UNIT II POWER FLOW ANALYSIS

Importance of power flow analysis in planning and operation of power systems.

Power flow study or load flow study

The study of various methods of solution to power system network is referred to as load flow
study. The solution provides the voltages at various buses, power flowing in various lines and
line losses.
Information’s that are obtained from a load flow study
The information obtained from a load flow study is magnitude and phase angle of voltages,
real and reactive power flowing in each line and the line losses. The load flow solution also
gives the initial conditions of the system when the transient behavior of the system is to be
studied.
Need for load flow study
The load flow study of a power system is essential to decide the best operation of existing
system and for planning the future expansion of the system. It is also essential for designing a
new power system.

Statement of power flow problem

Quantities associated with each bus in a system

Each bus in a power system is associated with four quantities and they are real power (P),
reactive power (Q), magnitude of voltage (V), and phase angle of voltage (δ).

Work involved (or) to be performed by a load flow study

(i). Representation of the system by a single line diagram
(ii). Determining the impedance diagram using the information in single line
diagram (iii). Formulation of network equation
(iv). Solution of network equations

Iterative methods to solve load flow problems

The load flow equations are non linear algebraic equations and so explicit solution as not
possible. The solution of non linear equations can be obtained only by iterative numerical
techniques.

Mainly used for solution of load flow study

The Gauss seidal method, Newton Raphson method and Fast decouple methods.

Flat voltage start

In iterative method of load flow solution, the initial voltages of all buses except slack bus
assumed as 1+j0 p.u. This is referred to as flat voltage start
Classification of buses
Bus
The meeting point of various components in a power system is called a bus. The bus is a
conductor made of copper or aluminum having negligible resistance .At some of the buses
power is being injected into the network, whereas at other buses it is being tapped by the
system loads.

Bus admittance matrix

The matrix consisting of the self and mutual admittance of the network of the power system
is called bus admittance matrix (Ybus).

Methods available for forming bus admittance matrix

Direct inspection method.
Singular transformation method.(Primitive network)

Different types of buses in a power system

Types of bus Known or specified quantities Unknown quantities or

quantities to be determined
Slack or Swing or Reference V, δ P,Q
bus
Generator or Voltage control P, V Q, δ
or PV bus
Load or PQ bus P, Q V, δ

Need for slack bus

The slack bus is needed to account for transmission line losses. In a power system the total
power generated will be equal to sum of power consumed by loads and losses. In a power
system only the generated power and load power are specified for buses. The slack bus is
assumed to generate the power required for losses. Since the losses are unknown the real and
reactive power are not specified for slack bus.

Effect of acceleration factor in load flow study

Acceleration factor is used in gauss seidal method of load flow solution to increase the rate
of convergence. Best value of A.F=1.6

Generator buses are treated as load bus

If the reactive power constraint of a generator bus violates the specified limits then the
generator is treated as load bus.
Iterative solution using Gauss-Seidel method - Algorithm

Algorithm of Gauss seidal method

Step1: Assume all bus voltage be 1+ j0 except slack bus. The voltage of the slack bus is
a constant voltage and it is not modified at any iteration

Step 2: Assume a suitable value for specified change in bus voltage which is used
to compare the actual change in bus voltage between K th and (K+1) th
iteration
Step 3: Set iteration count K = 0 and the corresponding voltages are V10, V 0, V 0, ……
2 3
Vn0 except slack bus

Step 4: Set bus count P = 1

Step 5: Check for slack bus. It is a slack bus then goes to step 12 otherwise go to
next step

Step 6: Check for generator bus. If it is a generator bus go to next step. Otherwise go to
step 9

Step 7: Set │VPK│= │V P│ specified and phase of │V PK│ as the K th iteration value if the
bus P is a generator bus where │VP│ specified is the specified magnitude of
voltage for bus P. Calculate reactive power rating
P-1 n
QP K+1 Cal = (-1) Imag [(VPK)A (∑ Y pq Vq k+1 +∑ Y pq Vq K
q=1 q =P

Step 8: If calculated reactive power is within the specified limits then consider the bus as
generator bus and then set QP = QP K+1 Cal for this iteration go to step 10

Step 9 : If the calculated reactive power violates the specified limit for reactive power
then treat
K+1
this bus as load bus
If QP Cal < QP min then QP = QP min
K+1
QP Cal > QP max then QP = QP max

