9

__________________________________________________________
Load Flow Analysis

Introduction
In the early days of the development of electrical power engineering, generators were physically close
to the load they supplied. Central generating stations were established in many towns and cities for
the purpose of supplying power to local customers. By the 1920s many different companies operated
such generating stations each with their own standard for supply. In some cities a direct current supply
was available, in others alternating current at 12.5 Hz, 25 Hz, or 50 Hz might be generated. At this
time there was no need for load flow analysis as the electrical power systems were small, isolated and
generation matched load. The lack of a standard supply severely hampered the development of the
domestic appliance market.

In 1926, a British Act of Parliament established the Central Electricity Board with the specific objective
of standardizing the supply voltage and frequency. The Board also oversaw the interconnection of the
most efficient generating stations by a 132 kV high-voltage network. Interconnection between
generating stations and controlling the real and reactive powers flowing along the interconnectors then
became a priority. This required an understanding of the basics of load flow analysis.

In 1948 the British Electricity Supply Industry was nationalized and two organizations were established
in England and Wales (1) the area boards which were mainly concerned with distribution and
customer services and (2) the generating boards which were responsible foe the generation of
electricity and the operation of the high voltage transmission network. In Scotland two boards were
established, the North of Scotland Hydro Electricity Board and the South of Scotland Electricity Board.
There boards were responsible for all aspects of generation, transmission and distribution within their
geographical areas. By the 1980s the electrical network in the United Kingdom was fully
interconnected with transmission voltages increased to 275 kV and 400 kV (the super grid). Electrical
power generation was concentrated at large power stations directly connected to either the 275 kV or
400 kV network and the direction of real power flow was from the H-V network down to the
distribution network (132 kV, 33 kV, 11 kV etc). Load flow analysis of this complex interconnected
network had become an essential tool to aid electrical power systems engineers in the daily running of
the power system.

At the present time (2001) the British electricity supply industry has been privatized the move away
from concentrated large scale coal, oil, and nuclear generation to small scale distributed generation
using renewable energy sources has led to an increase in the number of load flow studies performed.
Generation is now distributed at all voltage levels within the power system and out with the direct
control of a single grid control centre.

341
Load Flow

The load flow calculation is used to ensure that the steady state voltages and currents flowing the
interconnected network are within the continuous ratings of the equipment for both normal and
emergency operating conditions. Such calculations are performed on computer as hand calculations
are too time consuming and impractical for large systems.

In this Chapter we will consider relatively simple power systems with two, three, or four bus-bars to
illustrate the techniques used in practical load flow analysis. In a real load flow study an analysis
involving perhaps 2000 or more bus-bars is not uncommon.

Learning Objectives

To understand the term load flow analysis as applied to an interconnected power system

To understand the factors which influence the flow of real and reactive powers over an inter-connector

To understand that the flow of both real power and reactive power contribute to the real power losses
in the transmission network

To understand the principle aims and objectives of a load flow study

To understand that the load flow problem is non-linear and cannot be solved by classical circuit
analysis

To understand how to construct the bus admittance matrix for a simple power system

To understand the Gauss-Seidel Method as applied to the iterative solution of simultaneous equations

To understand the three different types of bus bar in a load flow study. To understand which
quantities are known and which are unknown for each type at the beginnings of the calculation

To understand how to calculate the voltages in a power system using the Gauss-Seidel Method

To understand how to calculate the real and reactive powers on a generator bus bar once all the bus
bar voltages have been found.

342
Load Flow

9.1 Power and VAr flow in an Inter-connector

Consider the two bus-bar power system shown in Figure 9.1. Two generators and their
local loads are connected by a short transmission line which interconnects them.
Without the transmission line connecting the bus-bars 1 and 2, the generator G1 could
only supply the local load L1 while the generator G2 could only supply the load L2. By
interconnecting the bus-bars, it is possible for the generators to share both real and
reactive powers.

The question then arises, how much real power and reactive power is transmitted over
the inter-connector?

Figure 9.1 Two interconnected generating stations

If we consider the injection of real power P1 and reactive power Q1 at bus-bar 1and the
delivery of real power P2 and reactive power Q2 at bus-bar 2, we can derive equations
of the real and reactive power flows along the transmission line.

Figure 9.2 Directions for P and Q at each end of the interconnector

343
Load Flow

Figure 9.3 Phasor diagram showing voltages at each end of the interconnector

Assuming the inter-connector can be regarded as a short transmission line we can

neglect the shunt capacitance and leakage resistance to earth. The current flowing
along the line per phase is therefore given by:

V1 − V2
I= (9.1)
Z

where Z is the series impedance per phase of the line Z = (R + jX) and V1 and V2 are
the phasors representing the phase voltages on bus-bar 1 and bus-bar 2 as shown in
Figure 9.3. Note that the angle between the phasors V1 and V2 which is denoted by δ
is referred to as either the transmission angle, the load angle, or the power angle. (It is
unfortunate that three different names are in common use to describe the same
quantity.)

The complex power per phase at bus-bar 1 is given by:

V1∗ − V2∗
S1 = P1 + jQ1 = V1 I * = V1 (9.2)
Z∗

If the line resistance is small then in the limit as R tends to zero, then Z → jx and
Z ∗ → − jx . The expression for the complex power per phase S1 can now be rewritten
as:

2
V1* − V2* V1 V1* − V1 V2* V1 − V1 V2 e
jδ

S1 = V1 = = (9.3)
Z* − jx − jx

Expanding using Euler's Rule e jδ = cos δ + j sin δ yields:

344
Load Flow

2
V1 − V1 V2 cosδ − j V1 V2 sinδ
S1 =
− jx

Rearranging this equation gives:

2
V1 V2 sinδ V1 − V1 V2 cosδ
S1 = +j (9.4)
x x

Equation (9.4) and equation (9.2) are both expressions for the complex power S1.
Hence we can write the following.

2
V1 V2 sinδ V1 − V1 V2 cosδ
S1 = P1 + jQ1 = +j (9.5)
x x

By inspection of equation (9.5) it becomes apparent that we can write the following
equations for the real power and the reactive power per phase injected into bus-bar 1
by generator 1:

V1 V2 sin δ
P1 = Watts/phase (9.6)
x
2
V1 − V1 V2 cosδ
Q1 = VArs/phase (9.7)
x
where V1 is the voltage per phase at bus-bar 1
V2 is the voltage per phase at bus-bar 2
X is the reactance of the transmission line per phase
δ is the transmission angle

In deriving equations (9.11) and (9.12) we assumed that the series resistance of the line
was zero. This implies that real power loss in the line is zero, so the real power
injected at bus-bar 1 must equal the real power received at bus-bar 2 i.e.:

V1 V2 sin δ
P = P1 = P2 = (9.8)
x

Also if the reactive power injected in the transmission line at bus 1 is

345
Load Flow

2
V1 − V1 V2 cosδ
Q1 = (9.9)
x

then by symmetry we see that the reactive power at bus 2 is

2
V1 V2 cosδ − V2
Q2 = (9.10)
x

Here we note that Q1 ≠ Q 2 . This is due to the reactive power required to magnetize
the series reactance of the transmission line.

Having derived equations (9.8 – 9.10) we can now attempt to understand the important
mechanism whereby the real power flow and the reactive power flow may be
controlled in both magnitude and direction.

9.1.1 Control of Real Power Flow over inter-connector

The reader must remember that the transmission angle δ was defined as the phase
angle between V1 and V2. Note that the voltages and reactance must be given in per-
phase values to yield per-phase values of power. The reactance of the line X is a fixed
parameter. If we assume automatic excitation control of the generators at each end of
the line so as to maintain a constant terminal voltage magnitude, i.e. |V1| and |V2| are
fixed, then equation (9.8) shows that the only way in which P can be made to change is
to change the transmission angle δ.

Figure 9.4 shows how equation (9.8) can be represented by the power-delta curve for
fixed magnitudes of V1 and V2. When the transmission angle δ is zero, no power is
transferred over the inter-connector. If the phase angle of V1 is advanced relative to
V2, the value of δ increases so the power flow increases up to a maximum value at
δ=90o given by

V1 V2
Pmax = (9.11)
X

If we were to attempt to increase δ further beyond 90o the power transfer will
decrease. At this point the transmission collapses; the two generators at each end of
the inter-connector would fall out of synchronism, this means that generator G1 and
the load on bus-bar 1 would run at a different frequency from generator G2 and the
load on bus-bar 2. A transmission angle of δ = 90o corresponds to the transmission
limit or static stability limit of the line. In practice, δ is normally limited to 300 to

346
Load Flow

ensure system stability during faults and other transients. (A 3rd power limit exists, this
is the level at which the line heating limit is reached, this is much higher than Pmax)

Figure 9.4 Power-Delta Curve for a short inter-connector

If the transmission angle δ is increased in the negative sense (the phase angle of V1 now
lags behind V2) then figure 9.4 shows that the power flow becomes negative. The
direction of power transmission has now reversed.

Note the important difference between ac and dc transmission. In a dc system the

sending-end voltage magnitude must exceed that of the receiving-end voltage in order
for power to flow in the proper direction. In an ac system the voltage magnitudes do
not determine the power direction but the voltage phase difference does. By definition
real power flows from leading to lagging phasor. This is illustrated in diagrams
9.5 and 9.6.

Figure 9.5 Real power flow Figure 9.6 Real power flow
from bus1 to bus 2 from bus 2 to bus 1
347
Load Flow

Example 9.1

A 132 kV three-phase transmission line is 50 miles long and has a reactance of 40 Ω per
phase. If the line is to be operated as a “flat line” at rated voltage (i.e. the sending
voltage magnitude is equal to the received voltage magnitude), determine (a) the
maximum possible steady state power flow and (b) the transmission angle δ if the real
power flow is 200 MW.

Solution 9.1

Part a)
As this is a three phase transmission line, the given voltage is the rated line to line
132
voltage. Working per phase, the rated phase voltage is therefore kV = 76.2 kV .
3
The sending voltage is therefore 76.2 kV and the received voltage is 76.2 kV since this
is a flat line. The maximum steady state power that can be transmitted per phase is
given by equation (9.11)

V1 V2 76.2 ⋅ 10 3 × 76.2 ⋅ 10 3
Pmax = = = 145 MW
X 40

The maximum steady state three-phase power is therefore:

Pmax = 3 × 145 MW = 435 MW

Part b)
200 MW
If the real power flow is 200 MW, the power per phase is = 66,67 MW
3
The voltage per phase at the sending and receiving ends is still 76.2 kV, the required
transmission angle is found from equation (9.8)

V1 V2 sin δ
P=
x
76.2 × 10 3 ⋅ 76.2 × 10 3
66.67 × 10 6 = ⋅ sinδ
40
δ = 27.3o
The transmission angle required to transfer 200 MW over the line is therefore 27.3o.

Example 9.2

Repeat Example 9.1, using a per-unit calculation.

348
Load Flow

Solution 9.2

Part a)

Choose 200 MVA as Sbase and 132 kV as Vbase and calculate Zbase.

