Chapter 1. (8 Hrs.

)
• Y bus formulation

• Modifications in Ybus to account for change in

networks.

• Power Flow Problem and Its Solution:

– The Power Flow Problem on a DC Network,
– Gauss-Seidel Method,
– Newton-Raphson Method,
– Decoupled Power Flow, DC Power Flow.
Power Flow Study (Load flow Study)
• Need to analyze performance of power systems both in normal
operating conditions and under fault (short-circuit) condition.
• Analysis in normal steady-state operation is called power-flow
study (load-flow study)
• Targets on determining:
– voltages,
– currents, and
– real and reactive power flows in a system under given load conditions.
• Purpose :
• To design a new power system
• Planning extension of existing one for increased load demand.
• To account for various hypothetical situations.
(If transmission line within the power system properly supplying loads
must be taken off line for maintenance then can the remaining lines in
the system handle required loads without exceeding their rated
parameters)
• Load-flow study analyses voltages, phase angle,
currents, and power flows in a power system
under steady-state conditions.

• In such study, assumption made are

– either a voltage at a bus or
– the power being supplied to the bus for each
bus in the power system and then determine
the magnitude and phase angles of the bus
voltages, line currents, etc.
 Applications
1. On-line analyses
– State estimation
– Security
– Economic analyses

2. Off-line analyses
– Operation analyses
– Plannig analyses
• Network expansion planning
• Power exchange planning
• Security and adecuacy analyses
– Faults
– Stability
 Steps in Load Flow solution

• Formulation of the network equations

• Solution of the equations using suitable
mathematical technique.

Network equations are simple algebraic

equations as system is under steady state .
Assumed: Loads & hence generation are fixed
at a particular value over a suitable period of
time.
 power system Bus classification
• In power system each bus or node is associated
with 4 quantities:
– Real Power (P)
– Reactive Power (Q)
– Bus Voltage magnitudes and
– Its Phase angle

• In load flow solutions two of the four quantities

are tobe specified & remaining two are to be
obtained through solution of the equations.
 Power System Bus classification
• Based on quantities specified, buses classified
as:

1. Load bus (PQ bus)

2. Generator bus or Voltage controlled bus (PV
bus)
3. Slack, Swing or Reference bus
 Load bus (PQ bus)
• At this bus quantities specified are:
– Real power P and Reactive power Q
• Quantities to find through load flow solution :
– voltage magnitude & phase angle.

• Voltage on a load bus may change with

changing loads.
• Therefore, load busses have specified values of
P and Q, while V varies with load conditions.
• Real and reactive powers supplied to power
system are defined positive,
• The powers consumed from the system are
defined negative.
• Buses having no generators are load buses.
Generator bus or Voltage controlled bus
(PV bus)
• Quantities specified are:
– Voltage magnitude corresponding to the generation voltage &
– Real power PG corresponding to its ratings

• Quantities to find through load flow solution :

– Reactive power generation QG and
– phase angle of the bus voltage.

• Voltage magnitude kept constant by adjusting field

current of a synchronous generator on bus (increasing
field current of generator increases both reactive power
supplied and the terminal voltage of the system). Cont..
• So, field current is adjusted to maintain a constant
terminal voltage VT
• Increasing prime mover’s governor set points increases
power supplied by generator to power system.
• Generators work most efficiently when running at full
load.

• So, it is desirable to keep all but one (or few) generators

running at 100% capacity, while allowing the remaining
(swing) generator to handle increases and decreases in
load demand.
• Most busses with generators will supply a fixed
amount of power and the magnitude of their
voltages will be maintained constant by field
circuits of generators.

• These busses have specific values of P and |Vi

 Slack, Swing or Reference Bus
• A special generator bus serving as the
reference bus for the power system.

• Its voltage is assumed to be fixed in both

magnitude and phase (10˚ pu).
• The real and reactive powers are uncontrolled
• The bus supplies whatever real or reactive
power is necessary to make the power flows
in the system balance.
Cont….
• The controls on the swing generator will be set
up to maintain a constant voltage and
frequency, allowing P and Q to increase or
decrease as loads change.
Constructing Ybus for power-flow analysis

• Load flow calculations can be made by using:

• Bus admittance matrix or
• Bus Impedance matrix

• Bus admittance matrix is simpler as can be generated

by inspection.
 Four Bus System
One Line Diagram
• SGi = 3-phase complex generator G4
SL4
power flowing into ith bus SG4

• SLi = 3-phase complex load power 4

flowing out of the ith bus SL1 SL3

• SGi=PGi+jQGi SG1 SG3

• SLi=PLi+jQLi G1 G3

• At each bus generator & load power

can be combined so, 1
3

• Net complex power flowing into ith

2
bus is SG2

• Si = SGi- SLi =(PGi-PLi)+j(QGi-QLi)=Pi+jQi

SL2
G2
Four Bus System
Equivalent Circuit

• Representing each G4

transmission line by π S4 I
4

circuit 4

• S1, S2, S3, S4 = 3 phase

complex power flowing S1 1 3 S3
G1 G3
into the buses I1 I3

• I1, I2, I3, I4 = currents

flowing into the buses 2

S2 I2

G2
 Four Bus System
Modified Equivalent Circuit

• Representing each G4

transmission line by π S4 I
4

circuit 4

• S1, S2, S3, S4 = 3 phase

complex power flowing S1 1 3 S3
G1 G3
into the buses I1 I3

• I1, I2, I3, I4 = currents

flowing into the buses 2

S2 I2

G2
 Four Bus System
Modified Equivalent Circuit
• All sources connected to S
1
S2
1 y12 2
common reference at ground I1 I2

potential
• Shunt admittances are lumped y14
y13 y23

• Since No. of buses=4,

• No. of nodes m=5
S4 S3
4 y34 3
y10 y20
• Yi0= admittance of ith node I4 I3

with reference= charging y40 y30

admittance
• Yip= admittance between ith & 0
pth nodes
• Vi= Voltage at ith node
 Four Bus System
Modified Equivalent Circuit
S1 S2
I1 1 y12 2 I2

y13
I1 =V1 y10 +(V1 -V2 )y12 +(V1 -V3 )y13 +(V1 -V4 )y14 y23
y14
I 2 =V2 y 20 +(V2 -V1 )y12 +(V2 -V3 )y 23
S4 S3
I3 =V3 y30 +(V3 -V1 )y13 +(V3 -V2 )y 23 +(V3 -V4 )y 34 4 y34 3
y10 y20
I4 I3
I 4 =V4 y 40 +(V4 -V1 )y14 +(V4 -V3 )y34 y40 y30

In matrix form 0

 I1   y10  y12  y13  y14  -y12 -y13 -y14   V1 

I    
 2   -y12  y 20  y12  y 23  -y 23 0   V2 
 I3   -y13 -y 23  y30  y13  y 23  y34  -y34   V3 
    
 I 4   -y14 0 -y34  y 40  y14  y34   V4 
 Four Bus System
Modified Equivalent Circuit
Where
Y11  y10  y12  y13  y14
Matrix can be written as, Y22  y 20  y12  y 23
Y33  y30  y13  y 23  y34
 I1   Y11 Y12 Y13 Y14   V1  Y44  y 40  y14  y34
I  Y Y22 Y23 Y24   V2  Y12  Y21  -y12
 2    21
 I3   Y31 Y32 Y33 Y34   V3  Y13  Y31  -y13
     Y14  Y41  -y14
 I 4   Y41 Y42 Y43 Y44   V4 
Y23  Y32  -y 23
Y34  Y43  -y34
Y24  Y42  0
Yii  Self admittance (or driving point admittance) of ith node
= Sum of the admittaneces connected to ith node
Yip  Mutual admittance (or transfer admittance) between ith & pth node
=Negative of Sum of all admittaneces connected directly
between ith & pth nodes
So,
I bus   Ybus  Vbus

For n bus system,

I busis n 1 matrix with entriy I i
Vbusis n 1 matrix with entriy V i
 Ybus  is n  n matrix with entries
yii  yii  ii  short circuit driving point admittance at ith node
yip  yip  ip  short circuit transfer admittance between ith & jth nodes
 Ybus   Bus admittance matrix
Also Vbus   Zbus  I bus
where Zbus  Y 1bus
=Bus impedance matrix
Zbus used for fault calculations using computer
 Example: A power system has 4 buses,
5 transmission lines, 1 generator, and 3
loads. Series per-unit impedances
listed below. Construct Ybus.
line Bus to Series Z Series Y (pu)
# bus (pu)
1 1-2 0.1+j0.4 0.5882-j2.3529
2 2-3 0.1+j0.5 0.3846-j1.9231
3 2-4 0.1+j0.4 0.5882-j2.3529
4 3-4 0.5+j0.2 1.734-j0.689
5 4-1 0.5+j0.2 1.734-j0.689
  1.7647  j 7.0588 0.5882  j 2.3529 0 1.1765  j 4.7059 
 0.5882  j 2.3529 1.5611  j 6.6290 0.3846  j1.9231 0.5882  j 2.3529 
Ybus  
 0 0.3846  j1.9231 1.5611  j 6.6290 1.1765  j 4.7059 
 
