http://nptel.ac.in/courses/Webcourse-contents/IIT-KANPUR/powersystem/ui/Course_home-4.

htm
CHAPTER 4
LOAD FLOW STUDIES
Load flow studies are one of the most important aspects of power system planning
and operation. The load flow gives us the sinusoidal steady state of the entire system
voltages, real and reactive power generated and absorbed and line losses. Since the load is a
static quantity and it is the power that flows through transmission lines, the purists prefer to
call this Power Flow studies rather than load flow studies. We shall however stick to the
original nomenclature of load flow.
Through the load flow studies we can obtain the voltage magnitudes and angles at
each bus in the steady state. This is rather important as the magnitudes of the bus voltages are
required to be held within a specified limit. Once the bus voltage magnitudes and their angles
are computed using the load flow, the real and reactive power flow through each line can be
computed. Also based on the difference between power flow in the sending and receiving
ends, the losses in a particular line can also be computed. Furthermore, from the line flow we
can also determine the over and under load conditions.
The steady state power and reactive powers supplied by a bus in a power network are
expressed in terms of nonlinear algebraic equations. We therefore would require iterative
methods for solving these equations. In this chapter we shall discuss two of the load flow
methods. We shall also delineate how to interpret the load flow results.
4.1 REAL AND REACTIVE POWER INJECTED IN A BUS
For the formulation of the real and reactive power entering a bus, we need to define
the following quantities. Let the voltage at the ith bus be denoted by
Vi Vi i Vi cos i j sin i

(4.1)

Also let us define the self admittance at bus-i as

Yii Yii ii Yii cos ii j sin ii Gii jBii

(4.2)

Similarly the mutual admittance between the buses i and j can be written as
Yij Yij ij Yij cos ij j sin ij Gij jBij

(4.3)

Let the power system contains a total number of n buses. The current injected at bus-i
is given as

2.2
I i Yi1V1 Yi 2V2 YinVn
n

YikVk
k 1

(4.4)
It is to be noted we shall assume the current entering a bus to be positive and that leaving the
bus to be negative. As a consequence the power and reactive power entering a bus will also
be assumed to be positive. The complex power at bus-i is then given by
n

Pi jQi Vi I i Vi YikVk
k 1

Vi cos i j sin i YikVk cos ik j sin ik cos k j sin k

(4.5)

k 1

YikViVk cos i j sin i cos ik j sin ik cos k j sin k

k 1

Note that

cos i j sin i cos ik j sin ik cos k j sin k

cos i j sin i cos ik k j sin ik k
cos ik k i j sin ik k i
Therefore substituting in (4.5) we get the real and reactive power as
n

Pi YikViVk cos ik k i

(4.6)

k 1

Qi YikViVk sin ik k i

(4.7)

k 1

4.2 CLASSIFICATION OF BUSES

For load flow studies it is assumed that the loads are constant and they are defined by
their real and reactive power consumption. It is further assumed that the generator terminal
voltages are tightly regulated and therefore are constant. The main objective of the load flow
is to find the voltage magnitude of each bus and its angle when the powers generated and
loads are pre-specified. To facilitate this we classify the different buses of the power system
as listed below.
1. Load Buses: In these buses no generators are connected and hence the generated real
power PGi and reactive power QGi are taken as zero. The load drawn by these buses are
defined by real power PLi and reactive power QLi in which the negative sign
accommodates for the power flowing out of the bus. This is why these buses are
sometimes referred to as P-Q bus. The objective of the load flow is to find the bus
voltage magnitude Vi and its angle i.

2.3
2. Voltage Controlled Buses: These are the buses where generators are connected.
Therefore the power generation in such buses is controlled through a prime mover while
the terminal voltage is controlled through the generator excitation. Keeping the input
power constant through turbine-governor control and keeping the bus voltage constant
using automatic voltage regulator, we can specify constant PGi and Vi for these buses.
This is why such buses are also referred to as P-V buses. It is to be noted that the
reactive power supplied by the generator QGi depends on the system configuration and
cannot be specified in advance. Furthermore we have to find the unknown angle i of
the bus voltage.
3. Slack or Swing Bus: Usually this bus is numbered 1 for the load flow studies. This bus
sets the angular reference for all the other buses. Since it is the angle difference between
two voltage sources that dictates the real and reactive power flow between them, the
particular angle of the slack bus is not important. However it sets the reference against
which angles of all the other bus voltages are measured. For this reason the angle of this
bus is usually chosen as 0. Furthermore it is assumed that the magnitude of the voltage
of this bus is known.
Now consider a typical load flow problem in which all the load demands are known.
Even if the generation matches the sum total of these demands exactly, the mismatch
between generation and load will persist because of the line I2R losses. Since the I2R loss
of a line depends on the line current which, in turn, depends on the magnitudes and
angles of voltages of the two buses connected to the line, it is rather difficult to estimate
the loss without calculating the voltages and angles. For this reason a generator bus is
usually chosen as the slack bus without specifying its real power. It is assumed that the
generator connected to this bus will supply the balance of the real power required and
the line losses.
4.3 PREPARATION OF DATA FOR LOAD FLOW
Let real and reactive power generated at bus-i be denoted by PGi and QGi respectively.
Also let us denote the real and reactive power consumed at the ith bus by PLi and QLi
respectively. Then the net real power injected in bus-i is
Pi , inj PGi PLi

(4.8)

Let the injected power calculated by the load flow program be Pi,calc. Then the mismatch
between the actual injected and calculated values is given by
Pi Pi , inj Pi , calc PGi PLi Pi , calc

(4.9)
In a similar way the mismatch between the reactive power injected and calculated values is
given by
Qi Qi ,inj Qi , calc QGi QLi Qi , calc

(4.10)

The purpose of the load flow is to minimize the above two mismatches. It is to be noted that
(4.6) and (4.7) are used for the calculation of real and reactive power in (4.9) and (4.10).

