulation of Y,,, and Z,.,
I+J=yAV,, (10.14)
‘Premultiplying Eq, (10.14) by 4' (.e,, transpose of the bus incidence matrix)
AT+AS=AYAV (10.15)
Since the matrix A shows the incidence of elements to buses, 4’ is a phasor in which each
clement is the algebraic sum of currents through the network elements terminating at a bus. As
‘per Kirehhoft’s current law, the algebraic sum of currents ata bus is zero. Thus
Ale (10.16)
Similarly, A‘ J gives the algebraic sum of source currents at each bus. We have designated
this as J, in writing Eq, 10.5 (b). Thus
AS Igy (10.166)
‘Substituting Eq. 10.16 (a) and (b) into Eq. (10.15)
ry 1 =A yA, (10.17)
Comparing Eq, (10.17) and Eq, 10.5 (6)
s Buses =A (10.18)
\ Eq, (10.18) suggests the formulation of ¥,,,. Since matrix A is singular, A‘ y A is a singular
transformation of y. The bus incidence matrix can be obtained through a computer programme.
" Stendard matrix multiplication and matrix transpose sub-routines can be used to compute Y,_,
ts sing Eq, (10.18). As mentioned earlier Z,.. is the inverse of Y,,,.
ches «ee
Formulate Y,_. for the network of Fig. 10.1 (c) using, singular transformation.
‘Solution
Yio 0
A Yn
Yn
hes (10.9) Yo
Ne
L Yu
branches) to the buses in th ms
written according to the rules f va
0 Mia. 246 Power systom Analysis and Design
Using A from Bg, (10.13) we have Solution
yo «0 0 0 (a) From elementary :
0 ym +0 0
For the given valu
0 0 ym 0 Line (bus to bus
0 0 0 vy G
0 0 ~% 2
Om Recalling that the elem«
ew bus and the element ¥, equ:
0 ys -y 0 the th and pth buses, [Y,,]
%3 9-H; 0
dat
Ye “m2 9 0 ~y.
=
100010011 ‘ol °
ya(Q 2900 0 1 0 => 7”
00100 11-10 382
oOo01-711 000 23
Wel 4194 ~
Oro +914 +943 +2) “a ys “Ya 4
_ ye O20 + 925+ 912) “Ys 0 (©) The admittance «
a “Yn Oso + 934+ ¥23 tia) “a nett
Ne ° yu w+ Hut 350) de
The elements of [Y,,.] as obtained above are (as it should be) identical with the elements of| Vos aw Yoyo
Uc] given in Eq, (10.4). = 4.704
rt Yaa ee Maat
The new ¥,,i8
Figure 10.6 shows a 4-bus system, The shunt admittances at the buses are negligible. The line 4A
impedances are as under:
Line (bus to bus) 1-2 23 34
R(pu) 0.025 0.02 00s
X(pu) 0.10 0.08 0.20
(@) Assume that the line shown dotted (from a
bus 1 to bus 3) is not present, Formulate 3
Y. The parameters of a 4-bu
(6) Which elements of the ¥,_ obtained above Bus Code
are affected when the line from bus 1 to
bus 3 is added? (The new line has no 12
‘mutual coupling with the other lines). If 23
the pu impedance ofthis line is 0.1 +0.4, 24
find the new ¥,,. 7 34
13
Fig. 10.6 4 Bus system for Example 103 Draw the network and Solution
R
From elementary ac circuit theory we know that G =
@ y ry x
For the given values of line impedances the values of G and B are:
the th and pth buses, [Y,,] ean be written as:
Load Flow Stuilios
Line (bus to bus) 12 23 34 14
G 2.35 2.94 1.176 147
B 9. 11.76 -4.706 5.88
Recalling that the element ¥, of the [Y,.] is the sum of all admittances connected to the ith
tus andthe element ¥, equals the negative of sum of all admittances connected directly between
247
Yatra 7 “ya
Yan Inti “Yn 0
Y=
! Ul 9 “Yn Yat Ye
“ya 0 Ye Yaa tDhs
° 382-1529 -2.35+ 9.41 0 -1.AT+ 5.88
_)-2354j941 5.29-j21.17 2.944 /11.76 o
0 2.944 j11.76 4.116-f16.466 -1.176+ j4.706
a H1.AT + 5.88 0 =1.176+ j4:106 2.646 ~ 10.586
° {b) The admittance of the new line between buses 1 and 3 is 0.588 ~ j 2.352 pu. The
+3) “Thy elements of ¥,, affected by the addition of this line are Y,,, Yyy ¥, and Yy..
Ono +a +954), Yiaaoe Yay, ot Yay = 3:82 —j 15.29 + 0.588 —j 2.352
1.408 —j 17.642,
ane Yo, ay Yyy 4.116 —j 16.466 + 0.588 - J 2.352
= 4.104—j 18.818
Ysa Yas Yip "0588 4/2352 = Ty,
tical with the elements of
The new ¥,, is
es are negligible. The line AAR fI7.642 -2.35+j9A1 0588+ j2.352 147+ 5.88
La y= | 2354/94 $.29- 721.47 2.944 j11.76 0
0.04 0588+ j2.352 2.944 j11.76 4.704— 18.818 1.176 + j4.706
016 L474 5.88 0 1.176 + 74.706 2.646 ~ 10.586
a The parameters of a 4-bus system are as under:
Bus Code Line impedance Charging Admittance (pu)
(pu) ypala
12 0.2+j08 j 0.02
23 03+j09 70.03
24 025+) 1 7 0.04
4 34 0.2+j 0.8 “7 0.02
13 0.14504 70.01
system for
‘stem for Example 108 Draw the network and find bus admittance matrix 248 Power System Aasiysis ant Design
Solution i ¥,
‘The network is shown in Fig, 10.7. The line series admittances are
= 0.294 —j 1.176 pu Se N hs
1 ‘0
Yo” Tig joa ~0588-J2.352 pu yoo Se SE jo03 0.02 y
1 > 4
1 - 3 Yq
Ya 934709 ~O333—/lew Th 23 joo y
1 235 -0.94 %
Ja” asf 70235-7094 pu
Yue" Gap jog” 028-76 pu
Fig, 10.8 (a) shows the singh
‘The shunt admittances at different buses are Fig. 10.7 tums ratio transformer. Find
Vig =J 0.03 pu, Ynq = J 0.09 pu circuit for the system.
Yon 77 0.06 pH, Vag
0.06 pu Bust,
Yi2 +3 +20 in is 0 >
yn] De aby tet Io a “yu ‘
“ Yo “Yn YatYa tute I
. o Yo “Sse Yost Yas + Yoo, ,
182 3.498 2944 j1.176 -S88 + 72.352 0 —
2944 f1.176 862—j3.026 -.3334+ jl 2354 5.94
~ 588+ 72352-3334 jl 1.215— 4.468 -.294+ j1.176 %
0 —.235+j.94 294+ f1.176 .529— j2.056
rr
EEEBEIOS mete w
An incomplete nodal admittance matrix for a 4-bus system with negligible charging admittance Solution
is given below. Find the missing terms.
” T-j8 -24fl -S+j2 Na The currents m fy %, ar
Ky iy 3472-547 Which can be written from tk
% hp hs IFA
nl J
Ya Yo Yas You [ a ] -[?
Solution
I-73 =4C2+/1-5+/2+Y,)orY, By elementary matrix o Load Flow Studies 249
Vag = Agy + Yoy
S4e24j1-34j2-.5+/3)=1-/6
5+ f2.
3472
Yet hth)
$4j2-34/2-14
nalts
Hy # Yat Yo)
=4O5+j3~1+s4)= 15-77
§ ratio transformer. Find bus admittance matrix. Use this matrix to draw equivalent
for the system,
out act
3
2
=
=
s
vs YW ‘e
7
Fig. 10.8 (a)
4 + 4
ee
w [pS 1s 4
¥ z
Fig. 10.8 (b) Equivalent circuit for off-nominal ratio transformer
gible charging admittance pn
‘The currents m1 J, and voltages 1 and Y, are related by bus admittance matrix
eee ac ‘
By elementary matrix operations, the above matrix can be rewriteen as 250. Powor system Analysis and Design
[i]
‘The matrix in Eq, (ji) is the bus admittance matrix for the system of Fig. 10.8. (a)
From Eq, (if) we can draw the equivalent circuit by following the basic principles
network theory. The equivalent circuit is shown in Fig. 10.8 (6)
10.5 INCLUSION OF MUTUALLY COUPLED BRANCHES IN Y,,,.
‘Transmission lines run in different directions from various buses in a power system. In ger
they do not have mutual coupling with each other. Nevertheless one ot two lines in
total system may have some mutual coupling with another one or two lines. It is
to include the effect of such mutual coupling for correct evaluation of system perf
The easiest approach of the formation of Y,,, in a system having mutually coupled
Proper care has tobe exerci
+ or ~sign to the z, terms
when currents in both ba
dot marked terminals. Ot
negative. But the primiti
switen by inspection. The
relation with impedances :
elements of [y] do not hay
‘The primitive admittance n
by taking inverse of primi
The matrixes (z] and
is no mutual coupling. I
self admittances are equal
off-diagonal terms in (2} 2
ic)
is the use of singular transformation. Eq (10.18), i, (¥,.) = 4'y is valid in this case alsq some mutual coupled bran
However, the primitive admittance matrix should be properly formulated by inctading inspection. The primitive
‘mutual coupling. ‘Once the matrix [has be
Fig, 10.9 shows two branches having self impedances z, and z,, mutually coupled
impedance z,, KVL equations are
v,=1,2,+1,2,
712,412,
%) [22 2m |[to
. lL SIE cas
‘The matrix formed by z,, 2, and z, in Eq, (10.19) is the primitive impedance matrix forthe
system of Fig. 10.9.
