Chapter 4

CONTROL THEORIES FOR LOAD

COMPENSATION
(Lectures 29-40)

4.1 Introduction

In the previous chapter, we studied the methods of load compensation. These methods can elimi-
nate only the fundamental reactive power and unbalance in the steady state. These kinds of com-
pensators can be realized using passive LC filters and thyristor controlled devices. However, when
harmonics are present in the system, these methods fails to provide correct compensation. To cor-
rect load with unbalance and harmonics, instantaneous load compensation methods are used. The
two important theories in this context are Instantaneous Theory of load compensation often known
as pq theory [1] and Instantaneous Symmetrical Component Theory for load compensation [2].
These theories will be discussed in this chapter. Their merits and demerits and applications will be
explored in detail.

To begin with pq theory, we shall first recall the αβ0 transformation, which was discussed in
Chapter 2. For three-phase system shown in Fig. 4.1, the αβ0 transformations for voltages and
currents are given below.

a a

b b

c c

Fig. 4.1 A three phase system

119
  
v0 r  √1 √1 √1
 
va
2 2 2 2
1 1  
vα  = 1 −√2
−√2 vb (4.1)
vβ 3 3 3 v
0 2
−2 c

 
i0 r  √1 √1 √1
 
ia
2 2 2
1
2
1  
iα  = 1 −
√2
−√2
ib (4.2)
iβ 3 3 3 ic
0 2
−2
The instantaneous active, p(t) and reactive, q(t) powers were defined in Chapter 2 through equa-
tions (2.14)-(2.15) respectively. For the sake of completeness these are given below.
p3φ (t) = va ia + vb ib + vc ic
= vα iα + vβ vβ + v0 i0
= pα + pβ + p0
= pαβ + p0 (4.3)
Where, pαβ = pα + pβ = vα iα + vβ iβ and p0 = v0 i0 .
In the instantaneous reactive power theory, as discussed in Chapter 2, the instantaneous reactive
power, q(t) was defined as,

q(t) = qαβ = vα × iβ + vβ × iα
= vα iβ − vβ iα
1
= − √ [vbc ia + vca ib + vab ic ] (4.4)
3
Therefore, powers po , pαβ and qαβ can be expressed in matrix form as given below.
    
p0 v0 0 0 i0
pαβ  =  0 vα vβ  iα  (4.5)
qαβ 0 −vβ vα iβ
From the above equation, the currents, i0 , iα and iβ are computed as given below.

   −1    2  
i0 v0 0 0 p0 vα + vβ2 0 0 p0
1
iα  =  0 vα vβ  pαβ  =
2 + v2 )
 0 v0 vα −v0 vβ  pαβ  (4.6)
iβ 0 −vβ vα qαβ v0 (vα β 0 v0 vβ v0 vα qαβ
In the above equation,

p0 (vα2 + vβ2 ) p0 v0 i0
i0 = = = = i0 (4.7)
v0 (vα 2 + vβ 2 ) v0 v0

vα −vβ
iα = p + 2
2 αβ
(qαβ )
vα2
+ vβ vα + vβ 2
= iαp + iαq (4.8)

120
 vβ vα
iβ = p + 2
2 αβ
qαβ
vα2
+ vβ vα + vβ 2
= iβp + iβq (4.9)
where
i0 = zero sequence instantaneous current
vα
iαp = α-phase instantaneous active current = 2 +v 2 p
vα β
vβ
iβp = β-phase instantaneous active current = v2 +v 2p
α β
vβ
iαq = α-phase instantaneous reactive current = − v2 +v 2q
α β
iβq = β-phase instantaneous reactive current = v2v+vα
2q
α β

Using above definitions of various components of currents, the three phase instantaneous power
can be expressed as,

p3φ = v0 i0 + vα iα + vβ iβ
= v0 i0 + vα (iαp + iαq ) + vβ (iβp + iβq )
" # " #
vα −vβ vβ vα
= v0 i0 + vα 2 pαβ + 2 qαβ + vβ 2 pαβ + 2 qαβ
vα + vβ2 vα + vβ2 vα + vβ2 vα + vβ2
= v0 i0 + vα iαp + vα iαq + vβ iβp + vβ iβq
= v0 i0 + (pαp + pαq ) + (pβp + pβq )
= v0 i0 + (pαp + pβp ) (4.10)
In the above equation,
pαq + pβq = vα iαq + vβ iβq = 0 (4.11)
If referred to compensator (or filter), the equation (4.6) can be written as,

   2  
if 0 vα + vβ2 0 0 pf 0
1
if α  =
2 2
 0 v0 vα −v0 vβ  pf αβ  (4.12)
if β v 0 (v α + vβ ) 0 v0 vβ v0 vα qf αβ
Since the compensator does not supply any instantaneous real power, therefore,

pf 3φ = pf 0 + pf αβ = 0 (4.13)
The instantaneous zero sequence power exchanges between the load and the compensator and
compensator reactive power must be equal to load reactive power. Therefore we have,
pf o = plo = vo ilo (4.14)
pf αβ = −plo = −vo ilo (4.15)
qf αβ = ql = vα ilβ − vβ ilα (4.16)

121
Since over all real power from the compensator is equal to zero, therefore the following should be
satisfied.
pf o + pf αβ = 0 (4.17)
The power flow description is shown in Fig. 4.2.

 pl


o
plo
P
p l  p lo
ql

Fig. 4.2 Power flow description of three-phase 4-wire compensated system

Also the zero sequence current should be circulated through the compensator, therefore,
if o = ilo (4.18)
Using the conditions of compensator powers as given in the above, the α and β components of
compensator currents can be given as following.
vα −vβ
if α = p
2 f αβ
+ 2 qf αβ
vα2
+ vβ vα + vβ2
1
= 2 [vα (−vo ilo ) − vβ (vα ilβ − vβ ilα )]
vα + vβ2
1
= 2 [−vα vo ilo − vβ vα ilβ + vβ2 ilα )] (4.19)
vα + vβ2
Similarly,

1
if β = [vβ ( pf αβ ) + vα ( qf αβ )]
+ vβ2
vα2
1
= 2 [vβ (−vo ilo ) + vα (vα ilβ − vβ ilα )]
vα + vβ2
1
= 2 [−vo vβ ilo − vα vβ ilα + vα2 ilβ ] (4.20)
vα + vβ2

122
The above equations are derived based on assumption that in general vlo 6= 0 If vo = 0 ,then

if 0 = ilo
1
if α = [v 2 ilα − vα vβ ilβ ] (4.21)
+ vβ2 β
vα2
1
if β = 2 2
[−vα vβ ilα + vα2 ilβ ]
vα + vβ

Once the compensator currents, if o , if α and if β are known, they are transformed back to the abc
frame in order to implement in real time. This transformation is given below.

1  
∗ 
if a r √
2
1 0 if o
√
 i∗f b  = 2  √1 1 3 
 2 − 2 2√  ifα
 (4.22)
∗
if c 3 1 1 3 if β
√
2
−2 − 2

These reference currents are shown in Fig. 4.3. Once reference compensator currents are known,
these are tracked using voltage source inverter (VSI). The other details of the scheme can be as
following.
Thus, compensator powers can be expressed in terms of load powers as following.

v sa i sa ila
LOAD

v sb ilb
N i sb
LOAD n
v sc i sc i lc
LOAD

i*fa i*fc
i*fb
n'

Fig. 4.3 A three-phase four-wire compensated system with ideal compensator

pf o = plo
pf αβ = (pl − plavg ) − plo = pel
pf = pf o + pf αβ = pel

123
The power components and various components of currents are related as following.
   
pα vα iα
=
pβ vβ iβ
 
vα (iαp + iαq )
=
vβ (iβp + iβq )
   
vα iαp vα iαq
= +
vβ iβp vβ iβq
   
pαp p
= + αq (4.23)
pβp pβq
The following quantities are defined.
α - axis instantaneous active power = pαp = vα iαp
α - axis instantaneous reactive power = pαq = vα iαq
β - axis instantaneous active power = pβp = vβ iβp
β - axis instantaneous reactive power = pβq = vβ iβq

It is seen that,

(vα )(vα ) (vβ )(vβ ) vα2 + vβ2

pαp + pβp = vα iαp + vβ iβp = 2 p+ 2 p=( 2 )p = p (4.24)
vα + vβ2 vα + vβ2 vα + vβ2

and

(−vβ )(vα ) (vβ )(vα )

pαq + pβq = vα iαq + vβ iβq = p+ p=0 (4.25)
vα2 + vβ 2 vα2 + vβ 2
Thus, it can be observed that the sum of pαp and pβp is equal to total instantaneous real power p(t)
and the sum of pαq and pβq is equal to zero. Therefore,

p3φ = p + po
= p α + pβ + po
= pαp + pβp + po (4.26)

For an ideal compensator,

pf o = plo = vo ilo = po
pf αβ = −plo (4.27)
qf αβ = ql

For practical compensator, the switching and ohmic losses should be considered. These losses
should be met from the source in order to maintain the dc link voltage constant. Let these losses
are denoted by Ploss , then the following formulation is used to include this term. Let the average

124
power that must be supplied to the compensator be ∆p, then ∆p is given as following.

