Chapter 3

FUNDAMENTAL THEORY OF LOAD

COMPENSATION
(Lectures 19-24)
3.1 Introduction
In general, the loads which cause uctuations in the supply voltage due to poor power factor,
unbalanced and harmonics, d.c components require compensation. Typical loads requiring com-
pensation are arc furnaces, induction furnaces, arc welders, steel rolling mills, winders, very large
motors, which start and stop frequently, high energy physics experiments,which require pulse high
power supplies. All these loads can be classied into three basic categories.
1. Unbalanced ac load
2. Unbalanced ac + non linear load
3. Unbalanced ac + nonlinear ac + dc component load.
The dc component is generally caused by the usage of have wave rectiers. These loads, par-
ticularly nonlinear loads generate harmonics as well as fundamental frequency voltage variations.
For example arc furnaces generate signicant amount of harmonics at the load bus.
Other serious loads which degrade power quality are adjustable speed drives which include power
electronic circuitry, all power electronics based converters such as thyristor controlled drives, rec-
tiers, cyclo converters etc.. In general, following aspects are important, while we do provide the
load compensation in order to improve the power quality [1].
1. Type of Load (unbalance , harmonics and dc component)
2. Real and Reactive power requirements (maximum, minimum and concurrence of maximum real
and reactive power requirements in multiple loads)
3. Rate of change of real and reactive power etc.
In this unit, we however, discuss fundamental load compensation techniques for unbalanced linear
loads such as combination of resistance, inductance and capacitance and their combinations. The
69
objective here will be to maintain currents balanced and unity factor with their voltages.
3.2 Fundamental Theory of Load Compensation
We shall nd some fundamental relation ship between supply system the load and the compensator.
We shall start with the principle of power factor correction, which in its simplest form, and can be
studied without reference to supply system [2][5].
The supply system, the load and the compensator can be modeled in various ways. The supply
system can be modeled as a Thevenins equivalent circuit with an open circuit voltage and a se-
ries impedance, (its current or power and reactive power) requirements. The compensator can be
modeled as variable impedance or as a variable source (or sink) of reactive current. The choice of
model varied according to the requirements. The modeling and analysis done here is on the basis
of steady state and phasor quantities are used to note the various parameters in system.
3.2.1 Power Factor and its Correction
Consider a single phase system shown in 3.1(a) shown below. The load admittance is represented
V
s
I
l
I
V
R
I
l
I
X
I
l

Fig. 3.1 (a) Single line diagram of electrical system (b) Phasor diagram
by Y
l
= G
l
+jB
l
supplied from a load bus at voltage V = V 0. The load current is I
l
is given as,
I
l
= V (G
l
+jB
l
) = V G
l
+jV B
l
(3.1)
= I
R
+jI
X
According to the above equation, the load current has a two components, i.e. the resistive or in
phase component and reactive component or phase quadrature component and are represented by
I
R
and I
X
respectively. The current, I
X
will lag 90
o
for inductive load and it will lead 90
o
from
the reference voltage. This is shown in 3.1(b). The load apparent power can be expressed in terms
70
of bus voltage V and load current I
l
as given below.
S
l
= V (I
l
)

= V (I
R
+j I
X
)

= V (I
R
j I
X
)
= V (I
l
cos
l
j I
l
sin
l
)
= V I
l
cos
l
j V I
l
sin
l
= S
l
cos
l
j S
l
sin
l
(3.2)
From (3.1), I
l
= V (G
l
+jB
l
) = V G
l
+jV B
l
, equation (3.2) can also be written as following.
S
l
= V (I
l
)

= V (V G
l
+jV B
l
)

= V (V G
l
jV B
l
)
= V
2
G
l
j V
2
B
l
= P
l
+jQ
l
(3.3)
From equation (3.129), load active (P
l
) and reactive power (Q
l
) are given as,
P
l
= V
2
G
l
Q
l
= V
2
B
l
(3.4)
Now suppose a compensator is connected across the load such that the compensator current, I

is
equal to I
X
, thus,
I

= V Y

= V (G

+jB

) = I
X
= j V B
l
(3.5)
The above condition implies that G

= 0 and B

= B
l
. The source current I
s
, can therefore
given by,
I
s
= I
l
+I = I
R
(3.6)
Therefore due to compensator action, the source supplies only in phase component of the load
current. The source power factor is unity. This reduces the rating of the power conductor and
losses due to the feeder impedance. The rating of the compensator is given by the following
expression.
S

= P

+j Q

= V (I

= V (j V B
l
)

= jV
2
B
l
(3.7)
Using (3.4), the above equation indicates the P

= 0 and Q

= Q
l
. This is an interesting
inference that the compensator generates the reactive power which is equal and opposite to the
load reactive and it has no effect on active power of the load. This is shown in Fig. 3.2.
71
X
I
I

s R
I I
V
s
I
l
I I

l
I
l

X
I
Fig. 3.2 (a) Single line diagram of compensated system (b) Phasor diagram
Using (3.2) and (3.7), the compensator rating can further be expressed as,
Q

= Q
l
= S
l
sin
l
= S
l
_
1 cos
2

l
VArs (3.8)
From (3.8),
|Q

| =
_
1 cos
2

l
(3.9)
If |Q

| < |Q
l
| or |B

| < |B
l
|, then load is partially compensated. The compensator of xed
admittance is incapable of following variations in the reactive power requirement of the load. In
practical however a compensator such as a bank of capacitors can be divided into parallel sections,
each of switched separately, so that discrete changes in the reactive power compensation can be
made according to the load. Some sophisticated compensators can be used to provide smooth and
dynamic control of reactive power.
Here voltage of supply is being assumed to be constant. In general if supply voltage varies, the
Q

will not vary separately with the load and compensator error will be there. In the following
discussion, voltage variations are examined and some additional features of the ideal compensator
will be studied.
3.2.2 Voltage Regulation
Voltage regulation can be dened as the proportional change in voltage magnitude at the load bus
due to change in load current (say from no load to full load). The voltage drop is caused due to
feeder impedance carrying the load current as illustrated in Fig. 3.3(a). If the supply voltage is
represented by Thevenins equivalent, then the voltage regulation (VR) is given by,
V R =

|V |
|V |
(3.10)
for V being a reference phasor.
In absence of compensator, the source and load currents are same and the voltage drop due to the
72
feeder is given by,
V = E V = Z
s
I
l
(3.11)
The feeder impedance, Z
s
= R
s
+ jX
s
. The relationship between the load powers and its voltage
and current is expressed below.
S
l
= V (I
l
)

= P
l
+jQ
l
(3.12)
Since V = V , the load current is expressed as following.
I
l
=
P
l
jQ
l
V
(3.13)
Substituting, I
l
from above equation into (3.11), we get
V = E V = (R
s
+jX
s
)
_
P
l
jQ
l
V
_
=
R
s
P
l
+X
s
Q
l
V
+j
X
s
P
l
R
s
Q
l
V
= V
R
+jV
X
(3.14)
Thus, the voltage drop across the feeder has two components, one in phase (V
R
) and another is
in phase quadrature (V
X
). This is illustrated in Fig. 3.3(b).
E
V
R
V
X
j V
l
I
l
R I
l l
jX I
l

l
I
l l l
S P jQ
l l l
Y G jB
L
o
a
d
s
I
V
E
F
e
e
d
e
r
s s s
Z R jX
Fig. 3.3 (a) Single phase system with feeder impedance (b) Phasor diagram
From the above it is evident that load bus voltage (V ) is dependent on the value of the feeder
impedance, magnitude and phase angle of the load current. In other words, voltage change (V )
depends upon the real and reactive power ow of the load and the value of the feeder impedance.
Now let us add compensator in parallel with the load as shown in Fig. 3.4(a). The question is:
whether it is possible to make

, in order to achieve zero voltage regulation irrespective of

change in the load. The answer is yes, if the compensator consisting of purely reactive components,
has enough capacity to supply to required amount of the reactive power. This situation is shown
using phasor diagram in Fig. 3.4(b).
The net reactive at the load bus is now Q
s
= Q

+ Q
l
. The compensator reactive power (Q

) has
to be adjusted in such a way as to rotate the phasor V until

.
73
E
V
l
I
l

l
I
l l l
S P jQ
l l l
Y G jB
L
o
a
d
s
I
V
E
F
e
e
d
e
r
s s s
Z R jX
C
o
m
p
.
I

V
I

s
I
s s
R I
s s
jX I
Fig. 3.4 (a) Voltage with compensator (b) Phasor diagram
From (3.14) and Fig. 3.3(b),
E =
_
V +
R
s
P
l
+X
s
Q
s
V
_
+j
_
X
s
P
l
R
s
Q
s
V
_
(3.15)
The above equation implies that,
E
2
=
_
V +
R
s
P
l
+X
s
Q
s
V
_
2
+
_
X
s
P
l
R
s
Q
s
V
_
2
(3.16)
The above equation can be simplied to,
E
2
V
2
= (V
2
+R
s
P
l
)
2
+X
2
s
Q
2
s
+ 2(V
2
+$
$
$
R
s
P
l
) X
s
Q
s
+X
2
s
P
2
l
+R
2
s
Q
2
s

@
@
@
@
@
@@
2X
s
P
l
R
s
Q
s
(3.17)
Above equation, rearranged in the powers of Q
s
, is written as following.
(R
2
s
+X
2
s
) Q
2
s
+ 2V
2
X
s
Q
s
+ (V
2
+R
s
P
l
)
2
+ (X
s
P
l
)
2
E
2
V
2
= 0 (3.18)
Thus the above equation is quadratic in Q
s
and can be represented using coefcients of Q
s
as given
below.
a Q
2
s
+b Q
s
+c = 0 (3.19)
Where a = R
2
s
+X
2
s
, b = 2V
2
X
s
and c = (V
2
+R
s
P
l
)
2
+X
2
s
P
2
l
E
2
V
2
.
Thus the solution of above equation is as following.
Q
s
=
b
_
(b
2
4ac)
2a
(3.20)
In the actual compensator, this value would be determined automatically by control loop. The
equation also indicates that, we can nd the value of Q
s
by subjecting a condition such as E = V
74
irrespective of the requirement of the load powers (P
l
, Q
l
). This leads to the following conclusion
that a purely reactive compensator can eliminate supply voltage variation caused by changes in
both the real and reactive power of the load, provided that there is sufcient range and rate of Q
s
both in lagging and leading pf. This compensator therefore acts as an ideal voltage regulator. It
is mentioned here that we are regulating magnitude of voltage and not its phase angle. In fact its
phase angle is continuously varying depending upon the load current.
It is instructive to consider this principle from different point of view. We have seen that com-
pensator can be made to supply all load reactive power and it acts as power factor correction device.
If the compensator is designed to compensate power factor, then Q
s
= Q
l
+ Q

= 0. This im-
plies that Q

= Q
l
. Substituting Q
s
= 0 for Q
l
in (3.14) to achieve this condition, we get the
following.
V =
(R
s
+jX
s
)
V
P
l
(3.21)
From above equation, it is observed that V is independent of Q
l
. Thus we conclude that a purely
reactive compensator cannot maintain both constant voltage and unity power factor simultaneously.
Of course the exception to this rule is a trivial case when P
l
= 0.
3.2.3 An Approximation Expression for the Voltage Regulation
Consider a supply system with short circuit capacity (S
sc
) at the load bus. This short circuit capac-
ity can be expressed in terms of short circuit active and reactive powers as given below.
S
sc
= P
sc
+jQ
sc
= E I

sc
= E
_
E
Z
sc
_

=
E
2
Z

sc
(3.22)
Where Z
sc
= R
s
+jX
s
and I
sc
is the short circuit current. From the above equation
|Z
sc
| =
E
2
S
sc
Therefore, R
s
=
E
2
S
sc
cos
sc
X
s
=
E
2
S
sc
sin
sc
tan
sc
=
X
s
R
s
(3.23)
Substituting above values of R
s
and X
s
, (3.14) can be written in the following form.
V
V
=
_
P
l
cos
sc
+Q
l
sin
sc
V
2
+j
P
l
sin
sc
Q
l
cos
sc
V
2
_
E
2
S
sc
V
V
=
V
R
V
+j
V
X
V
(3.24)
75
Using an approximation that E V , the above equation reduces to the following.
V
V
=
_
P
l
cos
sc
+Q
l
sin
sc
S
sc
+j
P
l
sin
sc
Q
l
cos
sc
S
sc
_
(3.25)
The above implies that,
V
R
V