 1.1765  j 4.7059 0.5882  j 2.3529 1.1765  j 4.7059 2.9412  j11.7647 
• The parameters of a 4-bus system are as under

Bus Code Line Charging Admittance

Impedance (pu)
(py) Ypq/2
1-2 0.2+j0.8 j0.02
2-3 0.3+j0.9 j0.03
2-4 0.25+j1 j0.04
3-4 0.2+j0.8 j0.02
1-3 0.1+j0.4 j0.01

• Draw the network and find bus admittance

matrix
Line series admittances of the network:
1 1
y12    0.294  j1.176 pu
z12 0.2  j0.8
1 1
y13    0.588  j2.352 pu
z13 0.1  j0.4
1 1
y 23    0.333  j1 pu
z13 0.3  j0.9
1 1
y 24    0.235  j0.94 pu
z 24 0.25  j1
1 1
y34    0.294  j1.176 pu
z 34 0.25  j0.8

Shunt admittances at different buses are:

y10  j(0.02  0.01)  j0.03 pu
y 20  j(0.02  0.04  0.03)  j0.09pu
y30  j(0.01  0.02  0.03)  j0.06pu
y 40  j(0.02  0.06)  j0.06 pu
  y10  y12  y13  -y12 -y13 0 
 
 -y12  y 20 +y12  y23  y24  -y 23 -y 24 
YBUS 
 -y13 -y 23  y30  y13  y 23  y34  -y34 
 
 0 -y 24 -y34  y 40 +y 24  y34  

 0.882  j3.498 0.294  j1.176 0.588  j2.352 0 

 0.294  j1.176 0.862  j3.498 0.3 33  j1 0.235  j0.94 
= 
 0.588  j2.352 0.333  j1 1.125  j4.468 0.294  j1.176 
 
 0 0.235  j0.94 0.294  j1.176 0.529  j2.056 
 Formation of Ybus using step by
step method
• Dimension of [Ybus] is nxn.
• Consider an element having admittance yip
connected between buses I & p.
• Four entries in [Ybus] are affected: Yii, Yip, Ypi, Ypp
• Entries modified as,

Yii,new  Yii,old  yip

yip,new  yip,old - yip
Ypi,new  Ypi,old - yip
Ypp,new  Ypp,old +yip
• Figure shows one-line diagram of a simple four-bus
system. Table 1 gives line impedances identified by the
buses on which these terminate. The shunt admittance at
all the buses is assumed negligible.
(a) Find Ybus. assuming that the line shown dotted is not
connected.
(b) What modifications need to be carried out in Ybus, if
the line shown dotted is connected.

1 2

3 4
• From table 1, table 2 is obtained from which Ybus
of the system is
• The elements of Ybus modified when line is
added between buses 1 & 2 are

• So modified Ybus is,

Example: Formulate Ybus using step by step method.

Bus From Bus To R+jX Ypq/2

1 2 0.2 + j0.8 j 0.02
1 3 0.1 + j0.4 j 0.01
2 3 0.3 + j0.9 j 0.03
2 4 0.25 + j1.0 j 0.04
3 4 0.2 + j0.8 j 0.02
Equivalent Circuit for 4 bus
Line series admittances of the network:
1 1
y12    0.294  j1.176 pu
z12 0.2  j0.8
1 1
y13    0.588  j2.352 pu
z13 0.1  j0.4
1 1
y 23    0.333  j1 pu
z13 0.3  j0.9
1 1
y 24    0.235  j0.94 pu
z 24 0.25  j1
1 1
y34    0.294  j1.176 pu
z 34 0.25  j0.8

Shunt admittances at different buses are:

y10  j(0.02  0.01)  j0.03 pu
y 20  j(0.02  0.04  0.03)  j0.09pu
y  j(0.02  0.06)  j0.06 pu
1. Adding 0.294  j1.176 between node 1 & 2
 Y11,new  Y11,old  y12  0  (0.294  j1.176)  0.294  j1.176
Y12,new  Y12,old  y12  0  (0.294  j1.176)  0.294  j1.176
Y21,new  Y21,old  y12  0  (0.294  j1.176)  0.294  j1.176
Y22,new  Y22,old  y12  0  (0.294  j1.176)  0.294  j1.176

 0.294  j1.176 0.294  j1.176 0 0

 0.294  j1.176 0.294  j1.176 0 0 
Ybus,new 
 0 0 0 0
 
 0 0 0 0
2. Adding  0.588  j2.352  between node 1 & 3

 Y11,new  Y11,old  y13  (0.294  j1.176)   0.588  j2.352   0.882  j3.528

Y13,new  Y13,old  y13  0   0.588  j2.352   0.588  j2.352
Y31,new  Y31,old  y13  0   0.588  j2.352   0.588  j2.352
Y33,new  Y33,old  y13  0   0.588  j2.352   0.588  j2.352

 0.882  j3.528 0.294  j1.176 0.588  j2.352 0

 0.294  j1.176 0.294  j1.176 0 0 
Ybus,new  
 0.588  j2.352 0 0.588  j2.352 0
 
 0 0 0 0
3. Adding  0.333  j1 between node 2 & 3

 Y22,new  Y22,old  y 23  (0.294  j1.176)   0.333  j1  0.627  j2.176

Y23,new  Y23,old  y 23  0   0.333  j1  0.333  j1
Y32,new  Y32,old  y 23  0   0.333  j1  0.333  j1
Y33,new  Y33,old  y 23   0.588  j2.352    0.333  j1  0.921  j3.532

 0.882  j3.698 0.294  j1.176 0.588  j2.352 0

 0.294  j1.176 0.627  j2.176 0.333  j1 0 
Ybus,new  
 0.588  j2.352 0.333  j1 .921  j3.532 0
 
 0 0 0 0
4. Adding  0.235  j0.94  between node 2 & 4

 Y22,new  Y22,old  y 24  (0.627  j2.176)   0.235  j0.94   0.862  j3.116

Y24,new  Y24,old  y 24  0   0.235  j0.94   0.235  j0.94
Y42,new  Y42,old  y 24  0   0.333  j1  0.235  j0.94
Y44,new  Y44,old  y 24  0   0.235  j0.94   0.235  j0.94

 0.882  j3.698 0.294  j1.176 0.588  j2.352 0 

 0.294  j1.176 0.862  j3.116 0.333  j1 0.235  j0.94 
Ybus,new  
 0.588  j2.352 0.333  j1 0.921  j3.532 0 
 
 0 0.235  j0.94 0 0.235  j0.94 
5. Adding (0.294  j1.176) between node 3 & 4

 Y33,new  Y33,old  y34  (0.921  j3.532)   0.294  j1.176   1.215  j4.708

Y34,new  Y34,old  y34  0  (0.294  j1.176)  0.294  j1.176
Y43,new  Y43,old  y34  0  (0.294  j1.176)  0.294  j1.176
Y44,new  Y44,old  y34   0.235  j0.94   (0.294  j1.176)  0.529  j2.116

 0.882  j3.698 0.294  j1.176 0.588  j2.352 0 

 0.294  j1.176 0.862  j3.116  0.333  j1  0.235  j0.94 
Ybus,new  
 0.588  j2.352 0.333  j1 1.215  j4.708 0.294  j1.176 
 
 0 0.235  j0.94 0.294  j1.176 0.529  j2.116 
Finally Adding charging admittances to get bus self admittances

Shunt admittances at different buses are:

y10  j(0.02  0.01)  j0.03 pu
y 20  j(0.02  0.04  0.03)  j0.09pu
y30  j(0.01  0.02  0.03)  j0.06pu
y 40  j(0.02  0.06)  j0.06 pu
Self admittances at 4 buses are:

 Y11,new  Y11,old  y10  0.882  j3.528   j0.03  0.882  j3.498

Y22,new  Y22,old  y 20  0.862  j3.116   j0.09   0.862  j3.026
Y33,new  Y33,old  y30  1.215  j4.708   j0.06   1.215  j4.648
Y44,new  Y44,old  y 40  0.529  j2.117   j0.06   1.215  j2.057

So, new Bus admittance matrix is

 0.882  j3.498 0.294  j1.176 0.588  j2.352 0 
 0.294  j1.176 0.862  j3.062  0.333  j1  0.235  j0.94 
Ybus,new  
 0.588  j2.352 0.333  j1 1.215  j4.648 0.294  j1.176 
 
 0 0.235  j0.94 0.294  j1.176 0.529  j2.057 
Ybus Formation using Singular
Transformation
• Uses a network graph theory

• Graph: A diagram in which each element of network

is represented just by a line.
• Each source & shunt admittance across it are taken
as single element.

