CONCEPTS AND APPLICATIONS: POWER QUALITY IN

POWER DISTRIBUTION SYSTEMS

Dr. Mahesh Kumar

Professor
Department of Electrical Engineering
Indian Institute of Technology Madras
Chennai - 600 036
Contents

1 1

2 3

3 5

4 7

5 SERIES COMPENSATION: VOLTAGE COMPENSATION USING DVR 9

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
5.2 Conventional Methods to Regulate Voltage . . . . . . . . . . . . . . . . . . . . . . 9
5.3 Dynamic Voltage Restorer (DVR) . . . . . . . . . . . . . . . . . . . . . . . . . . 10
5.4 Operating Principle of DVR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
5.4.1 General Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.5 Mathematical Description to Compute DVR Voltage . . . . . . . . . . . . . . . . 16
5.6 Transient Operation of the DVR . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.6.1 Operation of the DVR With Unbalance and Harmonics . . . . . . . . . . . 21
5.7 Realization of DVR voltage using Voltage Source Inverter . . . . . . . . . . . . . 22
5.8 Maximum Compensation Capacity of the DVR Without Real Power Support from
the DC Link . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

i
Chapter 5

SERIES COMPENSATION: VOLTAGE

COMPENSATION USING DVR

5.1 Introduction
Power system should ensure good quality of electric power supply, which means voltage and cur-
rent waveforms should be balanced and sinusoidal. Furthermore, the voltage levels on the system
should be within reasonable limits, generally within 100 ± 5% of their rated value. If the voltage is
more or less than this pre-specified value, performance of equipments is sacrificed. In case of low
voltages, picture on television starts rolling, the torque of induction motor reduces to the square of
voltage and therefore there is need for voltage compensation.

5.2 Conventional Methods to Regulate Voltage

In order to keep load bus voltage constant, many conventional compensating devices such as listed
below can be used. In general, these can be referred as VAR compensator.

1. Shunt Capacitors
2. Series Capacitors
3. Synchronous Capacitor
4. Tap Changing Transformer
5. Booster Transformer
6. Static Synchronous Series Capacitor
7. Dynamic Voltage Restorer
The first six methods are employed at transmission level while the last method is by employing
Dynamic Voltage Restorer (DVR), is mostly employed in power distribution network to protect any
voltage variation at the load bus connected to the sensitive and critical electrical units. The DVR

9
is a series connected custom power device used to mitigate the voltage unbalance, sags, swells,
harmonics and any abrupt changes due to abnormal conditions in the system. In the following
section, dynamic voltage restorer will be described in detail.

5.3 Dynamic Voltage Restorer (DVR)

A dynamic voltage restorer (DVR) is a solid state inverter based on injection of voltage in series
with a power distribution system [1], [2]. The DC side of DVR is connected to an energy source or
an energy storage device, while its ac side is connected to the distribution feeder by a three-phase
interfacing transformer. A single line diagram of a DVR connected power distribution system is
shown in the figure below. In this figure, vs (t) represents supply voltage, vt (t) represents terminal
voltage and vl (t) represents the load voltage. Since DVR is a series connected device, the source
current, is (t) is same as load current, il (t). Also note in the figure, vf (t) is DVR injected voltage
in series with line such that the load voltage is maintained at sinusoidal nominal value.

vs (t ) Rs  jX s vt (t ) vl (t )

is  il
v f (t )
L
O
A
D

Fig. 5.1 A single-line diagram of DVR compensated system

The three-phase DVR compensated system is shown in Fig. 5.2 below. It is assumed that the
transmission line has same impedance in all three phases. A DVR unit which is represented in Fig.
5.1, have following components [3]- [5].

vsa v fa vla
Rs  jX s
A
isa
vsb v fb vlb
Rs  jX s
N B n
isb
vs c v fc vlc
Rs  jX s
C
isc
Fig. 5.2 A single-line diagram of DVR compensated system

10
 1. Voltage Source Inverter
2. Filter capacitors and inductors
3. Injection transformer
4. DC storage system

These components are shown in Fig. 5.3. Some other important issues i.e., how much voltage
should be injected in series using appropriate algorithm, choice of suitable power converter topol-
ogy to synthesize voltage and design of filter capacitor and inductor components have to be ad-
dressed while designing the DVR unit.

PCC Cf
vt vl
- v +
f
Rs Ls Injecion
transformer Load
Lf
vs

VSI

Energy
Cdc
storage

Fig. 5.3 Schematic diagram of a DVR based compensation in a distribution system

5.4 Operating Principle of DVR

Consider a DVR compensated single phase system as shown in Fig. 5.4. Let us assume that source
voltage is 1.0 pu and we want to regulate the load voltage to 1.0 pu. Let us denote the phase angle
between V s and V l as δ. In this analysis, harmonics are not considered. Further we assume that
during DVR operation, real power is not required except some losses in the inverter and the non
ideal filter components. These losses for the time being are considered to be zero. This condition
implies that the phase difference between V f and I s should be 90o . Let us first consider a general
case to understand the concept.

The DVR equivalent circuit with fundamental voltages and current is shown in Fig. 5.4. Ap-
plying Kirchoff’s voltage law in the circuit,

V s + V f = I s (Rs + jXs ) + V l
= I s Zs + V l . (5.1)

11
 Vs Rs  jX s Vt Vl

Is
Vf
L
O
A
D

Fig. 5.4 Schematic diagram of a DVR based compensation in a distribution system

Note that in above curcuit I s = I l = I. The load voltage V l can be written in terms of load current
and load impedance as given below.

V s + V f = I (Zs + Zl ) (5.2)
Using (5.1), the source voltage can be expressed as in the following.
V s = V l + I Rs − (V f − jI Xs ) (5.3)
With the help of above equation, the relationship between load voltage and the source and DVR
voltages can be expressed as below.

 
Vs+Vf
Vl = Zl (5.4)
Zs + Zl
Example 6.1 Let us apply condition to maintain load voltage same as source voltage i.e., V l = V s .
Discuss the feasibility of injected voltage in series with the line as shown in Fig. 5.4, to obtain load
voltage same as source voltage. Consider the following cases.
(a) Line resistance is negligible with Zs = j0.25 pu and Zl = 0.5 + j0.25 pu.
(b) When the load is purely resistive with Zs = 0.45 + j0.25 pu and Zl = 0.5 pu.
Solution:

(a) When line resistance is negligible

The above condition implies that, Rs = 0. Without DVR, the load terminal voltage V l can be
given as following.

Vs 1.0∠0o
Il = = = 1.0 − j1.0 = 1.4142∠ − 45o pu
Zl + Zs 0.5 + j0.5

12
Therefore the load voltage is given as following. V l = Zl I l = 0.5590∠26.56o × 1.4142∠ − 45o =
0.7906∠ − 18.43o pu. This is illustrated in Fig. 5.5(a). Thus the load voltage has reduced by 21%.
Now it is desired to maintain load voltage same as supply voltage in magnitude and phase angle.
Thus, substituting V s = V l in equation (5.1), we get,

V s + V f = I (Rs + jXs ) + V l
⇒ V f = I (Rs + jXs )
jXs
Vf = V l , since Rs = 0 and I = V l /Zl
Zl
Neglecting resistance part of the feeder impedance, Zs = j0.25, the DVR voltage can be computed
as above.

j0.25
Vf = 1.0∠0o for V l = 1.0∠0o
0.5 + j0.25
= 0.4472∠63.4349o pu.

From the above the, line current is computed as following.