P
l
cos
sc
+Q
l
sin
sc
S
sc
V
X
V

P
l
sin
sc
Q
l
cos
sc
S
sc
Often (V
X
/V ) is ignored on the ground that the phase quadrature component contributes negli-
gible to the magnitude of overall phasor. It mainly contributes to the phase angle. Therefore the
equation (3.25) is simplied to the following.
V
V
=
V
R
V
=
P
l
cos
sc
+Q
l
sin
sc
S
sc
(3.26)
Implying that the major change in voltage regulation occurs due to in phase component, V
R
.
Although approximate, the above expression is quite useful in terms of short circuit level (S
sc
),
(X
s
/R
s
, active and reactive power of the load.
On the basis of incremental changes in active and reactive powers of the load, i.e., 0 P
l
and
0 Q
l
, the above equation can further be written as,
V
V
=
V
R
V
=
P
l
cos
sc
+ Q
l
sin
sc
S
sc
. (3.27)
Further, feeder reactance (X
s
) is far greater than feeder resistance (R
s
), i.e., X
s
>> R
s
. This
implies that,
sc
90
o
, sin
sc
1 and cos
sc
0. Using this approximation the voltage
regulation is given as following.
V
V

V
R
V

Q
l
S
sc
sin
sc

Q
l
S
sc
. (3.28)
That is, per unit voltage change is equal to the ratio of the reactive power swing to the short circuit
level of the supply system. Representing V approximately by E V , the equation (3.28) can be
writtwn as,
E V
V

Q
l
S
sc
. (3.29)
The above leads to the following expression,
V
E
(1 +
Q
l
Ssc
)
E(1
Q
l
S
sc
) (3.30)
76
with the assumption that, Q
l
/S
sc
<< 1. Although above relationship is obtained with approxi-
mations, however it is very useful in visualizing the action of compensator on the voltage. The
above equation is graphically represented as Fig. 3.5. The nature of voltage variation is drooping
with increase in inductive reactive power of the load. This is shown by negative slope E/S
sc
as
indicated in the gure.
The above characteristics also explain that when load is capacitive, Q
l
is negative. This makes
V > E. This is similar to Ferranti effect due to lightly loaded electric lines.
0
l
Q
V
sc
E
S

E
Fig. 3.5 Voltage variation with reactive power of the load
Example 3.1 Consider a supply at 10 kV line to neutral voltage with short circuit level of 250
MVA and X
s
/R
s
ratio of 5, supplying a star connected load inductive load whose mean power is
25 MW and whose reactive power varies from 0 to 50 MVAr, all quantities per phase.
(a) Find the load bus voltage (V ) and the voltage drop (V ) in the supply feeder. Thus determine
load current (I
l
), power factor and system voltage (E ).
(b) It is required to maintain the load bus voltage to be same as supply bus voltage i.e. V =10 kV.
Calculate reactive power supplies by the compensator.
(c) What should be the load bus voltage and compensator current if it is required to maintain the
unity power factor at the supply?
Solution: The feeder resistance and reactance are computed as following.
Z
s
= E
2
s
/S
sc
= (10 kV)
2
/250 = 0.4 /phase
It is given that, X
s
/R
s
= tan
sc
= 5, therefore
sc
= tan
1
5 = 78.69
o
. From this,
R
s
= Z
s
cos
sc
= 0.4 cos(78.69
o
) = 0.0784
X
s
= Z
s
sin
sc
= 0.4 sin(78.69
o
) = 0.3922
(a) Without compensation Q
s
= Q
l
, Q

= 0
To know V , rst the voltage at the load bus has to be computed. This is done by rearranging
77
(3.18) in powers of voltage V . This is given below.
(R
2
s
+X
2
s
) Q
2
l
+ 2 V
2
X
s
Q
l
+ (V
2
+R
s
P
l
)
2
+X
2
s
P
2
l
E
2
V
2
= 0
(R
2
s
+X
2
s
) Q
2
l
. .
III
+2 V
2
X
s
Q
l
. .
II
+ V
4
..
I
+R
2
s
P
2
l
. .
III
+2 V
2
R
s
P
l
. .
II
+X
2
s
P
2
l
. .
III
E
2
V
2
. .
II
= 0
Combining the I, II and III terms in the above equation, we get the following.
V
4
+
_
2(R
s
P
l
+X
s
Q
l
) E
2
_
V
2
+ (R
2
s
+X
2
s
)(Q
2
l
+P
2
l
) = 0 (3.31)
Now substituting values of R
s
, X
s
, P
l
, Q
l
and E in above equation, we get,
V
4
+
_
2 [0.0784 25 + 0.3922 50] 10
2
_
V
2
+ (0.0784
2
+ 0.3922
2
)(25
2
+ 50
2
) = 0
After simplifying the above, we have the following equation.
V
4
56.86V
2
+ 500 = 0
Therefore
V
2
=
56.86

56.86 4 500
2
= 45.985, 10.875
and V = 6.78 kV, 3.297 kV
Since rms value cannot be negative and maximum rms value must be a feasible solution, therefore
V = 6.78 kV.
Now we can compute V using (3.14), as it is given below.
V =
R
s
P
l
+X
s
Q
l
V
+j
X
s
P
l
R
s
Q
l
V
=
0.0784 25 + 0.392 50
6.78
+j
0.3922 25 0.0784 50
6.78
= 3.1814 +j0.8677 kV = 3.297615.25
o
kV
Now the line current can be found out as following.
I
l
=
P
l
Q
l
V
=
25 j50
6.782
= 3.86 j7.3746 kA
= 8.242 63.44
o
kA
The power factor of load is cos (tan
1
(Q
l
/P
l
)) = 0.4472 lagging. The phasor diagram for this
case is similar to what is shown in Fig. 3.3(b).
(b) Compensator as a voltage regulator
Now it is required to maintain V = E = 10.0 kV at the load bus. For this let their be reac-
tive power Q

supplied by the compensator at the load bus. Therefore the net reactive power at the
load bus is equal to Q
s
, which is given below.
Q
s
= Q
l
+Q

78
Thus from (3.18), we get,
(R
2
s
+X
2
s
)Q
2
s
+ 2V
2
X
s
Q
s
+ (V
2
+R
s
P
l
)
2
+X
2
s
P
2
l
E
2
V
2
= 0
(0.784
2
+ 0.3922
2
)
2
Q
2
s
+ 2 10
2
0.3922 Q
s
+
_
(10
2
+ 0.784 25)
2
+ 0.3922
2
25
2
10
4
_
= 0
From the above we have,
0.16 Q
2
s
+ 78.44 Q
s
+ 491.98 = 0.
Solving the above equation we get,
Q
s
=
78.44

78.44
2
4 0.16 491.98
2 0.16
= 6.35 or 484 MVAr.
The feasible solution is Q
s
= 6.35 MVAr because it requires less rating of the compensator.
Therefore the reactive power of the compensator (Q

) is,
Q

= Q
s
Q
l
= 6.35 50 = 56.35 MVAr.
With Q
s
= 6.35 MVAr, the V is computed by replacing Q
s
for Q
l
in (3.14) as given below.
V =
R
s
P
l
+X
s
Q
s
V
+j
X
s
P
l
R
s
Q
s
V
=
0.0784 25 + 0.39225 6.35
10
+j
0.39225 25 0.0784 (6.35)
10
=
1.96 2.4
10
+j
9.805 + 0.4978
10
= 0.0532 +j1.030 kV = 1.0313792.95
o
kV
Now, we can nd supply voltage E as given below.
E = V +V
= 10 0.0532 +j1.030
= 9.9468 +j1.030 = 105.91
o
kV
The supply current is,
I
s
=
P
l
jQ
s
V
=
25 j(6.35)
10
= 2.5 +j0.635 kA = 2.57914.25
o
kA.
This indicates that power factor is not unity for perfect voltage regulation i.e., E = V . For this
case the compensator current is given below.
I

=
jQ

V
=
j(56.35)
10
I

= j5.635 kA
79
The load current is computed as below.
I
l
=
P
l
jQ
l
V
=
25 j50
10
= 2.5 j5.0 = I
lR
+jI
lX
= 5.5963.44
o
kA
The phasor diagram is similar to the one shown in Fig. 3.4(b). The phasor diagram shown has
interesting features. The voltage at the load bus is maintained to 1.0 pu. It is observed that the
reactive power of the compensator Q

is not equal to load reactive power (Q

l
). It exceeds by 6.35
MVAr. As a result of this compensation, the voltage regulation is perfect, however power factor is
not unity. The phase angle between V and I
s
is cos
1
0.969 = 14.25
o
as computed above. There-
fore the angle between E and I
s
is (14.25
o
5.91
o
= 8.34
o
). Thus, source power factor (
s
) is
cos(8.34
0
) = 0.9956 leading.
(c) Compensation for unity power factor
To achieve unity power factor at the load bus, the condition Q

= Q
l
must be satised, which
further implies that the net reactive power at the load bus is zero. Therefore substituting Q
l
= 0 in
(3.31), we get the following.
V
4
+
_
2(R
s
P
l
E
2
)
_
V
2
+ (R
2
s
+X
2
s
)(P
2
l
+Q
2
l
) = 0
V
4
+ (2 0.0784 25 10
2
)V
2
+ (0.0784
2
+ 0.3922
2
) 25
2
= 0
From the above,
V
4
+ 96.08V
2
+ 99.79 = 0
The solution of the above equation is,
V
2
=
96.08 93.97
2
= 95.02, 1.052
V = 9.747 kV, 1.0256 kV.
Since rms value cannot be negative and maximum rms value must be a feasible solution, therefore
V = 9.747 kV. Thus it is seen that for obtaining unity power factor at the load bus does not ensure
desired voltage regulation. Now the other quantities are computed as given below.
I
l
=
P
l
jQ
V
=
25 j50
9.747
= 2.5648 j5.129 = 5.734563.43
o
kA
Since Q

= Q
l
, this implies that I

= jQ

/V = jQ
l
/V = j5.129 kA. The voltage drop across
the feeder is given as following.
V =
R
s
P
l
+X
s
Q
l
V
+j
X
s
P
l
R
s
Q
l
V
=
(0.784 25 +j0.3922 25)
9.747
= 0.201 +j1.005 = 1.02495.01
o
kV
80
The phasor diagram for the above case is shown in Fig. 3.6.

5
.
1
3
k
A
I
j

10 kV E
V =9.75 kV

5
.
1
3
k
A
l
X
I
j

=2.56kA
lR
I
R
V
X
V
5.73 63.43 kA
l
I


5.77

s
I
Fig. 3.6 Phasor diagram for system with compensator in voltage regulation mode
The percentage voltage change = (10 9.748)/10 100 = 2.5. Thus we see that power factor
improves voltage regulation enormously compared with uncompensated case. In many cases, de-
gree of improvement is adequate and the compensator can be designed to provide reactive power
requirement of load rather than as a ideal voltage regulator.
3.3 Some Practical Aspects of Compensator used as Voltage Regulator
In this section, some practical aspects of the compensator in voltage regulation mode will be dis-
cussed. The important parameters of the compensator which play signicant role in obtaining
desired voltage regulation are: Knee point (V
k
), maximum or rated reactive power Q
max
and the
compensator gain K

.
The compensator gain K

is dened as the rate of change of compensator reactive power Q

with change in the voltage (V ), as given below.