• Oriented Graph: If each element of graph is assigned

a direction it is an oriented graph.
• Direction assigned coincides with assumed positive
direction of current in the element.
Oriented Graph for 4-bus network

S1 S2
y12 1 9 2
I1 1 2 I2

y13 y23 8
y14 5 7

6
S4 S3
y34 4 3
4 3 2
y10 I4 I3 y20 1
4 3
y40 y30

0
0
 Cont..
• Tree of a graph: The part of
the graph which has sufficient
no. of branches to connect all
the nodes without forming a 1 2

closed path

• No. of branches to form a 4 3

tree= no. of buses 1

2

4 3

• Total no. of nodes, including

reference node = 1 + no. of 0

buses.
 Cont..
• Links: Those elements of
graphs which are not
included in the tree are 1 9 2

links
8
5 7

• Co-tree: A sub graph 6

formed by links 4 3

• No. of links l=no. of

elements-no. of tree
branches
No. of Independent variables
• No of branches=9
S1 S2
I1 1 y12 2 I2

• No. of variables=18(9 branch

y13
y14
y23
voltages & 9 branch currents )
S4 S3
4 y34 3
y10 I4 I3 • No. of independent variables
y20

y40 y30 are found from tree

• No of tree branches = no. of
0
independent voltages.
• No. of links= no. of
independent current
variables
 Primitive Network
• A set of unconnected elements is known as primitive
network
• Any branch of the network consists of passive and/or
active elements.
• Impedance and admittance representation of both active
& passive elements is as below
Impedance and admittance representation

Zip = Self impedance of element

i-p
Yip = Self admittance of element
i-p
Vip = voltage across element i-p
Eip = voltage source in series
with element i-p
Iip = Current through element i-p
Jip = Current source in parallel
with element i-p
Performance equation for impedence representation
Vip  E ip  z ip Iip -------------1

Performance equation for admittance represe

ntation
Iip  J ip  yip Vip ------------2
Impedance & admittance representations related by
From equation 1 Vip  z ip Iip  E ip and

Iip  J ip J ip
From equation 2 Vip   z ip Iip 
yip yip
J ip
So,  E ip 
yip
or J ip   yip E ip
Performance equations in admittance or impedance form
can be wtitten for all the branches.

Set of these equations in impedance form is V=E+zI

In admittance form is I+J=yV

V = Branch voltage matrix

E = Source voltage matrix
I = Branch current matrix
J = Branch current matrix
z= primitive impedance matrix
(a matrix whose elements are branch self impedances)
y= primitive admittance matrix
(a matrix whose elements are branch self admittance)
As no mutual coupling is assumed between lines,
 Vb1  V1
Bus Incidence matrix Vb2  V2
Vb3  V3
• In the graph 9-branch
Vb4  V4
voltages (Vb1,Vb2….Vb9) can
be expressed in terms of 4- Vb5  V1  V4
bus voltages (V1----V4) as Vb6  V4  V3
Vb7  V2  V3
1 9 2
Vb8  V1  V3
5
8
7 Vb9  V1  V2
6
4 3
2
1
4 3
In matrix form
0 V=AVbus
A  bus incidence matrix
 1 2 3 4
1 1 0 0 0   Buses 

2 0 1 0 0   Tree  
branches  
3 0 0 1 0   
   
4 0 0 0 1   Ab 
A  5 1 0 0 1   
   
6 0 0 1 1   
Link  Al 
7 0 1 1 0 
  Branches  
8 1 0 1 0   
9 1 1 0 0   
 
The element a ip if matrix is written as :
a ip  1 if ith element is incident to and oriented away from pth bus
a ip  1 if ith element is incident to and oriented towards pth bus
a ip  0 if ith element is not incident to the pth bus
Bus Incidence matrix
• Bus incidence matrix is rectangular and
therefore singular
• Its dimensions are e x n where
• E=no. of elements or branches
• N=no. of buses excluding reference)
• If rows of Bus incidence matrix are arranged
according to a particular tree, the matrix can
be partitioned into two sub matrices
– Ab i\of dimension b x n , b=no. of tree branches
– Al i\of dimension l x n , l=no. of link branches
Formulation of Ybus and Zbus
• Bus incidence matrix is rectangular and
therefore singular
• Its dimensions are e x n where
• E=no. of elements or branches
• N=no. of buses excluding reference)
• If rows of Bus incidence matrix are arranged
according to a particular tree, the matrix can
be partitioned into two sub matrices
– Ab is of dimension b x n , b=no. of tree branches
– Al is of dimension l x n , l=no. of link branches
Substituting V=AVbus in eq. I+J=yV

 I + J = y AVbus

Premultiplying by A t
A t I +A t J = A t y AVbus

A t I is a phasor in which each element is

the algebric sum of currents through the
network elements terminating at a bus.
As per KCL, algebric sum of currents at a bus =0
So A t I  0
A t J gives algebric sum of source current at each bus
and is designated as I bus
So, A t J  I bus

so, eq. A t I +A t J = A t y AVbus becomes

I bus  A t
y AVbus

Comparing with I bus   Ybus  Vbus

  Ybus   A t y A

Since matrix A is singular,

A t y A is a singular transformation of y.
Formulate Ybus using singular transformation
S1 S2
I1 1 y12 2 I2

y13 y23
y14

S4 S3
4 y34 3
y10 y20
I4 I3

y40 y30

0
1. formation of Primitive admittance matrix
 y10 0
 y 20 
 
 y30 
 
 y 40 
y=  y14 
 
 y34 
 y 23 
 
 y13 
0 y12 

2. Formation of Bus Incidence matrix 1 2 3 4
1 1 0 0 0 

2 0 1 0 0 
1 9 2

3 0 0 1 0 
5
8
7  
6
4 0 0 0 1 
4 3

A  5 1 0 0 1
2
1

 
4 3

6 0 0 1 1 
7 0 1 1 0 
0

 
8 1 0 1 0 
 
9 1 1 0 0 
  y10 0 0 0 
0 y 20 0 0 

0 0 y30 0 
 
0 0 0 y 40 
yA   y14 0 0  y14 
 
0 0  y34 y34 
0 y 23  y 23 0 
 
 y13 0  y13 0 
y  y12 
 12 0 0 
 1 0 0 1 0 1
0 0 1
0 1 0 0 0 1
0 1 0
A 
t

0 0 1 0 0 1 1 1 0 
 
0 0 0 1 1 1 0 0 0 

 Ybus   A t yA

 y10  y12  y13  y14  -y12 -y13 -y14 

 
-y12  y 20  y12  y 23  -y 23 0
= 
 -y13 -y 23  y30  y13  y 23  y34  -y34 
 
 -y14 0 -y34  40 14 34  
y  y  y
Formulation of Ybus and Zbus
including mutually coupled branches
• Some transmission lines
may have mutual
coupling with each Va
other
Ia Za
Zm
• So, necessary to include
Ib
effect of it for correct Zb
evaluation of system Vb
performance.
 Va va  Ia z a  I b z m
Ia Za v b  Ia z m  I b z b
Zm or
Ib
 va   z a z m   Ia 
 v   z
Zb
Vb  I 
 b  m z b  b