Vl 1.0∠0o
Is = = = 1.7889∠ − 26.56o pu.
Zl 0.5590∠26.56o

It is to be noted that, although V s = V l = 1.0∠0o pu, it does not imply that no power flows from
‘
source to load. In fact the total effective source voltage is V s = V s + V f = 1.2649∠18.8◦ pu.
Therefore it implies that the effective source voltage is leading the load voltage by an angle of
18.43o . This ensures the power flow from the source to load. This is illustrated by drawing phasor
diagram in the Fig. 5.5(b). below.

jX s I
Vs =1.00o Vs I
s
18.8o jX


 jX s I 26.56o Vl =1.00o
45o
f
V

Vl
I I

(a) (b)

Fig. 5.5 Terminal voltage (a) Without DVR (b) With DVR

(b) When load is purely resistive

13
For this case Xl = 0, therefore Zl = Rl = 0.5 pu. Substituting V s = V l in (5.1), we get the
following.

V f = (Rs + jXs )I

Substituting I = V l /Rl , we get,

   
Rs + jXs Vl Vl
Vf = V l = Rs + jXs = Rs I s + jXs I s .
Rl Rl Rl
From this above equation, it is indicated that DVR voltage has two components, the one is in phase
with I s and the other is in phase quadrature with I s . This implies that for purely resistive load, it
is not possible to maintain V l = V s without active power supplied from the DVR to the load. This
is due to the presence of in phase component of the DVR voltage in the above equation. This is
illustrated in the phasor diagram given in Fig. 5.6 below.

Vs
jX s I

I Vl I Rs

Fig. 5.6 When load is purely resistive

5.4.1 General Case

In general, it is desired to maintain the magnitude of the load voltage equal to the source voltage
i.e., 1.0 pu. The voltage equation in general relating the source, load and DVR has been expressed
in (5.3) and is given below.

V s = V l + I Rs − (V f − jI Xs )

The above equation is illustrated using phasor diagram description in Fig. 5.7 given below. Three
cases of voltage compensation are discussed below.

Case 1: When Rs I(= CD) < CD1

For this case, it is always possible to maintain load voltage same as source voltage i.e., Vl = Vs .
The DVR is expected to supply enough range reactive power to meet this condition. When Rs Is is
quite smaller than CD, the above condition can be met by supplying less reactive power from the
DVR. For this condition there are two solutions. Graphically, these solutions are represented by
points A an B in the Fig. 5.7.

Case 2: When Rs I(= CD) > CD1

For this condition, it is not possible to meet Vl = Vs . This is shown by lines passing through

14
 Vs'
A1 jX s I
O I Rs
Vl
l A
Vs
I

C
D Locus of

D1 Vf
D2
Vs

B1 B
Fig. 5.7 Compensation using DVR: General case

points between D1 and D2 . This may take place due to the higher feeder resistance or high current,
thus making product of I Rs relatively large.

Case 3: When Rs I(= CD) = CD1

This is limiting case of compensation to obtain Vs = Vl . This condition is now satisfied at only
one point when CD=Rs I. This is indicated by point D in the Fig. 5.7.

Now let us set the following objective for the load compensation.
Vl = Vs = V = 1.0 pu (5.5)
From Fig. 5.7, OC = V cos φl = cos φl . Therefore, CD = OD − OC = V (1 − cos φl ) =
(1 − cos φl ) pu. In order to meet the condition given by (5.5), the following must be satisfied.
Rs I ≤ V (1 − cos φl ) (5.6)
The above implies that
V (1 − cos φl )
Rs ≤ (5.7)
I
V (1 − cos φl )
or I≤ . (5.8)
Rs
15
Thus it is observed that for a given power factor, the DVR characteristics can be obtained by
varying Rs and keeping I constant or vice versa. This is described below. Let us consider three
conditions Rs = 0.04 pu, Rs = 0.1 pu and Rs = 0.4 pu. For these values of feeder resistance, the
line currents are expressed as following using (5.8).
I = 25 (1 − cos φl ) pu for Rs = 0.04 pu
I = 10 (1 − cos φl ) pu for Rs = 0.1 pu
I = 2.5 (1 − cos φl ) pu for Rs = 0.4 pu.
The above currents are plotted as function of load power factor and are shown in Fig. 5.8. Since
Rs I = Vl (1 − cosφl ), when Rs increases, I has to decrease to make Vl (1 − cosφl ) to be a constant
for a given power factor. Thus if the load requires more current than the permissible value, the
DVR will not be able to regulate the load voltage at the nominal value, i.e., 1.0 pu. However
we can regulate bus voltage less than 1.0 pu. For regulating the load voltage less than 1.0 pu the
current drawing capacity of the load increases.

18

16

14

12

10

8
-600 -300 00 300 600
6

0
-80 -60 -40 -20 0 20 40 60

Fig. 5.8 DVR characteristics for different load power factor and feeder resistance

5.5 Mathematical Description to Compute DVR Voltage

The previous section explains DVR characteristics and describes the feasibility of realizing DVR
voltage graphically under different operating conditions. In this section, a feasible solution for
the DVR voltage is presented with a mathematical description. This plays significant role while
implementing DVR on real time basis. Reproducing equation (??) for sake of completeness,

V s + V f = V l + (Rs + jXs ) I. (5.9)

16
Denoting, (Rs + jXs ) I = a2 + jb2 and V f = Vf ∠V f = Vf (a1 + jb1 ), the above equation can be
written as following.

V s = V l + (a2 + jb2 ) − V f
= V l + (a2 + jb2 ) − Vf (a1 + jb1 )
= 1.0∠0o + (a2 + jb2 ) − Vf (a1 + jb1 ) (5.10)

Since, source voltage and load voltage have to be maintained at nominal value i.e., 1.0 pu, therefore
V s = Vs ∠δ = 1.0∠δ. Substituting this value of V s in above equation, we get,

V s = 1.0∠δ = cos δ + j sin δ = {(1 + a2 ) − Vf a1 } + j(b2 − Vf b1 ) (5.11)

Squaring and adding the real and imaginary parts from both the sides of the above equation, we
get,

(1 + a2 )2 + Vf2 a21 − 2(1 + a2 ) a1 Vf + b22 + Vf2 b21 − 2 b1 b2 Vf − Vs2 = 0 (5.12)

Since a21 + b21 = 1, therefore summation of underlines terms, Vf2 (a21 + b21 ) = Vf2 . Using this and
rearranging above equation in the power of Vf , with Vs = 1.0 p.u., we get the following.

Vf2 − 2 {(1 + a2 ) a1 + b1 b2 } Vf + (1 + a2 )2 + b22 − 1.0 = 0 (5.13)

The above equation gives two solutions for Vf . These are equivalent to two points A and B shown
in the Fig. 5.7. However, the feasible value of the voltage is chosen on the basis of the rating of
the DVR.

Example 6.2 Consider a system with supply voltage 230 V = 1.0 pu, 50 Hz as shown in the Fig.
5.9. Consider feeder impedance as Zs = 0.05 + j0.3 pu and load impedance Zl = 0.5 + j0.3 pu.

a. Compute the load voltage without DVR.

b. Compute the current and DVR voltage such that Vl = Vs .

c. Compute the effective source voltage including DVR. Explain the power flow in the circuit.

d. Compute the terminal voltage with DVR compensation.

Solution: 1. When DVR is not connected.

The system parameters are given as following. The supply voltage Vs = 1.0∠0o pu, Zs = Rs +
j Xs = 0.05 + j0.3 and Zl = Rl + j Xl = 0.5 + j0.3 pu. The current in the circuit is given by,

Vs 1.0∠0
Is = I = Il = =
Zs + Zl 0.55 + j0.6
= 0.83 − j0.91 = 1.2286∠ − 47.49o pu

17
 vs (t ) Rs  jX s vt (t ) vl (t )

v f (t )

Rl  jX l
is  il

Fig. 5.9 A DVR compensated system

The load voltage is therefore given by,

V l = Zl I s = 1.2286∠ − 47.490 × 0.5 + j0.3

= 0.7164∠ − 16.53o pu
Thus we observe that the load voltage is 71% of the rated value. Due to reduction in the load
voltage, the load may not perform to the expected level.