∆p = po + Ploss (4.28)
Thus, the compensator powers can be expressed as,
pf 0 = pel0
pf αβ = pel − ∆p (4.29)
qf αβ = ql
Once these compensator powers are obtained, the compensator currents if 0 , if α and if β are com-
puted using (4.18), (4.19) and (4.20). Knowing these currents, we can obtain compensator currents
in abc frame using equation (4.22). These currents are realized using voltage source inverter (VSI).
One of the common VSI topology is illustrated in Fig. 4.4. This VSI topology is known as neutral
clamped inverter.

vsa isa PCC ila

LOAD

N vsb isb ilb

LOAD n
vsc isc ilc
LOAD

i1 i fa i fb i fc

Rf Rf
Cdc1 S1 S3 S5
Vdc1 io
Lf Lf
a
n b c
S4 S6 S2
Vdc2
Cdc2
i2

Fig. 4.4 Voltage Source Inverter

While realizing compensator using voltage source inverter, there are switching and other losses
in the inverter circuit. Therefore, a fraction of total power is required to maintain dc capacitor
voltage to a reference value by generating Ploss term. Once reference filter currents (if 0 , if α , if β )
are obtained, the filter currents in abc system are obtained as below.
1  
∗ 
if a r √
2
1 0 if o
√
 i∗f b  = 2  √1 1 3 
 2 −2 2√  if α

∗
if c 3 if β
√1
2
− 21 − 23
As discussed above, the compensator powers are substituted in equation (4.29), the compensator
currents are expressed as below.

125
    2  
if o vα + vβ2 0 0 plo
1
if α  =
2 2
 0 vo vα −v0 vβ  p̃l + p̃0 − ploss 
if β v0 (vα + v β ) 0 v0 vβ v0 vα q lαβ

   2  
if o vα + vβ2 0 0 plo
1
if α  =
2 2
 0 vo vα −v0 vβ  pl + p0 − ploss 
if β v0 (vα + v β ) 0 v0 vβ v0 vα q lαβ

Now once we know i∗f a , i∗f b , i∗f c signals, these have to be synthesized using voltage source
inverter. A typical voltage source inverter (VSI) along with a three-phase compensated system as
shown in Fig. 4.5.

isa ila
vsa L
isb
O
Source vsb isc A
v sc D
isn  0
i fa i fc
Vdc ref Vdc
Voltage source
Voltage regulator
inverter il n
Ploss
S1 S6
ila
ifa i fa
Dymanic i fc
ilb Active filter
controller hysteresis current
ilc
ifc control

vsa vsb vsc

Fig. 4.5 Control algorithm for three-phase compensated system

4.1.1 State Space Modeling of the Compensator

There are different VSI topologies to realize [3]. The most commonly used is neutral clamped
inverter topology as shown in Fig. 4.4. Since this is a three phase four-wire system each phase can
be considered independently. Therefore, to analyze above circuit, only one phase is considered,
which is shown in Fig. 4.6. The other phases work similarly. In Fig. 4.6(a), for switch S1 is closed
and switch S4 is open, the KVL can be written as below.

dif a
Lf + Rf if a + vsa − Vdc1 = 0 (4.30)
dt
From the above equation,
dif a Rf vsa Vdc1
= − if a − + (4.31)
dt Lf Lf Lf
126
Cdc1 + Cdc1 + S1 (OFF)
S1 (ON)
Vdc1 Lf  vsa  Vdc1 Rf Lf  vsa 
Rf i fa i fa
- -

+ +
Vdc2 S4 (OFF) Vdc2 S4 (ON)
Cdc 2 - Cdc 2 -

(a) (b)
Fig. 4.6 Equivalent circuit (a) S1 (ON), S4 (OFF) (b) S1 (OFF), S4 (ON)

Similarly, when S1 is open and S4 is closed as shown in Fig. 4.6 (b),

dif a Rf vsa Vdc2

= − if a − − . (4.32)
dt Lf Lf Lf

The above two equations can be combined into one by using switching signals Sa , S a , as given
below.

dif a Rf Vdc1 Vdc2 vsa

= − if a + Sa − Sa − (4.33)
dt Lf Lf Lf Lf
Similarly, for phases b and c, the first order derivative of filter currents can be written as following.
dif b Rf Vdc1 Vdc2 vsb
= − if b + Sb − Sb − (4.34)
dt Lf Lf Lf Lf
dif c Rf Vdc1 Vdc2 vsc
= − i f c + Sc − Sc − (4.35)
dt Lf Lf Lf Lf

where, Sa = 0 and S a = 1 implies that the top switch is open and bottom switch is closed and,
Sa = 1 and S a = 0 implies that the top switch is closed and bottom switch is open. The two logic
signals Sa and S a are complementary to each other. This logic also holds for the other two phases.
The inverter currents i1 and i2 as shown in Fig. 4.4, can be expressed in terms of filter currents
and switching signals. These are given below.
i1 = Sa if a + Sb if b + Sc if c
i2 = S a if a + S b if b + S c if c (4.36)
The relationship between DC capacitor voltages Vdc1 , Vdc2 and inverter currents i1 and i2 is given
as below.
dVdc1
Cdc1 = −i1
dt
dVdc2
Cdc2 = i2 (4.37)
dt
127
Considering Cdc1 = Cdc2 = Cdc and substituting i1 and i2 from (4.36), the above equations can be
written as,
dVdc1 Sa Sb Sc
= − if a − if b − if c (4.38)
dt Cdc Cdc Cdc
dVdc2 Sa Sb Sc
= if a + if b + if c (4.39)
dt Cdc Cdc Cdc
The equations (4.33), (4.34), (4.35), (4.38) and (4.39) can be represented in state space form as
given below.
 Rf 
Sa
− Lf 0 0 − LS fa    1 
− Lf 0 0
 
if a Lf i fa
R Sb Sb 
− L1f

− Lff − Lf 
 
 if b   0 0 0 0 
  if b    vsa
  
d    Lf
 if c  =  0 R
− Lff LSfc − LSfc   if c  +  0 0 − L1f   vsb  (4.40)
 
dt    0   
Vdc1   Sa  Vdc1    0 0 0  vsc

 − C − SCb − SCc 0 0 
Vdc2 Sa Sb Sc
Vdc2 0 0 0
C C C
0 0

The above equation is in the form,

ẋ = A x + B u. (4.41)

Where, x is a state vector, A is system matrix, B is input matrix and u is input vector. This state
space equation can be solved using MATLAB to implement the compensator for simulation study.

4.1.2 Switching Control of the VSI

In the equation (4.40), the switching signals Sa , S a , Sb , S b , Sc , and S c are generated using a hys-
teresis band current control. This is described as following. The upper and lower bands of the
reference filter current (say phase-a) are formed using hysteresis h i.e., i∗f a + h and i∗f a − h. Then,
following logic is used to generate switching signals.

If if a ≥ (i∗f a + h)
Sa = 0 and S a = 1
else if if a ≤ (i∗f a − h)
Sa = 1 and S a = 0
else if (i∗f a − h) < if a < (i∗f a + h)
Retain the current status of the switches.
end

The first order derivative of state variables can be easily solved using ‘c2d’ (continuous to dis-
crete) command in MATLAB. It is given below.

[Ad Bd ] = c2d(A, B, td ) (4.42)

128
The value of the state vector is updated using the following equation.

x[(k + 1)T ] = Ad x[kT ] + Bd u[kT ] (4.43)

Where x(k + 1) refers the value of the state vector at (k + 1)th sample. The Ad and Bd computed
by ‘c2d0 function as described above can be expressed as below.
The solution of state equation given by (4.41) is given as following [Nagrath].
Zt
x(t) = eA(t−t0 ) x(t0 ) + eA(t−τ ) B u(τ )dτ (4.44)
to

Where to represents initial time and t represents final time. The above equation can be re-written
as following.
Zt
x(t) = φ(t − to ) x(to ) + φ(t − τ ) B u(τ ) dτ (4.45)
to

Writing above equation for small time interval, kT ≤ t ≤ (k +1)T with to = kT and t = (k +1)T ,

(k+1)T
Z
x[(k + 1)T ] = eAT x[kT ] + eA {(k+1)T −τ } B u[kT ] dτ
kT
(k+1)T
Z
= eAT x[kT ] + eA {(k+1)T −τ } B dτ

u[kT ] (4.46)
kT

Comparing (4.43) and (4.46), the discrete matrices Ad and Bd computed by ‘c2d0 MATLAB func-
tion can be written as following.