Matrix formed by z a , z b and z m is primitive impedance matrix

 za zm 
z  z z 
 m b 

can be written as
 Ia   ya y m   va 
I    y  v 
 b  m y b  b

 ya ym 
 y   y y   primitive admittance matrix  y
 m b 

is inverse of primitive impedence matrix  z 

1
 za zm  1  zb z m 
 y  z    z z a 
 m zb  za zb  zm2  m
Example
• Figure shows a system with one source and four
lines. The reactances of source and lines are shown.
Lines c and b are mutually coupled through a
reactance of j0.2 pu. Formulate bus admittance
matrix. j0.2
d
3 4

c
j0.5 j0.25
j0.2
b

1 j0.4 2

a j1

0
 e / e a b c d e 
a  j1 0 0 
j0.2

3
d
4
0 0
b  0 j0.4 j0.2 0 0 
z  
c
j0.5
j0.2
j0.25

b
c  0 j0.2 j0.5 0 0 
j0.4

d 0 0 
1 2
0 0 j0.2
a j1  
e 0 0 0 0 j0.25
0

e/e a b c d e 
a   j1 0 0 0 0 
1 b  0  j3.125 j1.25 0 0 
 y   z  
c  0 j1.25  j2.5 0 0 
d  0 0 0  j5 0 
 
e  0 0 0 0  j4 
In a graph Elements a, b, c, d are tree branches and e is link

Bus incidence matrix A is

elements / bus  1 2 3 4
1 0  Va  V1 d
a  0 0 3 4
Vb  V2  V1
b  1 1 0 0
A    Vc  V3  V1 c e
c  1 0 1 0
Vd  V3  V4
d 0 0 1 1 1
b
  Ve  V4  V2 2
e  0 1 0 1
a

In matrix form 0

V=AVbus
A  bus incidence matrix

1 1 1 0 0 
0 1 0 0 1
 A    
t

0 0 1 1 0
 
 0 0 0 1 1 
   j1 0 0 0 
 j1.875  j3.125 j1.25 0 
 
yA   j1.25 j1.25  j2.5 0 
 
 0 0  j5 j5 
 0 j4 0  j4 

  j4.125 j1.875 j1.25 0 

 j1.875  j7.125 j1.25 j4 
 bus  yA  
Y  A t 
j1.25 j1.25  j7.5 j5 
 
 0 j4 j5  j9 
Load Flow problem
I bus   Ybus  Vbus

Current entering the ith bus of n bus system is

Ii  Yi1V1  Yi2 V2  .....  Yii Vi  ....  Yin Vn
n
=  Yip Vp
p 1
n
=  Yip Vp ( p  ip )
p 1

where, Vp  Vp  p and Yip  Yip  ip

The complex power injected into ith bus is Si  Pi  jQi  Vi Ii

Vi  Voltage at i th bus with respect to ground
Ii  Source current injected into the bus.

The load flow computations can be done more easily

if Ii (rather than Ii ) is used.
Taking the complex conjugate
Si  Pi  jQi  ViIi i=1,2,.....n
Substituting value of Ii
n

S  Pi  jQi  V
i i

Y
p 1
ip Vp

Since Vi  Vi  i , Vi  Vi    i and Yip  Yip  ip

n
Pi  jQi  Vi Yp 1
ip Vp ( p  ip   i )

n
 Vi Y
p 1
ip Vp   ( i  p  ip )

Seperating real and imaginary parts

n
Pi  Vi 
p 1
Yip Vp cos( i  p  ip )
n
Qi  Vi 
p 1
Yip Vp  sin( i  p  ip )

Static Load Flow Equations

• For n bus system 2n equations (n real power flow
equations and n reactive power flow equations)
• Each bus is characterizedb by 4 variables; Pi, Qi, lVil and
δi.
• So resulting in total 4n variables
• Equations can be solved for 2n variables if remaining 2n
variables are specified.
• Specify variables over which we have physical control.
• Solution for the remaining 2n bus variables is difficult as
equations are non-linear algebraic equations (as bus
voltages are involved in product form and sine and cosine
terms are present)
• Therefore, explicit solution is not possible
• Solutionc an only be obtained by iterative numerical
techniques.
Approximate Load Flow Solution
(A simple and approximate solution is possible if
following assumptions made in load flow analysis)
• Line resistances being small are neglected. Shunt
conductances of overhead lines is always negligible
– so active power loss PL of the system=0.
– This reduces equation complexity as,
active power generation=total active power demand.
– So, in load flow equations, γip=90° and γii=-90°
• Angle δi is small.
– so sin δi = δi .
– This approximation converts non-linear load flow equation
into linear so analytical solution is possible.
• All buses except swing bus are voltage controlled
buses.
– i.e. voltage of all buses are specified.
Modification in equations with assumptions
n
Pi  Vi Y
p 1
ip Vp ( i  p )            1

n
Qi   Vi  Yip Vp cos( i  p )  Vi Yii    2
2

p 1
p 1

Eq. 1 is linear as Vi is specified

Only unknowns are angles 
For bus1 (slack bus)  i  0
so we have (n-1) equations from which  at all buses can be found.

These  values when submitted in eq. 2 gives Qi at all buses.

 Types of Buses
Bus type Specified quantities Unknown quantities Approximate number

Load bus or PQ bus Pi Qi Vi δi 85%

Generator bus or voltage Pi Vi Qi δi 15%

Controlled bus or PV bus

Reference bus or slack bus Vi Pi Qi 1

or swing bus δi=0

1. The load buses are most common in power system, at these buses Pi and Qi are
known because Pli and Qli are known from load forecast data and PGi,QGi are either
zero or specified
2. At the generator buses, the governor control and excitation control allows us to
specify PGi and Vi. Since Pli is known (from load forecast data) Pi is known.
3. The need to designate one of the buses as swing or slack bus is evident from the fact
that the system power losses are not known initially
4. Therefore net power flow into the system cannot be fixed in advance.
5. The swing bus is a generator bus and the generators at this bus supply the difference
between the specified real power injected into the system at other buses and the total
system output plus losses
6. Since voltage through out the system must be close to 1 pu,specify voltage at swing
bus 1pu and δi=0
7. Bus of the largest generating station selected as swing bus and numbered as bus no 1
Assumptions made have decoupled 2 equations
so no need to solve simeltaneously but can be solved
sequencially.

Since the solution is non iterative and

the dimensions are reduced to n-1 from 2n,
it is computationally highly economical.
 Example • Figure shows a 4 bus system.
Line resistances may be
neglected and pu line
reactances are as shown. The
SG2=0+jQG2 SG3=3.5+jQG3 SG4=0+jQG4
voltage at all buses should be
SL2=2+j0.5 2
j0.1
3
j0.1
4 SL4=2+j0.8
1 pu. The pu values of loads
and active power generated
j0.2 j0.2 j0.2
at the different buses are also
shown. Find:
1 a. Reactive power generation
SL1=1.5+j1 SG1=PG1+jQG1 at buses
b. Active and reactive losses
c. Real and reactive power
transfer over the lines.
• Let bus 1 be the swing bus.

• Active and reactive powers at

this bus can be controlled.
SG2=0+jQG2 SG3=3.5+jQG3 SG4=0+jQG4

• At buses 2 and 4 there is no SL2=2+j0.5 2 3 4 SL4=2+j0.8

active power generation but j0.1 j0.1

have reactive power j0.2 j0.2 j0.2

compensation equipment.
1

• Bus 3 has a generator whose SL1=1.5+j1 SG1=PG1+jQG1

active power output is 3.5 pu

but reactive power output can
be controlled.
Genaration and loads at the buses can be combined,
So net power injected into the buses are,

S1  SG1  SL1  (PG1  1.5)  j(QG1  1)

S2  SG 2  SL2  2  j(QG 2  0.5)
S3  SG3  SL3  3.5  jQG3
S4  SG 4  SL4  2  j(QG 4  0.8)

With no active power loss,

net active power generated = active powers of all loads
i.e. PG1  3.5  2  2  1.5
So, PG1  2  2  1.5  3.5
 2pu
Thus, net active power injected into the buses are,
P1  (2  1.5)  0.5
P2  (0  2)  2 SG2=0+jQG2 SG3=3.5+jQG3 SG4=0+jQG4

P3  (3.5  0)  3.5 SL2=2+j0.5 2 3 4 SL4=2+j0.8

P4  (0  2)  2
j0.1 j0.1

j0.2 j0.2 j0.2

Admittances between different buses

y12  y13  y14   j5.0pu SL1=1.5+j1
1
SG1=PG1+jQG1

y 23  y34   j10.0pu

  j15.0 j5.0 j5.0 j5.0 

 j5.0  j15.0 j10.0 0 
Ybus  
 j5.0 j10.0  j25.0 j10.0 
 
 j5.0 0 j10.0  j15.0 
   j15.0 j5.0 j5.0 j5.0 
 j5.0  j15.0 j10.0 0 
Ybus  
 j5.0 j10.0  j25.0 j10.0 
Voltages at all buses = 1pu  
 j5.0 0 j10.0  j15.0 
1  0
n
Using equation Pi  Vi Y
p 1
ip Vp ( i  p )

net active power injected into the buses can be written as,

P1  0.5  5 2  5 3  5 4
P2  2  5 2  10( 2   3 )  15 2  10 3
P3  3.5  5 3  10( 3   2 )  10( 3   4 )
 10 2  25 3  10 4
P4  2  5 4  10( 4   3 )  10 3  15 4
Solving equations for P2 , P3 and P4