2.4
However since the magnitudes of all the voltages and their angles are not known a priori, an
iterative procedure must be used to estimate the bus voltages and their angles in order to
calculate the mismatches. It is expected that mismatches Pi and Qi reduce with each
iteration and the load flow is said to have converged when the mismatches of all the buses
become less than a very small number.
For the load flow studies we shall consider the system of Fig. 4.1, which has 2
generator and 3 load buses. We define bus-1 as the slack bus while taking bus-5 as the P-V
bus. Buses 2, 3 and 4 are P-Q buses. The line impedances and the line charging admittances
are given in Table 4.1. Based on this data the Ybus matrix is given in Table 4.2. This matrix is
formed using the same procedure as given in Section 3.1.3. It is to be noted here that the
sources and their internal impedances are not considered while forming the Ybus matrix for
load flow studies which deal only with the bus voltages.

Fig. 4.1 The simple power system used for load flow studies.
Table 4.1 Line impedance and line charging data of the system of Fig. 4.1.
Line (bus to bus)

Impedance

Line charging (Y/2)

1-2

0.02 + j0.10

j0.030

1-5

0.05 + j0.25

j0.020

2-3

0.04 + j0.20

j0.025

2-5

0.05 + j0.25

j0.020

3-4

0.05 + j0.25

j0.020

3-5

0.08 + j0.40

j0.010

4-5

0.10 + j0.50

j0.075

Table 4.2 Ybus matrix of the system of Fig. 4.1.

1

2.6923 j13.4115

1.9231 + j9.6154

0.7692 + j3.8462

1.9231 + j9.6154

3.6538 j18.1942

0.9615 + j4.8077

0.7692 + j3.8462

0.9615 + j4.8077

2.2115 j11.0027

0.7692 + j3.8462

0.4808 + j2.4038

0.7692 + j3.8462

1.1538 j5.6742

0.3846 + j1.9231

0.7692 + j3.8462

0.7692 + j3.8462

0.4808 + j2.4038

0.3846 + j1.9231

2.4038 j11.8942

2.5
The bus voltage magnitudes, their angles, the power generated and consumed at each
bus are given in Table 4.3. In this table some of the voltages and their angles are given in
boldface letters. This indicates that these are initial data used for starting the load flow
program. The power and reactive power generated at the slack bus and the reactive power
generated at the P-V bus are unknown. Therefore each of these quantities are indicated by a
dash (). Since we do not need these quantities for our load flow calculations, their initial
estimates are not required. Also note from Fig. 4.1 that the slack bus does not contain any
load while the P-V bus 5 has a local load and this is indicated in the load column.
Table 4.3 Bus voltages, power generated and load initial data.
Bus voltage
Bus no.

Power generated

Load

Magnitude (pu)

Angle (deg)

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

1.05

96

62

35

14

16

1.02

48

24

11

4.4 LOAD FLOW BY GAUSS-SEIDEL METHOD

The basic power flow equations (4.6) and (4.7) are nonlinear. In an n-bus power
system, let the number of P-Q buses be np and the number of P-V (generator) buses be ng such
that n = np + ng + 1. Both voltage magnitudes and angles of the P-Q buses and voltage angles
of the P-V buses are unknown making a total number of 2np + ng quantities to be determined.
Amongst the known quantities are 2np numbers of real and reactive powers of the P-Q buses,
2ng numbers of real powers and voltage magnitudes of the P-V buses and voltage magnitude
and angle of the slack bus. Therefore there are sufficient numbers of known quantities to
obtain a solution of the load flow problem. However, it is rather difficult to obtain a set of
closed form equations from (4.6) and (4.7). We therefore have to resort to obtain iterative
solutions of the load flow problem.
At the beginning of an iterative method, a set of values for the unknown quantities are
chosen. These are then updated at each iteration. The process continues till errors between all
the known and actual quantities reduce below a pre-specified value. In the Gauss-Seidel load
flow we denote the initial voltage of the ith bus by Vi(0), i = 2, , n. This should read as the
voltage of the ith bus at the 0th iteration, or initial guess. Similarly this voltage after the first
iteration will be denoted by Vi(1). In this Gauss-Seidel load flow the load buses and voltage
controlled buses are treated differently. However in both these type of buses we use the
complex power equation given in (4.5) for updating the voltages. Knowing the real and
reactive power injected at any bus we can expand (4.5) as
n

Pi , inj jQi , inj Vi YikVk Vi Yi1V1 Yi 2V2 YiiVi YinVn

k 1

(4.11)
We can rewrite (4.11) as

2.6

Vi

1 Pi , inj jQi ,inj

Yi1V1 Yi 2V2 YinVn

Yii
Vi

(4.12)
In this fashion the voltages of all the buses are updated. We shall outline this procedure with
the help of the system of Fig. 4.1, with the system data given in Tables 4.1 to 4.3. It is to be
noted that the real and reactive powers are given respectively in MW and MVAr. However
they are converted into per unit quantities where a base of 100 MVA is chosen.
4.4.1 Updating Load Bus Voltages
Let us start the procedure with bus-2. Since this is load bus, both the real and reactive
power into this bus is known. We can therefore write from (4.12)
1