Ths a-[% &]
im
Figure 10.10 (a) shows as
lines are as shown. Lines c:
bus admittance matrix,
Solution
‘The primitive impedance r
Equation (10.19) can be written as
(é)-Bs SIE]
41" bn ydbes.
where the matrix formed by y,, ,, and y, in the primitive admittance matrix
nw of %]
Ym Yb
‘The primitive admittance matrix [y] is the inverse of primitive impedance matrix [z]. Thus
o-[k s] tal’. 2]
Itis seen that the primitive impedance matrix [2] can be written from Fig. 10.9 by inspection
‘The graph is shown in I
istnk. The bus incidence 1 Load Flow Studies 251
care has to be exercised in assigning the proper
Proper | |
or~ sign to the z, terms, The z, terms are positive
vyhen currents in both branches enter or leave the
tot marked terminals. Otherwise the z, terms are
gotive. But the primitive [)] matrix’ cannot be
is sine witten by inspection, The elements of [z}have direct
rwing the basic principles if | aon with impedances z, z, and z,. However, the
) ts of [y] do not have any such direct relation.
RANCHES IN Y, imitive admittance matrix can only be obtained
7 A ~ ng inverse of primitive impedance matrix.
fae rad re ‘matrixes [2] and [)] are diagonal if there
or two lined. It is “ho mutual coupling, In this case the primitive
ation of system :
xg mutually coupled b gonal terms in [z] and (y] represent the mutual impedances. In a large network having
y Ais valid in this case mutual coupled branches, the primitive impedance matrix can be straightaway written by
formulated by including tf ippection. The primitive admittance matrix (y] is the inverse of primitive impedance matrix.
tive impedance matrix fort
elefa_ bc
ala 0 0 o
ee)
1 7
e]0 jo2 fos 0 0
ajo 0 0 jo2 0
elo 0 0 0 jos
sce matrix
eel ab e
a) 0
: ay 6/0 325 jl2s 0 0
npedance matrix (z]. Thus DI“ of} 9 nas -j25 0 0
~t_ ajo 0 0 ys 0
ze elL0 0 0 0 ~¥4
_ The graph is shown in Fig. 10.10 (b). Elements a, 5, c, dare tree branches while element e
Sink. The bus incidence matrix A is 252 Power System Analysis and Design
102
©
Fig. 10.10
element/busf 123
a 100
5b ja 10
4- 01
a oot
e 0-10
1-1-1 0 0
and m-(2 1 oo
oo1rtl.)
oo O- 1
“0 0
S875 -j3.125 71.25
yA=| 25 fl.25 = /25
0 0 ~ss
o f 0
—j4.125 j1.875
tr d=aya-| AES 17.28
AIS fl.2s
° J
j02
5888
q
3 4
jas
jas
i758
is
10.6 LOAD FLOW E
pecaling that = 71,
ingiven by:
where
‘The complex power
‘The load flow computatior
Toking the complex conju
Substituting the value of /,
; :
a4 Since
i
oA Poi
the real and im
Equations 10.24 (a) and (
on-linear and, therefore, on
Wehave two such equations
0, Vand 6, giving a total 0
Yariables at each bus so that
° ‘pecify the variables over wl
o by the devices which are cor
tbe buses can be classified it Load Flow Studies 253
10.6 LOAD FLOW EQUATIONS AND METHODS OF SOLUTION
Recalling that /,, ~ [Y,.] Figs (Bq, 10.5), the current entering the h bus of an 1 bus system
A isivenby:
: YVAN Vet ewe bY Vt ee HV, (10.208)
. - Sr, (10206)
po
se DWM, +19) (10.200)
V,=WJ 25, and ¥,=|0,2 7,
he complex power injected into ith bus is
S=P+IQ=V,0 0.21)
) load flow computations can be done more easily if J, (rather than J+) is used.
: g the complex conjugate of Eq. (10.21)
5 Si= P10) (1022s)
i ihstituting the value of J from Eq. (10.20b) into Eq. (10.228)
125 0 > -4O,=V; Lior (10.226)
ee V.= 26, V;=1"] £-6,and¥,=¥,=1% £7, wehave
js 8
is -79.
P,=10,= 10) D1% lI¥, 126, + % -5)
a
= LIF lI" |Z-(6,-6, ~%) (10.23)
ma
P= zy IY,IIV,| e08(8,- x, - 8) (10.24a)
nm
2-1" 2 IXIIF,) sin(6~ ¥, 3) (10.24)
m
ions 10.24 (a) and () are known as static load flow equations, These equations are
+1} ton-linear and, therefore, only a numerical solution is possible. For each of the n system buses
|f Wehave two such equations giving a total of 2» equations. At each bus we have 4 variables P,
|, ¥ and giving a total of 4n variables. To obtain a solution it is necessary to specify two
variables at each bus so that the number of unknowns is reduced to 2n. Evidently we should
Sih variables anc hich shave physical conta. The cole is ienced sie What
'y the devices which are connected to a particular bus. Depending on the quantities specified,
the buses can be classified into three types shown in Table 10.1. 254° Power system Analysis an Design
Table 10.1 Bus types for load flow studies
Bus type ‘Specified
quantities -
Load bus or PQ bus PQ °
Generator bus or voltage
controlled bus or PV bus PY, 5 15%
Reference bus or alack
since we have specified V, E
bus or swing bus
For bus I (slack bus) 5,
fa all buses can be found. The
fhe values of Q, at all buses
‘The load buses are the most common in the power supply network. At these buses P, and
are known because P,, 0, ate known from the load forecast data and P,Q, are either Zero
generation at these buses) or specified. At the generator buses, the governor control and exci
control allow us to specify P,, and V,, Since P,, is known (from load forecast data), P, is kn
‘The need to designate one of the buses as swing or slack bus is evident from the fact,
the system power losses are not known initially. Therefore, the net power flow into the sys
cannot be fixed in advance. The swing bus is a generator bus and the generators at this bus sup
the difference between the specified real power injected into the system at other buses and|
total system output plus losses, Since the voltages throughout the system must be close to 1
‘we generally specify the voltage at the swing bus equal to 1 pu. From elementary ac theory
‘know that any one phasor can be selected as reference and we select the voltage of the sw
bus as reference making its angle 6, zero. Generally, the bus of the largest generating station i
selected as swing bus and numbered as bus 1.
Figure 10.11 shows @ 4-bus sys
jgeas shown. The voltages at al
generated at the different buses
{py active and reactive line loss.
Solution
(@) Let bus 1 be the swing
controlled. At buses 2
reactive power compes
output is 3.5 pu but rea
The solution to the load flow problem consists in assuming a certain initial bus load) at the buses can be con
configuration, specifying the 2n known variables listed in Table 10.1 and using some numerical, 5
method to find the remaining 2n variables for the system with known [Y,,.]. The final solutia Ss,
‘must satisfy some constraints. These constraints are (1) voltage magnitude at various buss, 5
‘must be within limits (2) active and reactive generator power at different buses must be witha | Ss
the minimum and maximum limits (3) total generation must equal total load plus losses. (4) n | Since there is no active pow:
addition the system stability considerations impose a limit on the maximum value with § cao
have. The two most commonly used methods for load flow studies are Gauss-Seidel method
and Newton-Raphson method. An approximate but easy solution is possible if some simplifying
assumptions are made.
10.7 APPROXIMATE LOAD FLOW STUDY
A simple but approximate solution is possible if the following assumptions are made in losd
flow analysis.
1. Line resistances are neglected which means that the active power loss in the lines
zer0. This reduces the complexity of equations because the total active power genet
tion becomes equal to the total active power demand. The effect of this assumption i
Eqs. 9.24 (a and b) is that 7, = 90° and 7, = 90°.
2. The angle & is small so that sin 5 = 6, This approximation converts the non-linear loud
flow equations into linear ones so that analytical solution is possible.