Ad = eA T
(k+1)T
Z
Bd = eA {(k+1)T −τ } B dτ (4.47)
kT

4.1.3 Generation of Ploss to maintain dc capacitor voltage

The next step is to determine Ploss in order to maintain the dc link voltage close to its reference
value. In compensation, what could be an indication of Ploss to account losses in the inverter. The
average voltage variation of dc link may be an indicator of Ploss in the inverter. If losses are more
than supplied by the inverter, the dc link voltage, i.e., Vdc = Vdc1 + Vdc2 , will decline towards zero
and vice versa. For proper operation of compensator, we need to maintain dc capacitor voltage
to two times of the reference value of each capacitor voltage i.e., Vdc1 + Vdc2 = Vdc = 2 Vdcref .
Thus, we have to replenish losses in inverter and sustain dc capacitor voltage to 2Vdcref with each

129
capacitor voltage to Vdcref . This is achieved with the help of proportional integral (PI) controller
described below [4]. Lets define an error signal as following.
eV dc = 2 Vdcref − (Vdc1 + Vdc2 ) = 2 Vdcref − Vdc

Then, the term Ploss is computed as following.

Z Td
Ploss = Kp eV dc + Ki eV dc dt
0

This control loop need not be very fast and may be updated once in a voltage cycle, preferably at
the positive of phase-a voltage and generate Ploss term at these points. The above controller can be
implemented using digital domain as following.
k
X
Ploss (k) = Kp eVdc (k) + Ki eVdc (j) Td . (4.48)
j=0

In the above equation, k represents the k th sample of error, eVdc . For k = 1, the above equation can
be written as,
1
X
Ploss (1) = Kp eVdc (1) + Ki eVdc (j) Td .
j=0
= Kp eVdc (1) + Ki [eVdc (0) + eVdc (1)] Td . (4.49)
Similarly for k = 2, we can write,
Ploss (2) = Kp eVdc (1) + Ki [eVdc (0) + eVdc (1) + eVdc (2)] Td . (4.50)
Replacing Ki [eVdc (0) + eVdc (1)] from (4.49), we get,
Ki [eVdc (0) + eVdc (1)] Td = Ploss (1) − Kp eVdc (1) (4.51)
Substituting above value in (4.50),we obtain the following.
Ploss (2) = Kp eVdc (2) + Ploss (1) − Kp eVdc (1) + Ki eVdc (2) Td
= Ploss (1) + Kp [eVdc (2) − eVdc (1)] + Ki eVdc (2) Td (4.52)
In general, for k th sample of Ploss ,
Ploss (k) = Ploss (k − 1) + Kp [eVdc (k) − eVdc (k − 1)] + Ki eVdc (k) Td . (4.53)
The algorithm can be used to implement PI controler to generate Ploss . The control action can be
updated at every positive zero crossing of phase-a voltage for example.

4.1.4 Computation of load average power (Plavg )

In reference current expressions, the average load power (Plavg ) is required to be computed. Al-
though low pass filter can be used to find this, however the dynamic response is quite slow. The

130
dynamic performance of computation of Plavg plays significant role in compensation. For this
reason, a moving average algorithm can be used, which is described below.

hpl i = Plavg = hva ila + vb ilb + vc ilc i

1 T
Z
= (va ila + vb ilb + vc ilc ) dt (4.54)
T 0

The above equation can be written with integration from t1 to t1 + T as given in the following.
Z t1+T
1
Plavg = (va ila + vb ilb + vc ilc ) dt (4.55)
T t1

This is known as moving average filter (MAF). Any change in variables instantly reflected with
settling time of one cycle.

4.2 Some Misconception in Reactive Power Theory

The instantaneous reactive power theory, that has evolved from Fortesque, Park and Clarke Trans-
formations of voltage and current specified in phases-a, b and c coordinates [5]. In general, for
3-phase, 4-wire system,
 
vo r  √1 √1 √1
 
va
2 2 2 2
1 1  
vα  = 1 −√2
−√2 vb (4.56)
vβ 3 3 3 v
0 2
−2 c

Similarly, for currents, the αβ0 components are given as following.

 
io r  √1 √1 √1
 
ia
2 2 2 2
1 1  
iα  = 1 −√2
−√2
ib (4.57)
iβ 3 3 3 ic
0 2
−2
√
For balanced system v0 = (va +vb +vc )/ 3 = 0. For three-phase, three-wire system, ia +ib +ic =
0, which implies that i0 = 0. Using these details, the above transformations in equations (4.56)
and (4.57) result to the following.
 
1  va
  r  1
vα 2 1 − √2
−√2  
= vb (4.58)
vβ 3 0 2 − 23 v
3
c

 
1  ia
  r  1
iα 2 1 −
√2
−√2  
= ib (4.59)
iβ 3 0 2 − 23 i
3
c

131
From equation (4.58), and using va + vb + vc = 0, we get the following.
r  
2 vb + vc
vα = va −
3 2
r h
2 vb vc i
= va − −
3 2 2
r h
2 vb  va vb i
= va − − − −
3 2 2 2
r
3
= va (4.60)
2

r "√ √ #
2 3 3
vβ = vb − vc
3 2 2
r "√ √ #
2 3 3
= vb − (−va − vb )
3 2 2
r "√ √ √ #
2 3 3 3
= vb + va + vb
3 2 2 2
1 √
= √ va + 2 vb (4.61)
2
Writing equations (4.60) and (4.61) in matrix form we get,
"q # 
3
 
vα 2
0 va
= √ (4.62)
vβ √1
2 vb
2

Similary, using ia + ib + ic = 0 the following can be written.

"q # 
3
 
iα 2
0 ia
= √ (4.63)
iβ √1
2 ib
2

According to the pq theory, the abc components of voltages and currents are transformed to the α
and β coordinates and the instantaneous powers p and q of the load can be expressed as following.

p = vα iα + vβ iβ (4.64)
q = vα iβ − vβ iα (4.65)

The above equations representing instantaneous active and reactive powers can be expressed in
matrix form as following [1].
    
p vα vβ iα
= (4.66)
q −vβ vα iβ

132
Therefore from above equation (4.66), the α β components of currents can be expressed as follow-
ing.
   −1  
iα vα vβ p
=
iβ −vβ vα q
 −1
vα vβ
The matrix, is given as following.
−vβ vα
 −1  
vα vβ 1 vα −vβ
= 2 (4.67)
−vβ vα vα + vβ2 vβ vα
From the above equation,
vα vβ
iα = p − q (4.68)
vα2 + vβ2 vα2 + vβ2
vβ vα
iβ = 2 2
p+ 2 q (4.69)
vα + vβ vα + vβ2
Which can further be written as,
iα = iαp + iαq (4.70)
iβ = iβp + iβq (4.71)
In the above equation,
vα
iαp = p (4.72)
+ vβ2
vα2
vβ
iαq = − 2 q (4.73)
vα + vβ2
vβ
iβp = 2 p (4.74)
vα + vβ2
vα
iβp = 2 q (4.75)
vα + vβ2
The instantaneous active and reactive components of currents in supplying line can be calculated
from the α and β components of the current as given in the following.
"q #−1   " q #
3 2
  
ia 2
0 iα 3
0 iαp + i αq
= √ =
ib √1 2 iβ − √1 √1 iβp + iβq
2 6 2
"q #  "q # 
2 2
3
0 iαp 3
0 iαq
= +
− √16 √12 iβp − √16 √12 iβq

"q #
2
  
iap 3
0 iαp
= (4.76)
ibp − √16 √1 iβp
2

133
 and,
"q #
2
  
iaq 3
0 iαq
= (4.77)
ibq 1
− 6
√ √1 iβq
2

The active and reactive components of the line currents must be consistent to the basic defini-
tions. However, these components of currents have little in common with the reactive power of the
load as defined in [5]. This is shown in the following illustrations.

Example 4.1 Assume a resistive load connected as shown in the√4.7 below. It is supplied from
a symmetrical source of a sinusoidal balanced voltage with va = 2V sin ωt, with V = 230 Volts.
Express the voltage and currents for primary and secondary side of the transformer. Express the
active and reactive component of the currents, powers and discuss them.