V2

P2 ,inj jQ2 ,inj

1
Y22

V2

Y21V1 Y23V3

Y24V4

Y25V5

(4.13)
From the data given in Table 4.3 we can write
1

V2

1 0.96 j 0.62

1.05Y21 Y23 Y24 1.02Y25

Y22
1

It is to be noted that since the real and reactive power is drawn from this bus, both these
quantities appear in the above equation with a negative sign. With the values of the Ybus
elements given in Table 4.2 we get V2(1) = 0.9927 2.5959.
The first iteration voltage of bus-3 is given by

V3

1
Y33

P3, inj jQ3, inj

V3

Y31V1 Y32V2 Y34V4

Y35V5

(4.14)
Note that in the above equation since the update for the bus-2 voltage is already available, we
used the 1st iteration value of this rather than the initial value. Substituting the numerical data
we get V3(1) = 0.9883 2. 8258. Finally the bus-4 voltage is given by

V4

1 P4,inj jQ4,inj
1
1
0

Y
V

Y
V

Y
V

Y
V

41
1
42
2
44
3
45
5

Y44
V4

(4.15)
Solving we get V4(1) = 0. 9968 3.4849.
4.4.2 Updating P-V Bus Voltages

2.7

It can be seen from Table 4.3 that even though the real power is specified for the P-V
bus-5, its reactive power is unknown. Therefore to update the voltage of this bus, we must
first estimate the reactive power of this bus. Note from Fig. 4.11 that

Qi , inj Im Vi YikVk Im Vi Yi1V1 Yi 2V2 YiiVi YinVn

k 1

(4.16)
And hence we can write the kth iteration values as
k

Y V Y V

Qi ,inj Im Vi

k 1

i1 1

i2 2

YiiVi

k 1

YinVn

k 1

(4.17)
For the system of Fig. 4.1 we have
1

Q5,inj Im V1 Y51V1 Y52V2 Y53V3 Y54V4 Y55V5

(4.18)
This is computed as 0.0899 per unit. Once the reactive power is estimated, the bus-5 voltage
is updated as

V5

1 P5,inj jQ5,inj
1
1
0

Y
V

Y
V

Y
V

Y
V

51 1
52 2
53 3
54 4
0
Y55
V5

(4.19)
It is to be noted that even though the power generation in bus-5 is 48 MW, there is a local
load that is consuming half that amount. Therefore the net power injected by this bus is 24
MW and consequently the injected power P5,inj in this case is taken as 0.24 per unit. The
voltage is calculated as V4(1) = 1.0169 0.8894. Unfortunately however the magnitude of
the voltage obtained above is not equal to the magnitude given in Table 4.3. We must
therefore force this voltage magnitude to be equal to that specified. This is accomplished by

V5, corr

V5

V5
V5

1
1

(4.20)
This will fix the voltage magnitude to be 1.02 per unit while retaining the phase of 0.8894.
The corrected voltage is used in the next iteration.
4.4.3 Convergence of the Algorithm
As can be seen from Table 4.3 that a total number of 4 real and 3 reactive powers are
known to us. We must then calculate each of these from (4.6) and (4.7) using the values of the

2.8
voltage magnitudes and their angle obtained after each iteration. The power mismatches are
then calculated from (4.9) and (4.10). The process is assumed to have converged when each
of P2, P3, P4, P5, Q2, Q3 and Q4 is below a small pre-specified value. At this point
the process is terminated.
Sometimes to accelerate computation in the P-Q buses the voltages obtained from
(4.12) is multiplied by a constant. The voltage update of bus-i is then given by

Vi , acc

1 Vi , acc

k 1

Vi

Vi , acc

k 1

Vi

Vi , acc

k 1

(4.21)

where is a constant that is known as the acceleration factor. The value of has to be below
2.0 for the convergence to occur. Table 4.4 lists the values of the bus voltages after the 1 st
iteration and number of iterations required for the algorithm to converge for different values
of . It can be seen that the algorithm converges in the least number of iterations when is
1.4 and the maximum number of iterations are required when is 2. In fact the algorithm will
start to diverge if larger values of acceleration factor are chosen. The system data after the
convergence of the algorithm will be discussed later.
Table 4.4 Gauss-Seidel method: bus voltages after 1st iteration and number of iterations required for convergence
for different values of .
Bus voltages (per unit) after 1st iteration
V2

V3

V4

V5

No of iterations
for convergence

0.9927 2.6

0.9883 2.83

0.9968 3.48

1.02 0.89

28

0.9874 5.22

0.9766 8.04

0.9918 14.02

1.02 4.39

860

1.8

0.9883 4.7

0.9785 6.8

0.9903 11.12

1.02 3.52

54

1.6

0.9893 4.17

0.9807 5.67

0.9909 8.65

1.02 2.74

24

1.4

0.9903 3.64

0.9831 4.62

0.9926 6.57

1.02 2.05

14

1.2

0.9915 3.11

0.9857 3.68

0.9947 4.87

1.02 1.43

19

4.5 SOLUTION OF A SET OF NONLINEAR EQUATIONS BY

NEWTON-RAPHSON METHOD
In this section we shall discuss the solution of a set of nonlinear equations through
Newton-Raphson method. Let us consider that we have a set of n nonlinear equations of a
total number of n variables x1, x2, , xn. Let these equations be given by
f1 x1 ,, xn 1
f 2 x1 ,, xn 2

f n x1 ,, xn n

(4.22)

where f1, , fn are functions of the variables x1, x2, , xn. We can then define another set of
functions g1, , gn as given below