2. When DVR is connected

It is desired to maintain Vl = Vs by connecting the DVR. Taking V l as reference phasor i.e.,

V l = 1.0∠0o , The line current is computed as below.
1.0∠0
I = = 1.47 − j0.88 = 1.715∠ − 30.96o pu
0.5 + j0.3
Writing KVL for the circuit shown in (5.9),
V s + V f = V l + (Rs + jXs ) I.
The DVR voltage V f can be expressed as following.
V f = Vf ∠(∠I s + 90o )
The angle of V f is taken as (∠I s + 90o ) so that DVR do not exchange any active power with the
system.
V f = Vf ∠(∠I s + 90o )
= Vf ∠(−30.96o + 90o )
= Vf ∠59.04o = Vf (0.51 + j0.86) = Vf (a1 + jb1 ) pu
The above equation implies that a1 = 0.51, b1 = 0.86.
Let us now compute (Rs + jXs ) I.
(Rs + jXs ) I = (0.05 + j0.3) (1.46 − j0.86)
= 0.3041∠80.54o × 1.715∠ − 30.96o
= 0.5199∠49.58o
= 0.3370 + j0.3958 pu

18
The above implies that a2 = 0.337 and b2 = 0.3958. As discussed in previous section, the equation
V s + V f = V l + (Rs + jXs ) I can be written in following form.
Vf2 − 2 {(1 + a2 ) a1 + b1 b2 } Vf + (1 + a2 )2 + b22 − 1.0 = 0.
Substituting a1 , b1 , a2 , b2 in the above equation, we get the following quadratic equation for the
DVR.
Vf2 − 2.0463 Vf + 0.9442 = 0
Solving the above equation, we get Vf = 0.7028, 1.3434 pu as two values of the DVR voltage.
These two values correspond to the points A and B respectively in Fig. 5.7. However, the feasible
solution is Vf = 0.7028 pu, as it ensures less rating of the DVR.
Therefore,
V f = 0.7028∠59.04o
= 0.3614 + j0.6028 pu.
The source voltage can be computed using the following equation.
V s = V l + (Rs + jXs ) I − V f .
= 1.0∠0o + (0.05 + j0.3) 1.715∠ − 30.96o − 0.7028∠59.04o
= 0.9767 − j0.2056 = 1.0∠ − 11.89o pu
3. Effective source voltage

It is seen that the magnitude of V s is 1.0 pu which is satisfying the condition Vs = Vl . How-
ever the angle of V s is ∠ − 11.89o which implies that power is flowing from load to the source.
‘
This is not true because the effective source voltage is now V s = V s +V f . This is computed below.

‘
V s + V f = V s = 0.9767 − j0.2056 + 0.3614 + j0.6028
= 1.3382 + j0.397
= 1.3958∠16.52o pu
From above it is evident that the effective source voltage has magnitude of 1.3958 pu and an
angle of ∠16.52◦ which ensures that power flows from source to the load. For this the equivalent
circuit is shown in the Fig. 5.10 below.

4. Terminal voltage with DVR compensation

The terminal voltage can also be computed as following.

V t = V s − Zs I = V l − V f
= 1.0∠0o − 0.6983∠59.06o
= 0.8796∠ − 43.28o pu
This indicates that for rated current flowing in the load, the terminal voltage is less than the 1.0 pu
and needs compensation. After compensation the load voltage is 1.0 pu as shown in the Fig. 5.10.

19
 Vs'  Vs  V f
 1.3916.52o Rs  jX s Vl  1.00o

Rl  jX l
I s  Il

Fig. 5.10 A DVR compensated system

5.6 Transient Operation of the DVR

In the previous section the operation of the DVR in the steady state was discussed with assumption
that full system information is available. While implementing the DVR compensation scheme, the
above discussed method should be implemented on the real time basis. For the single phase DVR
operation, following steps are required.
1. Define a reference quantity such as the load voltage vl (t) and other quantities are synchro-
nized to it.
2. To compute phase angle of the DVR voltage, a fundamental of line current is extracted with
respect to reference quantity.
3. Then DVR voltage is computed using (5.12), which is reproduced below.
Vf2 − 2 {(1 + a2 ) a1 + b1 b2 } Vf + (1 + a2 )2 + b22 − Vs2 = 0

4. DVR voltage Vf is then synthesized using magnitude Vf from the above equation and phase
angle that leads the fundamental of the line current by 90o .
The above method can be refereed as Type 1 control [1]. The method assumes that all circuit
parameters are known along with the information of the source impedance. This however may not
be feasible in all circumstances. To solve this problem Type 2 control is suggested. In Type 2
control only local quantities are required to compute the DVR voltage. The method is described
below.
The terminal voltage, which is local quantity to the DVR as shown in Fig. 5.1 can be expressed as
following.
Vt = Vl−Vf
= Vl ∠0o − Vf (a1 + jb1 )
= (Vl − a1 Vf ) − jb1 Vf (5.14)
Since, V t = Vt ∠δt = Vt cos δt + jVt sin δt , the above equation is written as following.
Vt cos δt + jVt sin δt = (Vl − a1 Vf ) − jb1 Vf (5.15)

20
Squaring adding both sides we get,
Vt2 = (Vl − a1 Vf )2 + b21 Vf2
= Vl2 + a21 Vf2 + b21 Vf2 − 2a1 Vl Vf
= Vl2 + Vf2 − 2a1 Vl Vf (since a21 + b21 = 1). (5.16)
The above equation can be arranged in the powers of the DVR voltage as given below.
Vf2 − 2a1 Vl Vf + Vl2 − Vt2 = 0 (5.17)
To implement DVR for unbalanced three-phase system without harmonics, angles of DVR voltages
are found by shifting current angles by 90o i.e.,
∠V f a = ∠I a + 90o
∠V f b = ∠I b + 90o (5.18)
∠V f c = ∠I c + 90o
The magnitude of DVR voltage can be found using equations (5.13) and (5.17) for Type 1 and
Type 2 control respectively.
Based on above the DVR voltages vf a , vf b , vf c can be expressed in time domain as given below.
√
vf a = 2 Vf a sin(ω t + ∠V f a )
√
vf b = 2 Vf b sin(ω t − 120o + ∠V f b ) (5.19)
√
vf c = 2 Vf c sin(ω t + 120o + ∠V f c )

5.6.1 Operation of the DVR With Unbalance and Harmonics

In the previous analysis, it was assumed that the supply voltages are unbalanced without harmonics.
In this section the operation of the DVR with harmonics will be discussed. The terminal voltages
(vta , vtb and vtc ) are resolved into their fundamental voltages and the rest part, as given below.
vta = vta1 + vta rest
vtb = vtb1 + vtb rest (5.20)
vtc = vtc1 + vtc rest
The angles of fundamental DVR voltages (∠V f a1 , ∠V f b1 and ∠V f c1 ) are extracted as following.
∠V f a1 = ∠I a1 + 90o
∠V f b1 = ∠I b1 + 90o (5.21)
∠V f c1 = ∠I c1 + 90o
The angles of currents, i.e., ∠I a1 , ∠I b1 , ∠I c1 , are computed by extracting fundamental of the three
phase load currents using Fourier transform. The magnitudes of the fundamental DVR voltages
(Vf a1 , Vf b1 and Vf c1 ) can be computed using equations (5.13) and (5.17) for Type 1 and Type 2
control respectively. For example, using Type 2 control the fundamental phase-a DVR voltage is
computed as per following equation.

21
 Vf2a1 − 2 aa1 Vl Vf a1 + Vl2 − Vta1 2 = 0 (5.22)

In above equation aa1 + jba1 = ∠V f a1 and Vta1 is fundamental positive sequence phase-a terminal
voltage as given above in (5.20). Similar expression can be written for phase-b and phase-c. This
equation gives solution only for fundamental component of the DVR voltage. The rest of the DVR
voltages which consist of harmonics and unbalance must be equal and opposite to that of the rest
part of the terminal voltages i.e., vta rest , vtb rest and vtc rest . Therefore these can be given using
following equations.
vf a rest = −vta rest
vf b rest = −vtb rest (5.23)
vf c rest = −vtc rest
Thus, the total DVR voltage to be injected can be given as following.
vf a = vf a1 + vf a rest
vf a = vf b1 + vf b rest (5.24)
vf a = vf c1 + vf c rest
In above equation, vf a1 , vf b1 , vf c1 are constructed using equation (5.19). Once vf a , vf b and vf c are
known, these voltages are synthesized using suitable power electronic circuit. It will be discussed
in the following section.