Step10: For generator bus the magnitude of voltage does not change and so for all
iterations the magnitude of bus voltage is the specified value. The phase of
the
busK+1
voltage can be calculated usingK *
V temp = 1 / Y [(P –jQ / V ) - ∑Y V K+1 - ∑Y V K]
P PP P P P pq q pq q

Step 11: For load bus the (k+K1+)1th iteration value of load bus KP *voltage VPKK++11 can be K
calculated using V temp = 1 / Y [(P –jQ / V ) - ∑Y V - ∑Y V ]
P PP P P P
Page 41 of 122

Step 12: An acceleration factor α can be used for faster convergence. If

acceleration factor is specified then modify the (K+1)th iteration value of
bus PK+1
using K
V = V + α (V K+1 – V K) then
Pacc P P P
Set VPK+1 = V PaccK+1
 K+1 K+1
Step 13: Calculate the change in bus-P voltage using the relation Δ V =V
P P
-VPK

Step 14: Repeat step 5 to 12 until all the bus voltages have been calculated. For
this increment the bus count by 1 go to step 5 until the bus count is n

Step 15: Find the largest of the absolute value of the change in voltage
│ΔV1K+1│,│ΔV2K+1│,│ΔV3K+1│,.....................│ΔVnK+1│
Let this largest value be the │ΔV max│. Check this largest change │ΔV max│ is
less than pre specified tolerance. If │ΔV max│ is less go to next step. Otherwise increment
the iteration count and go to step 4

Step 16: Calculate the line flows and slack bus power by using the bus voltages

Gauss - Seidal method flow chart

Page 42 of 122
 X

Set iteration count K = 0

Set bus count P = 1

Temp Set │ Vp │K spec = │Vp│spec calculate reactive power Qpcal K+1

Page 43 of 122
 Check
for slack bus A

NO

Check for Generator bus

YES

Set Qp =Qp min Check if

treat as load bus QpcalK+1
< Qpmin

Calculate YpK+1, NO
Qp K+1 acc B

Page 44 of 122
 Y

YES NO

Qp = Qpmax treat as load bus Check if Qpcal Qpmax

K+1>

B Treat the buses as generator buses and calculate δ PK+1.

Set VPK+1=│VP│spec < δPK+1

Calculate ΔVPK+1

Set VPK = VPK+1

Advance bus count by set K = K+1

A

Set bus count P = 1

NO
YES
Checkif allthebus takeninto
Evaluate ΔVmax which largestaccount D
ΔVP
is
of
K+1

E
Page 45 of 122
 E

NO

Check if ΔVmax Set K = K+1

<ε

YES

Evaluate line flows and slack bus power

Stop

Advantages and disadvantages of Gauss-Seidel method

Advantages: Calculations are simple and so the programming task is lessees.

The memory requirement is less. Useful for small systems;

Disadvantages: Requires large no. of iterations to reach converge .Not suitable for
large systems. Convergence time increases with size of the system

Page 45 of 122
Iterative solution using Newton-Raphson method – Algorithm

Step 1: Assume a suitable solution for all buses except the slack bus. Let Vp = a+j0 for P

= 2,3,……n V1 = a+j0

Step 2 : Set the convergence criterion = ε0

Step 3 : Set iteration count K= 0

Step 4 : Set bus count P = 2

Step 5 : Calculate Pp and Qp using

Pp = ∑ { ep(epGpq+fpBqp)+fp(fpGpq – epBpq)}

q=1

Qp = ∑ { fp(epGpq+fpBqp)+ep(fpGpq – epBpq)}

q=1

K
Step 6 : Evaluate ΔP =P - P PK
P spec

Step 7 : Check if the bus is the question is a PV bus. If yes compare QPK with the limits.

If it exceeds the limit fix the Q value to the corresponding limit and treat the

bus

Page 46 of 122

as PQ for that iteration and go to next step (or) if the lower limit is not violated
P
 evaluate │ΔVP│2 = │ Vspec│2 - │V K│2 and go to step 9