K

=
dQ
r
dV
(3.32)
For linear relationship between Q

and V with incremental change, the above equation be written

as the following.
Q

= V K

(3.33)
81
Assuming compensator characteristics to be linear with Q

Q
max
limit, the voltage can be
represented as,
V = V
k
+
Q

(3.34)
This is re-written as,
Q

= K

(V V
k
) (3.35)
Flat V-Q characteristics imply that K

. That means the compensator which can absorb or

generate exactly right amount of reactive power to maintain supply voltage constant as the load
varies without any constraint. We shall now see the regulating properties of the compensator,
when compensator has nite gain K
r
operating on supply system with a nite short circuit level,
S
sc
. The further which are made in the following study are: high X
s
/R
s
ratio and negligible load
power uctuations. The net reactive power at the load bus is sum of the load and the compensator
reactive power as given below.
Q
l
+Q

= Q
s
(3.36)
Using earlier voltage and reactive power relationship from equation (3.30), it can be written as the
following.
V E(1
Q
s
S
sc
) (3.37)
The compensator voltage represented by (3.34) and systemvoltage represented by (3.37) are shown
in Fig. 3.7(a) and (b) respectively.
k
V
Q

k
Q
V V
K

= +
V
E
s
Q
1-
s
SC
Q
V E
S
| |
~
|
\ .
(a) (b)
Fig. 3.7 (a) Voltage characterstics of compensator (b) System voltage characteristics
Differentiating V with respect to Q
s
, we get, intrinsic sensitivity of the supply voltage with
variation in Q
s
as given below.
dV
dQ
s
=
E
S
sc
(3.38)
82
It is seen from the above equation that high value of short circuit level S
sc
reduces the voltage
sensitivity, making voltage variation at irrespective of Q
l
. With compensator replacing Q
s
=
Q

+Q
l
in (3.37), we have the following.
V E
_
1
Q
l
+Q

S
sc
_
(3.39)
Substituting Q

from (3.35), we get the following equation.

V E
_
1 +K

V
k
/S
sc
1 +E K

/S
sc

Q
l
/S
sc
1 +E K

/S
sc
_
(3.40)
Although approximate, above equation gives the effects of all the major parameters such as load
reactive power, the compensator characteristics V

and K

and the system characteristics E and

S
sc
. As we discussed, V-Q characteristics is at for high or innite value K

. However the higher

value of the gain K

means large rating and quick rate of change of the reactive power with varia-
tion in the system voltage. This makes cost of the compensator high.
The compensator has two effects as seen from (3.40), i.e., it alters the no load supply voltage
(E) and it modies the sensitivity of supply point voltage to the variation in the load reactive power.
Differentiating (3.40) with respect to Q
l
, we get,
dV
dQ
l
=
E/S
sc
1 +K
r
E/S
sc
(3.41)
which is voltage sensitivity of supply point voltage to the load reactive power. It can be seen that
the voltage sensitivity is reduced as compared to the voltage sensitivity without compensator as
indicated in (3.38).
It is useful to express the slope (E/S
sc
) by a term in a form similar to K

= dQ

/dV , as given
below.
K
s
=
S
sc
E
Thus,
1
K
s
=
E
S
sc
(3.42)
Substituting V from (3.39) into (3.35), the following is obtained.
Q

= K

_
E
_
1
Q
l
+Q

S
sc
_
V
k
_
(3.43)
Collecting the coefcients of Q

from both sides of the above equation, we get

Q

=
K

1 +K

(E/S
sc
)
_
E
_
1
Q
l
S
sc
_
V
k
_
(3.44)
83
Setting knee voltage V
k
of the compensator equal to system voltage E i.e., V
k
= E, the above
equation is simplied to,
Q

=
K

(E/S
sc
)
1 +K

(E/S
sc
)
Q
l
=
_
K

/K
s
1 +K

/K
s
_
Q
l
. (3.45)
From the above equation it is observed that, when compensator gain K

, Q

Q
l
. This
indicates perfect compensation of the load reactive power in order to regulate the load bus voltage.
Example 3.2 Consider a three-phase system with line-line voltage 11 kV and short circuit capacity
of 480 MVA. With compensator gain of 100 pu determine voltage sensitivity with and without
compensator. For each case, if a load reactive power changes by 10 MVArs, nd out the change in
load bus voltage assuming linear relationship between V-Q characteristics. Also nd relationship
between compensator and load reactive powers.
Solution: The voltage sensitivity can be computed using the following equation.
dV
dQ
l
=
E/S
sc
1 +K

E/S
sc
Without compensator K

= 0, E = (11/

3) = 6.35 kV and S
sc
= 480/3 = 160 MVA.
Substituting these values in the above equation, the voltage sensitivity is given below.
dV
dQ
l
=
6.35/160
1 + 0 6.35/160
= 0.039
The change in voltage due to variation of reactive power by 10 MVArs, V = 0.039 10 =
0.39 kV.
With compensator, K

= 100
dV
dQ
l
=
6.35/160
1 + 100 6.35/160
= 0.0078
The change in voltage due to variation of reactive power by 10 MVArs, V = 0.0078 10 =
0.078 kV.
Thus it is seen that, with nite compensator gain their is quite reduction in the voltage sensitivity,
which means that the load bus is fairly constant for considerable change in the load reactive power.
The compensator reactive power Q

and load reactive power Q

l
are related by equation (3.45) and
is given below.
Q

=
K

(E/S
sc
)
1 +K

(E/S
sc
)
Q
l
=
100 (6.35/160)
1 + 100 (6.35/160)
Q
l
= 0.79 Q
l
It can be observed that when compensator gain, (Q

) is quite large, then compensator reactive

power Q

is equal and opposite to that of load reactive power i.e., Q

= Q
l
. It is further observed
84
that due to nite compensator gain i.e., K

= 100, reactive power is partially compensated The

compensator reactive power varies from 0 to 7.9 MVAr for 0 to 10 MVAr change in the load
reactive power.
3.4 Phase Balancing and Power Factor Correction of Unbalanced Loads
So far we have discussed voltage regulation and power factor correction for single phase systems.
In this section we will focus on balancing of three-phase unbalanced loads. In considering unbal-
anced loads, both load and compensator are modeled in terms of their admittances and impedances.
3.4.1 Three-phase Unbalanced Loads
Consider a three-phase three-wire system suppling unbalanced load as shown in Fig. 3.8.
an V
a
Z
b
Z
c
Z
bn V
cn V
N
n
a I
1 I
2 I
b I
c I
Fig. 3.8 Three-phase unbalanced load
Applying Kirchoffs voltage law for the two loops shown in the gure, we have the following
equations.
V
an
+Z
a
I
1
+Z
b
(I
1
I
2
) +V
bn
= 0
V
bn
+Z
b
I
2
+Z
b
(I
2
I
1
) +V
cn
= 0 (3.46)
Rearranging above, we get the following.
V
an
V
bn
= (Z
a
+Z
b
) I
1
Z
b
I
2
V
bn
V
cn
= (Z
b
+Z
c
) I
2
Z
b
I
1
(3.47)
The above can be represented in matrix form as given below.
_
V
an
V
bn
V
bn
V
cn
_
=
_
(Z
a
+Z
b
) Z
b
Z
b
(Z
b
+Z
c
)
_ _
I
1
I
2
_
(3.48)
85
Therefore the currents are given as below.
_
I
1
I
2
_
=
_
(Z
a
+Z
b
) Z
b
Z
b
(Z
b
+Z
c
)
_
1
_
V
an
V
bn
V
bn
V
cn
_
=
1

Z
_
(Z
b
+Z
c
) Z
b
Z
b
(Z
a
+Z
b
)
_ _
V
an
V
bn
V
bn
V
cn
_
Therefore,
_
I
1
I
2
_
=
1

Z
_
(Z
b
+Z
c
) Z
b
Z
b
(Z
a
+Z
b
)
_ _
V
an
V
bn
V
bn
V
cn
_
. (3.49)
Where,
Z
= (Z
b
+Z
c
)(Z
a
+Z
b
) Z
2
b
= Z
a
Z
b
+Z
b
Z
c
+Z
c
Z
a
. The current I
1
is given below.
I
1
=
1

Z
_
(Z
b
+Z
c
)(V
an
V
bn
) +Z
b
(V
bn
V
cn
)

=
1

Z
_
(Z
b
+Z
c
)V
an
Z
c
V
bn
Z
b
V
cn

(3.50)
Similarly,
I
2
=
1

Z
_
Z
b
(V
an
V
bn
) + (Z
a
+Z
b
)(V
bn
V
cn
)

=
1

Z
_
Z
b
V
an
+Z
a
V
bn
(Z
a
+Z
b
)V
cn

(3.51)
Now,
I
a
= I
1
=
1

Z
_
(Z
b
+Z
c
)V
an
Z
c
V
bn
Z
b
V
cn

I
b
= I
2
I
1
=
1

Z
_
$
$
$$
Z
b
V
an
+Z
a
V
bn
(Z
a
+Z
b
)V
cn
(

Z
b
+Z
c
)V
an
+Z
c
V
bn
+Z
b
V
cn

=
_
(Z
c
+Z
a
)V
bn
Z
a
V
cn
Z
c
V
an

Z
=
(Z
c
+Z
a
)V
bn
Z
c
V
an
Z
a
V
cn

Z
(3.52)
and
I
c
= I
2
= I
b
I
a
=
(Z
a
+Z
b
)V
cn
Z
b
V
an
Z
a
V
bn

Z
(3.53)
Alternatively phase currents can be expressed as following.
I
a
=
V
an
V
Nn
Z
a
I
b
=
V
bn
V
Nn
Z
b
(3.54)
I
c
=
V
cn
V
Nn
Z
c
86
Applying Kirchoffs current law at node N, we get I
a
+ I
b
+ I
c
= 0. Therefore from the above
equation,
V
an
V
Nn
Z
a
+
V
bn
V
Nn
Z
b
+
V
cn
V
Nn
Z
c
= 0. (3.55)
Which implies that,
V
an
Z
a
+
V
bn
Z
b
+
V
cn
Z
c
=
_
1
Z
a
+
1
Z
b
+
1
Z
c
_
V
Nn
=
Z
a
Z
b
+Z
b
Z
c
+Z
c
Z
a
Z
a
Z
b
Z
c
V
Nn
(3.56)
From the above equation the voltage between the load and system neutral can be found. It is given
below.
V
Nn
=
Z
a
Z
b
Z
c