3. All buses except the swing bus are voltage controlled buses. This means that the vol
ages of all buses are specified.
With the above assumptions, Eq. 10.24 (a and b) get modified to the following equations:
active power of all loads. Thus,
P.
Sa=2+/05 +3,
sus
Fig. cow Studies 255,
6 10.25a)
‘Approximate | q ) (10.25a)
number
Q,= WI YW, I F,leostd,- 5) + WP (10.26)
pee
bat
1p we have specified V, Eq, (10.25a) is a linear equation, The only unknowns are angles
bus 1 (slack bus) 6,=0. Thus, we have n ~ | linear equations from which the values of
‘albuses can be found, These values of when substituted in the set of Eqs. (10.25b) give
vork. At these buses P, yalues of Q, at all buses.
and P,Q, are either ero
sis evident from the fact dif igure 10.11 shows a 4-bus system. Line resistances may be neglected and pu line reactances
et power flow into the shown. The voltages at all buses should be 1 pu. The pu values of loads and active power
1 generators atthis bus ed at the different buses are also shown. Find (a) reactive power generation at buses
system at other buses and ive and reactive line losses, (c) real and reactive power transfer over the lines.
system must be close to |
{f@) Let bus I be the swing bus. The active and reactive power generation at this bus can be
controlled. At buses 2 and 4 there is no active power generation but these buses have
reactive power compensation equipment, Bus 3 has @ generator whose active power
output is 3.5 pu but reactive power output can be controlled. The generation and loads
at the buses can be combined so that the net power injected into the buses are:
S$, S8o,-S,, = Po, ~ 1-5) +4 Qo)
elect the voltage of the
ie largest generating station
ng a certain initial bus I
0.1 and using some
nown [Y,,.]. The final soh S; 12-2 + HOeq ~ 0.5)
e magnitude at various bus 5,°5q7Sa7 35+) 2m
different buses must be 5.7 5c1- Siu 2 +4 Qo, 0.8)
al total load plus losses. (4) © there is no active power loss, the net total active power generated must be equal to the
Je maximum value with 6 ath ative power of all loads. Thus, P,, can be straightaway found.
dies are Gauss-Seidel meth Po =2+2415-3.5=2 pu
Sig 24/08
/ Sia=24/08
sn converts the non-linear!
n is possible. ]
uses. This means that the vo
Sut 4/1 a) hsq5=Fo1+ i001
tothe following equations |
ing opal Fig. 10.11 4 Bus system for example 10.8 256 Power system Anatysis aiid be
Thus, the net active power injected into the buses are
~2. The admittances between the different buses are substituting [Z|~ LX,) and ¢
VAN, A
7 a
eX!
75.0 pu
4g =~ J 10.0 pu
Using the method of section 10.3, ¥,,, can be formulated. The result is
=/15.0 50 j5.0 5.0
j.0 -/1S0 00 0
5.0 10.0 ~j25.0 0.0
J5.0 0 10.0 =/15.0
Voltages at all buses is 1 pu. 5, = 0. Using Eq, (10.25a) the net active powers injected i
the buses can be written as:
cere P, isthe real power tre
onnecting the ih and pth bus
reference (swing bus voltage)
Substituting the values of
PaaPy
5 =- 58, -58, -56,
2= 56, + 10(5, - ) = 156, - 106,
5 = 58, + 10(5, 6) + 10(6,- 6)
Only three of the above four equations are independent. If we add the equations for P,, P,
and P, we get the equation for P,, Solving the equations for P,, P, and P, we get
0857 rad, =—4.91°
.071428 rad. = 4.093°
Pur-PamG
‘The reactive power transf
form as:
‘Substituting these values in Eq. (10.25b) we get 2, HP
Q, =—$ c0s (4.91%) — § cos (4.093%) — 5 cos (4,919) + 15 = 0.0494 pu Z|
'— 5 cos (4.91°) — 10 cos (9.003°) + 15 = 0.1415 pu ‘Substituting @= 90° and |:
5 cos (4.093°) ~ 10 cos (9.003°) — 10 cos (9.003%) + 25 = 0.259 pu 0- we
5 c0s (4.91°) — 10 cos (9.003%) + 15 = 0.1415 pu * |Xpl
Reactive power generation at the buses is where Q, isthe reactive pow.
the values of X and 5
Qo, =Q, + 1 = 1.0494 pu
Qn =O, + 5=0.6415 pu
Qe, = Q, = 0.259 pu
Qqy = 2, + 0.8 = 0.9415 pu
(b) Since line resistance has been neglected, active power loss in the lines is zero.
Reactive ine losses = Q, +0, + 0, + 0, = 0.5914 pu
(c) Since the line shunt capacitances have been neglected, the lines behave as short lines.
Eq. (4.78a) gives the real power transfer over a short line. We can write this equation
in general form as:
aie AW
Pym Typ eos O- Ta cos (04 5-6) Load Flow Studies 257
ne different buses ore substituting [Z| =[X,] and @= 90°
p= He! in 6—6
sin(5,—
ral 4-6)
esult is PP is the real power transferred from ith bus to pth bus, X, is the total reactance of line
0 ‘the ith and pth buses and , 6, are the bus voltage angies with respect to a common
0
00
15.0,
t active powers injected i
ubstituting the values of Xand 6
é sin (6,— 6) =5 sin (4.91°) = 0.428 pu
PyanPy=
a= Py
a sin (6, — 6) =5 sin (4.91°) = 0.428 pu
108,
6)
= 0.357 pu
° \
“ :
WP gy all”
= sin sin (0+ 6,- 6,
+15 = 0.0488 pu eon a ry Ora
5 pu ‘Substituting @= 90° and |Z] = IX)
003°) + 25 = 0.259 pu IHR IAM |
EMO cos (5,-8
5 pu 1Xpl (Xyl OD
(0, isthe reactive power flow from ith busto pth bus. Its evident that 0, =O, Substituting
¢valugs of Xand 6
02" O° 5 Joos (6-5) = 5-5 00s (4.91°) = 0.0183 pu
11
sisi been y= Oy. = 5 608 (6,— 5) = 55 08 (- 4.0937) = 0.01275 pu
1¢ lines behave as short lin y= Qy= 5-4 eos (6-4) =5=5 cos (4.91°) = 0.0183 pu
e. We can write this eq
+ £08 (6, ~ 6) = 10-10 cos (~ 9.003") = 0.123 pu
+ cos (6, 8) = 10-10 08 9.0039) = 0.128 pu 8 Power Systein sia! ign
Pg =~ Py = 0.428 means that bus 1 exports 0.428 pn real pow
428 pa real power fon: bus 1.0, G., * 10183 pu means that bus | exports 0.0183 py
reactive power (lagging) 1o bus 2. Also bus 2 exports 0.0183 pu reactive power (lagging) to by
1 or imports ~ 0.0183 pu reactive power (leading) from bus 1. Therefore, the complex powe,
export from bus 1 fo bus 2 has been shown as 0.428 + j0.0183 pu and the complex power imp,
by bus 2 from bus 1 has been shown as 0.428 ~ /0.0183 pu. Following a similar reasoning fi
other lines, the real and reactive generation, loads and line flows can be shown as in Fig, 10,12,
10 bus 2 or bus 2 impor
é Sos
=35+j0259 Ot PSAs
2108 [r2eio
2 3 4
1.565 T
4565 +j.123 | 1.965 =j129
0.428 Leyes 0.428
=/o.018 10.0183
0.357
soos |+ jo0tar5
0.428
+ jo.018s
1
+ 7.0804
Fig. 10.12 Generation, load and line flows for example 10.7
10.8 GAUSS-SEIDEL METHOD-SOLUTION OF
ALGEBRAIC EQUATIONS
A set of simultaneous non-linear algebraic equations cannot be solved by the usual methods,
For the solution of these equations iterative techniques are necessary. In an iterative technique
wwe progressively compute more accurate estimates of the unknown until results are obtained to
a desired degree of accuracy in a finite number of iterations. When this happens the solution is
said to converge. The iterative methods ate applicable to those systems of equations in which the
diagonal elements ofthe coefficient matrix are arge in comparison with the off-diagonal elements.
Consider a system of non-linear equation having x,, x, ..x, a8 m unknowns. After some
manipulations these equations can be putin the form.
KL Opp oy oe) (10.26)
where Isisn
We start with initial known (or assumed) approximate values of x's writen as x, x,
Our first approximation for x, is
x
the first approximation for x, is
EL By Xe BS oe) (10.276)
Oy Boo) (10.272)
and the first approximation fo
In general the Ath approxin
‘We examine the change in
a.
When Ax, <¢ for i=1,2,..
practical problems the numbe
update the old value by some
«zis known as acceleration fa
a
Find the value of x, accurate
Seidel method.