 2   4  0.0857rad  4.91
 3  0.071428rad  4.093
Substituting these values in
n
Qi   Vi Y Vp cos( i  p )  Vi Yii we get
2
ip
p 1
p i

Q1  5cos(4.91)  5cos(4.093)  5cos(4.91)  15  0.0494 pu

Q 2  5cos(4.91)  10 cos(9.003)  15  0.1415pu
Q3  5cos(4.093)  10 cos(9.003)  10 cos(9.003)  25  0.259pu
Q 4  5cos(4.91)  10 cos(9.003)  15  0.1415pu
 SG2=0+jQG2 SG3=3.5+jQG3 SG4=0+jQG4

SL2=2+j0.5 2 3 4 SL4=2+j0.8
j0.1 j0.1

Reactive power generation at buses is j0.2 j0.2 j0.2

QG1  Q1  1  1.0494pu SL1=1.5+j1

1
SG1=PG1+jQG1

QG2  Q 2  0.5  0.6415pu

QG3  Q3  0.259pu
QG4  Q 4  0.8  0.9415pu

b. Since line resistances neglected, active power loss in line=0

Reactive line losses=Q1  Q 2  Q3  Q 4  0.5941pu
c. Real and Reactive power transfer over the lines
Since line shunt capacitances neglected, lines behaves as short lines.
Equation for Real power transfer over a short lines is

2
Vi Vi Vp
P ip  cos   cos(   i   p )
Z Z
Substituting Z  X ip and   90

Vi Vp
P ip  sin( i   p )
Z

P ip is power tramsferred from ith bus to pth bus

Xip is total reactance of line connecting ith and pth buses
δi and δ p are bus voltage angles with respect to a common
reference (swing bus voltage)
Substituting values of X and 

1
P 12  P 21  sin(1   2 )  5sin(4.91)  0.428pu
0.2
1
P 14  P 41  sin(1   4 )  5sin(4.91)  0.428pu
0.2
1
P 13  P 31  sin(1   3 )  5sin(4.093)  0.357pu
0.2
1
P 23  P 32  sin( 2   3 )  10sin(9.003)  1.565pu
0.1
1
P 34  P 43  sin( 3   4 )  10sin(9.003)  1.565pu
0.1
Reactive power transfer over a short lines is given by

2
Vi Vi Vp
Q ip  sin   cos(   i   p )
Z Z
Substituting Z  X ip and   90

2
Vi Vi Vp
Q ip   cos( i   p )
Z Z

Q ip is reavtive power flow from ith bus to pth bus

substituting values of X and 
 1 1
Q 12  Q 21   cos(1   2 )  5  5cos(4.91)  0.0183pu
0.2 0.2
1 1
Q 13  Q 31   cos(1   3 )  5  5cos( 4.093)  0.01275pu
0.2 0.2
1 1
Q 14  Q 41   cos(1   4 )  5  5cos(4.91)  0.0183pu
0.2 0.2
1 1
Q 23  Q 32   cos( 2   3 )  10  10 cos( 9.003)  0.123pu
0.1 0.1
1 1
Q 34  Q 43   cos( 3   4 )  10  10 cos(9.003)  0.123pu
0.1 0.1
Gauss-Seidel method for load flow studies

• One of the common methods used in load flow studies.

• Advantages:
– Simplicity of the technique
– Small computer memory requirement
– Less computational time per iteration

• Disadvantages:
– Slow rate of convergence, so large no. of iterations
– Increase in no. of iterations with increase in no. of buses
– Effect on convergence due to choice of slack bus

• So used only for system having small no. of buses.

G-S method when PV Buses are Absent

• For n bus system,

– one bus is slack bus
– n-1 buses are P-Q buses

• Initially assume voltage magnitude and angles at n-1 buses

• Update these voltages at every step of iteration.

Power Flow Equation

• Consider a typical bus of a power Vi

V1
yi1
system network
V2
Ii yi2

• Transmission lines are

yin Vn
represented by their equivalent
π model. yi0

• Impedances converted to pu
admittances on common MVA A typical bus of the power system

base.
 Vi

Power Flow Equation yi1

V1

V2
Ii yi2

Current entering the ith bus of n bus system is

Ii  yi0 Vi  yi1  Vi  V1   yi2  Vi  V2   ...   yin  Vi  Vn  yin Vn

=  yi0  yi1  yi2  ...  yin  Vi  yi1V1  yi2 V2  ....yin Vn yi0

or A typical bus of the power system

n n
Ii =Vi  yip   yip Vp
p 0 p 1
p 1

Real and reactive power at bus i is,

Pi  jQi  Vi I*i
Pi  jQi n n
or *
 Vi  yip   yip Vp
Pi  jQi
Vi p 0 p 1
Ii  p i
Vi*
Power Flow Equation
Pi  jQi n n

*
 Vi  yip   yip Vp
Vi p 0 p 1
p i

Rearranging equation,

Pi  jQi n

*
  yip Vp
Vi p 1
p i
Vi  n

y
p 0
ip
• Assume initial values of V2,V3…Vn.

• As voltages at different buses are close to 1 pu initially

assume V2,V3…Vn = 1+j0

• At every step of iteration most updated values of bus

voltages are used to compute new voltages.

• When (k+1)th iteration is performed to calculate Vi, the

most updated values of the bus voltage are the values
obtained in (k+1)th iteration for all buses previous to
ith bus (i.e. for p=1,2…i-1) and the values obtained in
kth iteration for ith bus and the buses after ith bus (i.e.
for p=i,i+1…n)
Finally calculate load flows.
From the circuit, current Iip and complex power Sip can be obtained as
Iip  Iip1  Iip0   Vi  Vp  yip  Vi yip0
Iip  I pi
Sip  Pip  jQip  Vi Iip  Vi Iipo

Similarly, power fed by pth bus into line connecting pth and ith bus is
Spi  Vp Ipi  Vp Ipio

Thus power flow over all lines can be computed.

Bus i Bus p
Vi Iip Iipi yip Ipi1 Ipi

Sip Spi
Iipo Ipio
yipo ypio
• Power loss in the line connecting ith and pth
bus is given by Sip+Spi.

• Total transmission loss= sum of losses over all

the lines.
Algorithm for G-S method with PV bus absent
1. Re ad system data and formulate YBus
2. Asume initial bus Voltage Vi0  1  j0, for p = 2 , 3,..., n
 1, slack bus
3. Set iteration count k=0 and Vmax  
4. Compute K i & Lip
5. Set bus count p=1
6. If bus is PQ bus, then
i. Compute Vik 1
ii. Compute Vik = Vik 1 -Vik
If Vik  Vmax
iii. Assign new voltage to old
i.e. Vik = Vik 1
otherwise (slack bus) go to next bus
7. Increment the bus count, i=i+1

8. Check all buses are taken into account. If i  1,

then go to step 6 and repeat.

9. Check convergence
If Vmax   , then increment iteration count
k=k+1, go to step 5 repeat.

10. Calculate line power flows and slack bus power.

 Start

Read: primitive y matrix, bus incidence

matrix, Slack bus voltage V1, real bus
power Pi, for i=2,3,...n, reactive bus
power Qi for i=m+1..n(PQ buses), Bus
voltage magnitude Vi for i=2,3..m (PV
buses), reactive power limits Qi,max. &
Qi,min for i=2,...3m

Form Ybus using any method

Assume initial values of

Vi &  i for i=2,3..n

Compute K i for i=m+1...n & Lip

for i=2,3..n, p = 1,2..n and i  p

Set bus count i=1

if
Yes
slack bus
i=1
No
Ki i1 n
Compute V
i
k+1
from V i
k 1
 k   Lip Vp   Lip Vpk
k 1

Vi p1 p i 1

Compute Vik 1  Vik 1  Vi k

Example
• Figure shows the one line diagram of a simple three
bus power system with generation at bus 1. The
magnitude of voltage at bus 1 is adjusted to 1.05 p.u.
The scheduled loads at buses 2 and 3 are as marked on
the diagram. Line impedances are marked in per unit
on a 100 MVA base and the line charging susceptances
are neglected.
1 2
0.02+j0.04
256.6
MW

0.01+j0.03 0.0125+j0.025 110.2

MVar
3
Slack Bus
V1=1.05<0
138.6
110.2
MW
MVar
Example cont..