Z
_
V
an
Z
a
+
V
bn
Z
b
+
V
cn
Z
c
_
=
1
1
Za
+
1
Z
b
+
1
Zc
_
V
an
Z
a
+
V
bn
Z
b
+
V
cn
Z
c
_
(3.57)
Some interesting points are observed from the above formulation.
1. If both source voltage and load impedances are balanced i.e., Z
a
= Z
b
= Z
c
= Z, then
V
Nn
=
1
3
_
V
an
+V
bn
+V
cn
_
= 0. Thus their will not be any voltage between two neutrals.
2. If supply voltage are balanced and load impedances are unbalanced, then V
Nn
=0 and is
given by the above equation.
3. If supply voltages are not balanced but load impedances are identical, then V
Nn
=
1
3
_
V
an
+V
bn
+V
cn
_
.
This equivalent to zero sequence voltage V
0
.
It is interesting to note that if the two neutrals are connected together i.e., V
Nn
= 0, then each
phase become independent through neutral. Such conguration is called three-phase four-wire
system. In general, three-phase four-wire system has following properties.
V
Nn
= 0
I
a
+I
b
+I
c
= I
Nn
= 0
_
(3.58)
The current I
Nn
is equivalent to zero sequence current (I
0
) and it will ow in the neutral wire.
For three-phase three-wire system, the zero sequence current is always zero and therefore following
properties are satised.
V
Nn
= 0
I
a
+I
b
+I
c
= 0
_
(3.59)
Thus, it is interesting to observe that three-phase three-wire and three-phase four-wire system have
dual properties in regard to neutral voltage and current.
87
c
I
b
I
a
I
ab
l
Y
bc
l
Y
c
a
l Y
c
a
b
c
I
b
I
a
I
c
l
Z
c
a
b
a
l
Z
b
l
Z
n
(a) (b)
Fig. 3.9 (a) An unbalanced delta connected load (b) Its equivalent star connected load
3.4.2 Representation of Three-phase Delta Connected Unbalanced Load
A three-phase delta connected unbalanced and its equivalent star connected load are shown in Fig.
3.9(a) and (b) respectively. The three-phase load is represented by line-line admittances as given
below.
Y
ab
l
= G
ab
l
+jB
ab
l
Y
bc
l
= G
bc
l
+jB
bc
l
Y
ca
l
= G
ca
l
+jB
ca
l
_

_
(3.60)
The delta connected load can be equivalently converted to star connected load using following
expressions.
Z
a
l
=
Z
ab
l
Z
ca
l
Z
ab
l
+Z
bc
l
+Z
ca
l
Z
b
l
=
Z
bc
l
Z
ab
l
Z
ab
l
+Z
bc
l
+Z
ca
l
Z
c
l
=
Z
ca
l
Z
bc
l
Z
ab
l
+Z
bc
l
+Z
ca
l
_

_
(3.61)
Where Z
ab
l
= 1/Y
ab
l
, Z
bc
l
= 1/Y
bc
l
and Z
ca
l
= 1/Y
ca
l
. The above equation can also be written in
admittance form
Y
a
l
=
Y
ab
l
Y
bc
l
+Y
bc
l
Y
ca
l
+Y
ca
l
Y
ab
l
Y
bc
l
Y
b
l
=
Y
ab
l
Y
bc
l
+Y
bc
l
Y
ca
l
+Y
ca
l
Y
ab
l
Y
ca
l
Y
c
l
=
Y
ab
l
Y
bc
l
+Y
bc
l
Y
ca
l
+Y
ca
l
Y
ab
l
Y
ab
l
_

_
(3.62)
Example 3.3 Consider three-phase system supply a delta connected unbalanced load with Z
l
a
=
R
a
= 10 , Z
l
b
= R
b
= 15 and Z
l
c
= R
c
= 30 as shown in Fig. 3.8. Determine the voltage
88
between neutrals and nd the phase currents. Assume a balance supply voltage with rms value of
230 V. Find out the vector and arithmetic power factor. Comment upon the results.
Solution: The voltage between neutrals V
Nn
is given as following.
V
Nn
=
R
a
R
b
R
c
R
a
R
b
+R
b
R
c
+R
c
R
a
_
V
an
R
a
+
V
bn
R
b
+
V
cn
R
c
_
=
10 15 30
10 15 + 15 30 + 30 10
_
V 0
o
10
+
V 120
o
15
+
V 120
o
30
_
=
4500
900
_
3V 0
o
+ 2V 120
o
+V 120
o
30
_
=
4500
900
1
30
V
_
3 + 2
_

1
2
j

3
2
_
+
_

1
2
+j

3
2
__
=
V
6
_
3 1
1
2
j2

3
2
+j

3
2
_
=
V
6
_
3
2
j

3
2
_
= V
_
1
4
j
1
4

3
_
= V [0.25 j0.1443]
V
Nn
=
V
2

3
30
o
= 66.3930
o
Volts
Knowing this voltage, we can nd phase currents as following.
I
a
=
V
an
V
Nn
R
a
=
V 0
o
V/(2

3)30
o
10
=
V [1 0.25 +j0.1443]
10
= 230 [0.075 +j0.01443]
= 17.5610.89
o
Amps
Similarly,
I
b
=
V
bn
V
Nn
R
b
=
V 120
o
V/(2

3)30
o
15
= 230 [0.05 j0.04811]
= 15.94136.1
o
Amps
and
I
c
=
V
cn
V
Nn
Z
c
=
V 120
o
V/(2

3)30
o
30
= 230 [0.025 +j0.03367]
= 9.64 126.58
o
Amps
89
It can been seen that I
a
+I
b
+I
c
= 0. The phase powers are computed as below.
S
a
= V
a
(I
a
)

= P
a
+jQ
a
= 230 17.56 10.81
o
= 3976.12 j757.48 VA
S
b
= V
b
(I
b
)

= P
b
+jQ
b
= 230 15.94(120
o
+ 136.1
o
= 3522.4 +j1016.69 VA
S
c
= V
c
(I
c
)

= P
c
+jQ
c
= 230 9.64(120
o
126.58
o
) = 2202.59 j254.06 VA
From the above the total apparent power S
V
= S
a
+ S
b
+ S
c
= 9692.11 + j0 VA. Therefore,
S
V
=

S
a
+S
b
+S
c

= 9692.11 VA.
The total arithmetic apparent power S
A
=

S
a

S
b

S
c

== 9922.2 VA. Therefore, the

arithmetic and vector apparent power factors are given by,
p
f
A
=
P
S
A
=
9692.11
9922.2
= 0.9768
p
f
V
=
P
S
V
=
9622.11
9622.11
= 1.00.
It is interesting to note that although the load in each phase is resistive but each phase has some
reactive power. This is due to unbalance of the load currents. This apparently increases the rating
of power conductors for given amount of power transfer. It is also to be noted that the net reactive
power Q = Q
a
+ Q
b
+ Q
c
= 0 leading to the unity vector apparent power factor . However the
arithmetic apparent power factor is less than unity showing the effect of the unbalance loads on the
power factor.
3.4.3 An Alternate Approach to Determine Phase Currents and Powers
In this section, an alternate approach will be discussed to solve phase currents and powers directly
without computing the neutral voltage for the system shown in Fig.3.9(a). First we express three-
phase voltage in the following form.
V
a
= V 0
o
V
b
= V 120
o
=
2
V (3.63)
V
c
= V 120
0
= V
Where, in above equation, is known as complex operator and value of and
2
are given below.
= e
j2/3
= 1120
o
= 1/2 +j

3/2

2
= e
j4/3
= 1240 = 1 120 = 1/2 j

3/2 (3.64)
Also note the following property,
1 + +
2
= 0. (3.65)
Using the above, the line to line voltages can be expressed as following.
90
V
ab
= V
a
V
b
= (1
2
)V
V
bc
= V
b
V
c
= (
2
)V (3.66)
V
ca
= V
c
V
a
= ( 1)V
Therefore, currents in line ab, bc and ca are given as,
I
abl
= Y
ab
l
V
ab
= Y
ab
l
(1
2
)V
I
bcl
= Y
bc
l
V
bc
= Y
bc
l
(
2
)V (3.67)
I
cal
= Y
ca
l
V
ca
= Y
ca
l
( 1)V
Hence line currents are given as,
I
al
= I
abl
I
cal
= [Y
ab
l
(1
2
) Y
ca
l
( 1)]V
I
bl
= I
bcl
I
abl
= [Y
bc
l
(
2
) Y
ab
l
(1
2
)]V (3.68)
I
cl
= I
cal
I
bcl
= [Y
ca
l
( 1) Y
bc
l
(
2
)]V
Example 3.4 Compute line currents by using above expressions directly for the problem in Exam-
ple 3.3.
Solution: To compute line currents directly from the above expressions, we need to compute Y
ab
l
.
These are given below
Y
ab
l
=
1
Z
ab
l
=
Z
c
l
Z
a
l
Z
b
l
+Z
b
l
Z
c
l
+Z
c
l
Z
a
l
Y
bc
l
=
1
Z
bc
l
=
Z
a
l
Z
a
l
Z
b
l
+Z
b
l
Z
c
l
+Z
c
l
Z
a
l
(3.69)
Y
ca
l
=
1
Z
ca
l
=
Z
b
l
Z
a
l
Z
b
l
+Z
b
l
Z
c
l
+Z
c
l
Z
a
l
Substituting, Z
l
a
= R
a
= 10 , Z
l
b
= R
b
= 15 and Z
l
c
= R
c
= 30 into above equation, we get
the following.
Y
ab
l
= G
ab
=
1
30

Y
bc
l
= G
bc
=
1
90

Y
ca
l
= G
ca
=
1
60

Substituting above values of the admittances in (3.68) , line currents are computed as below.
91
I
a
=
_
1
30
_
1 (
1
2
j

3
2
)
_

1
60
_
(
1
2
+j

3
2
) 1
__
V
= V (0.075 +j0.0144)
= 0.07637 V 10.89
o
= 17.5610.89
o
Amps, for V=230 V
Similarly for Phase-b current,
I
b
=
_
1
90
_
(
1
2
j

3
2
) (
1
2
+j

3
2
)
_

1
30
_
1 (
1
2
j

3
2
)
__
V
= V (0.05 j0.0481)
= 0.06933 V 136.1
o
= 15.94 136.91
o
Amps, for V=230 V
Similarly for Phase-c current,
I
c
=
_
1
60
_
(
1
2
+j

3
2
) 1)
_

1
90
_
(
1
2
j

3
2
) (
1
2
+j

3
2
)
__
V
= V (0.025 +j0.0336)
= 0.04194 V 126.58
o
= 9.64126.58
o
Amps, for V=230 V
Thus it is found that the above values are similar to what have been found in previous Example
3.3. The other quantities such as powers and power factors are same.
3.4.4 An Example of Balancing an Unbalanced Delta Connected Load
An unbalanced delta connected load is shown in Fig. 3.10(a). As can be seen from the gure that
between phase-a and b there is admittance Y
ab
l
= G
ab
l
and other two branches are open. This is
an example of extreme unbalanced load. Obviously for this load, line currents will be extremely
unbalanced. Now we aim to make these line currents to be balanced and in phase with their phase
voltages. So, let us assume that we add admittances Y
ab

, Y
bc

and Y
ca

between phases ab, bc and

ca respectively as shown in Fig. 3.10(b) and (c). Let values of compensator susceptances are given
by,
Y
ab

= 0
Y
bc

= jG
ab
l
/

3
Y
ca

= jG
ab
l
/

3
92
c
I
b
I
a
I
ab ab
l l
Y G
c
b
0
c
a l
Y

0
b
c
l
Y

a
c
I
b
I
a
I
ab ab
l l
Y G
c
a
b
/
3
c
a
a
b l
Y
j
G

/
3
b
c
a
b
l
Y
j
G

0
ab
Y

0
c
a l
Y

0
b
c
l
Y

c
I
b
I
a
I
c
a
b
/
3
c
a
a
b l
Y
j
G

/
3
b
c
a
b
l
Y
j
G

ab ab
l
Y G
(a)
(b) (c)
Fig. 3.10 (a) An unbalanced three-phase load (b) With compensator (c) Compensated system
Thus total admittances between lines are given by,
Y
ab
= Y
ab
l
+Y
ab