Solution
2x~log
o
‘We assume that x°= 1
Since the values of x* ant
continued further and 3.78:
Obtain the values of x and y
Assume
Solution Load Flow Studies 259
over © bus 2 ot bus 2 impgg | gad the frst approximation for x, is
that us ext 0018 HA flaley weekly Bax 2d) (w02te)
sactive power (lagging) tof PN eee
Sen rgercy In general the Ath approximation for x, is computed as:
and the complex power ing, af Ohxfett yh Mat xt) (10.274)
Axextaxi! (10.28)
‘When Ax, Q,n40 Set!" = Q,,.. find new value of K, and treat the ith bus as PQ bus and
continue the computations as in section (10.9.1). Similarly, if O'*' {__ Setbos count = 28 Vaan = 0
sot VI = |v, epee.)
‘rat % a8 the
phase angle of
Vi epee.
——
Fig. 10.14 (Continued on page 264) 264 Power System Analysis and Design
® solution
1 —__t__ (@ Foreach line y=
Find new vale of Find new vale of K
asd use to fnd
VE ua Ea. (8:34)
sat vi" T= spec}
‘Compute Vi Fung treat af" as te
£9. (034) phase angoofV spec
Similarly, the other €
Compute AVF using Ea (36)
ves
No ‘Set AVinox = JavA* 1]
Replace VF by VF* and advance bus count by one
(©). The power injected j
‘Compute slack bus power
Py and Q1 using £90.24 a & b)
‘and al ine flows using
Eq(0.37a 8b)
Using Eq. (10.40)
Fig. 10.14 Flow chart for oad flow studies using G-S method O5=- We
+1
eo 1021
Figure 10.15 shows a five bus power system. Each line has an impedance of 0.05 + j 0.15 pu +
The line shunt admittances may be neglected. The bus power and voltage specifications ar = 0.2448 ¢
given below: The value of O! is within
Bus PQ, Py & v Bus specification | we find V' from Eq. (10.32)
110 0.5 notapecified not epecified 1.0220 Slack bus fa
2 0 0 2 not specified 1.02 PV bus nals
3 05 on ° o not specified PQ bus Yal ¢
4 05 02 o 0 not specified PQ bus
5 05 02 0 0 not specified Pus ag
@) From ¥,, (b) Find 0,, 6, V,, V,and V, aftr the first iteration using Gauss-Seidel method.
Assume Q, = 0.2 pu and Q,,.. = 0.6 pu. Load Flow Studies 265,
) Solution 1
— (a) For each iney= 5555 55 yg ~ 2 ~/6) siemens
y+ 94g 4-12 siemens
2476 siemens
2+ j6 siemens
Similarly, the other elements of ¥,, can be calculated, The result is
4-2 24j6 0 0 24)6
24j6 6-flB -2+j6 0 -2456
0-24 j6 4-12 2+j6 0
0 0 -24j6 4-fl2 2456
c= lave}
= -24+j6 -2+j6 0 24/6 6-J18
by one (b) The power injected into the network by each bus is
5, =P,+)9, 5 .
5, =P, 430,
2 5
Also-2 + j6 = 6.32455 Z 108.43°
6—f18 = 1897366 2-71.57
Assume 2" V2=V,=120
and 8 = 5°=59= 59-0 '
1.02. Fig. 10.15 Five bus system for example 10.11
D+je 1.32455 Z 108.43°
Yyy 8.97366 Z-T1.57°
Using Eq. (10.40)
eS maior) Dy =~ Wp UMM sin 8° + Ya 829 * Wel IM sin (B+ Ye — 82)
“FY sin B+ 7, ~ 8) + in B+ 5 BY)
095+)015 =~ 1.02 {(6.32455) (1.02) sin (108.43°) + (18.97366) (1.02) sin (- 71.57%)
pedance of 0.05 + 0, :
nd voltage specifications ce “maine (1) sin (108.43*) + (6.32455) (1) sin (108.43°)]
"The value of Q! is within the limits imposed by @,,,, and O,,.,.- Using the above value of O',
_ We find V from Eq. (10.32). Putting = 2 and k= 0, this equation can be written as:
Yah —YaV ~Yos¥s
1 f2-s0.2448 4, Lc
cal Tan C2 ¥I9) (02 £0) (2 +96) (1 +70)
~E2+j9(1+j0)] 266 Power System Analysis and Design
8-18.36
= SLE = 1.0555 2 5.11", Therefore, 54 = 5.11° oF 0,0892 rad, ©
6-18 a
AAs discussed in section 10.9.2 we set [7] = [Vl and retain the phase angle 5, Thus, te .
revised value of V’; for subsequent iteration is = 804~ 18.4.
6-18
Vi = 1.02 25.11°
(4.0765 j1.714
4-fl2
= 0.98 £0.76
LAR 1 “
Me Fir} rh]
SRE EA (2+ J6)(0.98 £0.76") —(-2+ JOXLL |
= REA MN.654 0 963 21.53
a= jl2
ge ge] iO van rar Hah
is
Ws)*
p10. 10 NEWTON-R
OF ALGEB)
Consider the equation
Ifthe value of Q, ,,, is 0.3 pu in example 10.11, find the values of V,, ¥,, V, and V, after first
iteration using Gauss-Seidel method. Assume a fat voltage start at buses,
‘Solution
‘The value of Q! found in example 10.10 is 0.2448 pu. Since the minimum permissible value of
, is 0.3 pu, we fix Q, = 0.3 pu and treat bus 2 as P-Q bus with V.°=1 2.0.
1 | R~- iQ
Vis | = Ya Vi ~VagVs ~ Yas
Tal ye
Assume that xis the ini
Added to 2° gives the corre:
Therefore, fore,
Equation (10.42) can be Load Flow Studies. 267
or 0.0892 rad.
=—_ | 23 ae =e
= sal 19 OP FIOM £0)~(-24 J6N01 £0)~¢ 2+jo020)]
the phase angle 6. Thus, thy J
_ 8.04— 18.42 _ .
eis 7 1059-2 5:15
05+ j02
Ta P+ ION1.059. 25.159) (2+ OM zo|
+0420]
_ 418= 11.94 _ .
= 4 j12 120.864
1812p rr]
Z
as
2+ j6XL | (2+ JO Zoa6e)-(2+ 90.20]
aj 1 ens.
4-712
(— 3.5897 - j11.769
4= fiz
9727 L-1LA?
1 | B~jQs A
| Hh -YonVd -You¥
| Gye ah ‘|
2-155]
3 .10 NEWTON-RAPHSON METHOD-SOLUTION
fe Vy Vand V, afte: OF ALGEBRAIC EQUATIONS
t buses.
= nsider the equation
Sx) =0 (10.41)
imum permissible value ff Assume that x*is the initial estimate of the unknown x and A.x*is the correction which when
PHt£0. ded to x° gives the correct solution. In other words, x° + Ax®is the solution of Eq, (10.41).
fore, foe+ ax) =0 (00.42) *
Equation (10.42) can be expanded into Taylor's series. The result is 268 Power System Analysis and Design
Lae (AL) 6 caeey (PLY
POE + Ax) = fe) + Ax (<4) +5 ey (5) +000 (10.43)
e
The symbol { 2£ } stands for the vatue of 2) at x= x,
dx dx
A first estimate of the correction is made by truncating the series after the first two terms. Then
(10.44)
made an approximation in calculating Ax* it is necessary to make repeated use of Eq. (10.44)
till the solution converges.
The first estimate of the root is
savtavers|2oo (10.45)
‘The (K+ 1th estimate of the root is,
He.
(10.46)
‘The iterations are continued till
xioxtse
where e is the specified tolerance,
Now consider a set of two equations
Ai G%y%) =O and f(x, x5) =0 (a047)
Let-x? an x3 be the initial estimates of unknowns x, and x,..A.x°, and A x? are the corrections
which when added to x? and x*, give the correct solution. Then
AGHA, 2+ Ax) =
LOR + AX, + Ax)
Expanding Eq, (10.48) into Taylor’s series
AOR) + Ot (%J + Ax (#) +000=0
(10.48)
ax,
Sultyxy + art (# ae (2) +0000 1049)
:
sine (2) ses tee
ax,
The other partial derivativ,
sariable, we use partial deriva
Eq, (10.49) can be written as
[183]
‘The matrix containing part
written as
(283)
‘The values of Ax? and Ax’
At).
mit
Since we have used appro»
‘estimates of correction. The fi
x
x
The (k + I)th estimates of t
where
The iterative process is con
‘Now consider a system of
Let xf, x,° x, be the init
corrections. Procéeding in the Load Flow Studies 269
denotes the partial derivative
4 i" avative Lita)
(& a"
‘The other partial derivatives also have
D aatisbie, we use partial derivatives in this case. Truncating the series after the first two terms,
“9, (10.49) can be written as
fter the first two terms, a af, \°
[i 3 )+Yon pany La Pld oso
ay) (ae,
matrix containing partial derivatives is called Jacobian JP so that Eq. (10.50) can be
as
cact solution, Since we have) AGH OD] at :
repeated use of Eq (10.49) (Ke ¥) | wi [sd ] cosy
values of Ax and Ax can be found by rewriting Eq (10:51) as:
4 at]. [AGL
cong [a ic +r [ie » (10,52)
Since we have used approximation of finding A x7 and A x*,, these values are only initial
timates of correction. The fist estimates of the roots are
=stAx,
axtt Ax, (10.53)
fp Te (e+ 1h etimates ofthe rots are
xB textt Ax!
ables Axt (10.54)
at) -- [|
(S)-er [fe
The iterative process is continued till convergence is obtained.