K K
Step 8: Evaluate ΔQ =Q -Q
P spec P

Step 9 : Advance bus count P = P+1 and check if all buses taken in to account if not go

to step 5

Step 10 : Determine the largest value of │ΔVP│2

Step 11: If ΔVP < ε go to step 16

Step 12: Evaluate the element of Jacobin matrices J1, J2, J3, J4, J5 and J6

Step 13: Calculate ΔePK and Δf PK

K+1 K K K+1 K K
Step 14: Calculate e =e + Δe and f =f + Δf
P P P P P P

Step 15 : Advance count (iteration) K=K+1 and go to step 4

Step 16: Evaluate bus and line power and print the result

Page 47 of 122
Iterative solution using Newton-Raphson method – Flow chart

Start

Read the system date and formulate the nodal admittance matrix

Advance bus voltage Vp = 1+j0 for P = 2,3,……n V1 = a+j0

D
Set the convergence criterion = ε0

Set iteration count K= 0

Set bus count P = 2

Calculate Pp and Qp using

Pp = ∑ { ep(epGpq+fpBqp)+fp(fpGpq – epBpq)} q=1

Page 48 of 122
 E

Evaluate ΔPPK =
Pspec - PPK

NO
ChecK PV bus

YES
YES

Set Qpk = Qpmin If

Qp<Qmin

NO
YES

Evaluate │ΔVp│2 =
Set Qpk = Qpmax If
Qp>Qmax │Vps│2 - │Vp│2

Evaluate ΔQPK =
Advance bus count P= P+1
Qspec - QPK

Page 49 of 122

SCE Department of Electrical and electronics Engineering

 B

NO

Check
C P≥n

YES

Determine the largest value of │ΔVP│2

YES

IfΔVP
< ε0 Evaluate bus and line power

NO

Evaluate the element of Jacobin matrices

END
Calculate ΔePK and ΔfPK

Calculate ePK+1
= eP + KΔeP and
K
fP= fK+1 K
P + ΔfP
K

Advance count (iteration) K=K+1

Page 50 of 122

VMKVEC Department of Electrical and electronics Engineering

 B

Page 51 of 122

VMKVEC Department of Electrical and electronics Engineering

 Advantages and disadvantages of N.R method

Advantages: Faster, more reliable and results are accurate, require less number of
iterations;

Disadvantages: Program is more complex, memory is more complex.

Iterative solution using Fast Decoupled Load flow method – Algorithm

Step 1: Assume a suitable solution for all buses except the slack bus. Let Vp =1+j0

for P=2,3,.............n and V=a+j0

Step2: Set the convergence criterion = ε0

Step3: Set iteration count K = 0

Step 4: Set bus count P = 2

Step 5: Calculate Pp and Qp using

Pp = Σ |VpVqYpq| cos (θpq+δP-δq)

q=1

Qp = Σ |VpVqYpq| sin (θpq+δP-δq)

q=1

Step 6: Compute the real and reactive power mismatches ΔPK and ΔQK. If the

mismatches Are with in desirable tolerance the iteration end

Page 52 of 122

VMKVEC Department of Electrical and electronics Engineering

Step 7: Normalize the mismatches by dividing each entry by its respective bus
K K K
voltage magnitude ΔP
2 =ΔP
2 /V

ΔP 3K / V 3 K

ΔP nK / V n K

ΔQK =ΔQ 2K / V2 K

ΔQ 3K / V 3 K

ΔQ nK / V n K

Step 8: Solve for the voltage magnitude and the correction factors ΔVK and ΔδK by using
the constant matrices B’ and B” which are extracted from the bus admittance matrix Y
Bus

[B’] ΔδK = ΔPK

[B”]ΔQK = ΔQK

Step 9: Up date the voltage magnitude and angel vectors

δK+1 = δK +ΔδK

VK+1 = VK +ΔVK

Page 53 of 122

VMKVEC Department of Electrical and electronics Engineering

Step 10: Check if all the buses are taken into account if yes go to next step otherwise

go to next step. Otherwise go to step 4

Step 11: Advance iteration count K = K+1 go to step 3

Step 12: Evaluate bus and load powers and print the results

Iterative solution using Fast Decoupled Load flow method – Flow chart

Start

Read in system data and formulate nodal admittance matrix

Assume bus voltage Vp(0) = 1+j0 for P = 1,2,.n

Set the convergence = ε0

Set iteration count K = 0

Set bus count P = 2

Calculate Pp and Qp using formula

Page 54 of 122

VMKVEC Department of Electrical and electronics Engineering

 X

Calculate ΔPK =Pspec-Ppk

ES
Y

Test for
convergence A
is |ΔPK |<εp

NO

Normalizethemismatchesby
dividing each entry by its respective
bus voltage ΔP iK = ΔP Ki / V Ki