(Vbase ) 2 (132 ⋅ 10 3 ) 2
Z base = = = 87.12 Ω
S base 200 ⋅ 10 6

The reactance of the 40 Ω transmission line in per unit is therefore

40
Z pu = = 0.459 pu
87.12

The voltage at the sending end is 132 kV or 1 pu. The voltage at the receiving end is
also 1 pu. The maximum power transfer is given by equation (9.11)

V1 V2 1.0 × 1.0
Pmax = = = 2.178 pu
X 0.459

Part b)
200MW
If the real power flow is 200 MW, in per unit this is = 1.0pu
200MVA

The required transmission angle is found from equation (9.8)

V1 V2 sin δ
P=
x
1 .0 × 1 .0
1.0 = ⋅ sinδ
0.459
δ = 27.3 o

Examples 9.1 and 9.2 show that the equations derived above for real and reactive
powers along an inter-connector are valid for both per phase and per unit calculations.

9.1.2 Control of Reactive Power over inter-connector

In our discussions on the control of the real power flow over the inter-connector we
noted that in practice the transmission angle δ is usually limited to values less than 30o,
to ensure stability during transients and other disturbances. In general we can

349
Load Flow

therefore assume that the transmission angle is relatively small and cos δ equal to one
(cos 0o = 1, cos 30o = 0.866) Equations (9.9) and (9.10) can therefore be simplified:

2 2
V1 − V1 V2 cosδ V1 − V1 V2 V1
Q1 = = = (V
1 − V2 ) (9.12)
x x x

2 2
V1 V2 cosδ − V2 V1 V2 − V2 V2
Q2 = = = (V 1 − V2 ) (9.13)
x x x

The factor inside the parenthesis in the equations for Q1 and Q2 shows us that the
reactive power flow tends to be proportional to the difference in voltage
magnitudes.

By definition reactive power tends to flow in the direction of lowest voltage

amplitude. The larger the voltage difference, the stronger the reactive power flow.

This reactive power flow tendency is perhaps more clearly seen by considering the
average reactive power flow:

2 2
Q + Q2 V1 − V2
Q ave ≡ 1 = (9.14)
2 2x

Remember that Q1 is not equal to Q2 since reactive power is absorbed by the series
reactance of the transmission line.

Figure 9.7 shows the phasor relationship for a small power flow from 1 to 2, and a large
reactive power flow from 2 to 1.

Figure 9.7 Phasor relationship

Example 9.3

Consider again the 132 kV transmission line in examples 9.1 and 9.2. Calculate the
reactive power flow in the line for a real power flow of 200 MW with the following
voltage profiles:

350
Load Flow

Case a) V1 = V2 = 1.0pu
Case b) V1 = 1.2pu, V2 = 1.0pu
Case c) V1 = 1.0pu, V2 = 1.2pu

Solution 9.3

In example 9.2 using 200 MVA and 132 kV as base quantities, the per unit impedance of
the 40 Ω transmission line was shown to be 0.459pu.

Case a)
The transmission angle corresponding to 1pu power was calculated to be 27.3o.in
example 9.2. Using equations (9.9) and (9.10), the per-unit reactive powers can be
calculated.
(1.0)2 − (1.0) ⋅ (1.0)cos(27.3) = 0.243 pu
2
V1 − V1 V2 cosδ
Q1 = =
x 0.459

(1.0 ) ⋅ (1.0)cos(27.3) − (1.0)

2
V1 V2 cosδ − V2 2
Q2 = = = −0.243 pu
x 0.459

Note that when the voltage profile of the line is kept flat, reactive power flows into the
line from both ends. Also note that the simplified equations for Q1 and Q2 have not
been used for this calculation.

Case b)
First we must calculate the transmission angle δ. Using equation (9.8)

V1 V2 sin δ
P=
x
1.2 × 1.0
1.0 = ⋅ sinδ
0.459
δ = 22.5 o

Hence using equations (9.9) and (9.10), the new per-unit reactive powers can be
calculated.

(1.2 ) − (1.0 ) ⋅ (1.2)cos(22.5)

2
V1 − V1 V2 cosδ 2
Q1 = = = 0.721 pu
x 0.459

351
Load Flow

(1.2) ⋅ (1.0)cos(22.5) − (1.0)2

2
V1 V2 cosδ − V2
Q2 = = = 0.236 pu
x 0.459

Case c)
The transmission angle δ will be the same as in Case b), i.e. δ = 22.5o. Using equations
(9.9) and (9.10) as before:

(1.0) − (1.2) ⋅ (1.0)cos(22.5)

2
V1 − V1 V2 cosδ 2
Q1 = = = −0.236 pu
x 0.459

(1.0) ⋅ (1.2)cos(22.5) − (1.2)2

2
V1 V2 cosδ − V2
Q2 = = = -0.721 pu
x 0.459

In this case the reactive power flow is from bus 2 to bus 1. The results of these
calculations are shown in graphical form in Figure 9.8. In is interesting to note that by
changing the voltage levels of the two bus-bars, a strong effect is noted in the reactive
power flow but not effect whatsoever on the real power flow. There is however a
slight effect on the transmission angle δ. This example confirms the claim that the
magnitude of the reactive power flow depends on the voltage magnitude and that the
reactive power flows towards the low voltage bus-bar.

Figure 9.8 Varying reactive power flows

352
Load Flow

Real Power Losses in an Inter-connector

The real and reactive power flows shown in Figure 9.8 which refers to Example 9.3
show a substantial reactive power loss on the line. This reactive power is consumed by
the series reactance of the line. Equations (9.8), (9.9), and (9.10) were derived on the
assumption of zero resistance and the real power losses were zero. The effect of the
shunt capacitance of the line was also neglected.

In practice a transmission line will have a certain series resistance which will cause a
real power loss which can be expressed as:

2
PΩ = R I (9.15)

This power loss is of much more practical significance that the reactive power loss.
Due to the distributed transmission line series impedance and shunt capacitance the
voltage, current, and powers vary along the length of the line.

Let Vave, Iave, Pave, and Qave represent average values measured at the mid point of the
line. Theses variable are related by the following equation:

*
Pave + jQ ave = Vave ⋅ I ave

The complex conjugate of the average current can therefore be written as

* Pave + jQ ave
I ave =
Vave
and
Pave − jQ ave
I ave = *
Vave

Multiplying these equations yields

2 2
* 2 Pave − jQ ave Pave + jQ ave Pave + Q ave
I ave ⋅ I ave = I ave = *
⋅ = 2
Vave Vave Vave

Substituting this expression into equation (9.15) gives the following approximate loss
formula:

2 2
2 Pave + Q ave
PΩ ≅ R I = R 2
(9.16)
Vave

353
Load Flow

The formula is important because it shows us that the real and the reactive line powers
contribute equally to the real power loss in the line. From the point of view of power
loss we should therefore minimize the reactive power flow in the line.

Example 9.4

A three-phase 140kV transmission line has the following parameters per phase:
resistance = 9.10 Ω, inductive reactance = j50.5 Ω, total shunt admittance = j334 µΩ-1.
a) Using the π model of the line calculate the sending end conditions if the
magnitude of the receiving end voltage is 140 kV and the load complex power is
S = 100 + j60 MVA.
b) Use the sending and receiving end complex powers to calculate the power loss
in the line and show that equation (9.16) can be used to estimate the
approximate line loss.

Solution 9.4

Part a)

The solution of this problem requires a five step approach similar to that used in
Chapter 5.

Step 1 Draw the per-phase circuit diagram and mark all voltages, currents,
real and reactive powers.

Step 2 Calculate the receiving end current, remembering to work per phase

The complex power at the load is:

354
Load Flow

100 60
S 2 = V2 ⋅ I 2 * = + j = 33.33 + j20.00 MVA per phase
3 3
⎛ 140 ⎞
If V2 is the reference phasor then, V2 = ⎜⎜ + j0.0 ⎟⎟kV = (80.82 + j0.0)kV
⎝ 3 ⎠

The received current is therefore given by

33.33 + j20.00
I2 = = (412.4 − j247.4) A /phase
80.82

Step 3 Find the sending end voltage

The current Ish2 is given by

( )
I sh2 = V2 ⋅ Y2 = 80.82 × 10 3 j167 ⋅ 10 −6 = j13.5 A /phase

The line current IL is now calculated by adding the load current and the shunt
current.
I L = I 2 + I sh2 = (412.4 − j247.4 ) + j13.5 = (412.4 − j233.9 ) A /phase

The voltage drop across the series branch is

∆V = I L ⋅ (R + jX ) = (412.4 − 233.9) ⋅ (9.10 + 50.5) = (15.56 + j18.70) kV /phase

The sending end voltage is then calculated

V1 = V2 + ∆V = 80.82 + (15.56 + j18.70 ) = (96.39 + j18.70 ) kV /phase

Step 4 Find the sending-end current

The current Ish1 is given by

( )
I sh1 = V1 ⋅ Y1 = (96.39 + j18.70 ) ⋅ 10 3 ⋅ j167 ⋅ 10 −6 = (− 3.1 + j16.1) A /phase

The sending end current therefore equals

I1 = I sh1 + I L = (409.3 − j217.8) A /phase

Step 5 Find the sending end power

S1 = V1 ⋅ I 1 = (96.39 + j18.70) ⋅ 10 3 ⋅ (409.3 − j 217.8)

*

= (35.38 + j 28.64) ⋅ 10 6

355
Load Flow

The sending complex power is therefore 35.38 MW and 28.64 MVAr per phase.
The total three phase real and reactive powers are 106.14 MW and
85.93 MVAr.

Part b)

The sending real power is 106.14 MW and the received real power is 100.00MW. The
difference between these must be the power lost in transmission, i.e 6.14 MW.

To use equation (9.16) the average real power, reactive power and voltage must be
found.

The magnitude of the received voltage per phase is 80.82 kV.

The magnitude of the sending voltage per phase is

V1 = (96.39 + j18.70 ) 10 3 = 96.39 2 + 18.70 2 ⋅ 10 3 = 98.18 kV

The average voltage is therefore

80.82 + 98.18
Vave = = 89.5 kV per phase
2

The average real power per phase is

33.33 + 35.38
Pave = = 34.35 MW /phase
2

The average reactive power per phase is

20.00 + 28.64
Q ave = = 24.32 MW /phase
2

The power loss per phase in the transmission line is therefore

PΩ ≅ R I = R
2 P2 + Q2
= 9.10
(34.35 ⋅10 ) + (24.32 ⋅10 )
3 2 3 2
= 2.01 MW /phase
V
2
(89.5 ⋅ 10 )
3 2

The total three-phase power loss is therefore 2.01 x 3 = 6.03 MW.

356
Load Flow

The use of the average voltage, real power and reactive power to calculate the loss
returns a value of 6.03 MW, the accurate calculation returns a value of 6.14 MW. The
error in using the approximation is less that 2%.

Tutorial Questions

9.1.1 A 140 kV three-phase transmission line has a reactance of 50 Ω per phase. If the line is to be operated
as a “flat line” at rated voltage (i.e. the sending voltage magnitude is equal to the received voltage
magnitude), determine (a) the maximum possible steady state power flow and (b) the transmission angle
δ if the real power flow is 100 MW.