VA  2300 V
Ia Va

IA
VB Vb
Ib
IB
VC Vc
IC Ic
n

Fig. 4.7 An unbalanced resistive load supplied by three-phase delta-star connected transformer

Solution: With he above given values, the primary side phase voltages with respect to virtual
ground could be expressed as the following.
√ √
vA = 2 V sin ωt = 230 2 sin ωt
√ √
vB = 2 V sin(ωt − 120o ) = 230 2 sin(ωt − 120o ) (4.78)
√ √
vC = 2 V sin(ωt + 120o ) = 230 2 sin(ωt + 120o )
In phasor form,
VA = 230∠0◦ V
VB = 230∠−120◦ V
VC = 230∠120◦ V
Therefor the primary side line-to-line voltage are expressed as following.
√ √
vAB = 2 3 V sin(ωt + 30o )
√ √
vBC = 2 3 V sin(ωt − 90o ) (4.79)
√ √
vCA = 2 3 V sin(ωt + 150o )

134
In phasor form,
V AB = 398.37∠30◦ V
V BC = 398.37∠−90◦ V
V CA = 398.37∠150◦ V
These voltages are transformed to the secondaries and are expressed below.
√ √ √
va = 2 3 V sin ωt = 398.37 2 sin(ωt + 30◦ )
√ √ √
vb = 2 3 V sin(ωt − 120o ) = 398.37 2 sin(ωt − 90◦ ) (4.80)
√ √ √
vc = 2 3 V sin(ωt + 120o ) = 398.37 2 sin(ωt + 150◦ )
In phasor form,
Va = 398.37∠30◦ V
Vb = 398.37∠−90◦ V
Vc = 398.37∠150◦ V
Therefore the currents on the secondary side are given below.
√ √
2 3V
ia = cos ωt
R
ib = 0 (4.81)
ic = 0
Taking V = 230 V and R = 4 Ω, the currents on the secondary side of the transformer are given as
following.
√ √
va 2 3V √ √
ia = = sin ωt = 99.56 2 sin(ωt + 30◦ ) = 2I sin(ωt + 30◦ ) (4.82)
R 4
ib = 0
ic = 0
In phasor form, the above can be expressed as,
I a = 99.59∠30◦ A (4.83)
Ib = 0 A (4.84)
Ic = 0 A (4.85)
This phase-a current ia in the secondary side of transformer is transformed to the primary of the
delta connected winding, therefore the currents on the secondary side of transformer are given as
following.

√
iA = 2I sin(ωt + 30◦ ) (4.86)
√
iB = −iA = − 2I sin(ωt + 30◦ )
iC = 0

135
The above can be written in phasor form as given below.
I A = 99.59∠30o = I∠30o A
I B = −I A = 1 ∠−180o × 99.59∠30o
= 99.59∠ − 150o = I∠ − 150o A (4.87)
IC = 0 A
After, knowing the voltages and currents of the primary side of the transformer, their α and β
components are expressed as below.
  "q 3 # 
vα 0 vA
= 12 √ (4.88)
vβ √ 2 vB
2

Substituting vA and vB from (4.78) in the above equation, we get the following.
  "q 3 # √   √ 
vα 2
0 √ 2 V sin ωt √ 3 V sin ωt
= 1 √ = (4.89)
vβ √ 2 2 V sin(ωt − 120◦ ) − 3 V cos(ωt)
2

And,
  "q 3 #  "q # √
3 ◦

iα 2
0 iA 2
0 √ 2 I sin(ωt + 30 )
= 1 √ = √
iβ √ 2 iB √1 2 − 2 I sin(ωt + 30◦ )
2
√ 2 
3 I sin(ωt + 30◦ )
= (4.90)
−I sin(ωt + 30◦ )
Based on the above the active and reactive powers are computed as below.

p(t) = vα iα + vβ iβ
√ √ √
= 3V sin(ωt) 3I sin(ωt + 30◦ ) − 3V cos(ωt)(−I) sin(ωt + 30◦ ))
"√ #
√ 3 1
= 2 3V I sin(ωt + 30◦ ) sin(ωt) + cos(ωt)
2 2
√
= 3V I [2 sin(ωt + 30) sin(ωt + 30)]
√
3V I 2 sin2 (ωt + 30◦ )
 
=
√
= 3V I [1 − cos 2(ωt + 30◦ )] (4.91)

q(t) = vα iβ − vβ iα
√ √ √
= 3V sin(ωt) {−I sin(ωt + 30◦ )} − (− 3V cos(ωt)) 3I sin(ωt + 30◦ )
" √ #
√ 1 3
= − 3V I 2 sin(ωt + 30◦ ) sin ωt − cos ωt
2 2
√
= − 3V I 2 sin(ωt + 30◦ )(− cos(ωt + 30◦ ))
√
= 3V I sin 2(ωt + 30◦ ) (4.92)

136
Based on above values of p and q powers, the α and β components of active and reactive compo-
nents are given below.
   
iα iαp + iαq
=
iβ iβp + iβq
Where,
vα
iαp = p
vα2 + vβ2
√
3V sin(ωt)
= √ √ p
( 3V sin(ωt))2 + (− 3V cos(ωt))2
1 √
= √ sin(ωt) 3V I(1 − cos 2(ωt + 30◦ ))
3V
= I sin ωt {1 − cos 2(ωt + 30◦ )} (4.93)
Similarly,

vβ
iβp = p
vα2 + vβ2
√
− 3V cos(ωt)
= √ √ p
( 3V I sin(ωt))2 + (− 3V cos(ωt))2
√
− 3V cos(ωt) √
= 3V I(1 − cos 2(ωt + 30◦ ))
3V 2
= −I cos ωt {1 − cos 2(ωt + 30◦ )} (4.94)

−vβ
iαq = q
+ vβ2
vα2
√
−(− 3V cos ωt) √
= 3V I sin 2(ωt + 30◦ )
3V 2
= I cos ωt sin 2(ωt + 30◦ ) (4.95)

vα
iβq = q
vα2 + vβ2
√
3V sin ωt √
= 3V I sin 2(ωt + 30◦ )
3V 2
= I sin ωt sin 2(ωt + 30◦ ) (4.96)
Thus knowing iαp , iαq , iβp and iβq we can determine active and reactive components of currents on
the source side as giben below.
   
iap  −1 iαp
= C
ibp iβp

137
 Where,
"q #−1 "q #
3 2
0 0
[C]−1 = 2
√ = −1
3
√1 2 √ √1
2 6 2

Using above equation, we can find out the active and reactive components of the current, as given
below.

"q #
2
I sin ωt {1 − cos 2(ωt + 30◦ )}
  
iap 3
0
=
ibp −1
√ √1 −I cos ωt {1 − cos 2(ωt + 30◦ )}
6 2

From the above,

r
2
iap = I sin ωt {1 − cos 2(ωt + 30◦ )}
3
r
2I
= {2 sin ωt − 2 sin ωt cos 2(ωt + 30◦ )}
32
I
= √ {2 sin ωt − sin(3ωt + 60◦ ) − sin(−ωt − 60◦ )}
6
I
= √ {2 sin ωt + sin(ωt + 60◦ ) − sin(3ωt + 60◦ )}
6

1 1
ibp = − √ iαp + √ iβp
6 2
1 1
= − √ I sin ωt {1 − cos 2(ωt + 30◦ )} + √ (−I cos ωt) {1 − cos 2(ωt + 30◦ )}
6 2
 
◦ sin ωt cos ωt
= −I {1 − cos 2(ωt + 30 )} √ + √
6 2
" √ #
2I 1 3
= − √ {1 − cos 2(ωt + 30◦ )} sin ωt + cos ωt
6 2 2
2I
= − √ {1 − cos 2(ωt + 30◦ )} sin(ωt + 60◦ )
6
2I
= − √ {sin(ωt + 60◦ ) − sin(ωt + 60◦ ) cos 2(ωt + 30◦ )}
6
( √ )
2I 1 3 1 1
= −√ sin(ωt) + cos(ωt) + sin(ωt) − sin(3ωt + 120◦ )
6 2 2 2 2
I
= − √ {sin(ωt) + 2 sin(ωt + 60◦ ) − sin(3ωt + 120◦ )}
6
138
 r
2
iaq = iαq
3
r
2I
= {2 sin 2(ωt + 30◦ ) cos ωt}
32
I
= √ {sin(ωt + 60◦ ) + sin(3ωt + 60◦ )}
6

1 1
ibq = − √ iαq + √ iβq
6 2
1 1
= − √ I cos ωt {sin 2(ωt + 30◦ )} + √ (I sin ωt) {sin 2(ωt + 30◦ )}
6 2
( √ )
I 1 3
= − √ 2 sin 2(ωt + 30◦ ) − cos ωt + sin ωt)
6 2 2
I
= √ 2 sin 2(ωt + 30◦ ) sin(ωt − 30◦ )
6
I
= √ {cos(ωt + 90◦ ) − cos(3ωt + 30◦ )}
6
I
= √ {− sin(ωt) − cos(3ωt + 30◦ )}
6
Thus we have,
I
iap = √ {2 sin ωt + sin(ωt + 60◦ ) − sin(3ωt + 60◦ )} (4.97)
6
I
ibp = − √ {sin(ωt) + 2 sin(ωt + 60◦ ) − sin(3ωt + 120◦ )} (4.98)
6
I
iaq = √ {sin(ωt + 60◦ ) + sin(3ωt + 60◦ )} (4.99)
6
I
ibq = √ {− sin(ωt) − cos(3ωt + 30◦ )} . (4.100)
6
From the above equations, the names instantaneous active current and instantaneous reactive cur-
rent given in the pq theory do not have commonality with the notion of active and reactive currents
used in electrical engineering. Also, the reactive current iq occurs in supply lines of load in spite
of the absence of the reactive element of the in the load. Furthermore, the nature of load is linear
and harmonics are absent, still resolutions of active and reactive components of the current based
on pq theory gives harmonics. For example, in the above discussion,
I
iap = √ {2 sin ωt + sin(ωt + 60◦ ) − sin(3ωt + 60◦ )} (4.101)
6
is the active current component in the line ‘a’ and it contains the third order harmonic. This
contradicts the basic notion of the active current that was introduced to electrical engineering by

139
Fryze [5]. Thus, it seems a major misconception of electrical phenomenon in three phase circuits
with balanced sinusoidal voltages for linear load that do not have harmonics. Moreover, the active
current ip that results from pq theory is not the current that should remain in the supply lines after
the load is compensated to unity power factor as defined by Fryze. Therefore it can not be a com-
pensation goal.