5.7 Realization of DVR voltage using Voltage Source Inverter

In the previous section, a reference voltage of DVR was extracted using discussed control algo-
rithms. This DVR voltage however should be realized in practice. This is achieved with the help
of power electronic converter which is also known as voltage source inverter. Various components
of the DVR were listed in the beginning of chapter. They are shown in detail in the Fig. 5.11.
The transformer injects the required voltage in series with the line to maintain the load bus voltage
at the nominal value. The transformer not only reduces the voltage requirement but also provides
isolation between the inverters. The filter components of the DVR such as external inductance (Lt )
which also includes the leakage of the transformer on the primary side and ac filter capacitor on
the secondary side play significant role in the performance of the DVR [Sasitharan Thesis].
The same DC link can be extended to other phases as shown in Fig. 5.11. The single phase
equivalent of the DVR is shown in the Fig. 5.12.
In Figs. 5.11 and 5.12, vinv denotes the switched voltage generated at the inverter output termi-
nals, the inductance, Lt represents the total inductance and resistance including leakage inductance
and resistance of transformer. The resistance, Rt models the switching losses of the inverter and
the copper loss of the connected transformer. The voltage source inverter (VSI) is operated in a
switching band voltage control mode to track the reference voltages generated using control logic
as discussed below.
Let Vf∗ be the reference voltage of a phase that DVR needs to inject in series with the line with help
of the VSI explained above. We form a voltage hysteresis band of ±h over this reference value.

22
 vf
+
Rs Ls vt vl il
Cf

Load bus
vs

Load
vp
Lt
S1 S3 iinv
D1 Rt
D3
+
Vdc vinv
Cdc
S4 D4 S2 D2

To other
phases

Fig. 5.11 The DVR circuit details

Cf i fac
vt vl

vinv Xt Rt

Fig. 5.12 Equivalent circuit of the DVR

Thus, the upper and lower limits within which the DVR has to track the voltage can be given as
following.
vf up = vf∗ + h
vf dn = vf∗ − h (5.25)
The following switching logic is used to synthesize the reference DVR voltage.

If vf ≥ vf up
S1 − S2 OFF and S3 − S4 ON (‘-1’ state)
else if vf ≤ vf dn
S1 − S2 ON and S3 − S4 OFF (‘+1’ state)
else if vf dn ≥ vf ≤ vf up
retain the current switching status of switches
end.

23
It is to be noted that switches status S1 − S2 ON and S3 − S4 OFF is denoted by ‘+1’ state
and it gives vinv = +Vdc . The switches status S1 − S2 OFF and S3 − S4 ON corresponds to ‘-1’
state providing vinv = −Vdc as shown in Fig. 5.11. The above switching logic is very basic and
has scope to be refined. For example ‘0’ state of the switches of the VSI as shown in Fig. 5.11, can
also be used to have smooth switching and to minimize switching losses. In the zero state, vinv = 0
and refers switches status as S3 D1 or S4 D2 for positive inverter current (iinv > 0). Similarly, for
negative inverter current (iinv < 0), ‘0’ state is obtained through S1 D3 or S2 D4 . With the addition
of ‘0’ state, the switching logic becomes as follows.

If vf∗ > 0
if vf ≥ vf up
‘0’ state
else if vf ≤ vf dn
‘+1’ state
end
else if vf∗ < 0
if vf ≥ vf up
‘-1’ state
else if vf ≤ vf dn
‘0’ state
end
end.

In order to improve the switching performance one more term is added in the above equation
based on the feedback of filter capacitor current.

vf up = vf∗ + h + α if ac
vf dn = vf∗ − h + α if ac (5.26)

Where α is a proportional gain given to smoothen and stabilize the switching performance of the
VSI [2]. The dimension of α is Ω and is thus is equivalent to virtual resistance, whose effect to
damp out and smoothen the DVR voltage trajectory resulted from the switching of the inverter
[4]. The value of hysteresis band (h) should be chosen in such a way that it limits switching
frequency within the prescribed maximum value. This kind of voltage control using VSI is called as
switching band control. The actual DVR voltage is compared with these upper and lower bands of
the voltage (Vf up , Vf dn ) and accordingly switching commands to the power switch are generated.
The switching control logic is described in the Table 5.1. To minimize switching frequency of
the VSI, three level logic has been used. For this an additional check of polarity of the reference
voltage has been taken into consideration. Based on this switching status, the inverter supplies
+Vdc , 0 and −Vdc levels of voltage corresponding to the 1, 0 and -1 given in the table, in order to
synthesis the reference DVR voltage.
In addition to switching band control, an additional loop is required to correct the voltage in
the dc storage capacitor against losses in the inverter and transformer. During transients, the dc
capacitor voltage may rise or fall from the reference value due to real power flow for a short

24
 Table 5.1 Three level switching logic for the VSI

Conditions Switching value

Vf∗ ≥0 Vf > Vup 0
Vf∗ ≥0 Vf < Vdn 1
Vf∗ <0 Vf > Vup -1
Vf∗ <0 Vf < Vdn 0

duration. To correct this voltage deviation, a small amount of real power must be drawn from the
source to replenish the losses. To accomplish this, a simple proportional-plus-integral controller
(PI) is used. The signal uc is generated from this PI controller as given below.
Z
uc = Kp e V dc + Ki e V dc dt (5.27)

Where, e V dc = Vdc ref − Vdc . This control loop need not to be too fast. It may be updated once
in a cycle preferably synchronized to positive zero crossing of phase-a voltage. Based on this
information the variable uc will be included in generation of the fundamental of DVR voltage as
given below.
V f 1 = Vf 1 ∠(∠I s + 90o − uc ) = Vf 1 (ã1 + j b̃1 ) (5.28)
Then the equation (5.17), is modified to the following.
Vf21 − 2 ã1 Vl Vf 1 + Vl2 − Vt12 = 0 (5.29)
The above equation is used to find the DVR voltage. It can be found that the phase difference
between line current and DVR voltage differs slightly from 90o in order to account the losses in
the inverter.

5.8 Maximum Compensation Capacity of the DVR Without Real Power

Support from the DC Link
There is direct relationship between the terminal voltage, power factor of the load and the maxi-
mum possible achievable load voltage, with assumption that no real power is required from the dc
bus. Referring to quadratic equation in (5.29), for given value of Vt1 and a target load bus voltage
Vl , The equation gives two real values of Vf 1 for feasible solution. In case solution is not feasible,
the equation gives two complex conjugate roots. This concludes that the maximum voltage that
DVR can compensate corresponds to the single solution of the above equation, which is given
below. This solution corresponds to point ‘D’ in Fig. 5.7.
p
(2 ã1 Vl )2 − 4 (Vl2 − Vt12 )
2 ã1 Vl ±
Vf 1 = (5.30)
2
Since, voltage should not be complex number, the value of the terms within square root must not
be negative. Therefore
(2 ã1 Vl )2 ≥ 4 (Vl2 − Vt12 ) (5.31)

25
The above equation implies that

Vt1
Vl ≤ p . (5.32)
1 − ã21

And therefore, the DVR voltage is given by the following equation.

Vf 1 = ã1 Vl (5.33)

With no losses in the VSI, uc = 0,

Vt1 V
Vl ≤ p = p t1 (5.34)
1 − ã21 1 − a21

Since, ã1 + j b̃1 = 1∠(90o + φl ) p

= cos(90o +pφl ) + j sin(90o + φl ) = − sin φl + j cos φl . This
implies ã1 = − sin φl , therefore 1 − ã21 = 1 − (− sin φl )2 = cos φl . using this relation, the
above equation can be written as following.