9.1.2 A 220 kV, three-phase transmission line has a series reactance of 140 Ω per phase and negligible shunt
admittance. Calculate the transmission angle δ and the reactive power flow in the line for a real power
flow of 40 MW with the following voltage profiles:
Case a) V1 = V2 = 1.0pu

Case b) V1 = 1.20pu, V2 = 1.0pu

9.1.3 A three-phase 140kV transmission line has the following parameters per phase: resistance = 9.10 Ω,
inductive reactance = j50.5 Ω, total shunt admittance = j334 µΩ-1. The sending end voltage is kept
constant at 145 kV. The three-phase sending end power equals 120 MW at a power factor of 0.8 lagging.

Using the π model of the line calculate:

a) the current at each end of the line

b) the voltage at the receiving end of the line
c) the three-phase power at the receiving end of the line
d) the efficiency of the transmission.

9.1.4 Show that the power loss in a transmission line can be estimated using the following equation:
P2 + Q2
PL = R 2
V
where R is the series resistance of the line per phase
P is the average of the sending and receiving powers per phase
Q is the average of the sending and receiving reactive powers per phase
V is the average of the sending and receiving voltage magnitudes per phase

Use the calculated real and reactive powers and voltages per phase at the sending and receiving ends of
the transmission line in question 9.1.3. to estimate the power loss per phase and efficiency of the three-
phase transmission.

357
Load Flow

9.2 The Load Flow Problem

In the previous section, we have seen how the real power flowing over an inter-
connector is a function of the difference in voltage phase angle between the two ends
of the inter-connector. The reactive power flow is a function of the difference in
voltage magnitude between the two ends of the inter-connector.

In practice electrical power systems have many inter-connections and calculating the
real and reactive power flows in the system is not a simple task. The purpose of load
flow analysis is to determine the actual real and reactive power flow in a given system,
and more importantly how to control these power flows.

The primary aims and objectives of any load flow study are as follows:

1. Find the real and reactive power flows in the transmission lines based on certain
assumptions concerning the nature of the loads and the generators on the
system.

2. Calculate the voltages on all bus-bars in the power system.

3. Check that no cable, transmission line, transformer, or generator is overloaded.

4. Find the optimum configuration of the power system to minimise transmission

losses.

Any electrical engineer will immediately identify the load flow problem as an electrical
circuits problem. The standard approach to solving such problems is to first represent
the sources in the network as either voltage or current sources. Network equations
are then written down using Kirchhoff’s Laws in which either the voltages or the
currents take on the role of unknown quantities. The loads are represented by
impedances, and if these are assumed known and constant, the resultant equations are
linear and can be easily solved. In this problem the source e.m.f. is known as is the
circuit impedances.

In an electrical power system, the loads do not behave in a simple manner and cannot
generally be represented by constant impedances. The e.m.f.s of each generator is not
known and cannot be measured since it is internal to the generator. In the load flow
problem we therefore have to establish a system of equations which describe the
network in terms of quantities which are easily measured. These quantities are real
powers, reactive powers and voltage magnitudes.

A load flow study therefore involves network equations written in terms of voltages
and power, not voltages and currents as in a classical circuits problem. This difference
results in the load flow equations being non-linear. This eliminates the possibility of a

358
Load Flow

direct solution to the load flow equations in most cases. We can however always
arrive at an iterative numerical solution to solve this system of non-linear equations.

Load flow analysis of a power system containing hundreds of bus-bars and transmission
lines is therefore a rather complex task always performed using a computer program
and beyond the scope of the introductory text. It is possible however to consider
relatively simple power systems to illustrate the techniques required for the iterative
solution of the load flow problem.

Consider the three bus-bar power system shown in Figure 9.10, a generator is
connected to each of the first two buses and an electrical load is connected to the third
bus. If the voltages in bus-bars 1 and 2 where known, then the real and reactive power
transferred over the transmission line linking bus-bar 1 and bus-bar 2 could be
calculated using the relationships derived in the previous section.

Figure 9.10 Simple three bus-bar power system

The real and reactive power demanded at the load bus-bar is known, as are the
voltages at the generator bus-bars 1 and 2. The three transmission lines
interconnecting the buses contain both resistance and reactance, therefore there must
be losses in the system. The two generators must jointly supply the total load (including
losses) whilst staying within their individual power limits. Each line must operate within
its normal limits, i.e. δ the transmission angle across each line should not be too large.
The voltage on the load bus-bar 3 must be within acceptable limits. No part of the
power system must be over-loaded.

The first step in performing the load flow analysis is to take the single line diagram in figure 9.10 and form
the bus admittance matrix. This matrix contains all the information required to describe the power system.

359
Load Flow

9.2.1 The Bus Admittance Matrix

In order to form the bus admittance matrix, the generators in the network are
represented as current sources, the loads are represented as current sinks, and the
transmission lines interconnecting each bus-bar are represented by their π equivalent
circuit. This allows us to redraw the single line diagram shown in Figure 9.10 as the
network shown in Figure 9.11.

Normally we would represent each line using its series impedance Z and shunt
admittance Y. In this case it is more convenient to convert the series impedance Z to
its equivalent admittance Y.

1 1 1 R − jX R − jX
Y= = = ⋅ = 2 (9.17)
Z R + jX R + jX R − jX R + X 2

In this way we work with admittances only, having converted all impedances to
admittances.

Figure 9.11 Admittance Diagram

In Figure 9.11, the current source at bus-bar 1 represents generator 1, the current
source at bus-bar 2 represents generator 2, and the current sink on bus-bar 3
represents the load. The transmission line linking bus-bars 1 and 3, is represented by
the three admittances Ya, Ye, and Yf, the line linking bus-bars 3 and 2 is represented
by the three admittances Yb, Yg, and Yh, the line linking bus-bars 1 and 2 is
represented by the three admittances Yc, Yd, and Yi.

360
Load Flow

In circuit analysis we are used to using Ohm’s Law to relate the voltage across an
element to the current through the element and the impedance of the element:

V = I⋅Z
The current can be calculated from

V
I= = V⋅Y (9.18)
Z

1
where Y = is the admittance of the circuit element.
Z

If we apply Kirchoff’s current law at each bus-bar in turn, sum of the currents flowing
into a bus-bar must equal the sum of the currents flowing away from the bus-bar, the
following bus-bar voltages are obtained:

I1 = (V1 − V2 )YC + (V1 − V3 )Ya + V1(Yd + Ye )

I2 = (V2 − V3 )Yb + (V2 − V1 )Yc + V2 (Yh + Yi ) (9.19)
− I3 = (V3 − V1 )Ya + (V3 − V2 )Yb + V3 (Yf + Yg )
Rearranging the equations:

I1 = (Ya + Yc + Yd + Ye )V1 − Yc V2 − Ya V3
I2 = − Yc V1 + (Yb + Yc + Yh + Yi )V2 − Yb V3 (9.20)
(
− I3 = − Ya V1 − Yb V2 + Ya + Yb + Yf + Yg V3 )
These equations can be written in matrix format:

⎡ I1 ⎤ ⎡Y11 Y12 Y13 ⎤ ⎡V1 ⎤

⎢ I ⎥ = ⎢Y
⎢ 2⎥ ⎢ 21 Y22 Y23 ⎥⎥ ⎢V ⎥
⎢ 2⎥
⎢⎣− I 3 ⎥⎦ ⎢⎣Y31 Y32 Y33 ⎥⎦ ⎢⎣V3 ⎥⎦

Bus Injector Bus Admittance Bus Voltage

Current Vector Matrix Vector
[Ibus] [Ybus] [Vbus]

361
Load Flow

The bus admittance matrix is square, sparse and symmetrical. A diagonal element
Ykk is called the self-admittance of bus k and is found by summing the primitive
admittances of all lines and transformers connected to bus k, plus the admittances of all
shunt connections between bus k and ground.

An off diagonal element Ykm is the mutual admittance of the line or transformer
between buses k and m. It is equal to the negative of the admittance of the line or
transformer between buses k and m. It is therefore zero if there is no connection from
bus k to bus m.

The formation of the admittance, matrix for the power systems to be studied is the
first stage of the load flow analysis. In should be noted that in deriving the bus
admittance matrix we included the shunt admittance representing the capacitance to
ground of the transmission lines; in practice this admittance is often neglected. This
simplifies the formation of the bus admittance matrix without introducing a significant
error.

Example 9.5

For the three bus-bar system shown below, form the bus admittance matrix.

Solution 9.5

In this question, the shunt admittance of the line is neglected. The impedances in the
diagram are in per unit. As this is a three bus-bar power system, the bus admittance
matrix will be a 3 x 3 matrix.

362
Load Flow

⎡ Y11 Y21 Y31 ⎤

[Ybus ] = ⎢⎢Y12 Y22 Y32 ⎥⎥
⎢⎣ Y13 Y23 Y33 ⎥⎦

It is easier to start with the off diagonal elements in the bus admittance matrix. An off
diagonal element Ykm is the mutual admittance of the line or transformer connected
between buses k and m. It is equal to the negative of the admittance of the line or
transformer between buses k and m.

Y12 = Y21 = −
1
=
−1 (0.2 − j0.4) = (− 0.2 + j0.4) = (−1 + j2) pu
0.2 + j0.4 (0.2 + j0.4) (0.2 − j0.4) 0.2 2 + 0.4 2
Y23 = Y32 = −
1
=
−1 (0.1 − j0.2) = (− 0.1 + j0.2) = (−2 + j4) pu
0.1 + j0.2 (0.1 + j0.2) (0.1 − j0.2) 0.12 + 0.2 2
Y31 = Y13 = 0

Elements on the main diagonal of the bus admittance matrix Ykk are the self admittance
of bus k formed by summing all the primitive admittances connected to bus k.

Y11 = -Y12 = (1 - j2 )

Y22 = − Y21 − Y23 = (1 − j2 ) + (2 − j4 ) = (3 − j6 )

Y33 = − Y32 = (2 − j4 )

The bus admittance matrix of this three bus-bar power system is therefore:

⎡ (1 − j2 ) (− 1 + j2) (0) ⎤
[Ybus ] = ⎢⎢(− 1 + j2) (3 − j6) (− 2 + j4)⎥⎥
⎢⎣ (0 ) (− 2 + j4) (2 − j4) ⎥⎦
In practice an interconnected power system will have more than three busbars and the
admittance matrix must be derived from the system single line diagram. As a power
system will contains transformers an many different voltage levels, the load flow analysis
will always be performed in per unit. It is therefore necessary to take the system line
diagram and per-unit impedances and convert these to admittances on a common base
before forming the bus admittance matrix. This is best illustrated by example.

363
Load Flow

Example 9.6

For the six bus power system shown below, derive the per unit bus admittance matrix.
The reactance of each transmission line is 20 Ω, and the generators and transformers
are rated as follows:

Item MVA rating Voltage rating Impedance on rated MVA and Voltage
G1 20 MVA 12 kV X = 1.2 pu
G2 60 MVA 13.8 kV X = 1.4 pu
G3 50 MVA 13.2 kV X = 1.4 pu
T1 25 MVA 12/69 kV X = 0.08 pu
T2 75 MVA 13.8/69 kV X = 0.18 pu
T3 60 MVA 69/13.2 kV X = 0.14 pu
T4 75 MVA 69/13.8 kV X = 0.16 pu

Use a Sbase of 100 MVA and a voltage base of 12 kV in the circuit of generator 1.