Also, it is evident that the instantaneous reactive power q(t) as defined by pq theory does not
really identify the power properties of load instantaneously. For example for the above discussion,
the active and reactive power are given as following.
√
p(t) = 3V I {1 − cos 2(ωt + 30◦ )}
√
q(t) = 3V I sin 2(ωt + 30◦ )
The following points are noted.
1. The active components of currents (iap , ibp , icp ) and reactive components of currents, (iaq , ibq , icq )
contain third harmonic, which is not possible for a linear load as discussed above.
2. The sum of reactive components of currents, iaq and ibq is not equal to zero, i.e., iaq +ibq 6= 0),
even though no overall reactive power is required from the load.
3. The instantaneous reactive power q(t) defined by pq theory does not identify the power prop-
erties of the load instantaneously. Both powers p(t) and q(t) are time varying quantities, so
that a pair of their values at any single point of time does not identify the power properties
of load. The possibility of instantaneous identification of active and reactive power p(t) and
q(t) does not mean that power properties of load are identified instantaneously. For example,

 √ 
◦ ◦ p(t) = 2 3V I
At (ωτ + 30 ) = 90 ,
q(t) = 0
The above implies that as if it is resistive load.

 
◦ ◦ p(t) = 0
Similarly at (ωτ + 30 ) = 0 ,
q(t) = 0

Which implies as there is no load.

( √ q )
p(t) = 3V I (1 + 32 )
And when (ωτ + 30◦ ) = 105◦ , √
q(t) = − 3 V I( 21 )

implies as it is capacitive load.

( √ q
3
)
p(t) = 3V I (1 + )
Similarly when (ωτ + 30◦ ) = 75◦ , √ 2
q(t) = 3 V I( 12 )

implies as if the load is inductive.

140
 We therefore conclude that power properties cannot be identified without monitoring of the p(t)
and q(t) powers over the entire cycle period. For example, in above case, the instantaneous reactive
power q(t) has occurred is not because of load reactive power Q but because of voltage unbalance.
This unbalance nature of load can not be identified by instantaneous reactive power q(t) values.
Therefore, pq theory gives no advantage with respect to the time interval needed to identify the
nature of load and its property over the the over power theories based on time domain or frequency
domain approach that required the system to be monitored over one period.

Thus, we have seen that each phase has some reactive power. But there is no reactive element.
This reactive power appear because of unbalance in the system and not because of reactive compo-
nent. So this is an additional information what is required. From this illustration, it is evident that
the instantaneous reactive power current has commonality with the load reactive power Q. It also
appears that the instantaneous active current in pq theory (iap , ibp , icp ) have no commonality with
the load active power P .

Powers computation

The secondary side powers are given as following.

∗
S a = Pa + jQa = V a I a = 398.37∠30◦ 99.59∠ − 30◦ = 39675 VA
Thus, Pa = 39675 W, Qa = 0 VAr
∗
S b = Pb + jQb = V b I b = 398.37∠ − 90◦ × 0 = 0 VA
Pb = 0 W, Qb = 0 VAr
∗
S c = Pc + jQc = V c I c = 398.37∠150◦ × 0 = 0 VA
Pc = 0 W, Qc = 0 VAr
The total active and reactive powers on the secondary side are given as following.
P = Pa + Pb + Pc = 39675 W
Q = Qa + Qb + Qc = 0 VAr
Svect = Sarith = P = 39675 VA
pfvect = pfarith = P/S = 1.0
The primary side powers are given as following.
∗
S A = PA + jQA = V a I a = 230∠0◦ 99.59∠ − 30◦ = 19837.50 − j11453.16 VA
Thus, PA = 19837.50 W, QA = −11453.160 VAr
∗
S B = PB + jQB = V b I b = 230∠ − 120◦ (−99.59∠30◦ )∗ =
PB = 19837.50 W, QB = −11453.160 VAr
∗
S C = PC + jQC = V C I C = 230∠120◦ × 0
PC = 0 W, QC = 0 VAr

141
The total active and reactive powers on the primary side are given as following.
P = Pa + Pb + Pc = 39675 W
Q = Qa + Qb + Qc = 0 VAr
 
Svect = S A + S B + S C  = P = 39675 VA
     
Sarith = S A  + S B  + S C  = 22906 + 22906 + 0 = 45813 VA
pfvect = P/Svect = 1.0
pfarith = P/S = 39675/45813 = 0.866
Example 4.2: Assume an Inductive load connected as shown in the √ 4.7 below. It is supplied from
a symmetrical source of a sinusoidal balanced voltage with va = 2V sin ωt, with V = 230 Volts.
Express the voltage and currents for primary and secondary side of the transformer. Express the
active and reactive component of the currents, powers and discuss them.

VA  2300 V
Ia Va

IA
VB Vb
IB Ib
VC Vc
IC Ic
n X

Fig. 4.8 An unbalanced reactive load supplied by three-phase delta-star connected transformer

Solution: With the above given values, the primary side phase voltages with respect to virtual
ground could be expressed as the following.
√ √
vA = 2 V sin ωt = 230 2 sin ωt
√ √
vB = 2 V sin(ωt − 120o ) = 230 2 sin(ωt − 120o ) (4.102)
√ √
vC = 2 V sin(ωt + 120o ) = 230 2 sin(ωt + 120o )
Therefore, the primary side line-to-line voltage are expressed as following.
√ √
vAB = 2 3 V sin(ωt + 30◦ )
√ √
vBC = 2 3 V sin(ωt − 90◦ )
√ √
vCA = 2 3 V sin(ωt + 150◦ )
These voltages are transformed to the secondaries and are expressed below.
√ √ √
va = 2 3 V sin ωt = 398.37 2 sin(ωt + 30◦ )
√ √ √
vb = 2 3 V sin(ωt − 120o ) = 398.37 2 sin(ωt − 90◦ )
√ √ √
vc = 2 3 V sin(ωt + 120o ) = 398.37 2 sin(ωt + 150◦ )

142
In phasor form, the above voltages are expressed as below.
√
Va = 3 V ∠30◦
√
Vb = 3 V ∠ − 90◦
√
Vc = 3 V ∠150◦
Therefore, the currents on the secondary side are given below.
√ √
2 3V √
ia = sin(ωt − 60◦ ) = 99.59 2 sin(ωt − 60◦ )
X
ib = 0
ic = 0
In phasor form, the above can be expressed as,
I a = 99.59∠ − 60◦ A.
The above phase-a current (ia ) is transformed to the primary of the delta connected winding. Since
the currents should have 90◦ phase shift with respect to the voltages across the windings as given
by (4.103), therefore the currents on the secondary side of transformer are given as following.
√
iA = = iAB = ia = 2I sin(ωt − 60o )
√
iB = −iA = − 2I sin(ωt − 60◦ )
iC = 0
In phasor form, the above can be expressed as,
I A = I∠ − 60◦ = 99.59∠ − 60◦ A
I B = −I∠ − 60◦ = 99.59∠ − 60◦ A
I C = 0 A.
After, knowing the voltages and currents of the primary side of the transformer, their α and β
components are expressed as below.
  "q 3 # 
vα 0 vA
= 12 √
vβ √ 2 vB
2

Substituting vA and vB from (4.102) in the above equation, we get the following.
  "q 3 # √   √ 
vα 2
0 √ 2 V sin ωt √ 3 V sin ωt
= 1 √ = (4.103)
vβ √ 2 2 V sin(ωt − 120◦ ) − 3 V cos(ωt)
2

And,
  "q 3 #  "q # √
3 ◦

iα 2
0 iA 2
0 √ 2 I sin(ωt − 60 )
= 1 √ = √
iβ √ 2 iB √1 2 − 2 I sin(ωt − 60◦ )
2
√ 2 
3 I sin(ωt − 60◦ )
= (4.104)
−I sin(ωt − 60◦ )