1. Using Gauss-Seidal method, determine the

phasor values of the voltage at the load
buses 2 and 3 (P-Q buses) accurate to 4
decimal places.

2. Find the slack bus real and reactive power

3. Determine the line flows and line losses.

Construct a power flow diagram showing
the direction of line flow.
Step 1: Find Line admittances

1
y12   10  j20
0.02  j0.04

1
y13   10  j30
0.01  j0.03

1
y 23   16  j32
0.0125  j0.025
Step 2: Find complex power at bus 2 and 3
Complex power at bus 2 & 3,

S2  SG 2  SL2  (PG 2  PL2 )  j(QG 2  PL2 )

S3  SG3  SL3  (PG3  PL3 )  j(QG3  PL3 )

256.6  j110.2
S2    2.566  j1.102 pu
100

138.6  j45.2
S3    1.386  j0.452 pu
100
Step 3: Assuming initial values of voltages at bus 2
and 3
Bus 1 taken as slack bus.
So, start with initial estimate of V20 and V30
V20  1.0  j0.0 and
V30  1.0  j0.0
Voltage is obtained by
Pi  jQi n n

*
 Vi  yip   yip Vp
Vi p 0 p 1
p i

Rearranging equation,
Pi  jQi n
*
  Yip Vp
Vi p 1
p i
Vi  n

y
p0
ip
Step 4: Ist iteration to find voltages at bus 2
P2  jQ 2 n

*0
  yip Vp
V2 p 1
p 1
V 
1
2 n

y
p 0
ip

P2  jQ 2
0  y11V1  y12 V1  y 23V30
V2*
=
y12  y 23

2.566  j1.102
 10  j20 1.05  j0   16  j32 1.0  j0 
1.0  j0
=
26  j52

V21 = 0.9825-j0.0310
Step 4: Ist iteration to find reactive power at bus 3
P3  jQ3 n

*0
  yip Vp
V3 p 1
p 1
V 
1
3 n

y
p 0
ip

P3  jQ3
0  y13V1  y 23V21
V3*
=
y13  y 23

1.386  j0.452
 10  j30 1.05  j0   16  j32  0.9825  j0.0310 
1.0  j0
=
26  j62

V31 = 1.0011-j0.0353
Step 5: 2nd iteration to find voltages at bus 2 and 3

2.556  j1.102
 10  j20 1.05  j0   16  j32 1.0011-j0.0353
0.9825  j0.0310
V2 
2

26  j52

= 0.9816-j0.0520

1.386  j0.452
 10  j30 1.05  j0   16  j32  0.9816-j0.0520 
1.0011-j0.0353
V3 
2

26  j62

=1.0008-j0.0459
Process is continued and solution is converged

V  0.9808-j0.0578
3
2 V  1.0004-j0.0488
3
3

V24  0.9803-j0.0594 V34  1.0002-j0.0497

V  0.9801-j0.0598
5
2 V  1.0001-j0.0499
5
3

V26  0.9801-j0.0599 V36  1.0000-j0.0500

V27  0.9800-j0.0600 V37  1.0000-j0.0500

Final solution is
V2  0.9800-j0.0600=0.98183-3.5035 pu
V3  1.0000-j0.0500=1.00125-2.8624 pu
b. Slack bus real and reactive power

Pi  jQi n n

*
 Vi  yip   yip Vp pi
Vi p 0 p 1

P1  jQ1  V1  V1  y12  y13    y12 V2  y13V3  

=1.05 1.05  20  j50   10  j20    0.98  j0.06   10  j30 1.0  j0.05  

= 4.095-j1.890

So, Slack bus real power=P1  4.095 pu = 409.5MW

Slack bus real power=P2  1.890 pu = 189 Mvar.
c. Line flows
To find the line flows, first find line currents.

Neglecting charging capacitances

I12  y12  V1  V2   10  j20  1.05  j0    0.98  j0.06    1.9  j0.8

I 21 = -I12  1.9  j0.8

I13  y13  V1  V3   10  j30  1.05  j0   1.0  j0.05    2.0  j1.0

I31 = -I13  2.0  j1.0

I 23  y 23  V2  V3   16  j32   0.98  j0.06   1.0  j0.05    0.64  j0.48

I32 = -I 23  0.64  j0.48
Line flows
The line flows are

S12  V1I12  1.05  j0 1.9  j0.8   1.995  j0.84 pu
 199.5 MW  j 84 Mvar

S21  V2 I21   0.98  j0.06  1.9  j0.8   1.91  j0.67 pu

 191.0 MW  j 67 Mvar


S13  V1I13  1.05  j0  2.0  j1.0   2.1  j1.05 pu
 210.0 MW  j 105 Mvar
Line Losses

S31  V3I31  1.0  j0.05  2.0  j1.0   2.05  j0.9 pu

 205.0 MW  j 90 Mvar

and Line losses

S23  V I   0.98  j0.06  0.656  j0.48   0.656  j0.432 pu

2 23

 65.6 MW  j 43.2 Mvar

S32  V3I32  1.0  j0.05  0.64  j0.48   0.664  j0.448 pu

 66.4 MW  j 44.8 Mvar
Line Flows

SL12  S12  S21  8.5 MW  j 17 Mvar

and Line losses

SL13 =S13 + S31  5 MW + j 15 Mvar

SL23 =S23 + S32  0.8 MW + j 1.6 Mvar

Power Flow Diagram
Real power direction indicated by

Reactive power direction indicated by

Values within parentheses are real

and reactive losses
1 2
199.5 191
(8.5)
409.5 (17)
84.0 67 256.6

210 (5) 205 66.4 65.6

(0.8)
189
(15) (1.6) 110.2
90 44.8
105 43.2
3

138.8 45.2
G-S method when PV Buses are Present

• For n bus system, some buses are voltage controlled

buses where
– P and |V|specified and
– Q and δ are unknowns.

• Let buses numbered as,

• i=1, slack bus

• i=2,3,…,m PV buses
• i=m+1,m+2,…,n PQ buses
G-S method when PV (voltage controlled)
Buses are Present
• For voltage controlled buses :
– bus voltage=|Vi|specified.
– Maximum and minimum reactive powers also specified and
Qi (i=2,3,…m) must lie between these limits.

• Thus conditions to be met are

1. |Vi|=|Vi|specified. For i=2,3,…,m
2. Qi,min<Qi<Qi,max. for i=2,3,…,m

• Second condition is violated if |Vi|specified is either

too low or too high
• It is possible to control |Vi| only by controlling Qi
Cont…
• For PQ bus voltage calculations are usual.
• For PV bus we need to maintain voltage magnitude at
specified value.
• Voltage at PV bus is controlled by controlling Qi.
• So, Qi and δi are to be updated in every iteration.
• Found Vi(k+1) using Pi  jQi n
*k
  Yip Vp
k

Vi p 1
p i
Vik 1 
Yii
• Qi obtained by
  
 k V k Y  Y V k   p  i
n

 i ii 
k 1
Qi   Im Vi ip p 

 
p 1
 
 p i 
The complex power injected by the source into ith bus of a power system is
Si= Pi +jQi = ViIi*,i=1,2,…..,n
n
Ii =∑ (YipVp)
P=1
n
Pi-jQi = Vi*( ∑YipVp) , i=1,2…..n
P=1
n
Pi = Real Vi*( ∑YipVp) , i=1,2…..n
P=1
n
Qi = -imag Vi*( ∑YipVp) , i=1,2…..n
P=1

  
 k Vk Y  Y Vk   p  i
n

 i ii 
k 1
Pi  R Vi ip p 

 
p 1
 
 p i

and
  
 k V k Y  Y V k   p  i
n

 i ii 
k 1
Qi   Im Vi ip p 

 
p 1
 
 p i
•We first repeat the iteration for PQ buses as in case I, then continue the iteration for
PV buses.
•At PV buses and |v| are specified and Q and δ are unknowns to be determined.
•Therefore, the values of Q and δ are to be updated in every GS iteration through
appropriate bus equations
n
•Qi = -Im {Vi *∑ Yip Vp} , i=m+1,……n
p=1
•The revised value of Qi is obtained from the above equation by substituting most
updatated values of voltages on the right hand side
•For the(k+1)th iteration we can write
i-1 n
•Qi (k+1) = -Im{(Vi( k))* ∑ YipVp (k-1) + (Vi (k ) )* ∑ Yip Vp (k ) }, I = m+1,…n
p=1 P=i
•The revised value of δi is obtained from equation
i-1 n
Vi (k+1) =[ Ai/(Vik)* - ∑ (BipVp (k+1) ) - ∑ (BipVp (k))], i=2,….n
p=1 p=i+1
i-1 n
δi (k+1) = < Vi (k+1) = Angle[Ai (k+1) /(Vi (k) )* - ∑ BipVp (K+1) - ∑ Bip Vp (K)]
p=1 p=i+1
Cont…
• Since Vi is specified, only imaginary part of Vi(k+1) is retained
and its real part is selected to satisfy,
e   f 
k 1 2
i i
k 1 2
 Vi
2