2.9
g1 x1 ,, xn f1 x1 ,, xn 1 0
g 2 x1 ,, xn f 2 x1 ,, xn 2 0

(4.23)

g n x1 ,, xn f n x1 ,, xn n 0

Let us assume that the initial estimates of the n variables are x1(0), x2(0), , xn(0). Let us
add corrections x1(0), x2(0), , xn(0) to these variables such that we get the correct solution
of these variables defined by
0

x1 x1 x1

x2 x2 x2

(4.24)

0
0
xn xn xn
The functions in (4.23) then can be written in terms of the variables given in (4.24) as

g k x1 ,, xn g k x1 x1 ,, xn xn , k 1, , n

(4.25)

We can then expand the above equation in Taylors series around the nominal values of x1(0),
x2(0), , xn(0). Neglecting the second and higher order terms of the series, the expansion of gk,
k = 1, , n is given as

g k x1 ,, xn g k x1 ,, xn
0

x gx
0

x2

g k
x2

xn

g k
xn

(4.26)
where g k xi

is the partial derivative of gk evaluated at x2(0), , xn(0).

Equation (4.26) can be written in vector-matrix form as

g1 x1 g1 x2 g1 xn
g x g x g x
2
2
2
n
2 1

g n x1 g n x2 g n xn

x1 0
0

x2

0

xn

0 g x , , x

0 g1 x1 0 ,, xn 0

(4.27)

0 g n x1 ,, xn

The square matrix of partial derivatives is called the Jacobian matrix J with J(0) indicating that
the matrix is evaluated for the initial values of x2(0), , xn(0). We can then write the solution of
(4.27) as

2.10

x1 0

0
x2 J 0

0
xn

g1 0

0
1 g 2

0
g n

(4.28)

Since the Taylors series is truncated by neglecting the 2 nd and higher order terms, we cannot
expect to find the correct solution at the end of first iteration. We shall then have
1

x1 x1 x1
1
0
0
x2 x2 x2

(4.29)

xn xn xn

These are then used to find J(1) and gk(1), k = 1, , n. We can then find x2(1), , xn(1) from
an equation like (4.28) and subsequently calculate x2(1), , xn(1). The process continues till
gk, k = 1, , n becomes less than a small quantity.

Example 4.1: Let us consider the following set of nonlinear equations

g1 x1 , x2 , x3 x12 x22 x32 11 0

g 2 x1 , x2 , x3 x1 x2 x22 3 x3 3 0

g3 x1 , x2 , x3 x1 x1 x3 x2 x3 6 0

The Jacobian matrix is then given by

2 x1
J x2
1 x3

2 x2
x1 2 x2
x3

2 x3
3
x1 x2

The initial values to start the Newton-Raphson procedure must be carefully chosen. For
example if we choose x1(0) = x2(0) = x3(0) = 0 then all the elements of the 1st row will be zero
making the matrix J(0) singular. In our procedure let us choose x1(0) = x2(0) = x3(0) = 1. The
Jacobian matrix is then given by

Also the mismatches are given by

2
1
0

2
3
1

3
0

2.11

g1 0 0 g1 0 10
0
0

g
2
2

4
g3 0 0 g3 0 5

Consequently the corrections and updates calculated respectively from (4.28) and (4.29) are

x1 0

x1 1 5.75
4.75

1
0

x2 5.00 and x2 6.00

x 0 5.25
x 1 6.25

3
3
The process converges in 7 iterations with the values of
x1 2, x2 3 and x3 4

4.6 LOAD FLOW BY NEWTON-RAPHSON METHOD

Let us assume that an n-bus power system contains a total number of np P-Q buses
while the number of P-V (generator) buses be ng such that n = np + ng + 1. Bus-1 is assumed
to be the slack bus. We shall further use the mismatch equations of Pi and Qi given in (4.9)
and (4.10) respectively. The approach to Newton-Raphson load flow is similar to that of
solving a system of nonlinear equations using the Newton-Raphson method: at each iteration
we have to form a Jacobian matrix and solve for the corrections from an equation of the type
given in (4.27). For the load flow problem, this equation is of the form

2
P2

n
V2 Pn

V2 Q2

V1 n p
Q1 n p

V1 n p

(4.30)

where the Jacobian matrix is divided into submatrices as

J
J 11
J 21

J12
J 22

(4.31)

2.12
It can be seen that the size of the Jacobian matrix is (n + np 1) (n + np 1). For example
for the 5-bus problem of Fig. 4.1 this matrix will be of the size (7 7). The dimensions of the
submatrices are as follows:
J11: (n 1) (n 1), J12: (n 1) np, J21: np (n 1) and J22: np np
The submatrices are

P2
P2

n
2

J11

Pn Pn
2
n

P2
V2 V
2

V1 n p

J12

Pn
V2 V
2

V1 n p

(4.32)

P2

V1 n p

Pn

V1 n p

(4.33)

Q2

2

J 21
Q1 n p

2

Q2
n

Q1 n p

(4.34)

Q2
V2 V
2

J 22

Q1 n p

V2

V2

V1 n p

V1 n p

Q2

V1 n p

Q1 n p

(4.35)

V1 n p

4.6.1 Load Flow Algorithm

The Newton-Raphson procedure is as follows:
Step-1: Choose the initial values of the voltage magnitudes V(0) of all np load buses
and n 1 angles (0) of the voltages of all the buses except the slack bus.
Step-2: Use the estimated V(0) and (0) to calculate a total n 1 number of injected
real power Pcalc(0) and equal number of real power mismatch P(0).