143
Based on the above, the active and reactive powers are computed as below.

p(t) = pαβ = vα iα + vβ iβ
√ √ √
= 3V sin(ωt) 3I sin(ωt − 60◦ ) + (− 3V cos ωt)(−I sin(ωt − 60◦ ))
"√ #
√ 3 1
= 2 3V I sin(ωt − 60◦ ) sin ωt + cos ωt
2 2
√
= 3V I [2 sin(ωt − 60◦ ) cos(ωt − 60◦ )]
√
= 3V I sin 2(ωt − 60◦ ) (4.105)

q(t) = vα iβ − vβ iα
√ √ √
= 3V sin ωt {−I sin(ωt − 60◦ )} − (− 3V cos ωt) 3I sin(ωt − 60◦ )
" √ #
√ 1 3
= − 3V I 2 sin(ωt − 60◦ ) sin ωt − cos ωt
2 2
√
= − 3V I 2 sin(ωt − 60◦ )(sin(ωt − 60◦ ))
√
= − 3V I 2 sin2 (ωt − 60◦ )
√
= − 3V I {1 − cos 2(ωt − 60◦ )} (4.106)

Based on above values of p and q powers, the α and β components of active and reactive compo-
nents are given below.
   
iα iαp + iαq
=
iβ iβp + iβq

Where,

vα
iαp = p
vα2 + vβ2
√
3V sin ωt
= √ √ p
( 3V sin ωt)2 + (− 3V cos ωt)2
1 √
= √ sin ωt 3V I(sin 2(ωt − 60◦ ))
3V
= I sin ωt sin 2(ωt − 60◦ ) (4.107)

Similarly,

144
 vβ
iβp = p
vα2 + vβ2
√
− 3V cos ωt
= √ √ p
( 3V I sin ωt)2 + (− 3V cos ωt)2
√
− 3V cos ωt √
= 3V I sin 2(ωt − 60◦ )
3V 2
= −I cos ωt sin 2(ωt − 60◦ ) (4.108)

−vβ
iαq = q
+ vβ2
vα2
√
−(− 3V cos ωt) h √ ◦
i
= − 3V I {1 − cos 2(ωt − 60 )}
3V 2
= −I cos ωt(1 − cos 2(ωt − 60◦ )) (4.109)

vα
iβq = q
vα2 + vβ2
√
3V sin ωt h √ ◦
i
= − 3V I {1 − cos 2(ωt − 60 )}
3V 2
= −I sin ωt {1 − cos 2(ωt − 60◦ )} (4.110)

Thus, knowing iαp , iαq , iβp and iβq , we can determine active and reactive components of currents
on the source side as given below.
   
iap  −1 iαp
= C
ibp iβp

Where,
"q #−1 "q #
3 2
0 0
[C]−1 = 2
√ = −13 (4.111)
√1 2 √ √1
2 6 2

Using above equation, we can find out the active and reactive components of the current, as given
below.

"q #
2
I sin ωt sin 2(ωt − 60◦ )
  
iap 3
0
=
ibp −1
√ √1 −I cos ωt sin2(ωt − 60◦ )
6 2

From the above,

145
 r
2
iap = I sin ωt sin 2(ωt − 60◦ )
3
r
2I
= {2 sin ωt sin 2(ωt − 60◦ )}
32
I
= √ {cos(ωt − 2ωt + 120◦ ) − cos(3ωt − 120◦ )}
6
I
= √ {cos(ωt − 120◦ ) − cos(3ωt − 120◦ )} (4.112)
6

1 1
ibp = − √ iαp + √ iβp
6 2
1 1
= − √ I sin ωt {sin 2(ωt − 60◦ )} + √ (−I cos ωt) {sin 2(ωt − 60◦ )}
6 2
" √ #
I 1 3
= − √ {2 sin 2(ωt − 60◦ )} sin ωt + cos ωt
6 2 2
I
= − √ {2 sin 2(ωt − 60◦ ) sin(ωt + 60◦ )}
6
I
= − √ {cos(ωt − 180◦ ) − cos(3ωt − 60◦ )}
6
I
= √ {cos ωt + cos(3ωt − 60◦ )}
6

r
2
iaq = iαq
3
r
2
= − I {1 − cos 2(ωt − 60◦ )} cos ωt
3
I
= − √ {2 cos ωt − 2 cos ωt cos 2(ωt − 60◦ )}
6
I
= − √ {2 cos ωt − cos(3ωt − 120◦ ) − cos(ωt − 120◦ )}
6
I
= √ {−2 cos ωt + cos(ωt − 120◦ ) + cos(3ωt − 120◦ )} (4.113)
6
146
 1 1
ibq = − √ iαq + √ iβq
6 2
1 1
= − √ [−I cos ωt {1 − cos 2(ωt − 60◦ )}] + √ [−I sin ωt {1 − cos 2(ωt − 60◦ )}]
6 2
( √ )
I 1 3
= − √ 2 {1 − cos 2(ωt − 60◦ )} − cos ωt − sin ωt)
6 2 2
( √ )
I 1 3
= √ 2 {1 − cos 2(ωt − 60◦ )} cos ωt + sin ωt)
6 2 2
I
= √ 2 {1 − cos 2(ωt − 60◦ } cos(ωt − 60◦ )
6
I
= √ {2 cos(ωt − 60◦ ) − 2 cos(ωt − 60◦ ) cos 2(ωt − 60◦ )}
6
I
= √ {2 cos(ωt − 60◦ ) + cos 3 ωt − cos(ωt − 60◦ )} (4.114)
6
Thus we have,
I
iap = √ {cos(ωt − 120◦ ) − cos(3ωt − 120◦ )}
6
I
ibp = √ {cos(ωt) + cos(3ωt − 60◦ )}
6
I
iaq = √ {−2 cos ωt + cos(ωt − 120◦ ) + cos(3ωt − 120◦ )}
6
I
ibq = √ {2 cos(ωt − 60◦ ) + cos 3 ωt − cos(ωt − 60◦ )} .
6
From above equations, it is clear that there exist active components of current which also have third
harmonic component. Also, the reactive components too have third harmonics. This again does
not match with definitions of active and reactive components of currents proposed by Fryze [5].

Powers computation
The secondary side powers are given as following.
∗
S a = Pa + jQa = V a I a = 398.37∠30◦ 99.59∠60◦ = 39673.8361∠90◦ VA
Thus, Pa = 0 W, Qa = 39673.8361 VAr
∗
S b = Pb + jQb = V b I b = 398.37∠ − 90◦ × 0 = 0 VA
Pb = 0 W, Qb = 0 VAr
∗
S c = Pc + jQc = V c I c = 398.37∠150◦ × 0 = 0 VA
Pc = 0 W, Qc = 0 VAr

147
The total active and reactive powers on the secondary side are given as following.
P = P a + Pb + Pc = 0 W
Q = Qa + Qb + Qc = 39673.83 VAr
Svect = Sarith = Q = 39673.83 VA
pfvect = pfarith = P/S = 0
The primary side powers are given as following.
∗
S A = PA + jQA = V a I a = 230∠0◦ 99.59∠60◦ = 11452.85 + j19836.91 VA
Thus, PA = 11452.85 W, QA = 19836.91 VAr

∗
S B = PB + jQB = V b I b = 230∠ − 120◦ (−99.59∠ − 120◦ )∗ =
PB = −11452.85 W, QB = 19836.91 VAr

∗
S C = PC + jQC = V C I C = 230∠120◦ × 0
PC = 0 W, QC = 0 VAr
The total active and reactive powers on the primary side are given as following.
P = P a + Pb + Pc = 0 W
Q = Qa + Qb + Qc = 39673.82 VAr
 
Svect = S A + S B + S C  = 39673.82 VA
     
Sarith = S A  + S B  + S C  = 22906 + 22906 + 0 = 45811.4 VA
pfvect = P/Svect = 0
pfarith = P/S = 0
Example 4.3: Consider the star-delta connected ideal transformer with 1:1 turn ratio as shown in
Fig. 2. The secondary side of transformer, a load of 2 ohms is connected between the phase-a and
b. Compute the following.

(a) Time domain expressions of currents in each phase on both primary and secondary side.

(b) Does the load require reactive power from the source? If any, find its value. Also compute
the reactive power on each phase of either side of transformer.

(c) Also determine active powers on each phase and overall active power on either side of the
transformer.

(d) If you have similar arrangement with balanced load and same output power, comment upon
the rating of line conductors and transformers.

Solution In this example, we have three phase star-delta connected transformer of turns ratio 1:1
with star side connected to three phase balanced voltage source and neutral connected to ground.