eik 1  Re al component of Vik 1 and

f ik 1  Im aginary component of Vik 1
Or
k 1
 Vi   f 
k 1 2
2
e i i

• The updated voltage replace previous value in solution of

subsequent equation
• Process is continued till change in real and imaginary
components of bus voltages between successive iterations are
within specified accuracy.
Cont…
• Process is continued till change in real and imaginary
components of bus voltages between successive
iterations are within specified accuracy.
eik 1  eik   and
f ik 1  f ik  
• For small power mismatch, very tight tolerance must
be specified on both components of the voltage.
• A voltage accuracy in the range of 0.00001 to 0.00005
pu is satisfactory.
• Iterations continues until magnitude of largest element
in Δp and ΔQ column is < specified value.
• Typical power mismatch accuracy is 0.001 pu.
Cont…
• Ones solution is converged net real and reactive
powers at the slack bus are computed from

  
 k Vk Y  Y Vk   p  i
n

 i ii 
k 1
Pi  R Vi ip p 

 
p 1
 
 p i

and
  
 k V k Y  Y V k   p  i
n

 i ii 
k 1
Qi   Im Vi ip p 

 
p 1
 
 p i
Example 2
• Figure shows the one line diagram of a simple three
bus power system with generators at buses 1 and 3.
The magnitude of voltage at bus 1 is adjusted to 1.05
p.u. Voltage magnitude at bus 3 is fixed at 1.04 p.u with
a real power generation 200MW. A load consisting of
400 MW and 250 Mvar is taken from bus 2. Line
impedances are marked in per unit on a 100 MVA base
and the line charging susceptances are neglected.
1 2
0.02+j0.04
400M
W

0.01+j0.03 0.0125+j0.025 250M

Var
3
Slack Bus
V1=1.05<0 |V3|=1.04

200
MW
Example cont..

1. Obtain the power flow solution be the

Gauss-Seidel method including line flows
and line losses.
Step 1: Find Line admittances

1
y12   10  j20
0.02  j0.04

1
y13   10  j30
0.01  j0.03

1
y 23   16  j32
0.0125  j0.025
Cont. Step 2: Find complex power at bus 2 and 3

So, Complex power at bus 2 (load bus) & 3(generator bus) is,

400  j250
S2    4.0  j2.5 pu
100

200
P3   2.0 pu
100
Step 3: Assuming initial values of voltages at
bus 2 and 3
Bus 1 taken as slack bus.
So, starting with initial estimate of V20 and V30
V20  1.0  j0.0 and
V30  1.04  j0.0

V2 and V3 obtained by

Pi  jQi n
*
  Yip Vp
Vi p 1
p i
Vi  n

y
p0
ip
Step 4: Ist iteration to find voltages at bus 2
P2  jQ 2 n

*0
  yip Vp
V2 p 1
p i
V 
1
2 n

y
p 0
ip

P2  jQ 2
0  y11V1  y12 V1  y 23V30
V2*
=
y12  y 23

4  j2.5
 10  j20 1.05  j0   16  j32 1.04  j0 
1.0  j0
=
26  j52

V21 = 0.97462 - j0.042307

Step 4: Ist iteration to find reactive power at bus 3

  
 k V k Y  Y V k  
n

 i ii 
k 1
Qi   Im Vi ip p 

 
p 1
 
 p i

1
3  0
3 
Q   Im V  V30 (y13  y 23 )  y13V1  y 23V21 

 (1.04  j0)(26  j62)  (10  j30)(1.04  j0)  

  Im (1.04  j0)   
  (16  j32)(0.97462  j0.042307) 
= 1.16

• Q3(1) is used for finding voltage at bus3.

Step 5: Ist iteration to find complex voltage Vc3 at bus 3

P3  jQ3 n

*0
  yip Vp
V3 p 1
p i
V 
1
c3 n

y
p0
ip

P3  jQ3
0  y13V1  y 23V21
V3*
=
y13  y 23

2.0  j1.16
 10  j30 1.05  j0   16  j32  0.97462  j0.042307 
1.04  j0
=
26  j62

V31 = 1.03783-j0.005170
 1
V3 is held constane at 0.04 pu, only imaginary part of Vc3 is retained
i.e. f 31  0.005170, and
real part is obtained as

e  1.04    0.005170   1.039987

1 2 2
3

So
V31  1.039987  j0.005170
Step 6: 2nd iteration to find voltages at bus 2
i2
P2  jQ 2
1  y12 V1  y 23 V31
2 V2*
V =
y12  y 23
2

4  j2.5
 10  j20 1.05   16  j32 1.039987  j0.005170 
0.97462  j0.042307
=
26  j52

V21 = 0.971057 - j0.043432

Step 7: 2nd iteration to find reactive power at bus 3

 
Q32   Im V31  V31 (y13  y 23 )  y13V1  y 23V22 
 (1.039987  j0.005170)(26  j62)  
 
  Im (1.039987  j0.005170)  (10  j30)(1.05  j0) 
      
  (16 j32)(0.971057 j0.043432) 
= 1.38796
Step 5: Ist iteration to find complex voltage Vc3 at bus 3

P3  jQ3 n

*1
  yip Vp
V3 p 1
p i
V 
2
c3 n

y
p 0
ip

P3  jQ3
0  y13 V1  y 23 V22
V3*
=
y13  y 23

2.0  j1.38796
 10  j30 1.05   16  j32  0.971057  j0.043432 
1.039987  j0.00517
=
26  j62

V31 = 1.03908-j0.00730
V3 is held constane at 0.04 pu, only imaginary part of Vc32 is retained
i.e. f 32  0.00730,
and real part is obtained as

e  1.04    0.00730   1.039974

2 2 2
3

So
V32  1.039974  j0.00730
Process is continued and solution is converged
with an accuracy of 5x10-5 pu in 7 iterations
V23  0.97073-j0.05784479 Q33 =1.4294 V33  1.03996-j0.00833
V24  0.97065-j0.04533 Q33 =1.44833 V34  1.03996-j0.00873
V25  0.97062-j0.04555 V35  1.45621 V35  1.03996-j0.00893
V26  0.97061-j0.04565 V36  1.45947 V36  1.03996-j0.00900
V27  0.97061-j0.04569 V37  1.46082 V37  1.03996-j0.00903

Final solution is

V2  0.97061-j0.04569=0.97168-2.6948 pu
S3  2.0  j1.4617 pu

V3  1.03996-j0.00903=1.04-0.498 pu
S1  2.1842  j1.4085 pu
Newton – Raphson Power Flow Solution
• Suitable for load flow study on large systems.

• Advantages:
– more efficient and practical.
– More accurate and surety of convergence
– Only 3 iterations are required as compared to GS requiring more than 25.
– No. of iterations required to obtain solution is independent of system size .

• Disadvantages:
– Solution technique is difficult.
– Requires more functional evaluation at each iteration so computer time per
iteration is large.
– Computer memory requirement is large.

• Power flow equations are formulated in polar form.

The complex power injected by the source into ith bus of a power system is

Si= Pi +jQi = ViIi* ……………i=1,2,…..,n

n
Ii =∑ (YipVp)
p=1
n
Pi-jQi = Vi*( ∑YipVp) , …………. i=1,2…..n
p=1
n
Pi = Real Vi*( ∑YipVp) , ………….. i=1,2…..n
p=1
n
Qi = -imag Vi*( ∑YipVp) , . ………….. i=1,2…..n
p=1

Let Vi = |Vi| eδi , Vp = |Vp| eδp

Yip = | Yip| e jθip

 n
Pi – jQi = |Vi|< -δi ( ∑ |Yip| |Vp| < (θip +δp) , i=1,2…..n
p=1

n
Pi = ( ∑ |Vi| |Yip| |Vp| cos(θip +δp-δi) , i=1,2…..n
p=1
n
Qi = - ( ∑ |Vi| |Yip| |Vp| Sin(θip +δp-δi) , i=1,2…..n
p=1
Newton – Raphson Power Flow Solution
• Current entering bus i is given by
Vi
n n V1
Ii =Vi  yij   yijVj j  i
yi1
V2
j0 j1 Ii yi2

• In terms of Bus admittance matrix, Vn

yin

yi0

• In polar form
A typical bus of the power system

• Complex power at ith bus is

Cont…
• Substituting for Ii, Complex power is given by

• Separating real and imaginary parts,

• These equations gives set of non linear algebraic equations

in terms of independent variables:
– Voltage magnitude in per unit
– Phase angle in radians.
• Taylor’s series expansion about initial estimate and
neglecting higher order terms gives set of linear equations
given below
• Assuming bus 1 as slack bus.