Vt1
Vl ≤ (5.35)
cos φl

For this condition, the DVR voltage Vf is as following.

Vf = a Vl (5.36)

Example 6.3 A DVR is shown in Fig. 5.13. The feeder impedance of the line 0.1+j0.5 pu. Assume
il to be load current represented by square waveform approximated by the following expression.

il = 1.0 sin(ω t − 30o ) + 0.3 sin(3 ω t − 90o ) pu

a. Find the load voltage vl (t) with respect to source voltage without DVR compensation i.e., vf =
0.

b. Is it possible to maintain load voltage, Vl to be 1.0 pu sinusoidal waveform? If yes what is the
DVR voltage, vf (t)?

c. If no, how much maximum voltage can be maintained at load terminal with the DVR without
taking any real power from the dc bus?

26
 vs RT  jXT vt vf vl
is
il

Fig. 5.13 A DVR compensated system

Solution:

(a) When Vf = 0
X
vt = vs − Zsh ih
h=1,3

The impedance at the fundamental frequency, Zs1 = 0.1 + j0.5 = 0.51∠78.7o pu.
The impedance at third harmonic, Zs3 = 0.1 + j1.5 = 1.50∠86.18o pu.
Therefore the voltage drop due to fundamental component of the current,
Vzs1 = (0.1 + j0.5) × 0.707∠−30o
= 0.51∠78.69o × 0.707∠−30o
= 0.36∠48.69 pu.
The voltage drop due to third harmonic component of the current,
Vzs3 = (0.1 + j1.5) × 0.21∠−90o
= 0.31∠−3.820 pu.
The load voltage thus can be given by
vt = vs − (is1 Zs1 + is3 Zs3 )
= 1.0 sin ωt − 0.51 sin(ωt + 48.69o ) −0.45 sin(3 ωt − 3.82o )
| {z }
= 0.7659 sin(ωt − 30 ) − 0.45 sin(3 ωt − 3.82o ) pu
o

= vt1 (t) + vth (t).

Implying that,
0.7659
V t1 = √ ∠−30o = 0.5414∠−30o pu
2

27
(b) With DVR

From the above equation, Vt1 = 0.5415. With load voltage vl = 1.0 sin ωt, the DVR voltage
Vf 1 can be solved using quadratic equation as mentioned in Type 2 control. Further,
V f 1 = Vf 1 ∠(∠I s1 + 90o ) = Vf 1 ∠(−30o + 90o ) = Vf 1 ∠60o
= Vf 1 (cos 60o + j sin 60o ) = Vf 1 (0.5 + j0.8666) = Vf 1 (a1 + jb1 ) pu.
The above implies a1 = 0.5, b1 = 0.866.. Knowing this, we can solve Vf 1 using following
quadratic equation.
Vf21 − 2 a1 Vl Vf 1 + Vl2 − Vt12 = 0
From the above,
q
Vf 1 = a1 Vl ± a21 Vl2 − (Vl2 − Vt12 )
v
u  2 ( 2 )
1 u 1 1
= 0.5 × √ ± t(0.5)2 × √ − √ − (0.5415)2
2 2 2
√
= 0.35 ± 0.125 − 0.1842
The above solution is complex quantity, which implies that it is not possible to maintain load volt-
age at 1.0 sin ωt.

(c) Maximum possible load voltage

The maximum load voltage that can be obtained with the DVR, without any real power from
the dc bus can be given as following.
Vt1 0.5415
Vl = p 2
=√ = 0.6253 pu.
1 − a1 1 − 0.52
√
In the time domain the load voltage vl = vl1 = 2 × 0.6253 sin ωt = 0.8843 sin ωt. For this load
voltage the DVR voltage is given as following.
Vf 1 = a1 Vl = 0.5 × 0.6253 = 0.3126 pu.
This implies
V f 1 = 0.3126∠60o pu.
The time domain repression for the fundamental DVR voltage is given as,
√
vf 1 (t) = 2 × 0.3126 sin(ωt + 60o ) = 0.4421 sin(ωt + 60o ) pu.
The harmonic voltage that DVR compensates is as following.
vf h (t) = −vth = 0.45 sin (3 ωt − 3.82o ) pu.

28
The total DVR voltage is given as below.
vf (t) = vf 1 (t) + vth (t)
= 0.4421 sin(ωt + 60o ) + 0.45 sin (3 ωt − 3.82o ) p.u.
Example 6.4 A DVR compensated single phase system is shown in Fig. 5.14. Determine VD1 and
VD2 such that the voltage remains 1.0 pu at all three buses. Also, determine the value of δ1 and δ to
ensure power flow from source to load. The value of relevant parameters, all in per unit, are given
as follows.
Zs1 = 0.02 + j 0.5, Zs2 = 0.05 + j 0.3, Zl1 = 0.25 + j 0.4 and Zl2 = 0.3 + j 0.4.

Vs  1.0 VD1 V1  11 VD 2 V2  10

Il 2
Is Zs2
Z s1
Z l1 I l1
Zl 2

Fig. 5.14 A DVR compensated single phase system.

Solution:

The DVR2 is operated at nominal voltage, i.e., V 2 = 1∠0. Therefore, the current I l2 is given
as follows.
V2 1∠0
I l2 = = = 2.0∠ − 53.13o p.u.
Zl2 0.3 + j 0.4
Since, the DVR2 injects a voltage with an angle of 90o with respect to the current I l2 . Therefore,
∠V D2 = 90o − 53.13o = 36.87o .
The values of a1l and b1l is computed as follows:
a1l + j b1l = 1.0∠V D2 = 0.8 + j 0.6 p.u.
Voltage across the feeder 2 (Zs2 ) is given as,
∆V f 2 = Zs2 I l2 = (0.05 + j 0.3) × 2.0∠ − 53.13o = 0.54 + j 0.28 p.u. = a2l + j b2l .
The above equation implies that a2l = 0.54 and b2l = 0.28. Now, the voltage VD2 is computed by
solving following equation.
2
VD2 − 2 {(1 + a2l ) a1l + b1l b2l } VD2 + (1 + a2l )2 + b22l − 1.0 = 0.
Substituting variables, following quadratic equation is used to find out DVR voltage.
2
VD2 − 2.8 VD2 + 1.45 = 0.

29
Solving the equation, VD2 = 2.11 p.u. and 0.68 p.u. are obtained. However, VD2 = 0.68 p.u. is the
feasible solution.
Now, the voltage V 1 is computed as follows.
V 1 = Zs2 I l2 + V 2 − V D2
= 0.54 + j 0.28 + 1∠0 − 0.68∠36.87o
= 1∠ − 7.55o p.u.
The above implies that, δ1 = -7.55o .
Once value of V 1 is known, the current I l1 is given by
V1 1∠ − 7.55o
I l1 = = = 2.12∠ − 65.55o p.u.
Zl1 0.25 + j 0.4
Using KCL, source current is given as
I s = I l1 + I l2 = 2.0∠ − 53.13o + 2.12∠ − 65.55o = 4.09∠ − 59.52o p.u.
Again, DVR1 will inject a voltage perpendicular to current I l1 . Therefore,
∠V D1 = 90o − 59.52o = 30.48o .

a1s + j b1s = 1.0∠V D1 = 1.0∠30.48o = 0.86 + j 0.51 p.u.