‘Now considera system of n equations
(10.46
Alp 3) =0
LG p% ox) =O
7 (10.55)
o=0 AG Xo ¥) =O
be the initial estimates of unknowns x, x, ..%y Ax’, Ax... Ax? ar€ the
tions. Proceeding in the same manner we get an equation similar to Eq, (10.50), ie,
0=0 270 Power System Analysis and Design
ARYA) (AY
sos ax, | lan) “Las, -
lity 8 re | fas
Aes Ba) | wy(e (28) | | as
~ | Las
or Pr Par =O
The first estimate of correction is
aee-UTye
‘The first estimate of roots is
versa
The (K+ 1)th estimate of roots is
xe At
As before the iterations are continued till
wats e
Find the yalue of x in the equation 2x log x=7. Use NR method.
‘Solution
Changing the logarithm to the base e, the equation is
fla) = 2x ~ 0.43429 Inx—-7=0
af) _,_ 0.43429
dx x
Let wal
SeP=-5
of
(¥J = 2-0.43429 = 1.56571
5
Taes7i 741934304
sel
fer!) = 2(4,1934394) — 0.43429 In (4.1934394) - 7 = 0.7643149
= 1,896435847
apy a2 043429
dx
4.1934394
1643149
3 = 4,1934394 ~ aay
= 3.7904124
Se
ef
dx
___Itis seen that we have rea
+ iortions necessary in GS
~ are lengthy because calcula
(10.566)
(0.57) otve the equations using NI
i ti Sls,
0.5) tx,
‘Assume
(10.59) Solution
Se,
ele wie
a
F
ee 22
24,22
i Load Flow Studies. 271
‘fls*) = 2(3.1904124) ~ 0.43429 In (3.7904124) ~ 7 = 0.002144308
3429
Ay] fo] Sees
ax: |_lo 904124 ~ 85424
ALE on = 3.7904124- 2002144508
Ay} (0. ” 385424
It is seen that we have reached the correct estimate only in three iterations as compared to 6
(erations necessary in GS method (example 10.8). However, the calculations in each iteration
re Jengthy because calculations of and division by the derivative is required,
(LO.STIIRSoive the equations using NR method
° Sy) =x43 logx-y=0
‘gx, y) = 2 -xy-Sxt1=0
x =34andy~=22
Sey) =x+3 logx-y
= x43 x 043429 Inx—y? =x + 1.30287 Inx—y*
ar
Se 72430287
af
ay
(9) =2xt -ay—Se 41
2%
ax
2
ay
=-2y
s4x-y-5
934394) — 7 = 0.7643149
(zs) ta
ma Eel-Pee Sa] ea 272. Power System Analysis and Design
sax 4 A= 3.4+0157=3.557 (dz
yl ay + Ay? = 2,240,085 = 2.285 \ dx
yy [1367 457
YI=| 6543-3557
‘£(.557), 2.285] _ [-0.01097
(3.557), 2.285 | ~ | 0.39175 f
‘Ax'] __[l.367 4.57 T'7-0.01097] _ [-0.0685"
ay! |” [6.943 -3.587} 0.39175 0.0229,
557 — 0.0685 = 3.4885
Pay! + Ay! = 2.285 ~0.0229 = 2.2621
‘The values of Ax" and Ay! are small. The values of Ax? and A y*i'caleulated) will be found
to be still smaller. Hence, the iterative process can be stopped here. The solution is x = 3.4885,
y=2.2621.
‘Use NR method to find the value of x correct upto 4 decimal places in the equation,
3x—cosx-1=0.
‘The value of x correct to
10.11 NEWTON-R
‘Newton-Raphson (NR) me
advantages of this method
iterations are required as cc
umber of iterations is alm
factors like slack bus select
axe (1) The solution techniq:
and, thus, computer time pe
[NR method can be appli
being those using (a) rectar
10.11.1 NR Metho
In this formulation the quar
Let
The active and reactive
Solution
fle) = 3x— cos x1
dfx)
dx
=3+sinx
Let
SO) 2
Bee Grasp 9-3
0.214127
4f) <5 sen(2)=
() ->+ (3) 3.6183698
»__f@') 2 14127
© Cafldsy 6183698
fe) = 3(0.6074889) — cos (0.6074889)
For a system of n buses
relate the changes in active
= 0,6074889
1.38274 x 10° Load Flow Studies. 273
d,
( “ J =3 + sin (0.6074889) = 3.570807
2 3
pax (“} = 0.607489 — SETAE” _ 9 6071017
3.570807
dx
22) = 3(0.6071017) — 00s (0.6071017) ~ 1 = 1.853 x 107
_ [0.0685
= |-0.0229
85
21
(ifcatculated) will be found
. The solution is x= 3.4885)
(4)- 3 + sin (0.6071017) = 3.570489
SOY 1.853x107
fap¥ 06071017 a9
dx
s in the equation 14 NEWTON-RAPHSON METHOD FOR LOAD FLOW STUDIES
vege = 0.60710175
8 of this method are (1) More accuracy and surety of convergence (2) Only about 3
tions are required as compared to more than 25 or $0 as required by G-S method (3) The
et of iterations is almost independent of the system size (4) This method is insensitive to
ctors like slack bus selection, regulating transformers, etc. The disadvantages of this method
(1) The solution technique is dificult (2) The calculations in each iteration are more involved
‘computer time per iteration is large (3) The computer memory requirement is large.
[NR method can be applied to load flow problems in a number of ways, the most common
those using (a) rectangular co-ordinates (b) polar co-ordinates.
.11.1 NR Method using Rectangular Co-ordinates
th ‘this formulation the = are expressed in rectangular form,
=W125,=6,+if,
eves hia itn et
=u, f) (10.60)
Z wet)
‘or a system of m buses and bus 1 designated as slack bus, the difference equations which
ate the changes in active and reactive power to changes in e and ftake the form
(10.618)
= 0.607489
(10.616)
= 1.38274 x 10° 274 Power System Analysis and Design
AP, and AQ, represent the differences between the specified and the calculated values of p
and Q, There are two equations similar to Eq. (10.61) for each of the (m — 1) buses. In matty
form these equations can be written as
AP] _[ Jilye [ be |
AP) =| Zila |] de (uo,
[43] [% 19, |La7, 2)
Where J,,J>)Jy Jy are the elements ofthe Jacobian and are partial derivatives similar to those
appearing in Eq. (10.61). Eq. (10.62) can be used to determine the components of bus voltages
from an arbitrary set of values through an iterative procedure.
Fora PY bus the reactive power Q, is not specified but the voltage magnitude |V] is specified,
Thus,
Wet (10.63)
Therefore, the following equation replaces Eq. (10.61b) for each PY bus
awe aunt y, “
AIP = a be, + 4f, 9.64)
The total number of equations is 2 (7 ~ 1). As will be seen in the next section, the use of
polar form results in a smaller number of equations. Therefore, the NR method in rectangular
co-ordinates is rarely used. In actual practice only polar form is used.