= G
ab
l
+ 0 = G
ab
l
Y
bc
= Y
bc
l
+Y
bc

= 0 +jG
ab
l
/

3 = jG
ab
l
/

3
Y
ca
= Y
ca
l
+Y
ca

= 0 jG
ab
l
/

3 = jG
ab
l
/

3.
Therefore the impedances between load lines are given by,
Z
ab
=
1
Y
ab
=
1
G
ab
l
Z
bc
=
1
Y
bc
=
j

3
G
ab
l
Z
ca
=
1
Y
ca
=
j

3
G
ab
l
Note that Z
ab
+Z
bc
+Z
ca
= 1/G
ab
l
j

3/G
ab
l
+j

3/G
ab
l
= 1/G
ab
l
.
The impedances, Z
a
, Z
b
and Z
c
of equivalent star connected load are given as follows.
Z
a
=
Z
ab
Z
ca
Z
ab
+Z
bc
+Z
ca
= (
1
G
ab
l

3
G
ab
l
)/(
1
G
ab
l
)
=
j

3
G
ab
l
93
Z
b
=
Z
bc
Z
ab
Z
ab
+Z
bc
+Z
ca
= (
1
G
ab
l

3
G
ab
l
)/(
1
G
ab
l
)
=
j

3
G
ab
l
Z
c
=
Z
ca
Z
bc
Z
ab
+Z
bc
+Z
ca
= (
j

3
G
ab
l

3
G
ab
l
)/(
1
G
ab
l
)
=
3
G
ab
l
The above impedances seen from the load side are shown in Fig. 3.11(a) below. Using (3.57),
c
I
b
I
a
I
3
c
l ab
l
Z
G

c
a
b
3
a
l
ab
l
Z j
G

n
3
b
l
ab
l
Z j
G

c
I
b
I
a
I
c
a
b
N
1
ab
l
G
1
ab
l
G
1
ab
l
G
Source
Load
Fig. 3.11 Compensated system (a) Load side (b) Source side
the voltage between load and system neutral of delta equivalent star load as shown in Fig. 3.11, is
computed as below.
V
Nn
=
1
1
Za
+
1
Z
b
+
1
Zc
_
V
an
Z
a
+
V
bn
Z
b
+
V
cn
Z
c
_
=
1
1
(j

3/G
ab
l
)
+
1
(j

3/G
ab
l
)
+
3
G
ab
l
_
V 0
o
j

3/G
ab
l
+
V 120
o
j

3/G
ab
l
+
V 120
o
3/G
ab
l
_
=
3 V
G
ab
l
_
G
ab
l
3
j
G
ab
l

3
_
= 2 V
_
1
2
j

3
2
_
= 2 V 60
o
94
Using above value of neutral voltage the line currents are computed as following.
I
a
=
V
an
V
Nn
Z
a
=
[V 0
o
2 V 60
o
]
j

3
G
ab
l
= G
ab
l
V = G
ab
l
V
a
I
b
=
V
bn
V
Nn
Z
b
=
[V 120
o
2 V 60
o
]
j

3
G
ab
l
= G
ab
l
V 240
o
= G
ab
l
V
b
I
c
=
V
cn
V
Nn
Z
c
=
[V 120
o
2 V 60
o
]
3
G
ab
l
= G
ab
l
V 120
o
= G
ab
l
V
c
From the above example, it is seen that the currents in each phase are balanced and in phase with
their respective voltages. This is equivalently shown in Fig. 3.11(b). It is to be mentioned here that
the two neutrals in Fig. 3.11 are not same. In Fig. 3.11(b), the neutral N is same as the system
neutral as shown in Fig. 3.8, whereas in Fig. 3.11(a), V
Nn
= 2 V 60
o
. However the reader
may be curious to know why Y
ab

= 0, Y
bc

= jG
ab
l
/

3 and
ca
= jG
ab
l
/

3 have been chosen

as compensator admittance values. The answer of the question can be found by going following
sections.
3.5 A Generalized Approach for Load Compensation using Symmetrical
Components
In the previous section, we have expressed line currents I
a
, I
b
and I
c
, in terms load admittances
and the voltage V for a delta connected unbalanced load as shown in Fig 3.12(a). For the sake of
completeness, theses are reproduced below.
95
I
al
= I
abl
I
cal
= [Y
ab
l
(1
2
) Y
ca
l
( 1)]V
I
bl
= I
bcl
I
abl
= [Y
bc
l
(
2
) Y
ab
l
(1
2
)]V (3.70)
I
cl
= I
cal
I
bcl
= [Y
ca
l
( 1) Y
bc
l
(
2
)]V
c
I
b
I
a
I
c
a
Y
ab
l
Y
ab
Y

bc
Y
bc
l
Y
c
a
l Y
c
a
b
cl
I
bl
I
al
I
ab
l
Y
bc
l
Y
c
a
l Y
c
a
b
(a) (b)
Fig. 3.12 (a) An unbalanced delta connested load (b) Compensated system
Since loads are currents are unbalanced, these will have positive and negative currents. The zero
sequence current will be zero as it is three-phase and three-wire system. These symmetrical com-
ponents of the load currents are expressed as following.
_
_
I
0l
I
1l
I
2l
_
_
=
1

3
_
_
1 1 1
1 a a
2
1 a
2
a
_
_
_
_
I
al
I
bl
I
cl
_
_
(3.71)
From the above equation, zero sequence current is given below.
I
0l
=
_
I
al
+I
al
+I
al
_
/

3
The positive sequence current is as follows.
I
1l
=
1

3
_
I
al
+I
bl
+
2
I
cl

=
1

3
[ Y
ab
l
(1
2
) Y
ca
l
( 1) +
_
Y
bc
l
(
2
) Y
ab
l
(1
2
)
_
+
2
_
Y
ca
l
( 1) Y
bc
l
(
2
)
_
] V
=
1

3
[ Y
ab
l

2
Y
ab
l
+Y
ca
l
Y
ca
l
+
3
Y
bc
l

2
Y
bc
l
Y
ab
l

3
Y
ab
l
+
3
Y
ca
l

2
Y
ca
l

4
Y
bc
l
+
3
Y
bc
l
] V
=
_
Y
ab
l
+Y
bc
l
+Y
ca
l
_
V

3
96
Similarly negative sequence component of the current is,
I
2l
=
1

3
_
I
al
+
2
I
bl
+I
cl

=
1

3
[Y
ab
l
(1
2
) Y
ca
l
( 1) +
2
_
Y
bc
l
(
2
) Y
ab
l
(1
2
)
_
+
_
Y
ca
l
( 1) Y
bc
l
(
2
)
_
]

V
=
1

3
[ Y
ab
l

2
Y
ab
l
Y
ca
l
+Y
ca
l
+
4
Y
bc
l

3
Y
bc
l

2
Y
ab
l
+
4
Y
ab
l
+
2
Y
ca
l
Y
ca
l

3
Y
bc
l
+
2
Y
bc
l
] V
=
1

3
[ 3
2
Y
ab
l
3Y
bc
l
3Y
ca
l
] V
= [
2
Y
ab
l
+Y
bc
l
+Y
ca
l
]

3V
From the above, it can be written that,
I
0l
= 0
I
1l
=
_
Y
ab
l
+Y
bc
l
+Y
ca
l
_
3V (3.72)
I
2l
=
_

2
Y
ab
l
+Y
bc
l
+Y
ca
l
_
3V
When compensator is used, three delta branches Y
ab

, Y
bc

and Y
ca

are added as shown in Fig.

3.12(b). Using above analysis, the sequence components of the compensator currents can be given
as below.
I
0
= 0
I
1
=
_
Y
ab

+Y
bc

+Y
ca

_
3 V (3.73)
I
2
=
_

2
Y
ab

+Y
bc

+Y
ca

_
3 V
Since, compensator currents are purely reactive, i.e., G
ab

= G
bc

= G
ca

= 0,
Y
ab

= G
ab

+jB
ab

= jB
ab

Y
bc

= G
bc

+jB
bc

= jB
bc

(3.74)
Y
ca

= G
ca

+jB
ca

= jB
ca

.
Using above, the compensated sequence currents can be written as,
I
0
= 0
I
1
= j
_
B
ab

+B
bc

+B
ca

_
3 V (3.75)
I
2
= j(
2
B
ab

+B
bc

+B
ca

3 V
Knowing nature of compensator and load currents, we can set compensation objectives as fol-
lowing.
1. All negative sequence component of the load current must be supplied from the compensator
negative current, i.e.,
I
2l
= I
2
(3.76)
97
The above further implies that,
Re
_
I
2l
_
+j Im
_
I
2l
_
= Re
_
I
2
_
j Im
_
I
2
_
(3.77)
2. The total positive sequence current, which is source current should have desired power factor
from the source, i.e.,
Im(I
1l
+I
1
)
Re (I
1l
+I
1
)
= tan = (3.78)
Where, is the desired phase angle between the line currents and the supply voltages. The above
equation thus implies that,
Im(I
1l
+I
1
) = Re (I
1l
+I
1
) (3.79)
Since Re (I
1
) = 0, the above equation is rewritten as following.
Im(I
1l
) Re (I
1l
) = Im(I
1
) (3.80)
The equation (3.77) gives two conditions and equation (3.79) gives one condition. There are
three unknown variables, i.e., B
ab

, B
bc

and B
ca

and three conditions. Therefore the unknown

variables can be solved. This is described in the following section. Using (3.75), the current I
2
is
expressed as following.
I
2
= j[
2
B
ab

+B
bc

+B
ca

3 V
= j
__

1
2
j

3
2
_
B
ab

+B
bc

+
_

1
2
+j

3
2
_
B
ca

_

3 V
=
__

3
2
B
ab

3
2
B
ca

_
j
_

1
2
B
ab

+B
bc

1
2
B
ca

_
_

3 V (3.81)
= Re
_
I
2
_
j Im
_
I
2
_
Thus the above equation implies that
_

3
2
B
ab

3
2
B
ca

_
=
1

3V
Re
_
I
2
_
=
1

3V
Re
_
I
2l
_
(3.82)
and,
_

1
2
B
ab

+B
bc

1
2
B
ca

_
=
1

3V
Im
_
I
2
_
=
1

3V
Im
_
I
2l
_
(3.83)
Or
_
B
ab

+ 2 B
bc

B
ca

_
=
1

3V
2 Im
_
I
2l
_
(3.84)
98
From (3.75), Im (I
1
) can be written as,
Im(I
1
) =
_
B
ab

+B
bc

+B
ca

_
3 V. (3.85)
Substituting Im(I
1
) from above equation into (3.85), we get the following.
(B
ab

+B
bc

+B
ca

) =
1

3V
Im(I
1
) =
1

3V
_
Im(I
1l
) Re (I
1l
)
_
(3.86)
Subtracting (3.86) from (3.84), the following is obtained,
B
bc

=
1
3

V
[Im (I
1l
) 2 Im (I
2l
) Re (I
1l
)]. (3.87)
Now, from (3.82) we have

1
2
B
ab

1
2
B
ca

=
1

3V
Im (I
2l
) B
bc

=
1

3V
Im (I
2l
)
_
1
3

3V
_
Im (I
1l
) Re (I
1l
) 2 Im (I
2l
)
_
_
=
1
3

3V
_
Im (I
2l
) + Im (I
1l
) Re (I
1l
)
_
(3.88)
Reconsidering (3.88) and (3.82), we have

1
2
B
ab

1
2
B
ca

=
1
3

3V
_
Im (I
1l
) + Im (I
2l
) Re (I
1l
)
_
1
2
B
ab

1
2
B
ca

=
1
3

3V
[

3Re (I
2l
)]
Adding above equations, we get
B
ca

=
1
3

3V
[Im (I
1l
) + Im (I
2l
) +

3Re (I
2l
) Re (I
1l
)]. (3.89)
Therefore,
B
ab

= B
ca

+
2
3

3V
[

3Re (I
2l
)] (3.90)
=
1
3

3V
[Im (I
1l
) + Im (I
2l
) +

3Re (I
2l
) Re (I
1l
) 2

3Re (I
2l
)]
=
1
3

3V
[Im (I
1l
) + Im (I
2l
)