Thus, voltage across the feeder 1 (Zs1 ) is given as
∆V f 1 = Zs1 I s = (0.02 + j 0.5) × 4.09∠ − 59.52o = 1.8065 + j 0.968 pu = a2s + j b2s .
The above equation implies that a2s = 1.8065 and b2s = 0.968. Now, the voltage VD1 is computed
by solving following equation:
2
VD1 − 2 {(1 + a2s ) a1s + b1s b2s } VD1 + (1 + a2s )2 + b22s − 1.0 = 0.
Substituting variables, following quadratic equation is obtained:
2
VD1 − 5.8195 VD1 + 7.8137 = 0.
Solving the equation, VD2 = 2.1017 p.u. and 3.7178 p.u. are obtained. However, VD2 = 2.1017 p.u.
is the feasible solution. Now, the source voltage is given as following.
V s = Zs1 I s + V 1 − V D1
= 2.049∠28.18o + 1∠ − 7.55o − 2.1017∠30.48o
= 1∠ − 13.09o p.u.
Hence, δ = -13.09o .

30
Comment upon the results

Although it seems that the angle of source voltage is negative and, therefore, power should flow
from load to source. However, it is not true due to presence of DVR voltages V D1 and V D2 . This
can be seen by looking into the effective or equivalent source voltage i.e.,
0
V s = V s + V D1 = 1∠ − 13.09o + 2.107∠30.48o
= 2.9203∠16.64o p.u.
Similarly,
0
V s1 = V 1 + V D2 = 1∠ − 7.55o + 0.68∠36.87o
= 1.5652∠10.3o p.u.
The positive angles of the effective source voltages indicate that the power flows from source to
the load.

Example 6.5 For the circuit shown in Fig. 5.15, what will be the load voltages without DVR com-
pensation? Analyze whether DVR compensation is possible or not by satisfying the condition |Vl |
= |Vs | = 1.0 p.u. in all three phases. If yes, then find the voltages injected by the DVR in all three
phases. Also, calculate the terminal voltages (vta , vtb , and vtc ), source voltages (vsa , vsb , and vsc )
0 0 0
and effective source voltages (vsa , vsb , and vsc ) after compensation. Solve the above problem for
the following conditions.

Case 1: Vn = VN i.e., for three phase four wire system.

Case 2: Vn 6= VN i.e., for three phase three wire system.

Parameters of the circuit are given as follows:

Source voltage: 230 V rms per phase (1 p.u.), 50 Hz.
Feeder impedance (Zs ): 0.01 + j 0.1 p.u. in all three phases.
Load Impedance: Zla = 0.3 + j 0.4 p.u., Zlb = 1.2 + j 1.3 p.u. and Zlc = ∞.
Solution:

Case 1 Vn = VN

The circuit diagram of the system is shown in Fig. 5.16. Here, both the phases i.e., phase-a
and b behave like 2 individual phases as their neutral points are connected to supply neutral.

First, phase-a is considered. The load voltage V la without compensation is computed as follows.
Zla
V la = V sa
Zla + Zs
0.3 + j 0.4
= 1∠0
0.3 + j 0.4 + 0.01 + j 0.1
= 0.85∠ − 5.07o p.u.

31
 vsa v fa vla
Rs  jX s vta
Zla
isa
vsb v vlb
Rs  jX s vtb fb
N Zlb n
isb
vsc v fc vlc
Rs  jX s vtc
Zlc
isc

Fig. 5.15 A three phase DVR compensated system.

vsa v fa vla
Rs  jX s vta
Zla
isa
vsb v vlb
N Rs  jX s vtb fb n
Zlb
isb
vsc v fc inN
Rs  jX s vtc vlc

isc

Fig. 5.16 A three phase four wire DVR compensated system.

Now, the DVR is operated to maintain load voltage at nominal voltage. Therefore, the load current
I la is given as follows.

V la 1∠0
I la = = = 2.0∠ − 53.13o p.u.
Zla 0.3 + j 0.4

The DVR will inject a voltage with an angle of 90o with respect to the current I la . Therefore,

∠V f a = 90o − 53.13o = 36.87o .

The values of a1a and b1a is computed as follows:

a1a + j b1a = 1.0∠V D2 = 0.8 + j 0.6 p.u.

Above equation implies that a1a = 0.8 and b1a = 0.6. Also, voltage across the feeder (Zsa ) is given
as

Zsa I la = (0.01 + j 0.1) × 2.0∠ − 53.13o = 0.172 + j 0.104 p.u. = a2a + j b2a .

32
From the above equation, a2a = 0.172 and b2a = 0.104. Now, the voltage Vf a is computed by
solving following equation.

Vf2a − 2 {(1 + a2a ) a1a + b1a b2a } Vf a + (1 + a2a )2 + b22a − 1.0 = 0.

Substituting variables, following quadratic equation is used to find out DVR voltage.

Vf2a − 2 Vf a + 0.3844 = 0.

Solving the equation, Vf a = 1.7846 p.u. and 0.2153 p.u. are obtained. However, Vf a = 0.2153 p.u.
is the feasible magnitude. Hence, V f a = 0.2153∠36.87o p.u.

Terminal voltage after compensation:

V ta = −V f a + V la
= −0.2153∠36.87o + 1∠0
= 0.837∠ − 8.87o p.u.

Source voltage after compensation:

V sa = V ta + I la Zs
= 0.837∠ − 8.87o + [(2.0∠ − 53.13o ) × (0.01 + j 0.1])
= 1.0∠ − 1.43o p.u.

Effective source voltage after compensation:

0
V sa = V sa + V f a
= 1.0∠ − 1.43o + 0.2153∠36.87o
= 1.17∠5.1o p.u.

Now, phase-b is considered. The load voltage V lb without compensation is computed as follows.
Zlb
V lb = V sb
Zlb + Zs
1.2 + j 1.3
= 1∠ − 120o
1.2 + j 1.3 + 0.01 + j 0.1
= 0.956∠ − 121.87o p.u.

Here, the load voltage is assumed to be reference for calculation of the DVR voltage. Therefore,
the DVR maintains load voltage at 1∠0 p.u. In this case, the load current I lb is given as follows.

V lb 1∠ − 120
I lb = = = 0.565∠ − 47.29o p.u.
Zlb 1.2 + j 1.3
33
The load voltage V lb without compensation is computed as follows.
V lb = −Zs I lb + V sb
= − [(0.01 + j 0.1) × (2.0∠ − 53.13o )] + 1∠ − 120o
= 0.834∠ − 7.156o p.u.
The DVR will inject a voltage with an angle of 90o with respect to the current I lb . Therefore,
∠V f b = 90o − 47.3o = 42.7o .
The values of a1b and b1b is computed as follows:
a1b + j b1b = 1.0∠V f b = 0.735 + j 0.678 p.u.
Above equation implies that a1b = 0.735 and b1b = 0.678. Also, voltage across the feeder (Zsb ) is
given as
Zsb I lb = (0.01 + j 0.1) × 0.565∠ − 47.3o = 0.0453 + j 0.034 p.u. = a2b + j b2b .
Equating both sides of above equation, a2b = 0.0453 and b2b = 0.034. Now, the voltage Vf b is
computed by solving following equation.
Vf2b − 2 {(1 + a2b ) a1b + b1b b2b } Vf b + (1 + a2b )2 + b22b − 1.0 = 0.
Substituting variables, following quadratic equation is used to find out DVR voltage.
Vf2b − 1.582 Vf b + 0.094 = 0.
Solving the equation, Vf b = 1.53 p.u. and 0.062 p.u. are obtained. However, Vf b = 0.062 p.u. is the
feasible magnitude. Hence, V f b = 0.062∠42.7o p.u. Now with respect to phase, V b = Vb ∠ − 120o
p.u., all phase-b quantities will be shifted by −120o . Therefore phase-b filter voltage,
V f b = 0.062∠42.7o − 120o p.u
= 0.062∠ − 77.3o p.u.
Terminal voltage after compensation:
V tb = −V f b + V lb
= −0.062∠42.7o + 1∠0
= 0.955∠ − 2.52o p.u.
Source voltage after compensation:
V sb = V tb + I lb Zs
= 0.955∠ − 2.52o + [(0.565∠ − 47.29o ) × (0.01 + j 0.1])
= 1.0∠ − 0.448o p.u.
Effective source voltage after compensation:
0
V sb = V sb + V f b
= 1.0∠ − 0.448o + 0.955∠ − 2.52o
= 1.17∠5.1o p.u.