Solution 9.6

We are told to use a Sbase of 100 MVA and a voltage base of 12kV at generator 1. As
the transformer T1 has a voltage ratio of 12kV: 69kV, the voltage base on bus bar 5 is
69 kV (remember the voltage base changes in proportion to the turns ratio of each
transformer). Working through the system the voltage base at each bus bar is found to
be as follows:

Bus No 1 2 3 4 5 6
Voltage Base 12 kV 13.8 kV 13.2 kV 13.8 kV 69 kV 69 kV

364
Load Flow

The generator impedances are given in the question, but as we represent the generator
by a current source in the load flow model this information is not required.

The impedance of each transformer is given on its MVA rating and kV rating. As the kV
rating corresponds to the base voltages in the table above we only need to convert the
given per unit impedance to our chosen MVA rating.

100
For transformer 1: X T1 = (0.08) = 0.32 pu
25
100
For transformer 2: X T2 = (0.16) = 0.2133 pu
75
100
For transformer 3: X T3 = (0.14) = 0.2333 pu
60
100
For transformer 4: X T4 = (0.16) = 0.2133 pu
75

The impedance of each transmission line is given as 20 Ω per phase. To convert these
to per unit we must find the base impedance at 69 kV and 100 MVA.

Zbase =
Vbase 2 (
=
69 ⋅103 )
2
= 47.61 Ω
Sbase 100 ⋅106

20
The impedance of each transmission line in per unit is therefore = 0.42 pu .
47.61
Now that we have the per unit impedance for each element in the network we can
construct the admittance matrix using the rules outlined above.

The bus admittance matrix is square, sparse and symmetrical. A diagonal element
Ykk is called the self-admittance of bus k and is found by summing the primitive
admittances of all lines and transformers connected to bus k, plus the admittances of all
shunt connections between bus k and ground. An off diagonal element Ykm is the
mutual admittance of the line or transformer between buses k and m. It is equal to the
negative of the admittance of the line or transformer between buses k and m. It is
therefore zero if there is no connection from bus k to bus m.

Using these rules:

365
Load Flow

1 1
Y15 = Y51 = − =− = j3.125
jX T1 j0.32
1 1
Y25 = Y52 = − =− = j4.688
jX T2 j0.2133
1 1
Y36 = Y63 = − =− = j4.286
jX T3 j0.2333
1 1
Y46 = Y64 = − =− = j4.688
jX 4 j0.2133
1 1 1 1
Y56 = Y65 = − − =− − = j4.762 (The two transmission lines are in parallel)
jXL1 jXL2 j0.42 j0.42
Y11 = -Y15 = -j3.125
Y22 = -Y25 = -j4.688
Y33 = -Y36 = -j4.286
Y44 = -Y46 = -j4.688
Y55 = -Y51 - Y52 - Y56 = -3.125 - j4.688 - j4.762 = -j12.575
Y66 = -Y36 - Y46 - Y56 = -j4.286 - j4.688 - j4.762 = -j13.736

The bus admittance matrix can now be formed:

⎡− j3.126 0 0 0 j3.125 0 ⎤
⎢ 0 − j4.688 0 0 j4.688 0 ⎥
⎢ ⎥
⎢ 0 0 − j4.286 0 0 j4.286 ⎥
[Ybus ] = ⎢ ⎥
⎢ 0 0 0 − j4.688 0 j4.688 ⎥
⎢ j3.125 j4.688 0 0 − j12.575 j4.762 ⎥
⎢ ⎥
⎢⎣ 0 0 j4.286 j4.688 j4.762 − j13.736⎥⎦

Note that the bus admittance matrix is symmetrical about the main diagonal; this is a
useful check to ensure that the matrix is correct.

Students should ensure that they understand the formation of the bus admittance
matrix.

366
Load Flow

Tutorial Questions

9.2.1 The impedances of the transmission lines in a four bus-bar power system are listed below.

Line Impedance
1–2 0.20 + j0.6
1–3 0.10 + j0.3
2–3 0.30 + j0.9
2–4 0.10 + j0.3
3–4 0.20 + j0.6

Form the bus-admittance matrix.

9.2.2 The single line diagram of power system is shown below. The reactance of the transmission line is 30 Ω.
The generators and transformers are rated as follows:

Item MVA rating Voltage rating Impedance on rated MVA

and Voltage
G1 20 MVA 12 kV X = 1.20 pu
G2 60 MVA 13.8 kV X = 1.40 pu
G3 20 MVA 13.2 kV X = 1.20 pu
T1 25 MVA 12/69 kV X = 0.08 pu
T2 60 MVA 69/13.8 kV X = 0.14 pu
T3 75 MVA 13.2/69 kV X = 0.14 pu

Using an Sbase of 50 MVA and a voltage base of 12 kV in the circuit of generator 1, form the
bus admittance matrix.

367
Load Flow

The Jacobi and Gauss-Seidel Iterative Methods

Consider the following set of simultaneous equations:

0.6 x + 0.8 y + 0.1z = 1.0

1.1x + 0.4 y + 0.3z = 0.2
1.0 x + 1.0 y + 2.0 z = 0.5

To solve these equations using an iterative procedure we proceed as follows. Inspect

the equations and choose the equation with the largest coefficient multiplying the
variable x. In this case this is the second equation, which is divided throughout by the
coefficient of x, this gives:

1.1 0.4 0.3 0.2

x+ x+ z=
1.1 1.1 1.1 1.1
or
x + 0.3664 y + 0.27273z = 0.18182

Now inspect the remaining equations to find the one with the highest coefficient
multiplying the variable y. In this case it is the first equation which is divided
throughout by this coefficient to give:

0.6 0.8 0.1 1

x+ y+ z=
0.8 0.8 0.8 0.8

or
0.75 x + y + 0.125 z = 1.25

The third remaining equation is now used, the equation is divided by the coefficient of
z which yields:

1 1 2 0.5
x+ y+ z =
2 2 2 2
or
0.5 x + 0.5 y + z = 0.25

Writing these three new equations together:

368
Load Flow

x + 0.36364 y + 0.27273z = 0.18182

0.75 x + y+ 0.125 z = 1.25
0.5 x + 0.5 y + z = 0.25

These three equations can be solved using the Jacobi Iteration scheme. An initial
guess is made for the solution of the set of equations (x0, y0, z0) and a set of updated
estimates are found (x1, y1, z1) using the following equations in turn:

x(n + 1) = 0.18182 − 0.36364 y (n ) − 0.27273 z (n )

y (n + 1) = 1.25 − 0.75 x(n ) − 0.125 z (n )
z (n + 1) = 0.25 − 0.5 x(n ) − 0.5 y (n )

When the new values of x(n+1), y(n+1), z(n+1) are obtained these are then used to
calculate x(n+2), y(n+2), z(n+2) and so on. The solution continues until no further
improvement in the values returned is obtained.

Using this technique with an initial guess for the solution of (0,0,0) it takes 55
iterations for the solution to converge as shown below.

369
Load Flow

In practice the Jacobi Iteration method is not used because the rate of convergence
may be improved by using the new values as soon as they become available. The
previous equations then become:
x(n + 1) = 0.18182 − 0.36364 y (n ) − 0.27273 z (n )
y (n + 1) = 1.25 − 0.75 x(n + 1) − 0.125 z (n )
z (n + 1) = 0.25 − 0.5 x(n + 1) − 0.5 y (n + 1)

Using this improved technique with an initial guess for the solution of (0,0,0) it takes
only 11 iterations for the solution to converge as shown below

This improved iterative scheme known as the “Gauss Siedal” method converges to a
solution 5 times faster that the simple Jacobi method.

It is important to realize that the convergence of the iterative scheme is not

guaranteed, a poor choice of the initial guess may prevent the scheme converging on a
solution. The advantage of the Jacobi and Gauss Siedal methods is that they are straight
forward to program on a computer and no not require a large quantity of computer
memory. The Gauss Seidel method was the first to be applied to the load flow
problem.

370
Load Flow

Tutorial Questions

9.2.3 Find one root (solution) to the following quadratic equation using the Gauss-Seidel iteration scheme.
Use an initial guess of x(0) = 1.

x2 – 6x + 2 = 0
− b ± b 2 − 4ac
Use the quadratic formula x = to find both roots of the equation. Why does the
2a
Gauss-Seidel scheme return only one solution ?

9.2.4 Solve the following set of equations in x and y

y − 3.0 x + 1.9 = 0
y + x 2 − 1.8 = 0

using the Gauss-Seidel iteration scheme. Use an initial guess of x(0) = 1 and y(0) = 1.

Classification of bus-bars in a load flow study

In a load flow study, each bus bar can be classified as being one of three types.

1. Load Bus or P-Q bus

At the load bus k, the net real and reactive power demands are specified or
scheduled, i.e.

k = Pk + jQ k
S sp sp sp

= (Pgk − Pdk ) + j(Q gk − Q dk )

(9.22)

where:
Pksp , Q spk = specified real and reactive power at bus k.
Pgk , Q gk = real and reactive power generation at bus k.
Pdr , Q dk = real and reactive power demand at bus k.

2. Generator Losses or P-V bus

Generators are connected to P-V buses. At these buses, the real power
generation and the voltage magnitude are specified, i.e.

371
Load Flow

Pmsp = PGm − Pdm

sp
(9.23)
Vm = Vm

Vmsp is the specified voltage magnitude at bus m. The voltage magnitude at a P-

V bus is fixed at its specified or controlled values provided that the reactive
power generated is within operational limits. Otherwise, the voltage is allowed
to seek its proper level, and the bus is converted into a P-Q bus by specifying
the reactive power at the bus.

3. Slack Bus or Swing Bus

As the system losses are not known precisely before the load flow analysis is
performed, it is not possible to specify real power injected at every bus. Hence
the real power of one of the generator buses is allowed to swing, and it supplies
the differences between the scheduled real power generation and the sum of all
loads and system loses. The swing bus voltage magnitude is specified and its
voltage phase angle is usually chosen as the system reference and set equal to
zero, i.e.

Vs = Vssp
(9.24)
δs = 00

Figure 9.12 Four Bus Power System

Consider the four bus power system shown in Figure 9.12. All the load bus bars in the
system are P-Q buses. The power demanded at each load bus-bar is specified and the
voltage at each load bus is unknown. There are two generator buses on the system,

372
Load Flow

one of them is a P-V bus where the generator power and terminal voltage is specified.
The second generator bus is the slack or swing bus, its voltage is specified but its
power is unknown. These known and unknown quantities are summarised in the table
below.

Bus Type Known quantities Unknown quantities

No
1 Slack V1 , δ P1 , Q1
2 Load P2 , Q 2 V2 , δ 2
3 Load P3 , Q 3 V3 , δ 3
4 Generator P4 , V4 Q4 , δ4

At the end of the load flow study all the unknown quantities in this table will have been
found.

Equations to be solved in the load flow study

We have seen that the power system can be represented by following the matrix
equation:

[I bus ] = [Ybus ] [Vbus ]

At bus k, the phase current I k injected into the bus is given by:

Yk1V1 + Yk 2 V2 + Yk 3 + .....YkN VN = I k (9.25)

where:

Vm = phasor voltage at bus m

Ykm = element from the bus admittance matrix.
N = total number of buses in the system
Impedance
The specified, or scheduled complex power injection at bus k may be expressed in
terms of the current injected into the bus and the bus voltage phasor as follows:

Ssp sp sp *
k = Pk + jQ k = Vk I k (9.26)

Rearranging this equation:

373
Load Flow

Pksp + jQ sp
Ik* =
VK

form which the current injected at bus k is found to be:

Pksp − jQ sp
Ik = (9.27)
VK*

Substituting this into equation (9.25) gives I k in terms of the bus-bar voltages and the
elements of the bus admittance matrix.