2.13

Step-3: Use the estimated V(0) and (0) to calculate a total np number of injected
reactive power Qcalc(0) and equal number of reactive power mismatch Q(0).
Step-3: Use the estimated V(0) and (0) to formulate the Jacobian matrix J(0).
Step-4: Solve (4.30) for (0) and V(0)V(0).
Step-5: Obtain the updates from

1 0 0
V

(4.36)

(4.37)

Step-6: Check if all the mismatches are below a small number. Terminate the process
if yes. Otherwise go back to step-1 to start the next iteration with the updates given by (4.36)
and (4.37).
4.6.2 Formation of the Jacobian Matrix
We shall now discuss the formation of the submatrices of the Jacobian matrix. To do
that we shall use the real and reactive power equations of (4.6) and (4.7). Let us rewrite them
with the help of (4.2) as
n

Pi Vi Gii YikViVk cos ik k i

2

k 1
k i

(4.38)

Qi Vi Bii YikViVk sin ik k i

2

k 1
k i

(4.39)
A. Formation of J11
Let us define J11 as
L22
J11
Ln 2

L2 n

Lnn

(4.40)

It can be seen from (4.32) that Miks are the partial derivatives of Pi with respect to k. The
derivative Pi (4.38) with respect to k for i k is given by
Lik

Pi
YikViVk sin ik k i , i k
k

(4.41)

2.14
Similarly the derivative Pi with respect to k for i = k is given by

Lii

n
Pi
YikViVk sin ik k i
i k 1
k i

Comparing the above equation with (4.39) we can write

Lii

Pi
2
Qi Vi Bii
i

(4.42)

B. Formation of J21
Let us define J21 as

J 21

M 22

M
np 2

M 2n

M n p n

(4.43)

From (4.34) it is evident that the elements of J21 are the partial derivative of Q with respect to
. From (4.39) we can write
M ik

Qi
YikViVk cos ik k i , i k
k

(4.44)

Similarly for i = k we have

M ii

n
Qi
2
YikViVk cos ik k i Pi Vi Gii
i k 1
k i

(4.45)
The last equality of (4.45) is evident from (4.38).
C. Formation of J12
Let us define J12 as
N 22

J12

N 2n p

(4.46)

n 2 N nn p

As evident from (4.33), the elements of J21 involve the derivatives of real power P with
respect to magnitude of bus voltage V. For i k, we can write from (4.38)
N ik Vk

Pi
YikViVk cos ik k i M ik
Vk

ik

(4.47)

2.15

For i = k we have

n
Pi

N ii Vi
Vi 2 Vi Gii YikVk cos ik k i

Vi
k 1
k i

(4.48)

2 Vi Gii YikViVk cos ik k i 2 Vi Gii M ii

2

k 1
k i

D. Formation of J22
For the formation of J22 let us define
O22

J 22

O
np 2

O2 n p

(4.49)

On p n p

For i k we can write from (4.39)

Oik Vi

Qi
Vi YikViVk sin ik k i Lik ,
Vk

ik

(4.50)
Finally for i = k we have

Oii Vi

n
Qi
Vi 2 Vi Bii YikVk sin ik k i

Vk
k 1
k i

(4.51)

2 Vi Bii YikViVk sin ik k i 2 Vi Bii Lii

2

k 1
k i

We therefore see that once the submatrices J11 and J21 are computed, the formation of
the submatrices J12 and J22 is fairly straightforward. For large system this will result in
considerable saving in the computation time.
4.6.3 Solution of Newton-Raphson Load Flow
The Newton-Raphson load flow program is tested on the system of Fig. 4.1 with the
system data and initial conditions given in Tables 4.1 to 4.3. From (4.41) we can write
L23

Y23V2 V3

Similarly from (4.39) we have

sin 23 3 2 Y23 sin 23 B23 4.8077

2.16

Q2

V2

B22 Y2 kV2 Vk

k 1
k 2

sin 2 k k 2

B22 1.05 B21 B23 B24 1.02 B25 0.6327

Hence from (4.42) we get
L22

Q2

V2

B22 0.6327 B22 18.8269

In a similar way the rest of the components of the matrix J11(0) are calculated. This matrix is
given by

J11

4.8077
11.1058
3.8462
2.4519

18.8269
4.8077

3.9231

0
3.8462
5.8077
1.9615

3.9231
2.4519
1.9615

12.4558

For forming the off diagonal elements of J21 we note from (4.44) that
M 23

Y23V2 V3

cos 23 2 3 G23 0.9615

Also from (4.38) the real power injected at bus-2 is calculated as

P2

V2

G22 Y2 kV2 Vk
k 1
k 2

cos 2 k k 2

G22 1.05G21 G23 G24 1.02G25 0.1115

Hence from (4.45) we have
M 22 P2

V2

G22 3.7654

Similarly the rest of the elements of the matrix J21 are calculated. This matrix is then given as
0

J 21

3.7654
0.9615

0.9615

2.2212
0.7692

0.7692
1.1615

0.7846

0.4904
0.3923

For calculating the off diagonal elements of the matrix J12 we note from (4.47) that
they are negative of the off diagonal elements of J21. However the size of J21 is (3 4) while
the size of J12 is (4 3). Therefore to avoid this discrepancy we first compute a matrix M that
is given by

2.17

M 11
M
21
M 31

M 41

M 12
M 22

M 13
M 23

M 32
M 42

M 33
M 43

M 14
M 24
M 34

M 44

The elements of the above matrix are computed in accordance with (4.44) and (4.45). We can
then define
J 21 M 1 : 3, 1 : 4 and J12 M 1 : 4, 1 : 3