3Re (I
2l
) Re (I
1l
)] (3.91)
From the above, the compensator susceptances in terms of real and imaginary parts of the load
99
current can be written as following.
B
ab

=
1
3

3V
[Im (I
1l
) + Im (I
2l
)

3 Re (I
2l
) Re (I
1l
)]
B
bc

=
1
3

3V
[Im (I
1l
) Re (I
1l
) 2Im (I
2l
Re (I
1l
)] (3.92)
B
ca

=
1
3

3V
[Im (I
1l
) + Im (I
2l
) +

3 Re (I
2l
) Re (I
1l
)]
In the above equation, the susceptances of the compensator are expressed in terms of real and
imaginary parts of symmetrical components of load currents. It is however advantageous to express
these susceptances in terms of instantaneous values of voltages and currents from implementation
point of view. The rst step to achieve this is to express these susceptances in terms of load
currents, i.e., I
al
, I
bl
and I
cl
, which is described below. Using equation (3.71), the sequence
components of the load currents are expressed as,

I
0l
=
1

3
_

I
al
+

I
bl
+

I
cl

I
1l
=
1

3
_

I
al
+

I
bl
+
2

I
cl

(3.93)

I
2l
=
1

3
_

I
al
+
2

I
1l
+

I
cl

.
Substituting these values of sequence components of load currents, in (3.92), we can obtain com-
pensator susceptances in terms of real and inaginary components of the load currents. Let us start
from the B
bc

, as obtained following.
B
bc

=
1
3

3V
_
Im(I
1l
) 2Im(I
2l
) Re(I
1l
)

=
1
3

3V
_
Im
_
I
al
+I
bl
+
2
I
cl

3
_
2 Im
_
I
al
+
2
I
bl
+I
cl

3
_
Re
_
I
al
+I
bl
+
2
I
cl

3
__
=
1
9V
_
Im
_
(I
al
+ (2 + 3)I
bl
+ (2 + 3
2
)I
cl
_
Re (I
al
+I
bl
+
2
I
cl
)

=
1
9V
_
Im
_
I
al
+ 2I
bl
+ 2I
cl
+ 3I
bl
+ 3
2
I
cl
_
Re (I
al
+I
bl
+
2
I
cl
)

By adding and subracting I

bl
and I
cl
in the above equation we get,
B
bc

=
1
9V
_
Im
_
(I
al
I
bl
I
cl
) + 3I
bl
+ 3I
cl
+ 3I
bl
+
2
I
cl
_
Re(I
al
+I
bl
+
2
I
cl
)

We know that I
al
+I
bl
+I
cl
=0, therefore I
al
+I
bl
= I
cl
.
B
bc

=
1
3V
_
Im
_
I
al
_
+ Im
_
I
bl
_
+ Im
_

2
I
cl
_

3
Re
_
I
al
+I
bl
+
2
I
cl

_
(3.94)
100
Similarly, it can be proved that,
B
ca

=
1
3V
_
Im
_
I
al
_
Im
_
I
bl
_
+ Im
_

2
I
cl
_

3
Re
_
I
al
+I
bl
+
2
I
cl

_
(3.95)
B
ab

=
1
3V
_
Im
_
I
al
_
+ Im
_
I
bl
_
Im
_

2
I
cl
_

3
Re
_
I
al
+I
bl
+
2
I
cl

_
(3.96)
The above expressions for B
ca

and B
ab

are proved below. For convenience, the last term associated

with is not considered. For the sake simplicity in equations (3.95) and (3.96) are proved to those
given in equations (3.92).
B
ca

=
1
3V
_
Im
_
I
al
_
Im
_
I
bl
_
+ Im
_

2
I
cl
_
=
1
3V
_
Im
_
_
I
0l
+I
1l
+I
2l
_

_
I
0l
+
2
I
1l
+I
2l
_

3
+
2
_
I
0l
+I
1l
+
2
I
2l
_

3
__
Since I
0l
=0
B
ca

=
1
3

3V
Im
_
I
1l
2
2
I
2l

=
1
3

3V
Im
_
I
1l
2
_

1
2
j

3
2
_
I
2l
_
=
1
3

3V
Im
_
I
1l
+I
2l
+j

3I
2l
_
=
1
3

3V
_
Im
_
I
1l
_
+ Im
_
I
2l
_
+

3Re
_
I
2l
_
_
Note that Im(jI
2l
) = Re(I
2l
). Adding term, we get the following.
B
ca

=
1
3

3V
_
Im(I
1l
) + Im(I
2l
) +

3Re
_
I
2l
_
Re(I
1l
)
_
Similarly,
B
ab

=
1
3V
_
Im
_
I
al
_
+ Im
_
I
bl
_
Im
_

2
I
cl
_
=
1
3V
_
Im
_
_
I
0l
+I
1l
+I
2l
_

3
+
_
I
0l
+
2
I
1l
+I
2l
_

2
_
I
0l
+I
1l
+
2
I
2l
_

3
__
101
B
ab

=
1
3

3V
Im
_
I
1l
2I
2l

=
1
3

3V
Im
_
I
1l
2
_

1
2
+j

3
2
_
I
2l
_
=
1
3

3V
Im
_
I
1l
+I
2l
j

3I
2l
_
=
1
3

3V
_
Im
_
I
1l
_
+ Im
_
I
2l
_

3Re
_
I
2l
_
_
Thus, Compensator susceptances are expressed as following.
B
ab

=
1
3V
_
Im(I
al
) + Im(I
bl
) Im(
2
I
cl
)

3
Re(I
al
+I
bl
+
2
I
cl
)
_
B
bc

=
1
3V
_
Im(I
al
) + Im(I
bl
) + Im(
2
I
cl
)

3
Re(I
al
+I
bl
+
2
I
cl
)
_
(3.97)
B
ca

=
1
3V
_
Im(I
al
) Im(I
bl
) + Im(
2
I
cl
)

3
Re(I
al
+I
bl
+
2
I
cl
)
_
An unity power factor is desired from the source. For this cos
l
= 1, implying tan
l
= 0 hence
= 0. Thus we have,
B
ab

=
1
3V
_
Im
_
I
al
_
+ Im
_
I
bl
_
Im
_

2
I
cl
_
B
bc

=
1
3V
_
Im
_
I
al
_
+ Im
_
I
bl
_
+ Im
_

2
I
cl
_
(3.98)
B
ca

=
1
3V
_
Im
_
I
al
_
Im
_
I
bl
_
+ Im
_

2
I
cl
_
The above equations are easy to realize in order to nd compensator susceptances. As mentioned
above, sampling and averaging techniques will be used to convert above equation into their time
equivalents. These are described below.
3.5.1 Sampling Method
Each current phasor in above equation can be expressed as,
I
al
= Re(I
al
) +jIm(I
al
)
= I
al,R
+jI
al,X
(3.99)
102
An instantaneous phase current is written as follows.
i
al
(t) =

2 I
al
Im(e
jt
)
=

2 Im(I
al
e
jt
)
=

2 Im
_
(I
al,R
+jI
al,X
) e
jt

2 Im [(I
al,R
+jI
al,X
)(cos t +j sin t)]
=

2 Im [(I
al,R
cos t I
al,X
sin t) +j(I
al,R
sin t +I
al,X
cos t)]
=

2 [(I
al,R
sin t +I
al,X
cos t)] (3.100)
Im(I
al
) = I
al,X
=
i
al
(t)

2
at sin t = 0, cos t = 1 (3.101)
From equation (3.63), the phase voltages can be expressed as below.
v
a
(t) =

2 V sin t
v
b
(t) =

2 V sin(t 120
o
) (3.102)
v
c
(t) =

2 V sin(t + 120
o
)
From above voltage expressions, it is to be noted that, sin t = 0, cos t = 1 implies that the
phase-a voltage, v
a
(t) is going through a positive zero crossing, hence, v
a
(t) = 0 and
d
dt
v
a
(t) = 0.
Therefore, equation (3.101), can be expressed as following.
I
a,l
=
i
al
(t)

2
when, v
a
(t) = 0, dv
a
/dt > 0 (3.103)
Similarly,
I
b,l
= I
bl,R
+jI
bl,X
(3.104)
Therefore,
(I
b,l
) = (I
bl,R
+jI
bl,X
)
=
_

1
2
+j

3
2
_
(I
bl,R
+jI
bl,X
)
=
_

1
2
I
bl,R

3
2
I
bl,X
_
+j
_

3
2
I
bl,R

1
2
I
bl,X
_
(3.105)
From the above,
Im{(I
bl
)} =

3
2
I
bl,R

1
2
I
bl,X
(3.106)
103
Similar to equation, (3.100), we can express phase-b current in terms Im(I
bl
), as given below.
i
bl
(t) =

2 Im(I
bl
e
jt
)
=

2 Im(I
bl
e
jwt

1
)
=

2 Im(I
bl
e
j(wt120
0
)
)
=

2 Im
__

1
2
+j

3
2
_
(I
bl,R
+jI
bl,X
) e
j(wt120
0
)
_
=

2 Im[
_

1
2
I
bl,R

3
2
I
bl,X
_
+j
_

3
2
I
bl,R

1
2
I
bl,X
_
_
(cos(wt 120
0
) +j sin(wt 120
0
))
_
] (3.107)
=

2
_
(
1
2
I
bl,R

3
2
I
bl,X
) sin(wt 120
0
) + (

3
2
I
bl,R

1
2
I
bl,X
) cos(wt 120
0
)
_
From the above equation, we get the following.
Im(I
bl
) =
i
bl
(t)

2
when, v
b
(t) = 0, dv
b
/dt > 0 (3.108)
Similarly for Phase-c, it can proved that,
Im(
2
I
cl
) =
i
cl
(t)

2
when, v
c
(t) = 0, dv
c
/dt > 0 (3.109)
Substituting Im(I
al
) , Im(I
bl
) and Im(
2
I
cl
) from (3.103), (3.108) and (3.109) respectively, in
(3.98), we get the following.
B
ab
r
=
1
3

2V
_
i
a
|
(va=0,
dva
dt
>0)
+i
b
|
(v
b
=0,
dv
b
dt
>0)
i
c
|
(vc=0,
dvc
dt
>0)
_
B
bc
r
=
1
3

2V
_
i
a
|
(va=0,
dva
dt
>0)
+i
b
|
(v
b
=0,
dv
b
dt
>0)
+i
c
|
(vc=0,
dvc
dt
>0)
_
(3.110)
B
ca
r
=
1
3

2V
_
i
a
|
(va=0,
dva
dt
>0)
i
b
|
(v
b
=0,
dv
b
dt
>0)
+i
c
|
(vc=0,
dvc
dt
>0)
_
Thus the desired compensating susceptances are expressed in terms of the three line currents sam-
pled at instants dened by positive-going zero crossings of the line-neutral voltages v
a
, v
b
, v
c
. An
articial neutral at ground potential may be created measuring voltages v
a
, v
b
and v
c
to implement
above algorithm.
3.5.2 Averaging Method
In this method, we express the compensator susceptances in terms of real and reactive power terms
and nally expressed them in time domain through averaging process. The method is described
104
below.
From equation (3.98), susceptance, B
ab

, can be re-written as following.