Pksp − jQksp
Yk1V1 + Yk2 V2 + Yk3V3 + ....YkNVN =
Vk* (9.28)

Moving all terms except Vk Ykk to the RHS and dividing the resultant by Ykk gives
the following expression for the phasor voltage Vk .

1 ⎡ Psp − jQsp ⎤
Vk = ⎢ k * k − (Yk1V1 + Yk 2V2 + ... + Ykm Vm + ... + Yk N YN )⎥ (9.29)
Ykk ⎢⎣ Vk ⎥⎦

for m ≠ k

Remember the swing bus voltage is to be taken as the reference phasor, its voltage
magnitude is specified and its phase angle is set to zero. The above equation for the
voltage at bus k is used to write a system of (N-1) simultaneous algebraic equations
relating the phasor voltage at the individual buses with the corresponding power
injections at the bus and phasor voltages at all the buses. These equations are coupled
through the elements of the bus admittance matrix.

For example, consider the four bus-bar power system shown in Figure 9.12. The
voltage equations to be solved in the load flow analysis of this system are as follows:

374
Load Flow

V1 = V1 0 o
1 ⎡ P2SP − jQ SP ⎤
V2 = ⎢ *
2
− (Y12 ⋅ V1 + Y32 ⋅ V3 + Y42 ⋅ V4 )⎥
Y22 ⎣ V2 ⎦
1 ⎡ P3SP − jQ SP ⎤
V3 = ⎢ *
3
− (Y13 ⋅ V1 + Y23 ⋅ V2 + Y43 ⋅ V4 )⎥
Y33 ⎣ V3 ⎦
1 ⎡ P4SP − jQ SP ⎤
V4 = ⎢ *
4
− (Y14 ⋅ V1 + Y24 ⋅ V2 + Y34 ⋅ V3 )⎥
Y44 ⎣ V4 ⎦

The voltage on the generator bus bar 1 is known since this generator is the slack bus,
its voltage magnitude is specified and its phase angle set to zero. The voltage
equations for the load bus bars 2 and 3 use elements from the bus admittance matrix
and the specified real and reactive power at each bus. The voltage equation for the
voltage on the generator bus bar 4 cannot be solved immediately since the reactive
power on this bus is unknown (remember that on a generator bus the voltage
magnitude and the reactive power is specified). Rearranging equation (9.28), the
reactive power of the generator can be estimated:
Pksp − jQksp
Yk1V1 + Yk2 V2 + Yk3 V3 + ....YkN VN =
Vk*

Pksp + jQksp
Yk1V1 + Yk2 V2 + Yk3 V3 + ....YkN VN =
Vk

Vk (Yk1V1 + Yk2 V2 + Yk3 V3 + ....YkNVN ) = Pksp + jQksp

For a load bus (sink) the reactive power would be given by the imaginary part of the
expression on the LHS of the equation above. For a generator (source) the reactive
power generated with therefore be the negative of the imaginary part of the expression
of the LHS of the above equation, i.e. the calculated reactive power on the generator
bus is given by:

QKcalc = − Im Vk (Yk1V1 + Yk2 V2 + Yk3V3 + ....YkNVN )

This value is substituted into the equation for V4.

The load flow problem can be solved by finding the solution to these three
simultaneous equations for the voltages V2, V3, and V4. These equations can be
solved using the Gauss-Seidel iterative technique.

375
Load Flow

Gauss-Seidel Solution of the Load Flow Problem

The Gauss-Seidel method is an iterative technique for solving a system of non linear
algebraic equations, such as that formed by the (N-1) voltage equations derived for
the load flow analysis of an N bus power system. Using this technique, the phasor
voltage at a bus is found by using the latest computed values of the phasor voltage at
the other buses. This is identical to the procedure used to solve the three
simultaneous equations in the example above.

The Gauss-Seidel method applied to the load flow problem, produces the following
general result for the voltage at each bus k:

1 ⎡P sp − jQ sp w ⎤
(β ) ⎥
Vki +1 =
Ykk
⎢ k
⎢⎣ Vk(i )
k
− ∑ km m ⎥
Y V (9.30)
m =1,m #k ⎦

β=i for m > k

β = i + 1 for m < k
where i represents the iteration count.

In the solution of these equations, the most recently computed values for each voltage
are used. The solution procedure by passes the swing because the phasor voltage of
the swing bus is already known. For a P-Q or load bus, the procedure is applied
directly to obtain an improved estimate of the phasor voltage in as much as Pksp
and Q spk are both given. For a P-V or generator bus, the bus voltage Vk is specified;
hence, Q spk is not directly available. Therefore an estimate Q calc
k is computed by using
the current estimates of the phasor voltages.

Q calc
k [ ]
= − I m Vk* (Yk1 V1 + Yk 2 V2 + ...Ykm Vm + ...Ykn Vn ) (9.31)

This value of reactive power is used to replace Q spk and then Vk is recalculated. Then
the magnitude of Vk is reset to its specified value Vksp , but the new value of its phase
angle is retained. The Gauss-Seidel method has the attractions of simplicity,
comparatively good performance, and non-storage of previous values. It has a very
reliable convergence characteristic, but the rate of convergence is slow.

The iterative process is said to have converged when the process no longer yields any
improvement on the solution. At this point, the phasor voltages at all the busses have
been found and may be used to drive other information about the steady-state
operating characteristics of the power system. For example, the real and reactive

376
Load Flow

powers flowing through any transformer or transmission line, with series impedance
Z and shunt admittance Y, connected between buses k and m may be expressed as:

*
⎛ V − Vm Y ⎞
S km = Pkm + jQ km = Vk ⎜ k + Vk ⎟ (9.32)
⎝ Z 2 ⎠

Having read the above text, the reader may feel that the load flow problem and its
solution is a very complex task. By approaching the problem in a systematic manner
the calculations become very straight forward, this is best illustrated by an example.

Example 9.7

For the power system shown below, bus 1 is selected as the slack or swing bus and its
voltage is set to V1=(1.0 + j0.0) pu. The chosen power base is 100 MVA. Generator
2 delivers a real power of 0.75 pu at a voltage of 1.02 pu. The loads on buses 3 and 4
are Sd3 = (0.40 + j 0.30) pu and Sd4 = (0.80 + j 0.60) pu respectively. The impedance
parameters for the transmission lines referred to the given power base and a voltage
base of 115 kV are:

Line 1 - 2 Z = (0.01 + j0.05) pu Y = j0.30 pu

Line 2 – 3 Z = (0.03 + j0.15) pu Y = j0.90 pu

The transformer is connected between buses 3 and 4, and its reactance is 0.10 pu.
From a flat start, perform one iteration of the Gauss-Seidel iterative technique to find
the voltages at buses 2, 3, and 4.

Solution 9.7

First we need to calculate the elements in the bus admittance matrix as these are
required to form the iterative voltage equations. All impedances and admittances

377
Load Flow

parameters are given in per unit. Using the rules for building the bus admittance
matrix the following are calculated.

−1 − (0.01− j0.05)
Y12 = Y21 = = 2 2
= −3.846+ j19.23= 19.61∠101.3o
0.01+ j0.05 0.01 + 0.05
Y13 = Y31 = 0
Y14 = Y41 = 0
Y24 = Y42 = 0
−1
Y23 = Y32 = = −1.28 + j6.41= 6.54 ∠101.3o
0.03+ j0.15
−1
Y34 = Y43 = = j10
j0.1
j0.3
Y11 = −Y12 + = 3.846− j19.23+ j0.15 = 19.46 ∠ − 78.6o
2
j0.3 j0.9
Y22 = −Y21 − Y23 + + = 3.846− j19.23+1.28 − j6.41+ j0.15+ j0.45 = 25.56∠ − 78.4o
2 2
j0.9
Y33 = −Y23 + − Y34 = 1.28− j6.41+ j0.45− j10 = 16.0 ∠ − 85.4o
2
Y44 = −Y34 = −j10

As bus 1 is the swing bus the voltage on bus 1 is known, V1 = 1.0∠0o.

The voltage on bus 2 is found from the following equation iterative equation

1 ⎡ P2 − jQ 2 ⎤
V21 = ⎢ (0)*
− Y21 ⋅ V1 − Y23 ⋅ V3(0) − Y24 ⋅ V4(0) ⎥
Y22 ⎣ V2 ⎦

where V1 = 1.0∠ 0 o , V2(0) = 1.02 ∠0 o , V3(0) = V4(0) = 1.0 ∠0 o

A flat start is where the voltage on the swing bus is set at 1 pu and an angle of zero
degrees, the voltage on the generator at bus 2 is set to its specified magnitude at an
angle of zero degrees, and the completely unknown voltages on the load bus bars are
set to 1 pu at an angle of zero degrees as shown above.

The iterative equation for V2 cannot be solved since the reactive power at this
generator bus is unknown (remember only real power and voltage magnitude is
specified at a generator bus). The reactive power at the generator bus is calculated
using equation (9.31).

378
Load Flow

Q 2 = −Im[V2 * (Y21 ⋅ V1 + Y22 ⋅ V2 + Y23 ⋅ V3 + Y24 ⋅ V4 )]

o * ⎪(19.61 ∠101.3 ) ⋅ (1.0∠0 ) + (25.56 ∠ − 78.4 ) ⋅ (1.02 ∠0 )⎪

⎡ ⎧ o o 0 o
⎫⎤
= −Im⎢(1.02 ∠0 ) ⎨ ⎬⎥
⎣⎢ ⎪⎩ + (6.54 ∠101.3o ) ⋅ (1.0 ∠0 o ) + 0 ⎪⎭⎦⎥
= −0.62 pu

This value of reactive power is now substituted into the iterative equation for V2:

1 ⎡ P2 − jQ 2 ⎤
V2(1) = ⎢ (0)*
− Y21 ⋅ V1 − Y23 ⋅ V3(0) − Y24 ⋅ V4(0) ⎥
Y22 ⎣ V2 ⎦
⎡ 0.75 + j0.62
( )(
− 19.61 ∠101.3 o ⋅ 1.0 ∠0 o )⎤⎥
=
1 ⎢
25.56 ∠ − 78.4 o ⎢
(
1.02∠0 o *
) ⎥
⎢
⎣ ( )(
− 6.54 ∠101.3o ⋅ 1.0 ∠0 o )⎥
⎦
= 1.007 ∠1.6 o

Thus the 1st iteration returns a value of V2 equal to 1.007 ∠1.6o

We can now perform the 1st iteration for the voltage on bus bar 3. The scheduled real
and reactive powers on bus 3 are calculated using equation (9.22).