• Gives linearized relationship between small changes in

voltage angle Δδi(k) & voltage magnitude Δ|Vi(k)| with
small changes in real and reacive power ΔPi(k) & ΔQi(k)
• Elements of Jacobian matrix are partial derivatives of Pi &
Qi w.r.t. Δδi(k) & Δ|Vi(k)|
• In short form,

• For PV bus, voltage magnitude are known

• So if (m) voltage controlled buses, corresponding columns
of J matrix are eliminated.
• Accordingly there are n-1 real power constraints &
• n-1-m reactive power constraints.

• J1 matrix order : (n-1)x(n-1)

• J2 matrix order : (n-1)x(n-1-m)
• J3 matrix order : (n-1-m)x(n-1)
• J4 matrix order : (n-1-m)x(n-1-m)
• Diagonal and off diagonal elements of J1 are

• Diagonal and off diagonal elements of J2 are

• Diagonal and off diagonal elements of J3 are

• Diagonal and off diagonal elements of J4 are

• ΔPi(k) & ΔQi(k) are difference between scheduled
and calculated values known as power residuals,
given by,

• New estimates for bus voltages are,

Procedure for power flow solution
1. For load buses,
– Pi(sch) Qi(sch) are specified,
– Voltage magnitude and phase angles are set to slack
bus values ie |Vi(0)|=1 & δi(0)=0
For voltage controlled buses
– |Vi|& Pi(sch) specified
– Phase angle=0 ie δi(0)=0

2. For load buses

– Pi(k) & Qi(k) are calculated from power flow equation
– Δ Pi(k) & ΔQi(k) are calculated from
Cont..
3. For voltage controlled buses
– Pi(k) & Δ Pi(k) are calculated from Power Flow
equation &

4. Elements of Jacobian matrix J1,J2,J3,J4 calculated

5. Linear simultaneous eq. is solved

6. New voltage magnitudes and phase angles are

computed from
Cont..
7. Process is continued till the residuals Δ Pi(k)
& ΔQi(k) are less than the specifies accuracy
i. e. |Δ Pi(k)|< ε
|Δ Qi(k)|< ε
Example
• Obtain the power flow solution by Newton
Raphson method for the system shown below.

1 2
0.02+j0.04
400M
W

0.01+j0.03 0.0125+j0.025 250M

Var
3
Slack Bus
V1=1.05<0 |V3|=1.04

200
MW
Steps:
1. Find Bus Admittance matrix Ybus.
2. Convert Bus admittance matrix in polar form
3. Write equations for powers for all buses except
slack bus from the equations.

1. For Load bus we get two power equations (i.e. for

active and reactive power)
2. For PV bus only one power equation (Reactive
power)
Steps:
4. Find elements of Jacobian matrix by taking partial
derivatives of above power equations with respect to δ2,
δ3 and IV2I
5. Find scheduled powers at load and generator bus.
6. Starting with initial estimates for V and δ calculate power
residuals using equations .

7. Evaluate elements of Jacobian matrix with initial

estimates.
Steps:
8. Obtain set of linear equations in the first iteration.

9. Solving these linear equations will give small changes in

voltage magnitude and phase angles.

10. New bus voltage magnitudes and phase angles are

computed in first iteration using equations:

11. Evaluate ΔP2(1), ΔQ2(1), ΔP3(1) and Jacobian matrix

elements to Perform second iterations.
Steps:

12. Perform second and third iteration

13. Solution will converge in 3 iterations with max. accuracy.

14. Find powers at slack bus and other all.

Step 1: Find Bus Admittance matrix Ybus.

Step 2: Convert Bus admittance matrix in polar

form (angles in radians)
.
Step 3: Write equations for powers for all
buses except slack bus from the equations.

• Bus 2: Load bus

• So equations for Real (P2) and Reactive power (Q2) are
Step 3: Cont..

• Bus 3: PV bus
• So equations for Real (P3) is
Step 4: Find elements of Jacobian matrix by
taking partial derivatives of power
equations with respect to δ2, δ3 and
IV2I
• Taking partial derivative of P2 with respect to
δ2, δ3 and IV2I
Step 4: Cont…
• Taking partial derivative of P3 with respect to
δ2, δ3 and IV2I
Step 4: Cont…
• Taking partial derivative of Q2 with respect to
δ2, δ3 and IV2I
Step 5: Find scheduled powers at load
and generator bus

• Per Unit load and generation are:

Step 6: Starting with initial estimates for
V and δ calculate power residuals
• Slack bus voltage V1=1.05<0 pu
• Bus 3 voltage magnitude=IV3I=1.04 pu
• Starting with initial estimates of
– IV2(0)I=1.0,
– δ2(0)=0.0 and
– δ3(0)=0.0

• Power residuals computed from,

Step 6: Cont…
Step 7: Evaluate elements of Jacobian
matrix with initial estimates
Step 8: Obtain set of linear equations in
the first iteration

-2.866 = 54.28 Δδ (0)

2 - 33.28 Δδ (0)
3 + 24.86 Δ V(0)
2

1.4384 = -33.28 Δδ (0)

2 + 66.04 Δδ (0)
3 - 16.64 Δ V (0)
2

-0.22 = -27.14 Δδ (0)

2 + 16.64 Δδ (0)
3 + 49.72 Δ V (0)
2
Step 9: Solving linear equations to obtain
small changes in voltage magnitude
and phase angles
Step 10: Solving linear equations to obtain
small changes in voltage magnitude
and phase angles
• New bus voltages in first iteration are obtained by,
Step 10: Evaluate ΔP2(1), ΔQ2(1), ΔP3(1)
and Jacobian matrix elements to
Perform second iterations.

• For second iteration

Step 11: second iterations.
Step 12: For Third Iterations.
Step 13: Solution converge in 3 iterations
with max. accuracy.

• Solution converges in 3 iterations with maximum

power mismatch of 2.5x10(-4) with V2=0.97168<-
2.696° and V3=1.04<-0.4988°
Step 14: Find reactive power at bus 3 and
slack bus real and reactive powers.
Step 14: Cont..
Fast Decoupled power Flow Solution

• The off-diagonal submatrices J(Qδ) and J(PV) are usually

very small relative to the elements of the diagonal
submatrices J(P δ) and J(QV)

• These observations are consistent with our understanding

that P is not very sensitive to voltage magnitude

• and Q is not very sensitive to angle

Cont..
• Transmission lines have very high X/R ratio.
• For such systems, real power change ΔP are less
sensitive to changes in voltage magnitude and
most sensitive to changes in phase angle Δδ.
• Reactive power is less sensitive to changes in
angle and are mainly dependent on changes in
voltage magnitude.
• So, Jacobian matrix elements J2 and J3 =0
• DC load flow
• • Approximate solution.
• • Two simplifications:
• – In network model: do not consider series
• resistences and shunt admittances
• – Assume Vi=1 at all buses
 DC load flow
Approximate analytical solution
DC Power flow: example
 Comparisonof load flow methods
• 1. Gauss-Seidel (G-S)
• Simple technique
• Iteration time increases linearly with the
• number of buses. Lower iteration time than NR.
Seven times faster in large systems
• Linear rate of convergence. Many iterations
• required for getting close to the solution
• Number of iteration increases with the
• number of buses
• 2. Newton-Raphson (N-R)
• Widely used
• Iteration-time increases linearly with the number of
• buses
• Quadratic rate of convergency. A few iterations for
• getting close to the solution
• Number of iterations independent of the number of
• buses of the system
• The Jacobian is a very sparse matrix
• Method non-sensitive to slack bus choice and the
• presence of series capacitors
• Sensitive to initial solution
• 3. AC decoupled
• has to be computed and factorized only once
• It requires more iterations than NewtonRaphson
method
• Iteration time is 5 times lower than
NewtonRaphson´s iteration time
• Useful for analyzing topology changes because
• can be easily modified
• Used in planning and contigency analyses
• 4. DC Decoupled
• Analytical, approximate and non-iterative
method
• Good approximation for , not that good
• approximation for
• Used in reliability analyses
• Used in optimal pricing calculations
• Good for getting an initial point