Solve for ΔδK using [B’] ΔδK = ΔPK

Calculate δK+1 = δK +ΔδK

Calculate ΔQPK = Qpspec - QPK

Page 55 of 122

VMKVEC Department of Electrical and electronics Engineering

 Y

YES

Test for converge nce is

|ΔQK
A

NO

Normalizethemismatchesby
dividing each entry by its respective
bus voltage ΔP PK = ΔP PK / V PK

Solve for ΔVpK using [B”] ΔδK = ΔPK

Advance K=K+1

Page 56 of 122

VMKVEC Department of Electrical and electronics Engineering

 Calculate VpK+1 = VpK +ΔQpK

NO YES
Check if
P≥n A

Calculate slack bus power and line Stop

flows

Page 57 of 122

VMKVEC Department of Electrical and electronics Engineering

 Compare the Gauss seidel and Newton raphson methods of load flow study

S.No G.S N.R FDLF

1 Require large number Require less number Require more number

of iterations to reach of iterations to reach of iterations than N.R
convergence convergence. method

2 Computation time per Computation time per Computation time per

iteration is less iteration is more iteration is less

3 It has linear It has quadratic ….

convergence convergence
characteristics characteristics

4 The number of The number of The number of

iterations required for iterations are iterations are does not
convergence increases independent of the dependent of the size
with size of the size of the system of the system
system

5 Less memory More memory Less memory

requirements requirements. requirements than
N.R.method.

Page 58 of 122

VMKVEC Department of Electrical and electronics Engineering

 Two marks Questions

1. What is power flow study?

The study of various methods of solution to power system network is reffered to as power
flow or load flow study.

2. What are the informations that are obtained from a load flow study?
The information obtained from a load flow study are magnitude and phase of bus voltages,
real and reactive power flowing in each line and the line losses. The load flow solution also gives the
initial conditions of the system when the transient behaviour of the system is to be studied.

3. What is the need for load flow study?

The load flow study of a power system is essential to decide the best operation of existing
system and for planning the future expansion of the system. Also essential for designing a new
system.

4. What are the works involved in a load flow study?

(i)Representation of the system by single line
diagram.
(ii)Determining the impedance diagram using the information in single line diagram.
(iii)Formulation of network equations.
(iv)Solution of network equations.

5. What are the quantities that are associated with each bus in a system?
Each bus in a system are associated with four quantities and they are real power,reactive
power, magnitude of voltage and phase angle of voltage.

6. What are the different types of

buses? (i)Load bus or PQ bus
(ii)Generator bus or voltage controlled bus or PV bus
(iii)Slack bus or Swing bus or Reference bus.

7. Define voltage controlled bus.

A bus is called voltage controlled bus if the magnitude of voltage and real power are
specified. The magnitude of voltage is not allowed to change.

8. What is PQ bus?
A bus is called PQ are load bus when real and reactive components of power are specified for
the bus. In a load bus the voltage is allowed to vary within the permissible value.

9. What is swing bus?

A bus is called Swing bus when the magnitude and phase of the voltage are specified for it.
Page 59 of 122

VMKVEC Department of Electrical and electronics Engineering

Swing bus is the reference bus for load flow solution and it is required for accounting line losses.

Page 60 of 122

VMKVEC Department of Electrical and electronics Engineering

10. What is the need for slack bus?
The slack bus is needed to account for transmission losses.
11. What are the iterative methods used for solution of
load flow problems? Guass-Seidal (GS)
method and Newton Raphson(NR) method.

12. Why do we go for iterative methods to solve load problems?

The load flow equations are nonlinear algebraic equations and so
explicit solution is not possible. The solution of nonlinear equations can be
obtained only by iterative numerical techniques.

13. When the generator bus is treated as load bus?

If the reactive power of a generator bus violates the specified limits then
the generator bus is treated as load bus.

14. What are the advantages of GS method?

(i) Calculations are simple and so the
programmimg task is lesser. (ii)The memory
requirement is less.
(iii)Useful for small systems.

15. What are the disadvantages of GS method?

(i) Requires large number of iterations to reach convergence.
(ii) Not suitable for large systems.
(iii) Convergence time increases with size of the system.

16. What are the advantages of NR method?

(i) Faster, more reliable and the results are accurate.
(ii) Requires less number of iterations to reach convergence.
(iii) The numbers of iterations are independent of the size of the system.
(iv) Suitable for large systems.

17. What are the

disadvantages of
NR method?
(i)The
programming is
more complex.
(ii)Memory
requirement is
more.
(iii)Computation time per iteration is higher due to large number of calculations per
iterations.

Page 61 of 122
Page 62 of 122