Furthermore the diagonal elements of J12 are overwritten in accordance with (4.48). This
matrix is then given by

J12

3.5423
0.9615

0.7846

0.9615
2.2019
0.7692
0.4904

0.7692
1.1462

0.3923

Finally it can be noticed from (4.50) that J22 = J11(1:3, 1:3). However the diagonal
elements of J22 are then overwritten in accordance with (4.51). This gives the following
matrix

J 22

17.5615
4.8077

4.8077
10.8996
3.8462

3.8462
5.5408

From the initial conditions the power and reactive power are computed as

Pcalc 0.1115 0.0096 0.0077 0.0098

0

Qcalc 0.6327 0.1031 0.1335

0

Consequently the mismatches are found to be

P 0 0.8485
Q 0 0.0127

0.3404
0.0369

0.1523
0.0535

0.2302

Then the updates at the end of the first iteration are given as

2 0

3
0
3
4

4.91
V2 0 0.9864
6.95

deg V3 0 0.9817

7.19
V 0 0.9913

4
3.09

2.18

The load flow converges in 7 iterations when all the power and reactive power mismatches
are below 106.
4.7 LOAD FLOW RESULTS
In this section we shall discuss the results of the load flow. It is to be noted here that
both Gauss-Seidel and Newton-Raphson methods yielded the same result. However the
Newton-Raphson method converged faster than the Gauss-Seidel method. The bus voltage
magnitudes, angles of each bus along with power generated and consumed at each bus are
given in Table 4.4. It can be seen from this table that the total power generated is 174.6 MW
whereas the total load is 171 MW. This indicates that there is a line loss of about 3.6 MW for
all the lines put together. It is to be noted that the real and reactive power of the slack bus and
the reactive power of the P-V bus are computed from (4.6) and (4.7) after the convergence of
the load flow.
Table 4.4 Bus voltages, power generated and load after load flow convergence.
Bus voltage
Bus no.

Power generated

Load

Magnitude (pu)

Angle (deg)

P (MW)

Q (MVAr)

P (MW)

P (MVAr)

1.05

126.60

57.11

0.9826

5.0124

96

62

0.9777

7.1322

35

14

0.9876

7.3705

16

1.02

3.2014

48

15.59

24

11

The current flowing between the buses i and k can be written as

I ik Yik Vi Vk , i k

(4.52)

Therefore the complex power leaving bus-i is given by

Pi jQi Vi I i

(4.53)
Similarly the complex power entering bus-k is
Pk jQk Vk I k

(4.54)

Therefore the I2R loss in the line segment i-k is

Ploss , i k Pi Pk

(4.55)

The real power flow over different lines is listed in Table 4.5. This table also gives the I2R
loss along various segments. It can be seen that all the losses add up to 3.6 MW, which is the
net difference between power generation and load. Finally we can compute the line I2X drops
in a similar fashion. This drop is given by

2.19

Qdrop , i k Qi Qk

(4.56)

However we have to consider the effect of line charging separately.

Table 4.5 Real power flow over different lines.
Power dispatched

Power received

Line loss (MW)

from (bus)

amount (MW)

in (bus)

amount (MW)

101.0395

98.6494

2.3901

25.5561

25.2297

0.3264

17.6170

17.4882

0.1288

0.7976

0.7888

0.0089

15.1520

14.9676

0.1844

18.6212

18.3095

0.3117

15.4566

15.2112

0.2454
Total = 3.5956

Consider the line segment 1-2. The voltage of bus-1 is V1 = 1.050 per unit while
that of bus-2 is V2 = 0.98265.0124 per unit. From (4.52) we then have
I12 0.9623 j 0.5187 1.0932 28.33 per unit

Therefore the complex power dispatched from bus-1 is

S12 V1I12
100 101.0395 j 54.4645

where the negative signal indicates the power is leaving bus-1. The complex power received
at bus-2 is

S 21 V2 I12
100 98.6494 j 42.5141

Therefore out of a total amount of 101.0395 MW of real power is dispatched from bus-1 over
the line segment 1-2, 98.6494 MW reaches bus-2. This indicates that the drop in the line
segment is 2.3901 MW. Note that
I12

R12 100 1.0932 2 0.02 100 2.3901

MW

where R12 is resistance of the line segment 1-2. Therefore we can also use this method to
calculate the line loss.
Now the reactive drop in the line segment 1-2 is
2

I12 X 12 100 1.0932 2 0.1 100 10.9508

MVAr

2.20
We also get this quantity by subtracting the reactive power absorbed by bus-2 from that
supplied by bus-1. The above calculation however does not include the line charging. Note
that since the line is modeled by an equivalent-, the voltage across the shunt capacitor is the
bus voltage to which the shunt capacitor is connected. Therefore the current I12 flowing
through line segment is not the current leaving bus-1 or entering bus-2 it is the current
flowing in between the two charging capacitors. Since the shunt branches are purely reactive,
the real power flow does not get affected by the charging capacitors. Each charging capacitor
is assumed to inject a reactive power that is the product of the half line charging admittance
and square of the magnitude of the voltage of that at bus. The half line charging admittance of
this line is 0.03. Therefore line charging capacitor will inject
0.03 100 V1

3.3075

MVAr

at bus-1. Similarly the reactive injected at bus-2 will be

0.03 100 V2

2.8968

MVAr

The power flow through the line segments 1-2 and 1-5 are shown in Fig. 4.2.