P3 = PG3 − PD3 = −0.4 pu

Q 3 = Q G3 − Q D3 = −0.3 pu

The iterative equation for the voltage on bus 3 is:

1 ⎡ P3 − jQ 3 ⎤
V3(1) = ⎢ (0)
− Y31 ⋅ V1 − Y32 ⋅ V2(1) − Y34 ⋅ V4(0) ⎥
Y33 ⎣ V3 ⎦
⎡ − 0.4 + j0.3
− 0 − (6.54 ∠101.3o ) ⋅ (1.007 ∠1.6 o )⎥
⎤
=
1 ⎢
16.0 ∠ − 85.4 o ⎢
(1.0 ∠0 )
o *
⎥
⎢
⎣ − ( j10 ) ⋅ (
1.0 ∠0 )
o ⎥
⎦
= 1.009 ∠ − 0.8 pu
o

Notice that the updated value for V2 is used in this calculation i.e. 1.007 ∠1.6o and not
the initial value of 1.0 ∠0o. It is a requirement of the Gauss-Seidel technique that
updated values of voltages are used as soon as they become available.

In a similar way the voltage at bus bar 4 can be calculated.

379
Load Flow

P4 = PG4 − PD4 = −0.8 pu

Q 4 = Q G4 − Q D4 = −0.6 pu

1 ⎡ P4 − jQ4 ⎤
V4(1) = ⎢ ( 0 )*
− Y41 ⋅ V1 − Y42 ⋅ V2(1) − Y43 ⋅ V3(1) ⎥
Y44 ⎣ V4 ⎦
1 ⎡ − 0.8 + j 0.6 ⎤
= ⎢ − 0 − 0 − ( j10 ) ⋅ 1 .(009 ∠ − 0 . 8 o
)
⎥
(
− j10 ⎣⎢ 1.0 ∠0 o * ) ⎦⎥
= 0.953 ∠ − 5.7 o

After the first iteration the calculated bus bar voltages are therefore:

V1(1) = 1.0 ∠0 o
V2(1) = 1.007 ∠1.6 o
V3(1) = 1.009 ∠ − 0.8 o
V4(1) = 0.953 ∠ − 5.7 o

These updated voltages are used as the input to the next iteration and the whole
process is repeated until additional iterations fail to produce a change in the calculated
voltages. At this point the solution is said to have converged and the analysis is almost
complete. Once all the bus bar voltages are found, the currents in the net work can be
calculated and the powers flowing along all transmission lines can be found.
Transmission losses can then be calculated.

380
Load Flow

Tutorial Questions

9.2.5 For the two busbar power system shown below:

i) Form the bus admittance matrix (all impedances shown are in per-unit)
ii) If the load on bus 2 is S = (0.5 + j0.3) derive an iterative expression using the Gauss-
Seidel method for the voltage bus 2
iii) Find the voltage on bus 2
iv) Find the real and reactive powers on the generator bus 1
v) Estimate the efficiency of the transmission line.

9.2.6 For the three busbar power system shown below:

i) Form the bus admittance matrix (all impedances shown are in per-unit)
ii) For the given loads on buses 2 and 3 find the voltages on buses 2 and 3 using the
Gauss-Seidel method. Assume a flat start.
iii) Find the apparent power injected into the system by the generator on bus 1

9.2.7 The three busbar power system of question 9.2.6 is modified by the addition of a transmission line
between buses 2 and 3 of series impedance Z = (0.1 + j0.3).
i) Modify the bus admittance matrix to allow for this extra line
ii) Derive an iterative expression for the voltages on buses 2 and 3.
Perform one iteration of the Gauss-Seidel method to update the voltages on buses 2 and 3 from a flat
start.

381
Load Flow

Final Comments on Load Flow Analysis

Many students find the concepts introduced in this chapter difficult to grasp at the first
attempt for the following reasons

1. The load flow problem is not a difficult problem to solve, but it requires the
power system description to be set up in a very specific way using the bus-
admittance matrix. The rules for constructing the matrix are easy to remember
and with practice become routine.

2. The need to specify real and reactive powers on load bus bars, and voltage and
power on generator bus bars is very different from classical circuit analysis
where loads are represented by impedances supplied at specified voltages. This
results in the load flow problem being non-linear.

3. As the load flow problem is non-linear iterative techniques must be used to

solve for the system voltages. A common technique described in the text is the
Gauss-Seidel Method. Other techniques such as the Newton-Raphson
algorithm and Stott’s algorithm provide faster solutions but are more
mathematically intensive and beyond the scope of this text.

4. In practice the load flow calculation will involve many hundreds, perhaps
thousands of bus bars in an interconnected power system. The load flow
calculation will then be performed using a computer program. This chapter has
provided a introduction to the terminology and techniques used in a load flow
study.

382
Load Flow

Chapter Summary

To understand the term load flow analysis as applied to an interconnected power system

The load flow calculation is used to ensure that the steady state voltages and currents
flowing the interconnected network are within the continuous ratings of the equipment for
both normal and emergency operating conditions. Such calculations are performed on
computer as hand calculations are too time consuming and impractical for large systems.

To understand the factors which influence the flow of real and reactive powers over an inter-
connector

The power flow between two bus bars is a function of the transmission angle δ across the
transmission line and is not a function of voltage magnitude. Real power flows from leading
V1 V2 sin δ
to lagging phasor. For a loss less line P = P1 = P2 = . The reactive power flow
x
is a function of the difference in voltage magnitude. The larger the difference in magnitude,
the stronger the reactive power flow. Reactive power flows from the high voltage bus-bar
to the low voltage bus-bar. The reactive power flow at 1 into the line is
2
V1 − V1 V2 cosδ
Q1 = and the reactive power out of the line at 2 is
x
2
V1 V2 cosδ − V2
Q2 =
x

To understand that the flow of both real power and reactive power contribute to the real power
losses in the transmission network

The power loss in a transmission link can be estimated from the line average power,
2 2
2 P + Q ave
averages reactive power, and average voltage thus PΩ ≅ R I = R ave 2
Vave

To understand the principle aims and objectives of a load flow study

383
Load Flow

To find the real and reactive power flows in the transmission lines based on certain
assumptions concerning the nature of the loads and the generators on the system. To
calculate the voltages on all bus-bars in the power system. Check that no cable,
transmission line, transformer, or generator is overloaded. To find the optimum
configuration of the power system to minimise transmission losses.

To understand that the load flow problem is non-linear and cannot be solved by classical circuit
analysis

A load flow study network equations written in terms of voltages and power, not voltages
and currents as in a classical circuits problem. This difference results in the load flow
equations being non-linear. This eliminates the possibility of a direct analytical solution to
the load flow equations in most cases.

To understand how to construct the bus admittance matrix for a simple power system

The bus admittance matrix is square, sparse and symmetrical. A diagonal element Ykk is
called the self-admittance of bus k and is found by summing the primitive admittances of all
lines and transformers connected to bus k, plus the admittances of all shunt connections
between bus k and ground. An off diagonal element Ykm is the mutual admittance of the
line or transformer between buses k and m. It is equal to the negative of the admittance of
the line or transformer between buses k and m. It is therefore zero if there is no
connection from bus k to bus m.

To understand the Gauss-Seidel Method as applied to the iterative solution of simultaneous

equations

The Gauss-Seidel Method is used to solve sets of simultaneous equations. The equations
are rewritten in the form X(n + 1) = X(n ) + Y(n ) + ....... , one equation for each unknown.
Initial guesses are made for each unknown which are then updated by substituting these
values in the iterative equations. Convergence is not guaranteed and an intelligent first
guess is required.

To understand the three different types of bus bar in a load flow study. To understand which
quantities are known and which are unknown for each type at the beginnings of the calculation

At a Load Bus or P-Q bus the net real and reactive power demands are specified and the
voltage magnitude and phase angle are unknown. Generators are connected to P-V buses
at which the real power generation and voltage magnitude are specified; the reactive power
and voltage angle are unknown. The swing bus voltage magnitude is specified and its

384
Load Flow

voltage phase angle is usually chosen as the system reference and set equal to zero; the
power on the swing bus is unknown.

To understand how to calculate the voltages in a power system using the Gauss-Seidel Method

The Gauss-Seidel method applied to the load flow problem, produces the following general
result for the voltage at each bus k:

1 ⎡P sp − jQ sp w ⎤
(β ) ⎥
Vki +1 =
Ykk
⎢ k
⎢⎣ Vk(i )
k
− ∑ km m ⎥
Y V
m =1,m #k ⎦
β=i for m > k
β = i +1 for m < k
where i represents the iteration count.

To understand how to calculate the real and reactive powers on a generator bus bar once all the
bus bar voltages have been found.

For a generator the reactive power generated is given by:

Q Kcalc = − Im Vk (Yk1V1 + Yk2 V2 + Yk3 V3 + ....YkN VN )

385
Load Flow

Chapter 9 Tutorial Solutions

9.1.1
(a) The voltage given in the question is the line to line voltage. This calculation must be
performed working per phase.
V 140
Vph = L = kV = 80.8 kV
3 3
For flat line operation V1 = V2 = 80 kV per phase.
V1 V2
Power transfer is P = sinδ per phase
X
Maximum power transfer occurs when sinδ = 1 , i.e. δ = 90 o
V1 V2 (80.8 ⋅ 10 3 )(80.8 ⋅ 10 3 )
Pmax = = = 130 MW/ph
X 50
Maximum three-phase power is 3 x 130 MW = 390 MW

(b) If the power flow is 100 MW, then the power per phase is given by
100
= 333.33 MW/ph
3
V1 V2
Power transfer is P = sinδ per phase
X
33.33 ⋅ 10 =
6 ( )( )
80.8 ⋅ 10 3 80.8 ⋅ 10 3
sinδ
50
solving for δ yields 14.78o

The transmission angle is 14.78o

386
Load Flow

9.1.2
The voltage profile is given in per unit, so perform all calculations in per unit.

Choose 220 kV as Vbase, 40 MVA as Sbase and workout Zbase

Z base =
(Vbase )2
=
(220 ⋅10 ) 3 2
= 1210 Ω
S base 40 ⋅ 10 6

140 Ω
The impedance of the line in per unit is therefore = 0.115 pu
1210 Ω
Case (a) V1 = 1.0 pu and V2 = 1.0 pu
40
P = 40 MW = = 1.0 pu
40
V1 V2
Power transfer is P = sinδ per phase
X
1.0 =
(1.0) ⋅ (1.0) sinδ
0.115

solving for the transmission angle δ yields 6.6o

The reactive power at the sending end is:

2
V1 − V1 ⋅ V2 cosδ
Q1 =
X

=
(1.0 ) − (1.0 ) ⋅ (1.0 ) ⋅ cos6.6 o
2

0.115
= 0.057 pu

The reactive power at the receiving end is

2
V1 ⋅ V2 cosδ − V2
Q2 =
X

=
(1.0 ) ⋅ (1.0 ) cos6.6 o − (1.0 )
2

0.115
= −0.057pu

Case (b) V1 = 1.2 pu and V2 = 1.0 pu

40
P = 40 MW = = 1.0 pu
40

387
Load Flow

V1 V2
Power transfer is P = sinδ per phase
X

1.0 =
(1.2) ⋅ (1.0) sinδ
0.115

solving for the transmission angle δ yields 5.49o

The reactive power at the sending end is:

2
V1 − V1 ⋅ V2 cosδ
Q1 =
X

=
(1.0) − (1.2) ⋅ (1.0) ⋅ cos5.49 o
2

0.115
= 2.13 pu

The reactive power at the receiving end is

2
V1 ⋅ V2 cosδ − V2
Q2 =
X

=
(1.2) ⋅ (1.0) cos5.49 o − (1.0)2
0.115
= +1.69 pu

The results for cases (a) and (b) are summarized in the table below:

V1 V2 δ P1 P2 Q1 Q2
Case 1 1.0 1.0 6.60 1.0 1.0 0.057 -0.057
Case 2 1.2 1.0 5.49 1.0 1.0 2.13 1.69

388
Load Flow

The case (a) the reactive power required to energize the series impedance of the line
comes from both ends of the line. In case (b) there is a strong reactive power flow
from bus bar 1 to bus bar 2. This example clearly shows that raising the sending end
voltage increases the reactive power flow in the system.