10.11.2 NR Method using Polar Co-ordinates
‘The Newton Raphson method can be applied to the load flow studies when the bus voltages are
expressed in polar form. The active and reactive power at each bus are functions of magnitudes
and phase angles of bus voltages. Thus
PAK EIVD (10.658)
Q,=h(6. 1) (10.656)
For a system of m buses and bus 1 designated as slack bus, the equations which relate the
Sere ace ate ert Sea eae eee ey
AW, (10.668)
(10.660)
Fal
- 2a pt
AP, and AQ, represent the differences between the specified and the calculated values of P,
Equation (10.66) can be used to determine the bus voltage magnitudes and angles from
an arbitrary set of values through an iterative procedure, For each load bus both P and @ are
specified and, therefore, Eqs. (10.66 a and b) exist. However, for a PV bus, @ is not specified
and, therefore, there is no equation corresponding to Eq, (10.666) for PY bus. For an n bus
system having 1 slack bus and g voltage controlled (PY) buses, the total number of equations is
(2n-2~g), Thus, the use of the polar form results in lesser number of equations and a smaller
size of Jacobian as compared to the rectangular form. This is a definite advantage of polar form
‘over the rectangular form and, therefore, only polar form is used.
thas been found conveni
gs (10.66) become
a
‘The set of Eq. (10.67) for
AF
Ac
The partial derivatives 1,
Ithas been found that com
used. Therefore, the partial
From Eq, (10.226)
Since
also
‘Substituting these values
Ps
Differentiating partially
wi
Let wm
7 Ky
Vs
‘Then Eq, (10.71) takes t
aA 20
2, 738, Load Flow Studies 275
d the calculated values of p
‘the (n ~ 1) buses. In mati
AV
Irhas been found convenient to replace A |V| by ar in Eq, (10.66) as a result of which
gs. (10.66) become
10, e es Aly,
“4 ane SM ase Sst
cates 2B? pen Kol Al
I derivatives similar to those
components of bus voltages! * : Aly,
i 40, 52 ag t4 iw i (10.67)
xe magnitude |?) is spec pre parel'e °
“the set of Eq, (10.67) for all the n ~ 1 buses can be written inthe following matrix form
5
h PV bus AP) _[H NAIM (10.68)
sol"ly2 it .
an, aR,
B= x. N, Wy
‘the next section, the use of vO,” OP AIM! ‘
¢ NR method in rectans
ed. y=, 1-22 Wy
"3, air,"
“The partial derivatives HN, J, L are real functions of admittance matrix and the bus voltages,
‘been found that computer computations are faster if rectangular complex arithmetic is
‘Therefore, the partial derivatives are computed by rectangular complex arithmetic.
| From Eq, (10.22b)
es when the bus voltages an
are functions of magnitude
) P-i9,-V, 3 %%, (10.69)
(10.656), ie 2 vr
+ equations which relate tk V=ire%Vi= Iyer
tude and angles take the for V.=IVeMrand ¥, =f, 6%
"Substituting these values, Eq. (10.69) takes the form
| (10.665 ; ‘
P-i9,-Wie" > (rem) (ve) (10.70)
pat
l (10.660) "P" _Differentiating partially with respect to 4, p a i, we have
J the calculated values of P, BF, j 2. ejay emarjemav)e% 0.71)
magnitudes and angles from. ; 3, 3,
1 Joad bus both P and 0 até Wet =e-if
a PV bus, Q is not specified es
) for PY bus. For an n bus WiJerr = Gy +B,
‘total number of equations is We*=e,+5f,
+ of equations and a smaller ‘Then Eq. (10.71) takes the form
nite advantage of polar form 2F _ ,2Q)
€,-1) (Gy + 1B) (, +18) 276 Power System Analysis and Design
=F, -i) (a, +3) (10.7%, ag,
where a, +/b,=(G,,+/B,)(e, +f) ani!
From Eq. (10.72) the values of H,, and J,, when ia p, are .
ap, ‘The equations for calcula.
H, 35, be, (10.73) fF henp wi
22, = (a 6,4
a, @e+8,/) (10.74) N,=-J,=4,
For the case of f= p we have where 1-6, +38,
Seer 2 Vnetih
- payee ,
BaF age MY Mle «AK, Le” 4/17 Le %y LeU, le) (a, +16) =(G, +18
5” 85, ms
=~JP,-JQ) +i WP G,+5B)
From Eq. (10.75)
H,= 2h a
Eh ~ WEB,
i,
20. = p,- W796,
9,
Differentiating Eq, (10.70) with respect to |V,), p w i, we have
OF, 2 Liven
ar,i Jarl [Wie (Y,] x27” a)
Multiplying both sides by |7,]
sh Wy-s al I= IM] e** YL e™% (7%)
YI
=€,-i1) @, +7)
From Eq. (10.79) the values Of N, and L,, for p wi, are
Ia
V, +),
f am Wl= a
90 V)=a,f,-b,¢,=
ayy, toe
For the case of i=p, we have
ae a =e SF (yer He) +H] 4) (ne 0)
A
Maliping bt sides by |r]
aR, 20,
PL yj 2
ah Ml an
=P 10,490,418)
From Eq, (10.83), we a
N, wh in- P,+G,. ive
ami
=e" x dee” WV) 2%) + FF TL 2”)
(10.79)
(10.80)
(10.81)
(10.83),
oss) W git)
ive rem)
{fee ona ators 2.8
Load Flow Studies 277
(10.85)
‘The equations for calculating the elements of Jacobian are summarized below:
Hy Ly a f- 8,6,
J,=4,¢,+b, (10.86)
tiS
Form Yous using Eq, (9.18)
‘or step by stop mathod (Ea. 9.7)
t
‘Assume §P for /= 2,3 ...0
‘and VP fot = m+... 8, PO buses
——
Sat tration count k= 0
|_Setterton aunts
T
Find AP for/= 2, 9
Find AQ# for i= m+
a
t
Find max. A Pf and max. AGH
Find siack
‘bus power
line flows
Fig. 10.16 Continued on page 278 278 Power System Analysis and Design
z oN
at Ne a yes
Sots 0) wae OF 2 Q, min,
NS ve
Qimin= OP
—w
Calle He No
tor p
@Q4
Solve for
‘agang 4M
(€9.9.68) Mi
‘Advance
bus count
byone
|
Find new values of
Improve V,
of PV buses.
whore OF
crossed mit
eration
‘count
when
(10.87)
If buses i and p are not connected, ¥,= 0 and hence
N,=Ny=J,“J,=L, =1,,=0. Ths the Jacobian is as sparse as the ¥,, matrix.
The NR algorithin for lod flow studies canbe outlined asunder:
1, Assume |? and 8 at all PQ buses and 6, at all PY buses. In the absence of any other
information assume |V] at all PQ buses equal to 1 pu and 6 at all buses equal to zer0.
2. Calculate P, and Q, for all PQ buses and P, for all PV buses (except the slack bus) using
Eqs. (10.24'a and b).
3. Calculate AP, and AQ, for all PO buses and AP, for all PY buses (except the slack-bus)
by using the equations
AP,
40,
4, Caleulate the elements
5. Solve Eq, (10.68) for:
6, Update the values of
values of phase angles
7. Calculate @, for all Pr
start the next iteration.
this bus as a PQ bus, r
8. Continue till AP, and é
(or assumed) tolerance
9. Calculate line flows a
method using polar co
igure 10.17 (a) shows a 6-t
£q, (10.68) for this system.
21P0) wey)
(Siac)
(Po) sey
@
Solution
The complete set of equations
totfind a place in these equati
Jjand L, values forthe buses
‘two equations for each of these
for each of these buses. Since t
corresponding elements in the
Figure 10.18 shows a 3-bus s!
‘¥e 0.026 + j0.11 pu and j 0.0
the buses is as under: Lang Row Studies. 279
AP, =P, (specified) ~ P, (calculated)
AQ, = O, (specified) ~ Q, (calculated)
4, Calculate the elements of Jacobian using Eqs. (10.86 and 10.87)
a 5, Solve Eq, (10.68) for A Sand ALY,
2 O, ain IVI
4 Update the values of voltage magnitudes and phase angles forall PQ buses and the
‘0 values of phase angles for all PY buses.
- 4, Calculate O, for all PY buses and check Q, 4. < Q, 3 7A Be ao.)
%,
As seen in Eq. (10.86) L, = H,. Therefore,
1-H, =~ (WAI 1By| (10.92)
In the expressions for L, and H,, (Eq. 10.87) O, is generally very small as compared to |?
B,, Therefore,
L,=-WF B, (10.3)
In view of Eqs. (10.92) and (10.93), the equation for AP and AQ (Eq. 10.88 and b) can be
written as
[P] = 07118117) (43) (10.94)
(49) = (71811701 4771 (10.940)
‘The elements of matrices Band B” are the elements of [- B,] matrix (See Example 10.17).
The final algorithm is obtained by making the following approximations:
(Omit from [B” the representation of those network elements which predominantly af-
fect Mvar flow only and do not affect MW flow significantly, i., shunt rectances and
off-nominal in phase transformer taps.
i) Omit from [B”] the angle shifting effects of phase shifters.
(iii) Take the left hand ¥ terms in Eq, (10.94a and b) to the left hand side of equations and
set the other V term in Eq. (10.94a) to 1 pu.
(iv) Neglect the series resistances in calculating the elements of [8]. With these modifica-
tions Eqs. (10.94a and b) becomes
ap
= 12") [06 (10.958)
[im] =e
[42]- (B47) (10.95) Find now vues of |V
IF, fori=m
i 7 in rs
In the above equations {8”] and [B”] are real and sparse. Since (B’] and (8”] contain the
network admittance terms only, they are constant and need to be evaluated only once in the
beginning. Eqs. (10.95a and b) can be solved through an iterative procedure independently and | "™® 10.19 Flow chart fr!
altemately. Each cycle of iteration consists of one solution for [A] to update Sand one solution fears
of Vis not neet
for [/AM/] to update [|]. The iterations are continued till AP and AQ at all PQ buses and AP
for improvernet Load Flow Studies 283
(10.88)
(10.88) J
tall PY buses are within prescribed (or assumed) tolerances. Fig. 10.19 shows a flowchart for
joad flow studies using the above technique.
sumptions in deriving th
Read data,
See Fig 9.13,
a=
SotKP= 1, KQ= 1
Find APA for 7= 2...