(a)

(b)
Fig. 4.2 Real and reactive power flow through (a) line segment 1-2 and (b) line segment 1-5. The thin lines
indicate reactive power flow while the thick lines indicate real power flow.

4.8 LOAD FLOW PROGRAMS IN MATLAB

The load flow programs are developed in MATLAB. Altogether there are 4 mfiles that
are attached with this chapter. The program listings and descriptions of these mfiles are given
below. It must however be emphasized that these are not general purpose programs and are
written only for the examples of this chapter.
4.7.1 Forming Ybus Matrix

2.21
This is a function that can be called by various programs. The function can be invoked
by the statement
[yb,ych]=ybus;
where yb and ych are respectively the Ybus matrix and a matrix containing the line charging
admittances. It is assumed that the system data of Table 4.1 are given in matrix form and the
matrix that contains line impedances is zz, while ych contains the line charging
information. This program is stored in the file ybus.m. The program listing is given below.
% Function ybus
% THIS IS THE PROGRAM FOR CREATING Ybus MATRIX.
function [yb,ych]=ybus
% The line impedances are
zz=[0 0.02+0.1i 0 0 0.05+0.25i
0.02+0.1i 0 0.04+0.2i 0 0.05+0.25i
0 0.04+0.2i 0 0.05+0.25i 0.08+0.4i
0 0 0.05+0.25i 0 0.1+0.5i
0.05+0.25i 0.05+0.25i 0.08+0.4i 0.1+0.5i 0];
% The line chargings are
ych=j*[0 0.03 0 0 0.02
0.03 0 0.025 0 0.020
0 0.025 0 0.02 0.01
0 0 0.02 0 0.075
0.02 0.02 0.01 0.075 0];
% The Ybus matrix is formed here
for i=1:5
for j=1:5
if zz(i,j) == 0
yb(i,j)=0;
else
yb(i,j)=-1/zz(i,j);
end
end
end
for i=1:5
ysum=0;
csum=0;
for j=1:5
ysum=ysum+yb(i,j);
csum=csum+ych(i,j);
end
yb(i,i)=csum-ysum;
end

4.7.2 Gauss-Seidel Load Flow

The Gauss-Seidel program is stored in the file loadflow_gs.m. This calls the ybus.m
function discussed above. The program allows the selection of the acceleration factor. The

2.22
program lists the number of iterations required to converge, bus voltages and their
magnitudes and real and reactive power. The program listing is given below.
% Program loadflow_gs
% THIS IS A GAUSS-SEIDEL POWER FLOW PROGRAM
clear all
d2r=pi/180;w=100*pi;
% The Y_bus matrix is
[ybus,ych]=ybus;
g=real(ybus);b=imag(ybus);
% The given parameters and initial conditions are
p=[0;-0.96;-0.35;-0.16;0.24];
q=[0;-0.62;-0.14;-0.08;-0.35];
mv=[1.05;1;1;1;1.02];
th=[0;0;0;0;0];
v=[mv(1);mv(2);mv(3);mv(4);mv(5)];
acc=input('Enter the acceleration constant: ');
del=1;indx*=0;
% The Gauss-Seidel iterations starts here
while del>1e-6
% P-Q buses
for i=2:4
tmp1=(p(i)-j*q(i))/conj(v(i));
tmp2=0;
for k=1:5
if (i==k)
tmp2=tmp2+0;
else
tmp2=tmp2+ybus(i,k)*v(k);
end
end
vt=(tmp1-tmp2)/ybus(i,i);
v(i)=v(i)+acc*(vt-v(i));
end
% P-V bus
q5=0;
for i=1:5
q5=q5+ybus(5,i)*v(i);
end
q5=-imag(conj(v(5))*q5);
tmp1=(p(5)-j*q5)/conj(v(5));
tmp2=0;
for k=1:4
tmp2=tmp2+ybus(5,k)*v(k);

2.23
end
vt=(tmp1-tmp2)/ybus(5,5);
v(5)=abs(v(5))*vt/abs(vt);
% Calculate P and Q
for i=1:5
sm=0;
for k=1:5
sm=sm+ybus(i,k)*v(k);
end
s(i)=conj(v(i))*sm;
end
% The mismatch
delp=p-real(s)';
delq=q+imag(s)';
delpq=[delp(2:5);delq(2:4)];
del=max(abs(delpq));
indx=indx+1;
if indx==1
pause
end
end
'GS LOAD FLOW CONVERGES IN ITERATIONS',indx,pause
'FINAL VOLTAGE MAGNITUDES ARE',abs(v)',pause
'FINAL ANGLES IN DEGREE ARE',angle(v)'/d2r,pause
'THE REAL POWERS IN EACH BUS IN MW ARE',(real(s)+[0 0 0 0
0.24])*100,pause
'THE REACTIVE POWERS IN EACH BUS IN MVar ARE',(-imag(s)+[0 0 0 0
0.11])*100

4.7.3 Solving Nonlinear Equations using Newton-Raphson

This program gives the solution of the nonlinear equations of Example 4.1 using the
Newton-Raphson method. The equations and the Jacobian matrix are explicitly entered in the
program itself. The program gives the number of iterations and the final values of x1, x2 and
x3. The program listing is given below.
%
%
%
%
%
%
%
%
%
%
%

Program nwtraph
THIS IS A NEWTON-RAPHSON PROGRAM
We have to solve three nonlinear equations given by
g1=x1^2-x2^2+x3^2-11=0
g2=x1*x2+x2^2-3x3-3=0
g3=x1-x1*x3+x2*x3-6=0
Let us assume the initial conditions of x1=x2=x3=1
The Jacobian matrix is