B
ab

=
1
3V
_
Im
_
I
al
_
+ Im
_
I
bl
_
Im
_

2
I
cl
_
=
1
3V
2
_
V Im
_
I
al
_
+V Im
_
I
bl
_
V Im
_

2
I
cl
_
=
1
3V
2
_
Im
_
V I
al
_
+ Im
_
V I
bl
_
Im
_
V
2
I
cl
_
(3.111)
Note the following property of phasors and applying it for the simplication of the above expres-
sion.
Im
_
V I
_
= Im
_
V

_
(3.112)
Using above equation (3.111) can be written as,
B
ab

=
1
3V
2
_
Im(V I
al
)

+ Im(V
a
I
bl
)

Im(
2
V I
cl
)

=
1
3V
2
_
Im(V

a
I

al
) + Im(

a
I

bl
) Im((
2
)

a
I

cl
)
_
=
1
3V
2
_
Im(V

a
I

al
) + Im(
2
V

a
I

bl
) Im(V

a
I

cl
)
_
Since V
a
= V 0
o
is a reference phasor, therefore V
a
= V

a
= V ,
2
V
a
= V

b
and V
a
= V

c
.
Using this, the above equation can be written as following.
B
ab

=
1
3V
2
_
Im(V
a
I

al
) + Im(V
b
I

bl
) Im(V
c
I

cl
)
_
Similarly, B
bc

=
1
3V
2
_
Im(V
a
I

al
) + Im(V
b
I

bl
) + Im(V
c
I

cl
)
_
(3.113)
B
ca

=
1
3V
2
_
Im(V
a
I

al
) Im(V
b
I

bl
) + Im(V
c
I

cl
)
_
It can be further proved that,
Im(V
a
I

a
) =
1
T
_
T
0
v
a
(t)(/2) i
al
(t) dt
Im(V
b
I

b
) =
1
T
_
T
0
v
b
(t)(/2) i
bl
(t) dt (3.114)
Im(V
c
I

c
) =
1
T
_
T
0
v
c
(t)(/2) i
cl
(t) dt
105
Since,
v
a
(t)(/2) = v
bc
(t)/

3
v
b
(t)(/2) = v
ca
(t)/

3
v
c
(t)(/2) = v
ca
(t)/

3
Equation (3.115) can be written as,
Im(V
a
I

a
) =
1

3 T
_
T
0
v
bc
(t) i
al
(t) dt
Im(V
b
I

b
) =
1

3 T
_
T
0
v
ca
(t) i
bl
(t) dt (3.115)
Im(V
c
I

c
) =
1

3 T
_
T
0
v
ab
(t) i
cl
(t) dt
Substituting above values of Im(V
a
I

a
), Im(V
b
I

b
) and Im(V
c
I

c
) into (3.113), we get the fol-
lowing.
B
ab

=
1
(3

3V
2
)
1
T
_
T
0
(v
bc
i
al
+v
ca
i
bl
v
ab
i
cl
) dt
B
bc

=
1
(3

3V
2
)
1
T
_
T
0
(v
bc
i
al
+v
ca
i
bl
+v
ab
i
cl
) dt (3.116)
B
ca

=
1
(3

3V
2
)
1
T
_
T
0
(v
bc
i
al
v
ca
i
bl
+v
ab
i
cl
) dt
The above equations can directly be used to know the the compensator susceptances by performing
the averaging on line the product of the line to line voltages and phase load currents. The term
_
T
0
=
_
t
1
+T
t
1
can be implemented using moving average of one cycle. This improves transient
response by computing average value at each instant. But in this case the controller response
which changes the susceptance value, should match to that of the above computing algorithm.
3.6 Compensator Admittance Represented as Positive and Negative Sequence
Admittance Network
Recalling the following relations from equation (3.92) for unity power factor operation i.e. = 0,
we get the following.
B
ab

=
1
3

3V
[Im (I
1l
) + Im (I
2l
)

3 Re (I
2l
)]
B
bc

=
1
3

3V
[Im (I
1l
) Re (I
1l
) 2Im (I
2l
)] (3.117)
B
ca

=
1
3

3V
[Im (I
1l
) + Im (I
2l
) +

3 Re (I
2l
)]
106
From these equations, it is evident the the rst terms form the positive sequence suceptance as they
involve I
1l
terms. Similarly, the second and third terms in above equation form negative sequence
susceptance of the compensator, as these involve I
2l
terms. Thus, we can write,
B
ab

= B
ab
1
+B
ab
2
B
ab

= B
ab
1
+B
ab
2
(3.118)
B
ab

= B
ab
1
+B
ab
2
Therefore,
B
ab
1
= B
bc
1
= B
ca
1
=
1
3

3V
_
Im(

I
1l
)

(3.119)
And,
B
ab
2
=
1
3

3V
_
Im(I
2l
)

3Re(I
2l
)
_
B
bc
2
=
1
3

3V
_
2 Im(I
2l
)
_
(3.120)
B
ca
2
=
1
3

3V
_
Im(I
2l
) +

3Re(I
2l
)
_
Earlier in equation, (3.121), it was established that,
I
0l
= 0
I
1l
=
_
Y
ab
l
+Y
bc
l
+Y
ca
l
_
3V
I
2l
=
_

2
Y
ab
l
+Y
bc
l
+Y
ca
l
_
3V
Noting that,
Y
ab
l
= G
ab
l
+jB
ab
l
Y
bc
l
= G
bc
l
+jB
bc
l
Y
ca
l
= G
ca
l
+jB
ca
l
Therefore,
Im(I
1l
) = Im((Y
ab
l
+Y
bc
l
+Y
ca
l
)

3V )
= (B
ab
l
+B
bc
l
+B
ca
l
)

3V (3.121)
Thus equation (3.119) is re-written as following.
B
ab
1
= B
bc
1
= B
ca
1
=
1
3
_
B
ab
l
+B
ab
l
+B
ca
l
_
(3.122)
107
Now we shall compute B
ab
2
, B
bc
2
and B
ca
2
using equations (3.120) as following. Knowing that,

I
2l
=
_

2
Y
ab
l
+Y
bc
l
+Y
ca
l
_
3 V
=
__
1
2

j

3
2
_
_
G
ab
l
+jB
ab
l
_
+
_
G
bc
l
+jB
bc
l
_
+
_
1
2
+
j

3
2
_
(G
ca
l
+jB
ca
l
)
_

3V
=
_

G
ab
l
2
+

3
2
B
ab
l
+G
bc
l

G
ca
l
2

3
2
B
ca
l
j
_

3
2
G
ab
l
+
B
ab
l
2
B
bc
l

3
2
G
ca
l
+
B
ca
l
2
__

3V
=
_
G
ab
l
2

3
2
B
ab
l
G
bc
l
+
G
ca
l
2
+

3
2
B
ca
l
+j
_

3
2
G
ab
l
+
B
ab
l
2
B
bc
l

3
2
G
ca
l
+
B
ca
l
2
__

3V
(3.123)
The above implies that,
Im(I
2l
) =
_

3
2
G
ab
l
+
B
ab
l
2
B
bc
l

3
2
G
ca
l
+
B
ca
l
2
_

3 V

3 Re(I
2l
) =
_

3
2
G
ab
l
+
3
2
B
ab
l
+

3G
bc
l

3
2
G
ca
l

3
2
B
ca
l
_

3 V
Thus, B
ab
2
can be given as,
B
ab
2
=
1
3

3V
_
2B
ab
l
B
bc
l
B
ca
l
+

3 G
bc
l

3 G
ca
l
_

3 V
=
1
3
_
2B
ab
l
B
bc
l
B
ca
l
+

3
_
G
bc
l
G
ca
l
_
_
=
1

3
_
G
ca
l
G
bc
l
_
+
1
3
_
B
bc
l
+B
ca
l
2B
ab
l
_
(3.124)
Similarly,
B
ca
2
=
1
3

3V
_
Im(

I
2l
) +

3 Re(

I
2l
)
_
=
1
3

3V
_

3
2
G
ab
l
+
B
ab
l
2
B
bc
l

3
2
G
ca
l
+
B
ca
l
2
+

3
2
G
ab
l

3
2
B
ab
l

3G
bc
l
+

3
2
G
ca
l
+
3
2
B
ca
l
_

3 V
=
1
3
_

3G
ab
l

3G
bc
l
B
ab
l
B
bc
l
+ 2B
ca
l
_
=
1

3
_
G
bc
l
G
ab
l
_
+
1
3
_
B
ab
l
+B
bc
l
2B
ca
l
_
(3.125)
And, B
bc
2
is computed as below.
108
B
bc
2
=
1
3

3V
_
2Im(

I
2l
)

=
2
3

3V
_

3
2
G
ab
l
+
B
ab
l
2
B
bc
l

3
2
G
ca
l
+
B
ca
l
2
_
V

3
=
1

3
_
G
ab
l
G
ca
l
_
+
1
3
_
B
ab
l
+B
ca
l
2B
bc
l
_
(3.126)
Using (3.119), (3.124)-(3.126), We therefore can nd the overall compensator susceptances as
following.
B
ab

= B
ab
1
+B
ab
2
=
1
3
_
B
ab

+B
bc

+B
ca

_
+
1

3
_
G
ca
l
G
bc
l
_
+
1
3
_
B
ca
l
+B
bc
l
2B
ab
l
_
= B
ab
l
+
1

3
_
G
ca
l
G
bc
l
_
Similarly,
B
bc

= B
bc
1
+B
bc
2
=
1
3
_
B
ab

+B
bc

+B
ca

_
+
1

3
_
G
ab
l
G
ca
l
_
+
1
3
_
B
ab
l
+B
ca
l
2B
bc
l
_
= B
bc
l
+
1

3
_
G
ab
l
G
ca
l
_
And,
B
ca

= B
ca
1
+B
ca
2
=
1
3
_
B
ab

+B
bc

+B
ca

_
+
1

3
_
G
bc
l
G
ab
l
_
+
1
3
_
B
bc
l
+B
ab
l
2B
ca
l
_
= B
ca
l
+
1

3
_
G
bc
l
G
ab
l
_
Thus, the compensator susceptances in terms of load parameters are given as in the following.
B
ab

= B
ab
l
+
1

3
_
G
ca
l
G
bc
l
_
B
bc

= B
bc
l
+
1

3
_
G
ab
l
G
ca
l
_
(3.127)
B
ca

= B
ca
l
+
1

3
_
G
bc
l
G
ab
l
_
109
It is interesting to observe above equations. The rst parts of the equation nullies the effect of the
load susceptances and the second parts of the equations correspond to the unbalance in resistive
load. The two terms together make source current balanced and in phase with the supply voltages.
The compensators positive and negative sequence networks are shown in Fig. 3.13 .
What happens if we just use the following values of the compensator susceptances as given below?
B
ab

= B
ab
l
B
bc

= B
bc
l
(3.128)
B
ca

= B
ca
l
In above case, load suceptance parts of the admittance are fully compensated. However the source
currents after compensation remain unbalanced due to unbalance conductance parts of the load.
Load
1,abc I
sa
i
sc
i
sb
i
la
i
lc
i
lb
i
2,abc I
1
ab
B

1
ca
B

1
bc
B

2
ab
B

2
ca
B

2
bc
B

S
o
u
r
c
e
, ,
ab bc ca
B B B

Fig. 3.13 Sequence networks of the compensator
Example 3.5 For a delta connected load shown in 3.14 , the load admittances are given as follow-
ing,
Y
ab
l
= G
ab
l
+jB
ab
l
Y
bc
l
= G
bc
l
+jB
bc
l
Y
ca
l
= G
ca
l
+jB
ca
l
Given the load parameters:
Z
ab
l
= 1/Y
ab
l
= 5 +j12
Z
bc
l
= 1/Y
bc
l
= 3 +j4
Z
ca
l
= 1/Y
ca
l
= 9 j12
Determine compensator susceptances (B
ab