148
Thus, in this three phase star delta connected transformer, delta side phase voltages equal the star
side line voltages. The primary side instantaneous phase voltages are given by,
√
vA = 230 2 sin(ωt) = 325.27 sin(ωt)
√
vB = 230 2 sin(ωt − 120◦ ) = 325.27 sin(ωt − 120◦ ).
√ ◦
vC = 230 2 sin(ωt + 120 ) = 325.27 sin(ωt + 120◦ )

Therefore, instantaneous line to line voltages at delta side (secondary) are given by,
√
vab = 230 2 sin(ωt) = 325.27 sin(ωt)
√
vbc = 230 2 sin(ωt − 120◦ ) = 325.27 sin(ωt − 120◦ )
√
vca = 230 2 sin(ωt + 120◦ ) = 325.27 sin(ωt + 120◦ ).

Therefore, instantaneous phase voltages with respect to ground are given as follows.
√ √
va = √ 2
230
3
sin(ωt − 30◦ ) = 132.79 2 sin(ωt − 30◦ )
√ √
vb = 230
√ 2
3
sin(ωt − 150◦ ) = 132.79 2 sin(ωt − 150◦ )
√ √
vc = 230
√ 2
3
sin(ωt + 90◦ ) = 132.79 2 sin(ωt + 90◦ )

(a) On delta side, we have a resistive load of R = 3 Ω connected between terminals a and b. Thus,
expression for instantaneous currents flowing out of terminals a, b and c of the transformer are
given by,
√
vab 230 2 √
ia = = sin ωt = 76.67 2 sin ωt
R 3 √
vab 230 2 √
ib = −ia = − =− sin ωt = −76.67 2 sin ωt
R 3
ic = 0

Therefore, the winding currents on the secondary side are given as below.

√
iab = 76.67 2 sin ωt = 108.41 sin ωt
ibc = 0
ica = 0
These currents are transformed to the primary windings. Thus, the time domain expressions of
these currents are given as below.

149
 √
iA = 76.67 2 sin ωt
iB = 0
iC = 0
(b) Load does not require any reactive power from the source because it is a purely resistive load.
This fact can also be verified by the looking at the expressions for instantaneous phase voltages
and currents on star side.

(c) Similarly, no current (and hence power) is drawn by the load from phases-B and C. Thus,
various powers on the primary side are as follows.

For phase-A,
∗
SA = PA + jQA = V A I A = 230∠0◦ × 76.67∠0◦ = 17634.1 VA
PA = 17634.1 W, QA = 0 VAr
For phase-B,
∗
SB = PB + jQB = V B I B = 230∠ − 120◦ × 0 = 0 VA
PB = 0 W, QB = 0 VAr
For phase-C,
∗
SC = PC + jQC = V C I C = 230∠120◦ × 0 = 0 VA
PC = 0 W, QC = 0 VAr
Thus, various powers on the secondary side are as follows.

For phase-a,
∗
Sa = Pa + jQa = V a I a = 132.79∠ − 30◦ × 76.67∠0◦ = 8817.05 − j5090.53 VA
Pa = 8817.05 W, Qa = −5090.53 VAr
For phase-b,
∗
Sb = Pb + jQb = V b I b = 132.79∠ − 150◦ × (−76.67∠0◦ ) = 8817.05 + j5090.53 VA
Pb = 8817.05 W, Qb = 5090.53 VAr
For phase-c,
∗
Sc = Pc + jQc = V c I c = 132.79∠90◦ × 0 = 0 VA
Pc = 0 W, Qc = 0 VAr
For the above analysis, it is observed that the total active power, P = PA + PB + PC = Pa +
Pb + Pc = 17634.1 W and total reactive power Q = QA + QB + QC = Qa + Qb + Qc = 0 VAr.
However, due to unbalanced load, on delta side of the transformer phase-b and phase-c experience
reactive power as calculated above. This creates some power factor in each phase.

(d) Because, the load which is currently getting power from one phase on delta side will be shared

150
by all the three phases phase voltages being the same. Thus, required current rating of the line
conductors and transformers will be reduced.

Example 4.4
Consider the a three-phase balanced system shown in Fig. 4.9. The supply voltages are: V a =
220∠0◦ , V b = 220∠ − 120◦ , V c = 220∠120◦ with balance load resistance of 1.32 Ω. The output
power Po is 100 kW and the losses in the feeder are 5% of te output power. The reactance of feeder
is not considered in the study.
(a) Determine voltages at the load points, phase currents and feeder resistance.
(b) Now let us say that the load is unbalanced by connecting a resistive load between phases-a and
b with same power output. With this configuration, compute active, reactive, various apparent
powers and power factor based on them. Compute losses in the system and comment upon the
result.
(c) Now, with same losses and power output, find an equivalent three-phase balanced circuit and
repeat (b). Comment on the result.

Va '  2300o
r Ia Va Ra

Vb ' r Ic Vb Rb
N

Vc ' r Ib Vc Rc

Fig. 4.9 Study on three-phase unbalanced system

Solution:
(a) For the given system the phase voltages are V a = 220∠0◦ , V b = 220∠−120◦ , V c = 220∠120◦ .
The load active power is 100 kW with 5% losses in the feeder, i.e., ∆Ps = 5 kW. Using above
parameters the RMS value of the phase voltage is computed as following.

V2
Po = 100 × 1000 = 3 ×
1.32
From the above equation, value of V is given as following.
r
100000 × 1.32
V = = 209.76 V
3

Thus, the rms value of phase currents is given by,

V 209.76
I= = = 159.9 A
R 1.32

151
The voltage drop across the feeder is found using following equation.
∆V = 230 − 209.76 = 20.23 V

For 5% losses in the feeder, the value of feeder resistance can be computed as below.
3 × (158.9)2 × r = 0.05 × 100 × 1000
which implies, r = 0.066 Ω
Since the load is balanced, the various apparent powers (arithmetic, vector and effective) are same.
These are computed below.
SA = Sa + Sb + Sc
= Va Ia + Vb Ib + Vc Ic
= 3 Va Ia
= 3 × 209.5 × 158.9
= 100 kVA
Now we compute the vector apparent power (Sv ) as given below.
p
Sv = P 2 + Q2
p
= (Va Ia cos φa + Vb Ib cos φb + Vc Ic cos φc )2 + (Va Ia sin φa + Vb Ib sin φb + Vc Ic sin φc )2
p
= (3 Va Ia cos φa )2 + (3 Va Ia sin φa )2
= 3 Va Ia
= 100 kVA
Similarly, the effective apparent power can be computed as below.
Se = 3 Ve Ie = 3Va Ia = 100 kVA
The power factors based on above apparent powers are given below.
pfA = pfv = pfe = 1.0
Now an unbalanced circuit with same power output is considered. This is achieved by placing a
resistance of 1.17 Ω between any two phases (say between phases a and b). This is shown in Fig.
4.10.
For this circuit, the line currents are computed as below. Let I be the RMS value of the line
current, then following equation must be satisfied.
I 2 × R = 100 × 1000
(4.115)
From the above equation, I is given as below.
r
100000
I = = 292.35 A
1.17
152
 a' a
r Ia
Va R
' b
r Ib
Vb c
c'
r Ic
Vc
'
n
r n
In

Fig. 4.10 Three-phase unbalanved circuit

Thus, phase currents are as following.

I a = 292.35∠30◦ A
I b = −I a = −292.35∠30◦ A
Ic = 0 A
Knowing the currents in the circuit we can compute the voltages V a , V b and V c as following.
Va = 220∠0◦ − 292.25∠30◦ × 0.066
= 220 − (16.67 + j9.9296)
= 223.33 − j9.6296
= 203.55∠ − 2.70◦ V
Similarly, voltages V b and V c can be computed which are given below.

V b = 220∠ − 120◦ − (−292.25∠30◦ ) × 0.066

= 203.55∠ − 117.29◦ V

V c = 220∠120◦ − 0 × 0.066 = 220∠120◦

Thus, Va = Vb = 203.55 V and Vc = 0 V. Knowing three-phase voltages and currents, the active
and reactive powers are computed as following.
∗
Sa = V a I a = 203.55∠ − 2.70◦ × 292.35∠ − 30◦
= 50088.79 − j 32156.42
= P + jQa
pa
Thus, Sa = Pa2 + Q2a = 59522.46 VA

∗
S b = V b I b = 203.55∠ − 117.3◦ × 292.35∠150◦
= 50088.79 + j 32156.42
= Pb + jQb
q
Thus, Sb = Pb2 + Q2b = 59522.46 VA

153
 ∗
Sc = V c I c = 220∠120◦ × 0
= 0+j0
= Pc + jQc
p
Thus, Sc = Pc2 + Q2c = 0 VA

Based on the computations of active and reactive powers in the above, we shall compute arithmetic,
vector and effective apparent powers and corresponding power factors.

SA = Sa + Sb + Sc = 59522.46 + 59522.46 + 0 = 119094.92 VA

Sv = |S v | = |S a + S b + S c | = 100177.58 VA

The effective apparent power is computed as following.