2.24
%
% J=[2x1 -2x2
%
x2
x1+2x2
%
1-x3
x3

2x3
-3
-x1+x2];

clear all
x=[1;1;1];
% The Newton-Raphson iterations starts here
del=1;
indx=0;
while del>1e-6
g=[x(1)^2-x(2)^2+x(3)^2-11;x(1)*x(2)+x(2)^2-3*x(3)3;x(1)-x(1)*x(3)+x(2)*x(3)-6];
J=[2*x(1) -2*x(2) 2*x(3);x(2) x(1)+2*x(2) -3;1-x(3)
x(3) -x(1)+x(2)];
delx=-inv(J)*g;
x=x+delx;
del=max(abs(g));
indx=indx+1;
end
'NEWTON-RAPHSON SOLUTION CONVERGES IN
ITERATIONS',indx,pause
'FINAL VALUES OF x ARE',x

4.7.1 Newton-Raphson Load Flow

The Newton-Raphson load flow program is stored in the files loadflow_nr.m. The
outputs of the program can be checked by typing
indx
v
abs(v)
angle(v)/d2r
preal
preac
pwr
qwr
q
pl
ql

the number of iterations

bus voltages in Cartesian form
magnitude of bus voltages
angle of bus voltage in degree
real power in MW
reactive power in MVAr
power flow in the various line segments
reactive power flow in the various line segments
reactive power entering or leaving a bus
real power losses in various line segments
reactive drops in various line segments

It is to be noted that in calculating the power and reactive power the conventions that the
power entering a node is positive and leaving it is negative are maintained. The program
listing for the Newton-Raphson load flow is given below.
% Program loadflow_nr
% THIS IS THE NEWTON-RAPHSON POWER FLOW PROGRAM

2.25
clear all
d2r=pi/180;w=100*pi;
% The Y_bus matrix is
[ybus,ych]=ybus;
g=real(ybus);b=imag(ybus);
% The given parameters and initial conditions are
p=[0;-0.96;-0.35;-0.16;0.24];
q=[0;-0.62;-0.14;-0.08;-0.35];
mv=[1.05;1;1;1;1.02];
th=[0;0;0;0;0];
del=1;indx=0;
% The Newton-Raphson iterations starts here
while del>1e-6
for i=1:5
temp=0;
for k=1:5
temp=temp+mv(i)*mv(k)*(g(i,k)-j*b(i,k))*exp(j*(th(i)th(k)));
end
pcal(i)=real(temp);qcal(i)=imag(temp);
end
% The mismatches
delp=p-pcal';
delq=q-qcal';
% The Jacobian matrix
for i=1:4
ii=i+1;
for k=1:4
kk=k+1;
j11(i,k)=mv(ii)*mv(kk)*(g(ii,kk)*sin(th(ii)-th(kk))b(ii,kk)*cos(th(ii)-th(kk)));
end
j11(i,i)=-qcal(ii)-b(ii,ii)*mv(ii)^2;
end
for i=1:4
ii=i+1;
for k=1:4
kk=k+1;
j211(i,k)=-mv(ii)*mv(kk)*(g(ii,kk)*cos(th(ii)-th(kk))b(ii,kk)*sin(th(ii)-th(kk)));
end
j211(i,i)=pcal(ii)-g(ii,ii)*mv(ii)^2;
end
j21=j211(1:3,1:4);
j12=-j211(1:4,1:3);
for i=1:3

2.26
j12(i,i)=pcal(i+1)+g(i+1,i+1)*mv(i+1)^2;

end

j22=j11(1:3,1:3);
for i=1:3
j22(i,i)=qcal(i+1)-b(i+1,i+1)*mv(i+1)^2;
end
jacob=[j11 j12;j21 j22];
delpq=[delp(2:5);delq(2:4)];
corr=inv(jacob)*delpq;
th=th+[0;corr(1:4)];
mv=mv+[0;mv(2:4).*corr(5:7);0];
del=max(abs(delpq));
indx=indx+1;
end
preal=(pcal+[0 0 0 0 0.24])*100;
preac=(qcal+[0 0 0 0 0.11])*100;
% Power flow calculations
for i=1:5
v(i)=mv(i)*exp(j*th(i));
end
for i=1:4
for k=i+1:5
if (ybus(i,k)==0)
s(i,k)=0;s(k,i)=0;
c(i,k)=0;c(k,i)=0;
q(i,k)=0;q(k,i)=0;
cur(i,k)=0;cur(k,i)=0;
else
cu=-(v(i)-v(k))*ybus(i,k);
s(i,k)=-v(i)*cu'*100;
s(k,i)=v(k)*cu'*100;
c(i,k)=100*abs(ych(i,k))*abs(v(i))^2;
c(k,i)=100*abs(ych(k,i))*abs(v(k))^2;
cur(i,k)=cu;cur(k,i)=-cur(i,k);
end
end
end
pwr=real(s);
qwr=imag(s);
q=qwr-c;
% Power loss
ilin=abs(cur);
for i=1:4
for k=i+1:5
if (ybus(i,k)==0)
pl(i,k)=0;pl(k,i)=0;

2.27
ql(i,k)=0;ql(k,i)=0;
else
z=-1/ybus(i,k);
r=real(z);
x=imag(z);
pl(i,k)=100*r*ilin(i,k)^2;pl(k,i)=pl(i,k);
ql(i,k)=100*x*ilin(i,k)^2;ql(k,i)=ql(i,k);
end
end

end