34
In phase-c, no load is connected. Therefore, current is this phase will be zero. Hence, the load
voltage will be same as source voltage i.e., Vlc = 1.0∠120o p.u. Here, voltage drop in feeder is
zero. Thus, DVR in phase-c does not inject any voltage.

Case II Vn 6= VN

vsa v fa vla
Rs  jX s vta
Zla
isa
vsb v vlb
N Rs  jX s vtb fb n
Zlb
isb
vsc v fc vlc
Rs  jX s vtc
isc

Fig. 5.17 A three phase three wire DVR compensated system.

The circuit diagram of the three phase three wire system is shown in Fig. 5.17. In this system,
different phases are dependent on each other. Therefore, it is required to find the voltage VN n . This
voltage is computed as follows.
 
1 V sa V sb V sc
V Nn = 1 + +
Zla
+ Z1lb + Z1lc Zla Zlb Zlc
1∠ − 120o
 
1 1∠0
= 1 1 + +0
0.3+j0.4
+ 1.2+j1.3 + 0 0.3 + j0.4 1.2 + j1.3
= 0.685 − j0.216 p.u. = 0.718∠ − 17.54o p.u.
Now, consider the phase-a. The load voltage V la without compensation is computed as follows.

Zla
V 1a = V sa − V N n
Zla + Zs
0.3 + j 0.4
= [(1∠0) − (0.718∠ − 17.54o )]
0.3 + j 0.4 + 0.01 + j 0.1
o
= 0.325∠29.38 p.u.
The load current in phase-a is given by
V la − V N n 1∠0 − 0.718∠ − 17.54o
I la = = = 0.764∠ − 18.61o p.u.
Zla 0.3 + j 0.4
35
The angle of voltage injected by the DVR in this case will be

∠V f a = 90o − 18.61o = 71.39o .

The values of a1a and b1a is computed as follows:

a1a + j b1a = 1.0∠V f a = 0.319 + j 0.9477 p.u.

Above equation implies that a1a = 0.319 and b1a = 0.9477. Further, voltage across the feeder (Zsa )
in phase-a is given as

Zsa I la = (0.01 + j 0.1) × 0.764∠ − 18.61o = 0.031 + j 0.07 p.u. = a2a + j b2a .

From above equation, a2a = 0.031 and b2a = 0.07. Now, the voltage Vf a is computed by solving
following equation.

Vf2a − 2 {(1 + a2a ) a1a + b1a b2a } Vf a + (1 + a2a )2 + b22a − 1.0 = 0.

Substituting variables, following quadratic equation is obtained.

Vf2a − 0.79 Vf a + 0.68 = 0.

Solving the equation, Vf a = 0.69 p.u. and 0.099 p.u. are obtained. However, Vf a = 0.099 p.u. is
the appropriate solution. Hence, V f a = 0.099∠78.39o p.u.

Terminal voltage after compensation:

V ta = −V f a + V la
= −0.099∠78.39o + 1∠0
= 0.9848∠ − 5.65o p.u.

Source voltage after compensation:

V sa = V ta + I la Zs
= 0.9848∠ − 5.65o + [(0.764∠ − 18.61o ) × (0.01 + j 0.1])
= 1.0∠ − 1.52o p.u.

Effective source voltage after compensation:

0
V sa = V sa + V f a
= 1.0∠ − 1.52o + 0.099∠78.39o
= 1.022∠3.95o p.u.

36
Now, phase-b is considered. The load voltage V lb without compensation is computed as follows.

Zlb
V 1b = V sb − V N n
Zlb + Zs
1.2 + j 1.3
= [(1∠ − 120o ) − (0.718∠ − 17.54o )]
1.2 + j 1.3 + 0.01 + j 0.1
o
= 1.29∠ − 153.13 p.u.
The current drawn by load in this phase will be
V sb − V N n 1∠ − 120o − 0.718∠ − 17.54o
I lb = = = 0.764∠161.42o p.u.
Zlb 1.3 + j 1.2
0
Considering that the phase-b load voltage is reference voltage, i.e., V sb = 1.0∠0. Similarly, load
0
current I lb will be 0.764∠ (161.42o + 120o ) or 0.764∠ − 78.58o p.u. Therefore, the angle of DVR
voltage will be
∠V f b = 90o − 78.58o = 11.42o .
The values of a1b and b1b is computed as follows:
a1b + j b1b = 1.0∠V f b = 0.98 + j 0.198 p.u.
Above equation implies that a1b = 0.98 and b1b = 0.198. Further, voltage across the feeder (Zsb ) is
given as
0
Zsb I lb = (0.01 + j 0.1) × 0.764∠ − 78.58o = 0.0764 + j 0.0076 p.u. = a2b + j b2b .
The above equation implies that a2b = 0.0764 and b2b = 0.0076 . Now, the voltage Vf b is computed
by solving following equation.
Vf2b − 2 {(1 + a2b ) a1b + b1b b2b } Vf b + (1 + a2b )2 + b22b − 1.0 = 0.
Substituting variables, following quadratic equation is used to find out DVR voltage.
Vf2b − 2.113 Vf b + 0.158 = 0.
Solving the equation, Vf b = 2.03 p.u. and 0.0776 p.u. are obtained. However, Vf b = 0.077 p.u. is
the least voltage and is the feasible solution. Hence, V f b = 0.0776∠11.42o p.u.

Terminal voltage after compensation:

V tb = −V f b + V lb
= −0.0776∠11.42o + 1∠0
= 0.924∠ − 0.953o p.u.
Source voltage after compensation:
V sb = V tb + I lb Zs
= 0.924∠ − 0.953o + [(0.764∠ − 78.58o ) × (0.01 + j 0.1])
= 1.0∠ − 0.443o p.u.

37
Effective source voltage after compensation:
0
V sb = V sb + V f b
= 1.0∠ − 0.443o + 0.0776∠11.42o
= 1.076∠0.406o p.u.

In phase-c, the load is open circuited. Therefore, current is this phase will be zero. Hence, the load
voltage will be same as source voltage i.e., Vlc = 1.0∠120o p.u. Consequently, voltage drop in
feeder will be zero, and DVR in phase-c does not inject any voltage.

Problems

P 6.1 Explain the need for voltage compensation. What are different methods to regulate the volt-
age in power distribution system?

P 6.2 Explain the concept of dynamic voltage restorer (DVR). Draw the single line diagram for
DVR compensated system. Discuss it various components.

P 6.3 Discuss the operating principle of DVR. Explain its operation for the following cases.

(a) When line resistance is neglible.

(b) When load is purely resistive.
(c) General case with mention of DVR characteristics

P 6.4 A single phase load is supplied from a supply at given voltage (Vs ) through a feeder impedance,
Zs = Rs + jXs . The load voltage is specified as V t = 1.0∠0o . Draw the circuit diagram for DVR
compendated system and explain mathematically, how would you compute DVR voltage for a
given load.

P 6.5 Explain the transient operation of the DVR for (a) single phase with supply votage with
harmonics (b) three-phase balanced system without harmonics and (c) Three-phase unbalanced
system with harmonics.

P 6.6 Explain how woould you realize DVR circuit with voltage source inverter (VSI) to compen-
sate the estimated voltage. Draw its equivalent circuit. Discuss the switching operaton of the VSI
to synthesize the computed voltage.

P 6.7 Is there any maximum compensation capacity for the DVR without real power support? If
yes, how do you tackle the situation in case you are not able to meet the required compensation?