9.1.3

This problem is very similar to Example 9.4. Instead of the receiving end conditions
being specified, the sending end is specified. All information should be transferred to
the per phase equivalent circuit and the voltages and currents are then calculated using
straight forward circuit analysis.

145
Sending voltage = 145 kV line to line or = 83.7 kV per phase.
3
Sending three phase power is 120 MW at 0.8 pf

120
Sending three phase apparent power is = 150 MVA
0.8
Sending three phase reactive power is 150 2 − 120 2 = 90 MVA

Per phase values P = 40 MW

Q = 30 MVAr
S = 40 MVA

The complex power per phase is S1 = P1 + jQ1 = (40 + j30 ) MVA

389
Load Flow

Taking V1 as the reference phasor, V1 = (83.7 + j0.0 )kV

The complex power is also S1 = V1 ⋅ I 1* , hence the complex conjugate of the current
can be calculated, then the current I1 can be found:

S1 (40 + j40 ) 10 6
I =
*
1 = = (477 + j358) A
V1 83.7 10 3
I1 = (477 − j358) A

The current flowing into the shunt capacitance at the sending end can be found from
the shunt admittance and the sending end line voltage:

I sh1 = V1 ⋅ Y1 = 83.7 ⋅ 10 3 × j167 ⋅ 10 −6 = j13.97 A

The line current can now be calculated by applying Kirchhoff’s 1st Law I L + I sh 1 = I1

I L = I1 − I sh1 = (477 − j358) − j13.97 = (477 − j371) A

The voltage drop across the series impedance of the line can now be calculated

∆V = I L ⋅ (R + jX L )
= (477 − j371) ⋅ (9.10 + j50.5)
= 4340 + j24088 − j3376 + 18735
= (23075 + j20712) kV

The sending voltage is V1 = V2 + ∆V , hence we can write

V2 = V1 − ∆V
= 83700 − 23075 − j20712
= (60625 − j20712 ) V

It is convenient to calculate the magnitude of the received voltage:

V2 = (60625)2 + (20712)2 = 64 kV per phase

The current flowing into the shunt capacitance at the receiving end can be found from
the shunt admittance and the receiving end line voltage:

I sh2 = V2 ⋅ Y2 = (60625 − j 20712) × j167 ⋅ 10 −6 = (3.44 + j10.12) A

390
Load Flow

The received current I2 can now be calculated by applying Kirchhoff’s 1st Law
I L = I 2 + I sh 2

I 2 = I L = I sh2 = (477 − j371) − (3.45 + j10.12) = (473.5 − j360.8) A

As V2 and I2 are now known the complex power S2 can be determined:

S 2 = V2 ⋅ I *2
= (60625 − j20712 ) ⋅ (473.5 + j360.8)
= (28.7 + j21.8 − j9.80 + 7.47 ) ⋅ 10 6
= (36.17 + j12 ) ⋅ 10 6 VA

This is the complex power per phase.

The three-phase real power is therefore 3 × 36.17 MW = 108.5 MW and the three-
phase reactive power is 3 × 12 MVAr = 36 MVAr

The sending real power is 120 MW and the receiving power is 108.5 MW. The
efficiency of the transmission is therefore:

Po 108.5
η= = = 0.90 or 90%.
Pin 120

Summary Sending current = (477 – j 358) A per phase

Sending voltage = 83.7 kV per phase
= 145 kV line to line
Sending complex power = (120 +j 90) MVAr (three phase)

Received current = (474 – j 360) A per phase

Received voltage = 64 kV per phase
= 110 kV line to line
Received complex power =(108.5 +j 36) MVA (three phase)

9.1.4

The equation for the approximate power loss in a transmission line was derived in the
main text.

Using the summary table at the end of the solution to 9.1.3 the average per phase real
power, reactive power and voltage can be determined.

391
Load Flow

1
Pave = (40 + 36.17 ) = 28.08 MW
2
1
Q ave = (30.0 + 12.0 ) = 21.00 MVAr
2
1
Vave = (83.7 + 64 ) = 73.8 kV
2

Power loss = R
P2 + Q2
= 9.1
(28.08 ⋅10 ) + (21.00 ⋅ 10 )
6 2 6 2
= 3.15 MW/phase
V
2
(73.8 ⋅10 )
3 2

Total three phase loss is 3 x 3.15 MW = 9.45 MW.

120 − 9.45
Efficiency = = 0.92 or 92%
120

The approximate calculation is seen to give a reasonable estimate of the power loss in
the transmission line.
9.2.1

From the given line impedances a single line diagram of the power system can be
drawn. This can aid the formation of the bus admittance matrix.

It is easier to form the off diagonal elements first, i.e. the mutual admittances. Ykn is the
negative of the admittance connecting bus k to bus n.

392
Load Flow

1
Y12 = Y21 = − = −0.5 + j1.5
0.2 + j0.6
1
Y23 = Y32 = − = −0.333 + j1.0
0.3 + j0.9
1
Y13 = Y31 = − = −1.0 + j3.0
0.1 + j0.3
1
Y24 = Y42 = − = −1.0 + j3.0
0.1 + j0.3
1
Y34 = Y43 = − = −0.5 + j1.5
0.2 + j0.16
Y41 = Y14 = 0 + j0

The main diagonal elements can now be found, i.e. the self admittances. Ykk is the sum
of all the primitive admittances connected to bus k.

Y11 = − Y12 − Y13 − Y14 = (0.5 − j1.5) + (1.0 − j3.0) + (0 + j0) = 1.5 − j4.5
Y22 = − Y21 − Y23 − Y24 = (0.5 − j1.5) + (0.333 − j1.0) + (1.0 − j3.0) = 1.833 − j5.5
Y33 = −Y31 − Y32 − Y34 = (1.0 − j3.0) + (0.333 − j1.0) + (0.5 − j1.5) = 1.833 − j5.5
Y44 = − Y41 − Y42 − Y43 = (0 + j0) + (1.0 − j3.0) + (0.5 − j1.5) = 1.5 − j4.5

The bus admittance matrix can now be formed

⎡ (1.5 − j4.5) (− 0.5 + j1.5) (− 1.0 + j3.0) (0 + j0) ⎤

⎢(− 0.5 + j1.5) (1.833 − j5.5) (− 0.333 + j1.0) (− 1.0 + j3.0)⎥⎥
Ybus =⎢
⎢ (− 1.0 + j3.0 ) (− 0.333 + j1.0) (1.833 − j5.5) (− 0.5 + j1.5)⎥
⎢
⎣ (0 + j0 ) (− 1.0 + j3.0) (− 0.5 + j1.5) (1.5 − j4.5) ⎥⎦
9.2.2

We are told to use an Sbase of 50 MVA and a voltage base of 12 kV at generator 1.

Redraw the system showing the voltage bases in each part of the network. These are
found using the transformer turns ratio.

393
Load Flow

The impedance of each transformer is given on its own MVA rating and kV rating.
As the kV rating corresponds to the base voltages in the diagram above we need only
convert the given per unit impedance to our chosen MVA base.

⎛ 50 ⎞
For transformer 1: X T1 = (0.08)⎜ ⎟ = 0.16 pu
⎝ 25 ⎠
⎛ 50 ⎞
For transformer 2: X T2 = (0.14)⎜ ⎟ = 0.1167 pu
⎝ 60 ⎠
⎛ 50 ⎞
For transformer 3: X T3 = (0.16)⎜ ⎟ = 0.1066 pu
⎝ 75 ⎠

The impedance of the transmission line is given as 30 Ω per phase. To convert to per
unit find base impedance at 69 kV and 50 MVA.

Z base =
(Vbase )2
=
(69 ⋅10 ) = 95.22 Ω
3

S base 50 ⋅ 10 6

30
The impedance of the transmission line in per unit is therefore = 0.315 pu .
95.22

Now that the per unit impedance of every element in the network is known we
construct the bus admittance matrix. Starting with the off diagonal elements, i.e. the
mutual admittances between buses.

394
Load Flow

1
Y14 = Y41 = − = j6.25
j0.16
1
Y25 = Y52 = − = j8.57
j0.1167
1
Y34 = Y43 = − = j9.375
j0.1066
1
Y45 = Y54 = − = j3.174
j0.315

The self admittance of each bus can now be calculated

Y11 = − Y14 = − j6.25

Y22 = − Y25 = − j8.57
Y33 = −Y34 = − j9.375
Y44 = − Y41 − Y43 − Y45 = − j6.25 − j9.375 − j3.174 = − j18.8
Y55 = − Y52 − Y54 = − j8.57 − j3.174 = − j11.74

The bus admittance matrix can now be formed.

⎡− j6.25 0 0 j6.25 0 ⎤
⎢ 0 − j8.57 0 0 j8.57 ⎥⎥
⎢
Ybus =⎢ 0 0 − j9.375 j9.375 0 ⎥
⎢ ⎥
⎢ j6.25 0 j9.375 − j18.8 j3.174 ⎥
⎢⎣ 0 j8.57 0 j3.174 − j11.72⎥⎦

395
Load Flow

9.2.3

To solve the equation using an iterative approach it must be written in an alternative

form.

1 1 2
x 2 − 6x + 2 = 0 can be written as x= + x
3 6

The iterative scheme to solve this equation is therefore:

1 1 2
x (n +1) = + x (n)
3 6

Initial guess x(0) = 1

1 1 2
Iteration 1 x(1) = + (1) = 0.5
3 6
1 1
x(2) = + (0.5) = 0.375
2
Iteration 2
3 6
1 1
x(3) = + (0.375) = 0.3568
2
Iteration 3
3 6
1 1
x(4) = + (0.3568) = 0.3534
2
Iteration 4
3 6

To two decimal places the solution is 0.35

The quadratic formula

− b ± b 2 − 4ac 6 ± 6 2 − 4 ⋅ 1 ⋅ 2 6 ± 28
x= = =
2a 2 2

gives the solutions as x = 0.354 and x = 5.64

The Gauss-Seidel iterative technique finds the root closest to the initial guess. A guess
close to 5 would find the 2nd root of the equation. The success of the Gauss-Seidel
method in finding the roots of the equation is not guaranteed. An intelligent initial
guess is usually required to aid the convergence towards a solution.