, B
bc

, B
ca

) so that the supply sees the load as balanced

and unity power factor. Also nd the line currents and source active and reactive powers before
and after compensation.
110
ca
Y

ab
Y

a
b
c
bc
Y

ab
B

bc
B

ca
B

N
a
v
c
v
b
v
c
i
a
i
b
i
Fig. 3.14 A delta connected load
Solution:
Z
ab
l
= 5 +j12 Y
ab
l
= 0.03 j0.0710
Z
bc
l
= 3 +j4 Y
bc
l
= 0.12 j0.16
Z
ca
l
= 9 j13 Y
ca
l
= 0.04 +j0.0533
Once we know the admittances we know,
G
ab
l
= 0.03, B
ab
l
= 0.0710
G
bc
l
= 0.12, B
bc
l
= 0.16
G
ca
l
= 0.04, B
ca
l
= 0.0533
B
ab

= B
ab
l
+
1

3
(G
ca
l
G
bc
l
) = 0.0248
B
bc

= B
bc
l
+
1

3
(G
ab
l
G
ca
l
) = 0.1540
B
ca

= B
ca
l
+
1

3
(G
bc
l
G
ab
l
) = 0.0011
Total admittances are:
Y
ab
= Y
ab
l
+Y
ab

= 0.03 j0.0462
Y
bc
= Y
bc
l
+Y
bc

= 0.12 j0.006
Y
ca
= Y
ca
l
+Y
ca

= 0.04 +j0.0522
Knowing these total admittances, we can nd line currents using following expressions.
Current Before Compensation
111
I
a
= I
abl
I
cal
=
_
(1
2
)Y
ab
l
( 1)Y
ca
l

V = 0.2150 V 9.51
o
A
I
b
= I
bcl
I
abl
=
_
(
2
)Y
bc
l
(1
2
)Y
ab
l

V = 0.4035 V 161.66
o
A
I
c
= I
cal
I
bcl
=
_
( 1)Y
ca
l
(
2
)Y
bc
l

V = 0.2358 V 43.54
o
A
Powers Before Compensation
S
a
= V
a
(I
al
)

= P
a
+jQ
a
= V (0.2121 +j0.0355)
S
b
= V
b
(I
bl
)

= P
b
+jQ
b
= V (0.3014 +j0.2682)
S
c
= V
c
(I
cl
)

= P
c
+jQ
c
= V (0.0552 +j0.2293)
Total real power, P = P
a
+P
b
+P
c
= V 0.5688 W
Total reactive power, Q = Q
a
+Q
b
+Q
c
= V 0.5330 VAr
Power factor in phase-a, p
fa
= cos
a
= cos(9.51
o
) = 0.9863 lag
Power factor in phase-b, p
f
b
= cos
b
= cos(41.63
o
) = 0.7471 lag
Power factor in phase-c, p
fc
= cos
c
= cos(76.45
o
) = 0.2334 lag
Thus we observe that the phases draw reactive power from the lines and currents are unbalanced
in magnitude and phase angles.
After Compensation
I
a
= I
ab
I
ca
=
_
(1
2
)Y
ab
( 1)Y
ca

V = 0.1896 V 0
o
A
I
b
= I
bc
I
ab
=
_
(
2
)Y
bc
(1
2
)Y
ab

V = 0.1896 V 120
o
A
I
a
= I
ca
I
bc
=
_
( 1)Y
ca
l
(
2
)Y
bc

V = 0.1896 V 120
o
A
Powers After Compensation
S
a
= V
a
(I
al
)

= P
a
+jQ
a
= V (0.1986 +j0.0)
S
b
= V
b
(I
bl
)

= P
b
+jQ
b
= V (0.1986 +j0.0)
S
c
= V
c
(I
cl
)

= P
c
+jQ
c
= V (0.1986 +j0.0)
Total real power, P = P
a
+P
b
+P
c
= (V 0.5688 )W
Total reactive power, Q = Q
a
+Q
b
+Q
c
= 0 VAr
Power factor in phase-a, pf
a
= cos
a
= cos(0
o
) = 1.0
Power factor in phase-b, pf
b
= cos
b
= cos(0
o
) = 1.0
Power factor in phase-c, pf
c
= cos
c
= cos(0
o
) = 1.0
From above results we observe that after placing compensator of suitable values as calculated
above, the line currents become balanced and have unity power factor relationship with their volt-
ages.
112
Example 3.6 Consider the following 3-phase, 3-wire system. The 3-phase voltages are balanced
sinusoids with RMS value of 230 V at 50 Hz. The load impedances are Z
a
= 3 + j 4 , Z
b
=
5 +j 12 , Z
c
= 12 j 5 . Compute the following.
1. The line currents I
la
, I
lb
, I
lc
.
2. The active (P) and reactive (Q) powers of each phase.
3. The compensator susceptance ( B
ab

, B
bc

, B
ca

), so that the supply sees the load balanced and

unity power factor.
4. For case (3), compute the source, load, compensator active and reactive powers (after com-
pensation).
N
a
c b
Z
la
I
lb
I
lc
I
sa
V
sc
V
sb
V
Fig. 3.15 An unbalanced three-phase three-wire star connected load
Solution:
Given that Z
a
= 3 +j 4 , Z
b
= 5 +j 12 , Z
c
= 12 j 5 .
1. Line currents I
la
, I
lb
, andI
lc
are found by rst computing neutral voltage as given below.
V
nN
=
1
(
1
Za
+
1
Z
b
+
1
Zc
)
(
V
a
Z
a
+
V
b
Z
b
+
V
c
Z
c
)
= (
Z
a
Z
b
Z
c
Z
a
Z
b
+Z
b
Z
c
+Z
c
Z
a
)(
V
a
Z
a
+
V
b
Z
b
+
V
c
Z
c
)
=
Z
abc
Z
(
V
a
Z
a
+
V
b
Z
b
+
V
c
Z
c
)
=
5097.82
o
252.4155.5
o
(24.1299.55
o
)
= 43.79 j 67.04 V
= 80.7557.15
o
V
Now the line currents are computed as below.
113
I
la
=
V
a
V
nN
Z
a
=
2300
o
80.7557.15
o
3 +j 4
= 33.2 j 21.65
= 39.6333.11
o
A
I
lb
=
V
b
V
nN
Z
b
=
230120
o
80.7557.15
o
5 +j 12
= 15.85152.21
o
A
I
lc
=
V
c
V
nN
Z
c
=
230120
o
80.7557.15
o
12 j 5
= 23.89143.35
o
A
2. Active and Reactive Powers
For phase a
P
a
= V
a
I
a
cos
a
= 230 39.63 cos(33.11
o
) = 7635.9 W
Q
a
= V
a
I
a
sin
a
= 230 39.63 sin(33.11
o
) = 4980 VAr
For phase b
P
b
= V
b
I
b
cos
b
= 230 15.85 cos(152.21
o
120
o
) = 140.92 W
Q
b
= V
b
I
b
sin
b
= 230 15.85 sin(152.21
o
120
o
) = 3643.2 VAr
For phase c
P
c
= V
c
I
c
cos
c
= 230 23.89 cos(143.35
o
+ 120
o
) = 5046.7 W
114
Q
c
= V
c
I
c
sin
c
= 230 23.89 sin(143.35
o
+ 120
o
) = 2179.3 VAr
Total three phase powers
P = P
a
+P
b
+P
c
= 12823 W
Q = Q
a
+Q
b
+Q
c
= 6443.8 VAr
3. Compensator Susceptance
First we convert star connected load to a delta load as given below.
Z
ab
l
=
ZaZ
b
+Z
b
Zc+ZcZa
Zc
= 4 +j 19 = 19.4277.11
o

Z
bc
l
=
Z
Za
= 50.44 +j 2.08 = 50.422.36
o

Z
ca
l
=
Z
Z
b
= 19.0 j 4.0 = 19.42 11.89
o

The above implies that,

Y
ab
l
= 1/Z
ab
l
= G
ab
l
+j B
ab
l
= 0.0106 j0.050
Y
bc
l
= 1/Z
bc
l
= G
bc
l
+j B
bc
l
= 0.0198 j0.0008
Y
ca
l
= 1/Z
ca
l
= G
ca
l
+j B
ca
l
= 0.0504 +j0.0106
From the above, the compensator susceptances are computed as following.
B
ab

= B
ab
l
+
(G
ca
l
G
bc
l
)

3
= 0.0681
B
bc

= B
bc
l
+
(G
ab
l
G
ca
l
)

3
= 0.0222
B
ca

= B
ca
l
+
(G
bc
l
G
ab
l
)

3
= 0.0053
Z
ab

= j 14.69 (capacitance)
Z
bc

= j 45.13 (inductance)
Z
ca

= j 188.36 (inductance)
115
4. After Compensation
Z
ab

l
= Z
ab
l
|| Z
ab

= 24.97 j 41.59 = 48.52 59.01

o

Z
bc

l
= Z
bc
l
|| Z
bc

= 21.52 j 24.98 = 32.9749.25

o

Z
ca

l
= Z
ca
l
|| Z
ca

= 24.97 j 41.59 = 19.7332 6.0

o

Let us convert delta connected impedances to star connected.

Z
a
=
Z
ab
Z
ca

Z
ab
+Z
bc
+Z
ca

= 9.0947 j 10.55
= 13.93 49.25
o

Z
b
=
Z
bc
Z
ab

Z
ab
+Z
bc
+Z
ca

= 23.15 +j 2.43
= 23.286.06
o

Z
c
=
Z
ca
Z
bc

Z
ab
+Z
bc
+Z
ca

= 4.8755 +j 8.12
= 9.4759.01
o

The new voltage between the load and system neutral after compensation is given by,
V

nN
=
1
(
1
Za
+
1
Z
b
+
1
Zc
)
(
V
a
Z
a
+
V
b
Z
b
+
V
c
Z
c
)
= 205.4257.15
o
V
Based on the above, the line currents are computed as following.
I
a
=
V
a
V

nN
Z
a
= 18.580
o
A
I
b
=
V
b
V

nN
Z
b
= 18.58 120
o
A
I
c
=
V
c
V

nN
Z
c
= 18.58120
o
A
116
Thus, it is seen that after compensation, the source currents are balanced and have unity power
factor with respective supply voltages.
Source powers after compensation
P
a
= P
b
= P
c
= 230 18.58 = 4272.14 W
P = 3P
a
= 12820.2 W
Q
a
= Q
b
= Q
c
= 0
Q = 0 VAr
Compensator powers
S
ab

= V
ab
(I
ab

)
= V
ab
(V
ab

.Y
ab

)
= V
2
ab
Y
ab

= (230

3)
2
(j 0.0068)
= j 10802 VA
S
bc

= V
2
bc
Y
bc

= (230

3)
2
(j 0.0222)
= j 3516 VA
S
ca

= V
2
ca
Y
ca

= (230

3)
2
(j 0.0053)
= j 842 VA
References
[1] A. Ghosh and G. Ledwich, Power quality enhancement using custom power devices. Kluwer
Academic Pub, 2002, vol. 701.
[2] T. J. E. Miller, Reactive power control in electric systems. Wiley, 1982.
[3] L. Gyugyi, Reactive power generation and control by thyristor circuits, IEEE Transactions
on Industry Applications, no. 5, pp. 521532, 1979.
117
[4] R. Otto, T. Putman, and L. Gyugyi, Principles and applications of static, thyristor-controlled
shunt compensators, IEEE Transactions on Power Apparatus and Systems, no. 5, pp. 1935
1945, 1978.
[5] N. Hingorani and L. Gyugyi, Understanding FACTS: Concepts and Technology of Flexible AC
Transmission Systems, 2000. Wiley-IEEE Press, 1999.
118