P 6.8 Consider a system with rms supply voltage of 230 V (1 p.u.), 50 Hz as shown in Fig. 5.9.
The feeder and load impedances of the system are 0.1 + j 0.2 p.u. and 1 + j 0.6 p.u. respectively.
For this system

38
(a) Compute load current and load voltage when DVR is not in operation.
(b) Compute load current and DVR injected voltage such that |Vl | = |Vs |.
(c) Compute load current and DVR injected voltage such that |Vl | = |Vs |, whereas load is replace
by an inductive load of j 2 p.u.

[Ans.: (a) I l = 0.735∠−36.027o A, V l = 0.857∠−5.063o V,(b) I l = 0.857∠−30.96o A, V Dv = 0.3539∠59.036o V,

(c) I l = 0.5∠−90o A, V Dv = 0.1013∠0o V]

P 6.9 For the system given in Fig. 5.9, the feeder and load impedances are 0.1 + j 0.3 p.u. and
0.8 + j 0.9 p.u. respectively. The DVR injects a voltage such that load voltage magnitude is
maintained at 1 p.u. For this compensation system, compute rating of DVR for following cases.
(a) Source voltage magnitude is maintained at 1 p.u. i.e., |Vs | = 1 p.u.
(b) During voltage sag of 20% i.e., |Vs | = 0.8 p.u.
(c) During voltage swell of 20% i.e., |Vs | = 1.2 p.u.

[Ans.: (a) S Dv = 0.2759 p.u., V Dv = 0.3322∠41.63o V,(b) S Dv = 0.5907 p.u., V Dv = 0.7113∠41.63o V, (c) S Dv =
0.0479 p.u., V Dv = 0.0577∠41.63o V]

P 6.10 As shown in Fig. 5.13, the source is supplying to a nonlinear load through a feeder impedance
of 0.1 + j 0.5 p.u. The load current, il , is represented by the following expression.
il = 1 sin(ω t − 30o ) + 0.3 sin(3 ω t − 90o ) p.u.
(a) Compute load voltage when DVR is not operated i.e., vf = 0.
(b) If possible, compute DVR injected voltage such that |Vl | = |Vs | = 1 p.u. If not possible, compute
maximum possible voltage that can be maintained at load terminal by the DVR without taking
any real power from the dc bus?
(c) Compute voltage injected by DVR if there is voltage sag of 20% i.e., |Vs | = 0.8 p.u.
(d) Compute voltage injected by DVR if there is a voltage swell of 20% i.e., |Vs | = 1.2 p.u.

[Ans.: (a) V l = 0.6748∠−5.7193o V,(b) V Dv = 0.4129∠30o V, (c) V Dv = 0.662∠30o V, (d) V Dv = 0.1489∠30o V]

P 6.11 A three phase balanced system is shown in Fig. 5.15. The source voltages are balanced
sinusoids with rms value of 230 V (1.0 p.u.). The feeder and load impedances in each phase are
0.01 + j 0.1 and 3 + j 1 p.u. respectively. Compute the currents and DVR injected voltages in all
three phases such that |Vl | = |Vs | = 1.0 p.u. in all three phases.

[Ans.: I la = 0.316∠−18.43o A, V Dva = 0.0413∠71.56o V, I lb = 0.316∠−138.43o A, V Dvb = 0.0413∠−48.44o V, I lc =

0.316∠101.57o A, V Dvc = 0.0413∠11.56o V]

39
P 6.12 A three phase four wire balanced system with a feeder impedance of 0.01 + j 0.1 p.u. in
all three phases is supplying to a star connected unbalanced load of Ra + j Xa = 0.6 + j 0.3 p.u.,
Rb + j Xb = 0.4 + j 0.4 p.u. and Rc + j Xc = 0.2 + j 0.5 p.u. Compute the voltages injected by
DVR in terms of magnitude and phase angle in all three phases to maintain |Vl | = |Vs | = 1.0 p.u.

[Ans.: V Dva = 0.1802∠63.43o V, V Dvb = 0.1949∠−75o V, V Dvc = 0.1933∠141.8o V]

P 6.13 A three phase four wire balanced system with feeder impedances of Rsa + j Xsa = 0.01 +
j 0.1 p.u., Rsb + j Xsb = 0.005 + j 0.12 p.u. and Rsc + j Xsc = 0.008 + j 0.09 p.u. is supplying to
a star connected unbalanced load of Ra + j Xa = 0.6 + j 0.3 p.u., Rb + j Xb = 0.4 + j 0.4 p.u. and
Rc + j Xc = 0.2 + j 0.5 p.u. Compute magnitude and phase angle of voltages injected by DVR in
all three phases to maintain |Vl | = |Vs | = 1.0 p.u.

[Ans.: V Dva = 0.1802∠63.43o V, V Dvb = 0.2211∠−75o V, V Dvc = 0.1732∠141.8o V]

P 6.14 A three phase four wire balanced source, 1.0 p.u. rms per phase, is feeding to three single
phase full wave rectifiers. The three phase load currents are given as follows.

ila = 1.0 sin(ω t − 30o ) + 0.3 sin(3 ω t − 90o ) p.u.

ilb = 1.0 sin(ω t − 150o ) + 0.3 sin(3 ω t − 90o ) p.u.
ilc = 1.0 sin(ω t − 270o ) + 0.3 sin(3 ω t − 90o ) p.u.

The feeder impedances in all three phases are 0.05 + j0.2 p.u.

(a) Compute load voltage without DVR compensation.

(b) If possible, compute DVR injected voltage such that |Vl | = |Vs | = 1.0 p.u. in three respective
phases. If not possible, compute maximum possible voltage that can be maintained at load
terminal with the DVR without taking any real power from the dc bus?

(c) Compute voltage injected by DVR if there is balanced voltage sag of 20% i.e., |Vs | = 0.8 p.u.

(d) Compute voltage injected by DVR if there is balanced voltage swell of 20% i.e., |Vs | = 1.2 p.u.

[Ans.: (a) V la = 0.9047∠−6.65o V, V lb = 0.9047∠−126.65o V, V lc = 0.9047∠113.35o V (b) V Dv1a = 0.2381∠60o V, V Dv1b =

0.2381∠−60o V, V Dv1c = 0.2381∠180o V, V Dv3abc = 0.1277∠−175o V (c) V Dv = 0.662∠30o V, (d) V Dv = 0.1489∠30o V]

P 6.15 A DVR compensated single phase system is shown in Fig. 5.14. Determine VD1 and VD2
such that the voltage remains 1.0 pu at all three buses. Also, determine the value of δ1 and δ to
ensure power flow from source to load. The value of relevant parameters, all in per unit, are given
as follows.
Zs1 = 0.01 + j 0.2, Zs2 = 0.02 + j 0.5, Zl1 = 0.35 + j 0.6 and Zl2 = 0.2 + j 0.35.

[Ans.: δ1 = −3.33o and δ = −5.916o ]

40
References
[1] A. Ghosh and G. Ledwich, “Compensation of distribution system voltage using dvr,” IEEE
Transactions on Power Delivery, vol. 17, no. 4, pp. 1030–1036, Oct. 2002.
[2] A. Ghosh and G. Ledwich, “Structures and control of a dynamic voltage regulator (dvr),” in
IEEE Power Engineering Society Winter Meeting, vol. 3. IEEE, 2001, pp. 1027–1032.
[3] A. Ghosh, A. Jindal, and A. Joshi, “Design of a capacitor-supported dynamic voltage restorer
(dvr) for unbalanced and distorted loads,” IEEE Transactions on Power Delivery, vol. 19, no. 1,
pp. 405–413, Jan. 2004.
[4] S. Sasitharan and Mahesh K. Mishra, “Constant switching frequency band controller for dy-
namic voltage restorer,” IET Power Electronics, vol. 3, no. 5, pp. 657–667, Sept. 2010.
[5] S. Sasitharan, Mahesh K. Mishra, B. Kalyan Kumar, and V. Jayashankar, “Rating and design
issues of DVR injection transformer,” International Journal of Power Electronics, vol. 2, no. 2,
pp. 143–163, 2010.

41