396
Load Flow

9.2.4

The equations to be solved are:

y − 3x + 1.9 = 0
y + x 2 − 1.8 = 0
These equations can be rewritten as:
y
x = + 0.6333
3
y = 1.8 − x 2
This suggest the following iterative scheme:
y (n)
x (n +1) = + 0.6333
3
y (n +1) = 1.8 − x (n2 +1)

Using the initial guess of x(0) = 1 and x(y) = 1 , the solution proceeds as follows:

1
x(1) = + 0.6333 = 0.9666
3
= 1.8 − (0.9666) = 0.8656
2
y (1)
0.8656
x( 2 ) = + 0.6333 = 0.9218
3
= 1.8 − (0.9218) = 0.9502
2
y ( 2)
0.9502
x ( 3) = + 0.6333 = 0.9500
3
= 1.8 − (0.9500) = 0.8975
2
y ( 3)
0.8975
x( 4 ) = + 0.6333 = 0.9321
3
y ( 4) = 1.8 − (0.9321) 2 = 0.9311
0.9311
x( 5) = + 0.6333 = 0.9436
3
y ( 5) = 1.8 − (0.0463) 2 = 0.9045

The subsequent iterations are listed (0.9348, 0.9261), (0.9420, 0.9126), (0.9375,
0.9210), (0.9403, 0.9158), (0.9385, 0.9192), (0.9397, 0.9169), (0.9389, 0.9184).

The Gauss-Seidel Method slowly approaches the correct solution of (0.93926,0.9178).

397
Load Flow

9.2.5

I) Form the bus admittance matrix using the rules for mutual and self admittances.

1
Y12 = Y21 = − = (− 2.0 + j4.0 )
(0.1 + j0.2)
Y11 = − Y12 = (2.0 − j4.0 )
Y22 = −Y21 = (2.0 − j 4.0)

So the bus admittance matrix is therefore

⎡ (2.0 − j4.9 ) (− 2.0 + j4.0)⎤

Ybus = ⎢
⎣(− 2.0 + j4.0 ) (2.0 − j4.0) ⎥⎦
ii) The load on bus bar 2 is S = 0.5 + j 0.3 pu. The scheduled real and reactive powers on
bus 2 are therefore P2sp = −0.5 and Q sp2 = −0.3

The iterative equation for the voltage on bus 2 is

(i +1) 1 ⎡ P sp − jQ sp ⎤
V 2 = ⎢ 2 2
− Y12 ⋅ V1 ⎥
Y22 (i) *
⎣⎢ V2 ⎦⎥

Bus bar 1 is the swing or slack bus, the voltage on bus 1 is therefore equal to 1 pu at an
angle of zero degrees. Substituting known values:

1 ⎡ − 0.5 + j0.3 ⎤
V2(i +1) = ⎢ − (− 2.0 + j4.0 ) ⋅ (1.0 + j0.0 )⎥
(2.0 − j4.0) ⎢⎣ V2(i) * ⎥⎦

=
(− 0.11 − j0.07 ) + 1.0
*
V2(i)

This equation is the iterative equation for the voltage on bus bar 2. Initial guess for V2
is 1 pu at an angle of zero degrees. The solution proceeds as follows:

398
Load Flow

V2(0) = 1.0 + j0.0

V2(1) = 0.89 − j0.07
V2(2) = 0.8710 − j0.068
V2(3) = 0.8682 − j0.070
V2(4) = 0.8676 − j0.069

So after four iterations the solution has converged to V2 = (0.86 − j0.07) pu .

iii) The real and reactive powers on 1 can be determined using the following equation:

P1 − jQ1 = V1* (Y11 ⋅ V1 + Y12 ⋅ V2 )

= (1.0 + j0.0 ) ((2.0 − j4.0 )(1.0 + j0.0 ) + (− 2.0 + j4.0 )(0.86 − j0.7 ))
*

= (2.0 − j4.0 ) + (−1.44 + j3.58)

= (0.56 − j0.42)

The generator on bus bar 1 is therefore delivering 0.56 pu real power and 0.42 pu
reactive power to the network.

iv) Input power on bus 1 is 0.56 pu. Output power on bus 2 is 0.5 pu.

0.5
Efficiency = = 0.9
0.56

9.2.6

Form the bus admittance matrix using the rules for mutual and self admittances.

1
Y12 = Y21 = − = (− 1.0 + j3.0 )
(0.1 + j0.3)
1
Y13 = Y31 = − = (− 0.5 + j1.5)
(0.2 + j0.6)
Y23 = Y32 = 0
Y11 = − Y12 − Y13 = (1 − j3) + (0.5 − j1.5) = (1.5 − j4.5)
Y22 = − Y21 = (1 − j3.0) )
Y33 = −Y31 = (0.5 − j1.5)

So the bus admittance matrix is therefore

399
Load Flow

⎡ (1.5 − j4.5) (− 1 + j3) (− 0.5 + j1.5)⎤

Ybus = ⎢⎢ (− 1 + j3) (1 − j3) (0) ⎥⎥
⎢⎣(− 0.5 + j1.5) (0) (0.5 − j1.5) ⎥⎦
ii) The load on bus bar 2 is S = 0.2 + j 0.1 pu. The scheduled real and reactive powers on
bus 2 are therefore P2sp = −0.2 and Q sp2 = −0.1

The iterative equation for the voltage on bus 2 is

1 ⎡ P sp − jQ sp ⎤
V2(i +1) = ⎢ 2 *
2
− Y12 ⋅ V1 − Y23 ⋅ V3⎥
Y22 ⎢⎣ V2(i) ⎥⎦

Bus bar 1 is the swing or slack bus, the voltage on bus 1 is therefore equal to 1 pu at an
angle of zero degrees. Substituting known values gives the following iterative
procedure for the voltage on bus 2.

1 ⎡ - 0.2 + j0.1 ⎤
V2(i +1) = ⎢ − (− 1 + j3 ) ⋅ (1.0 + j0.0 )⎥
(1 - j3) ⎢⎣ V2(i) * ⎥⎦

=
(− 0.05 − j 0.05) + 1.0
*
V2(i)

Initial guess for V2 is 1 pu at an angle of zero degrees. The solution proceeds as

follows:

V2(0) = 1.0 + j0.0

V2(1) = 0.94999 - j0.0500
V2(2) = 0.94475 - j0.4972
V2(3) = 0.94444 - j0.0500
V2(4) = 0.94441 − j0.0499

So after four iterations the solution has converged to V2 = (0.94 − j0.05) pu .

The load on bus bar 3 is S = 0.2 + j 0.1 pu. The scheduled real and reactive powers on
bus 3 are therefore P3sp = −0.2 and Q sp3 = −0.1

The iterative equation for the voltage on bus 3 is

400
Load Flow

(i +1) 1 ⎡ P sp − jQ sp ⎤
V3 = ⎢ 3 3
− Y31 ⋅ V1 − Y32 ⋅ V3 ⎥
Y33 (i) *
⎣⎢ V3 ⎦⎥

Substituting known values gives the following iterative procedure for the voltage on bus
3.
1 ⎡ − 0.2 + j 0.1 ⎤
V3(i +1) = ⎢ − (− 0. 5 + j1.5 )(1. 0 + j 0. 0 )⎥
(0.5 - j1.5) ⎢⎣ V3(i) * ⎥⎦

=
(− 0.1 − j 0.1) + 1.0
*
V3(i)

Initial guess for V3 is 1 pu at an angle of zero degrees. The solution proceeds as

follows:

V3(0) = 1.0 + j0.0

V3(1) = 0.89999 − j0.10000
V3(2) = 0.87804 − j0.97560
V3(3) = 0.87500 − j1.00000
V3(4) = 0.87429 − j0.99919
V3(5) = 0.87419 − j0.10000

So after five iterations the solution has converged to V3 = (0.87 − j0.10) pu .

iii) The real and reactive powers on 1 can be determined using the following equation:

P1 − jQ1 = V1* (Y11 ⋅ V1 + Y12 ⋅ V2 + Y12 ⋅ V3 )

* ⎛ (1.5 - j4.5)(1.0 + j0.0 ) + (− 1.0 + j 3.0 )(0.9444 - j0.005 )⎞
= (1.0 + j0.0 ) ⎜⎜ ⎟⎟
⎝ + (− 0.5 + j1.5)(0.874 − j 0.1) ⎠
0.4186 − j 0.2558

The generator on bus bar 1 is therefore delivering 0.418 pu real power and 0.255 pu
reactive power to the network.

9.2.7

i) A line is added linking bus 2 and bus 3 so the mutual admittance Y23 is no longer zero
1
Y23 = Y32 = − = (− 1 + j3)
(0.1 + j0.3)

401
Load Flow

The self admittance of buses 2 and 3 both change due to the addition of the line 2-3

Y33 = −Y31 − Y32

= (0.5 − j1.5) + (1.0 − j3.0)
= (1.5 − j4.5)

Y22 = −Y21 − Y23

= (1.0 − j3.0) + (1.0 − j3.0)
= (2.0 − j6.0)

The modified bus admittance matrix is now

⎡ (1.5 − j4.5) (− 1 + j3) (− 0.5 + j1.5)⎤

Ybus = ⎢⎢ (− 1 + j3) (2 − j6) (− 1 + j3) ⎥⎥
⎢⎣(− 0.5 + j1.5) (− 1 + j3) (1.5 − j4.5) ⎥⎦
ii) The iterative equation for the voltage on bus 2 is

(i +1) 1 ⎡ P sp − jQ sp ⎤
V 2 = ⎢ 2 2
− Y12 ⋅ V1 − Y23 ⋅ V3 ⎥
Y22 (i) *
⎣⎢ V2 ⎦⎥
1 ⎡ (− 0.2 + j0.1) ⎤
= ⎢ − (− 1 + j3 )(1.0 + j0.0 ) − (− 1 + j3 ) ⋅ V3
(i)
⎥
(2 − j6) ⎣⎢ V2(i)
*
⎦⎥
=
(− 0.025 − j0.025) + 0.5 ⋅ V (i) + 0.5
* 3
V2(i)

The iterative equation for the voltage on bus 3 is

1 ⎡ P sp − jQ sp ⎤
V3(i +1) = ⎢ 3 *
3
− Y31 ⋅ V1 − Y32 ⋅ V2⎥
Y33 ⎢⎣ V3(i) ⎥⎦

1 ⎡ (− 0.2 + j0.1) ⎤
= ⎢ − (− 0.5 + j1.5)(1.0 + j0.0 ) − (− 1 + j3) ⋅ V2(i) ⎥
(1.5 − j4.5) ⎢⎣ V3(i)
*
⎥⎦

=
(− 0.033 − j0.033) + 0.6666 ⋅ V (i) + 0.3333
* 2
V3(i)

402
Load Flow

The expression for the voltage on bus 2 is now a function of the voltage on bus 3 and
visa versa. These two equations must now be iterated together to find both voltages.

iii) Assuming a flat start, the first iteration for V2 yields

V2(i +1) =
(− 0.025 − j0.025) + 0.5 ⋅ V (i) + 0.5
* 3
V2(i)

V2(1) =
(− 0.025 − j0.025) + 0.5 ⋅ (1.0 + j0.0) + 0.5
(1.0 + j0.0)
= (0.9750 - j0.2500)

This updated value of V2 is now used to estimate V3.

V3(i +1) =
(− 0.033 − j0.033) + 0.6666 ⋅ V (i) + 0.3333
* 2
V3(i)

V3(1) =
(− 0.033 − j0.033) + 0.6666 ⋅ (0.9750 - j0.025) + 0.3333
(1.00 + j0.0)
= (0.9493 - j0.0496)

A further 7 iterations would be required for the solution to converge to

V2 = (0.9272 − j0.0622 ) V3 = (0.9116 − j0.0744 )

403