¥
Find max. APE
1=2.n
(Eq. 884a)
r No Find tino flows
Find now valuos yy cote
Of for T= 2.
T
Mektiet
i
Flod QP2 for = m+ Aon
(PO buses) ‘STOP
i
Find AQP? for = m+ toni
Find new values of | Vil
for/=mn+1 ne
forces
vette getenty anf P1019 Bow dat S aeeaetnenes open oat
|to update Sand one solutio koe ‘I means that improvement of V is needed. KQ = 0 means that improvement
| AQ at all PQ buses and Pye not pended. 1 is iteration count for improvement of 6. k2 is iteration count
for improvement of V. 284 Power System Analysis and Design
(a The general equatic
In the 3-bus system of example 10.17, neglect the line resistances and shunt admittances. The weiss
rest of the data remains the same. a [ i
& | [-9.091
a) Form (Yj. ae,
( : 1%.
(b) Find P,P,*, 0,°
(©) Write the matrices [B’] and [B”] for the fast decoupled load flow method.
(@)_ Form the general equation for the load flow using fast decoupled method.
(©) Perform the iterations and obtain the values of P,, P, and Qy,
(® Find real and reactive line losses.
(g)_ Find real and reactive power generation at buses 1 and 2.
Solution
[2] _ fis.i82
[48 | |-9.091
1 0.073;
(@) Neglecting the line resistances, the series admittance of each line is ——— = —j 9.09 [
jo. 0.0366
Yaa Yay Ya Ya = Yn Fn = — Cs 9.091) =F 9.091
‘Neglecting the shunt admittances [2] = [18.187]
y, =2 (59.091) =—7 18.182 A
= 18.182 j9.091 9.091 yor air) ~00ss
W%.]=| 49.091 18.182 9.091 -
§9.091— j9.091 = 18.182. (©) First iteration
(b) V,=1.03 20, ie. |V,| = 1.03, 6, = 0, [V,| = 1.03
‘From Eq. (10.70) the equations for power at the 3 buses of this system can be written as
P,-JQ,=IMle* [le Ve! + [FeV Je + ¥ le |V,Je"] 0)
P,-j O,= We (Ye WV Je + [V1 eV Je + she IVJe"] ay a
P,-JQ,= (le (ye Vie + [Fa lem iV Jer + [¥le™ Wie] (ay Substituting the values in
‘Substituting the values of |V,|, |V,| and admittances in Eqs. (ii) and (iii) and separating the AB), 0.073.
real and imaginary parts Asi [c 0366
P,=-9.645 sin (6, ~ 5) - 9.364 7, sin (5, -8,) @
P,=~9.364 7, {sin (6, ~ 8) + sin (6,~ 5)] 06276.
Q, = - 9.364 |V,| [00s (6, - 65) + cos (6, - 6))] + 18.182 |7,] (i) 46! =~ 0.0345
Assume |V,¢|= 1.0, 6°= 0, §°=0. Then AIVi|=0.055 0
Ps=0,P=0,0°=-0.546 __ Second Iteration
(©) The elements of [B”} are the elements of [- B,] matrix. The terms to be included av J TH updated values of 6,
B,,, B,,, B,, and B,,, Thus 6 = S,+ AE
y= [18182 2.091 aan
I=! 9001 18.182 Wl= "3+ 4l
Substituting the above va
the matrix (B”] contains only one term, ie.,—B,,
[B”] = [18.182}
Pi = 1.51737. {@). The general equations for load flow using fast decoupled method are
AB
1¥_1 | _ [18.182 -9.091] 748,
9.091 18.182 || 485
and shunt admittances, Ty
jad flow method. An
-oupled method. ‘1 . 18.182 -9.0917"| [03
2 48; | ~ [9.091 18.182] | 4B
I¥s1
4h
et 1.03
achlin is 5 =-I90 B
V;
091 I%I
Fall = [18.182] [AIV,|]
A
A
5-O=15 pu
s system can be written a 12-0=-12pu
Y Je He!) (0.5) - (0.56) = 0.046 pu
W,le”™ We)
YJe™ WV,le*]
) and (ii) and separating th
g the values in Eqs. (vii and viii)
| _ [om ves
8} ~ [0.03665 0.0733 |] 13
yhence A 8! = 0,06276 radians
182 [VP A.6! =— 0.03459 radians
A|F}| = 0.055 (0.046) = 0.00253 pu
Second Iteration
The terms to be included he updated values of 6, 6, and | are
81 = 53+ A 5! = 0.06276 radians
81 = 5} + A 6! =— 0.03459 radians
[Vi = V3 +A |V}| = 1.00253 pu
Pi = 1.51737, P) =— 1.234, Q! =-0.404
AP! = 1.5 - 1.51737 =- 0.01737
Load Flow Studies
285
(wit)
(iii) 286 Power System Analysis and Design
Modern load flow studies tak
apd save computer time.
(ty Compact storage scheme
anditis ensured that full posi
API =~ 1.2-( 1.234) = 0.034
AQ} =—0.5 ~(-0.404) = -0.096
Substituting the values of Eqs. (vii and viii)
0.01737
[22] -(ome vo 1.03
a8} 0.03665 0.0733 |) 0.034
1.00253 Yo
whence A 83 = 6,82 x 10° radians
‘4.63 = 1,868 x 10° radians
A |V3|= 0.055 x (- 0.096) = 0.00528 pu define two veto
We ro vectors
The updated values of 6, 6, and |V,| are yoIAG
83 = 8! + A5% = 0.062768 radians YOrFD
8 = 51 + AS! =— 0.032722 radians ‘The vector YDIAG contair
[V3] = 0.99725 pu diagonal elements of ¥,.. To p
‘Substituting the values in Eqs. (iv, v, vi) JROW
2 a JCOL
; P= 1499, =~ 1199, 0) 0.497 sect The vector JROW identifie
Since the values of P3, P? and Q% are very close to the specified values, the iterative proces } gas with off diagonal elemes
need not be continued further. HM which is third element is
Hence, 6, = 0.0627668 radians = 3.598° YOFFD and so on,
‘The vector JCOL identifies
the element ~ 12 is in second
isin fourth column and so on.
(© Gauss elimination, Trie
inverse of a matrix. When a mi
Instead, it is better to use G
substitution. In triangularizatic
‘emade zero and the diagonz
that row is completed. After t
ton, We illustrate the proced
txample 10.16, This equation
0.032722 radians =~ 1.876°
99725 pu
13, 0, = 0.33, 0, = 0.37
(This has to be so since line resistances have been neglected)
Reactive line losses = Q, + , + Q,=0.2 pu
(g) Generation at the buses is
P,+P,,= 0.7 pu
83 pu
10.13 LOAD FLOW COMPUTATIONS IN LARGE SYSTEMS
(@) Sparsity: In large power systems each bus is connected to only a small number of other
bbuses (may be two or three). Therefore, bus admittance matrix of a large power system is very
sparse (i.e, it has a large number of zero elements). The sparsity feature of ¥,,, matrix extends
to Jacobian matrix also, The sparsity of an m x n matrix is defined as
pasty = Tottl mmber of zero clement
—_ “Ina large power system sparsity may be as high as 97%. It may be noted that though ¥,.i#
sparse, Z,, is full.
Divide the eleme: Load Flow Studies 287
Modern load flow studies take advantage of sparsity to reduce computer memory requirement
nd save computer time,
(0) Compact storage scheme: In.a compact storage scheme only non-zero elements are stored
adits ensured that full postion identification ean take place, Consider a sparse matrix.
si 2-12 0 8 0 0
“212 0 0 0 0
0 0 17-07 0
8 0-10 18 0 0
0 0 7 06 9
0 0 0 0-9 8
= 8,-10,-7,-9,-9]
the diagonal elements while vector YOFFD contains the off
elements of Y,, To position the elements in YOFFD, we define two more vectors
JROW =[1, 3,4, 6, 8, 10]
ICOL=[2, 4, 1, 4,5, 1,3, 3,6, 5]
-vector JROW identifies the start of different rows in the vector YOFED, e.g. first row
is with off diagonal element ~ 12 which is first element in YOFFD, second row starts with
which is third element in YOFFD, fourth row starts with — 8 which is sixth element in
D and so on.
vector JCOL identifies the column position of each element in the vector YOFFD, e.¢.,
ment — 12 is in second column, ~ 8 is in the fourth column, — 12 is in first column, ~ 10
fourth column and so on,
2) Gauss elimination, Triangular factorization: The load flow solution requires obtaining
‘of a matrix. When a matrix is sparse, the matrix inversion is a very inefficient approach.
istances have been neglects] is better to use Gauss elimination which uses triangular factorization and back
tion. In triangularization the off diagonal elements in each row below the main diagonal
zero and the diagonal element of each row is normalized as soon as the processing of
row is completed. After triangularization itis easy to obtain the solution by back substitu-
‘We illustrate the procedure by the solution of equation obtained in the first iteration of
Je 10.16, This equation is
| values, the iterative proce
18.002 -8.868 -2.035]] 43) 1437
8.868 17.736 3.948 |] Ad, | =| -1.078
GE SYSTEMS 2.096 -4.192 16.623 |} Aiv4| 0.057
ily a small number of of TAI.
al er system is
feature of Y,, max ext 1 Divide the elements of first row by a, (ie., 18.002). The result is
las 1 0.4926 -.11304][ 485 “07982
8.868 17.736 3.948] Ad, 1.078
2096 4.192 16.623 pr} | 9.087
xy be noted that though Y, Alt
Kl 288
Power System Analysis and Design
Step 2: Divide the elements of second row by 1, (ie.,~ 8.868) and subtract it from finy
row, This makes a, 2er0, The result is
1 0.4926 0.113047] 45) 07982]
0 15074 33216 |} aah -[-e
2.096 4.192 16,623 0.057
AlKL
Al
Step 3: Divide the elements of third row by a,,,(ie., 2,096) and subtract it from first roy,
‘This makes a,, zero. The result is
1 0.4926 -0.11304]] 48) 0.07982,
0 1.5074 0.33216]] Ads | =|-0.04174 j
0 1.5074 -8.04386] ajy,"'| | 0.05268,
IKI
Step4: Divide the elements of second row by the new value of a,, (ie., 1.5074). The resus
1 0.4926 -0.11304]) Ao 0.079827
0 1 022035|} a8} | = |-0.02769
0 1.5074 -8.04386}] atv, | 0.05263,
Kl
Step 5: Divide the elements of third row by new value of a;, (ie., 1.5074) and subtractit
from second row. This makes a,, zero. The result is
q
1 -0.4926 -0.11304]] 45 0.07982
oo. 0.22035] a8} | =| -0.02769
ara =G = 0626,
5.5566 |} a 17,
Al
Step 6: The triangularization is now complete. From the last equation
5.5566 [A |VJMIV,] =~ 0626
Since IJ =1.0, 47, =— 0.011266
On back substitution of this value in second equation we get, Ad! ~~ 0.0252
(On back substitution of above values in first equation we get, Ad, = 0.06613
Inthe above example we have proceeded by rows for elimination. Its also possible to process
by columns. However, the processing by rows leads to lesser computer memory requirement
and more efficient computations.
‘The use of Gauss elimination for solution is advantageous only if Jacobian is sparse (asi
‘generally is) and the rows are ordered properly
(4) Optimal ordering: To obtain full advantage of Gauss elimination, the rows of matrix
should be numbered properly. Optimal ordering leads to saving in computer memory and time.
Some schemes for optimal ordering are:
1, Number the rows sta
one having the most
2, Number the rows at ¢
one having the fewest
3. Number the rows at ¢
new non-zero terms,
‘The benefit of scheme 3 h
scaight solution is possible ang
‘The line admittances of a 4-b.
Bus code
12
1B
23
24
34
The schedule of active and
Bus code P
1 =
2 0s
3 04
4 03
Form Y,,, compute the vo
method. Table or= 1.6
Solution
The system is shown in Fig. 1
are shown, The elements of b
¥,,2¥y 7-2 458
Yya-1+j4
Yq, = 3.666 ~j 14.664
Wyn Vqt- 24/8
Y,, = 3.666 -j 14.664
¥,=3-J12
The bus admittance matt: Load Flow Studies. 289
1, Number the rows starting with that having the least non-zero terms and ending with
one having the most non-zero terms.
2. Number the rows at each step of elimination such that the row to be eliminated is the
one having the fewest non-zero terms,
3. Number the rows at each step so that the next row to be eliminated introduces fewest
‘new non-zero terms,
‘The benefit of scheme 3 has not been fully established as yet. Scheme I is used where a
368) and subtract it from fy
1
Abs 07982
48, | =|-04174
4! 0.057
Is it solution is possible and iteration isnot needed, Scheme 2is the most commonly used one,
and subtract it from first
line admittances of a 4-bus system are as under:
0.07982 an “adi
=|-0.04174 7 2-98
i] | 0.0s263 13 1/4
23 0,666 — 2.664
i 24 1-j4
fa, (ie, 1.5074). The res
‘az, (ie, 1.5074). The x se io.
0.07982" ‘schedule of active and reactive powers is
=|-0.02769 pus code P Q v Bus specification
I} Bases, 1 - - 1.06 20 slack
| 2 0s 02 not specified PO
3 04 03 not specified PO
1.5074) and subt ,
3 et S074) 4 03 ot not specified PO
Form ¥,,,, compute the voltages at buses 2, 3 and 4 at the end of frst iteration using G—S
d. Table a= 1.6
0.07982"
_ | = |-0.02769
= 0626
system is shown in Fig. 10.20. The line admittances and power infected into various buses
shown. The elements of bus admittance matrix are:
{equation 2438 Pl 2-8 tee
pt hye l+s4 eye
—j8+1-j4=3-/12
n= 0 &
1) = Yqq =~ 666 + 2.664 4a16 wy
y, je & 1-14
on Its also possible to proc a= Yael ti4 a
omputer memory requir Voy = 3.666 ~j 14.664
nly if Jacobian i ane
iy if Jeoobiin is pare (a 3.666 —} 14.664 Pet
imination, the rows of mut Wasi 290 Power System Analysis and Design
3-2-2478 “144 0
24 j8 3.666-j14.664 -666+ 2.664 1+ j4
~ | -14 j4 =6664 72.664 3.666- 714.664 2+ 78
o ~14i4 24j8 3-2
Assume V9=V,2=02=
Using Eq. (10.32) with i= 2
20
Yahi Halt rt]
« "Dar,
= _[ 5294 891.06) - (666+ 12.6641) — (14+ 740)
3.666 14.664 _ 4.1288
1 1237
oe b-. = 8:8 + 666 —/2.664 + 1-4
FET peas IS) 2+ 212/848 + 666 2.6644 1-4] aa.
1
1 (4286-15344
15931.2-7439° Find bus admittance matri
or 15931.2- 7439" 054 71
15.115 2-75.96 eer inFig, 1021.
AV, = V3—V,2=.054+ 70.029 Solutton ,
AAV, = 1.6054 +j.029) = 0864 + j 0464 ls yoda,
For calculation of V, we have
V+ VE + a(AV,) = 1.0864 +; .0464
Ve
Bry 17 |
(-1+ j4)(1.06) ~ (-.666 + 2.664) x
(1.0864 + J, 0468) (-2+ 80)
[4-j 3+ 1.06 ~/4.24 + 7235
1511527596
Figure 10.22 shows a power
+1236 +) 031 ~/2.894 +2~j8] load flow. Limit for generat
4307 /15.403 _ 1599427438" Bus1 Slack bus
15.115 2-75.96 15.115 275.96 Bus2 P¥bus I
Bus3 PO bus
058 x 1.58°= 1.058 +/ 029
AV, = Vi-Vo= 058+ j.029
G(AP,) = 1.6 (058 +/ 029) =.0928 +/ 0464 d Flow Studies 291
Vi=1+ a(Ar,) = 1.0928 + j 0464
ie 1 A
fot Yah Ya
= (14 j4)(1.0864 + j.0464)
(22+ 381.0928) + 4.0464)
3 —j 1+ 1.0864 + 1856 ~ 4.3456 + .0464
+2.1856 + 3712+ j.0928 -j8.7424]
666 — j2.664 + 1 —j4}
2
x za] | 2s
e: Es
o
Fy MW tr ystNe
Vy tyy tte =
(-.666 + j2.664) x Yg=¥qu=-Ys-% Fig. 10.21
NtVt Is +¥6 Ys ¥6
aan =e2 y=
364+ /.0464) —(-2+ /8N()] [ I-A TI +I.
igure 10.22 shows a power system. Determine load flows after first iteration using fast decoupled
load flow. Limit for generator at bus 2 is 0 521, |sin(6,-5)~10.0|¥, | sin (6,~6)
Solution is
Updated values are
o
In
Il,
Aly,
BRE
Ab
8;
Ae
ay, Load Flow Studies 293
Q, =~ 5.2 c0s (5, - 5) ~ 10| ¥, | cos (6,- 5) + 151, F
Assume
Ti [e e
ps| [10 15 }[.8,
Vs
A ase
‘A8;] _ [16.67 -10 1
48] | -10 15} | 4B
1%.
[aos on]
{0.0667 o.111 |]
IF.
90°
405)