Se = 3 × Ve × Ie
r r
Va2 + Vb2 + Vc2 Ia2 + Ib2 + Ic2
= 3× ×
q 3 q 3
= Va2 + Vb2 + Vc2 ×
Ia2 + Ib2 + Ic2
√ √
= 203.62 + 203.62 + 2202 × 292.352 + 292.352 + 02
= 149816.05 VA

Based on above apparent powers, the power factor are as following.

pfA = 100000/119044.92 = 0.84

pfv = 1000000/100000 = 1.0
pfe = 100000/149816.05 = 0.667

The power loss in the feeders is computed as below.

Ploss = Ia2 r + Ib2 r + Ic2 r

= 292.352 × 0.066 + 292.352 × 0.066 + 0
= 11260 = 11.26 kW

Thus it is observed that when the energy is delivered to the unbalanced load, the power loss in
the supply increases from 5 to 11.26 kW. It means that currents on the lines have increased and this
implies that an unbalanced purely resistive load cannot be considered as unity power factor load.
The calculation of the power factor using the vector apparent power leads to a value which is equal
to unity. This disqualifies, the vector apparent power Sv as an acceptable definition of the apparent
power in the presence of the load unbalance.
Since the value of the apparent power of balanced load is independent of this power definition,
we could ask the question: what is apparent power of a balanced load with the active power P =
100 kW , that causes same power loss of 11.26 kW, as the unbalanced load discussed earlier. For

154
power loss of 11.26 kW the following equation holds true.
3 × I 2 × r = 11260
r
11260
Therefore, I = = 283.01 A
3 × 0.066
Now we use the condition of output power as below.
P0 = 100000 = 3 × I 2 × R
r
100000
Implying that, R = = 0.59 Ω
3 × 238.015
The total input power must be equal to output power + losses. Therefore,
Pi = 11260 = 3 × 220 × I × cos φ
112600
Implying cos φ =
220 × 238.05
= 0.7078 =⇒ φ = 44.94◦
From the above equation,
X
tan φ = tan 44.94◦ = 0.9978 =
R+r
Therefore, X = 0.9978 × (0.59 + 0.0659) = 0.65 Ω.
Therefore, load impedance (Z) is given as below.
Z = 0.59 + j0.65 Ω
(4.116)
Knowing above parameters the voltages at the load points can be computed which are given below.
Va = 220∠0◦ − 238.015∠44.94◦ × 0.0659
= 220 − 11.1 − j11.07 = 208.89 − j 11.07
= 209.18∠ − 3.03◦ V
Similarly, Vb = 209.18∠ − 123.03◦ V
Vc = 209.18∠116.97◦ V
The phase currents are as following.
I a = 238.015∠ − 44.94◦ A
I b = 238.015∠ − 164.94◦ A
I c = 238.015∠ − 75.06◦ A
For this equivalent balanced circuit (with same output power and power loss), the three apparent
powers i.e., arithmetic, vector and effective are same and these are as following.
SA = Sv = Se = 3 × Va × Ia = 3 × 209.18 × 238.015 = 149.363 kVA
and P = Pa + Pb + Pc = 100 kW
Thus the power factor based on the above apparent powers will also be same. Therefore,
pfa = pfv = pfe = 100/149.363 = 0.67

155
4.3 Theory of Instantaneous Symmetrical Components
The theory of instantaneous symmetrical components can be used for the purpose of load balanc-
ing, harmonic suppression, and power factor correction [2], [4]. The control algorithms based on
instantaneous symmetrical component theory can practically compensate any kind of unbalance
and harmonics in the load, provided we have a high band width current source to track the filter
reference currents. These algorithms have been derived in this section. For any set of three-phase
instantaneous currents or voltages, the instantaneous symmetrical components are defined by,

    
ia0 1 1 1 ia
ia+  = 1 1 a a2   ib  (4.117)
3 1 a2 a ic
ia−
Similarly for three-phase instantaneous voltages, we have,
    
v a0 1 1 1 va
v a+  = 1 1 a a2   vb  (4.118)
v 3 1 a2 a v
a− c

In the above equations, a is a complex operator and it is given by a = ej 2π/3 and a2 = ej 4π/3 .
It is to be noted that the instantaneous components of currents, ia+ and ia− are complex time
varying quantities also they are complex conjugate of each other. This same is true for v a+ and
v a− quantities. The terms ia0 and v a0 are real quantities, however (-) has been used as upper script
for the sake of uniformity of notation. These instantaneous symmetrical components are used to
formulate equations for load compensation. First a three-phase, four-wire system supplying star
connected load is considered.

4.3.1 Compensating Star Connected Load

A three-phase four wire compensated system is shown in Fig. 4.11. In the figure, three-phase load
currents (ila , ilb and ilc ), can be unbalanced and nonlinear load. The objective in either three or
four-wire system compensation is to provide balanced supply current such that its zero sequence
component is zero. We therefore have,
v sa i sa ila
LOAD

N
v sb
i sb i lb
LOAD n
v sc
i sc i lc
LOAD

Ideal
compensator
i*fa i*fb i*fc

Fig. 4.11 A three-phase four-wire compensated system

156
 isa + isb + isc = 0 (4.119)

Using equations (4.117)-(4.118), instantaneous positive sequence voltage (v a+ ) and current (ia+ )
are computed from instantaneous values of vsa , vsb , vsc and isa , isb , isc respectively. To have a
predefined power factor from the source, the relationship between the angle of v a+ and ia+ is given
as following.

∠v a+ = ∠ia+ + φ+ (4.120)

Where φ+ is desired phase angle between v a+ and ia+ . The above equation is rewritten as
follows.
   
1 2 1 2
∠ [vsa + a vsb + a vsc ] = ∠ [isa + a isb + a isc ] + φ+
3 3
L.H.S = R.H.S

L.H.S of the above equation is expressed as below

" ( √ ! √ ! )#
1 1 3 1 3
L.H.S = ∠ vsa + − + j vsb + − − j vsc
3 2 2 2 2
" ( √ )#
1  vsb vsc  3
= ∠ vsa − − +j (vsb − vsc )
3 2 2 2
√
−1 ( 3/2) (vsb − vsc )
= tan
(vsa − vsb /2 − vsc /2)
K 1
= tan−1 (4.121)
K2
√
Where, K1 = ( 3/2) (vsb − vsc ) and K2 = (vsa − vsb /2 − vsc /2). Similarly R.H.S of the equa-
tion is expanded as below.
" ( √ ! √ ! )#
1 1 3 1 3
R.H.S = ∠ isa + − + j isb + − − j isc + φ+
3 2 2 2 2
" (  √ )#
1 isb isc 3
= ∠ isa − − +j (isb − isc ) + φ+
3 2 2 2
√
−1 ( 3/2) (isb − isc )
= tan + φ+
(isa − isb /2 − isc /2)
K3
= tan−1 + φ+ (4.122)
K4
157
 √
Where, K3 = ( 3/2) (isb − isc ) and K4 = (isa − isb /2 − isc /2). Equating (4.121) and (4.122),
we get the following.
K1 K3
tan−1 = tan−1 + φ+
K2 K4
Taking tangent on both sides, the following is obtained.
   
−1 K1 −1 K3
tan tan = tan tan + φ+
K2 K4
K1 (K3 /K4 ) + tan φ+
Therefore, =
K2 1 − (K3 /K4 ) × tan φ+
K1 K3 + K4 tan φ+
=
K2 K4 − K3 × tan φ+
The above equation implies that,
K1 K4 − K1 K3 tan φ+ − K2 K3 − K2 K4 tan φ+ = 0
Substituting the values of K1 , K2 , K3 , K4 in the above equation, the following expression is ob-
tained.

√  
3 isb isc 3
(vsb − vsc ) isa − − − (vsb − vsc ) (isb − isc ) tan φ+
2 2 2 4
√   
3 vsb vsc   vsb vsc  isb isc
− vsa − − (isb − isc ) − vsa − − isa − − tan φ+ = 0
2 2 2 2 2 2 2
Above equation can be arranged with terms associated with isa , isb and isc . This is given below.
(√ )
3 tan φ+
(vsb − vsc ) + (vsb + vsc − 2 vsa ) isa
2 2
(√ )
3 tan φ+
+ (vsc − vsa ) + (vsc + vsa − 2 vsb ) isb
2 2
(√ )
3 tan φ+
+ (vsa − vsb ) + (vsa + vsb − 2 vsc ) isc = 0
2 2
√
Dividing above equation by 23 , it can be written as follows.
 
tan φ+
(vsb − vsc ) + √ (vsb + vsc − 2 vsa ) isa
3
 
tan φ+
+ (vsc − vsa ) + √ (vsc + vsa − 2 vsb ) isb
3
 
tan φ+
+ (vsa − vsb ) + √ (vsa + vsb − 2 vsc ) isc = 0
3
158
 √
Assume β = tan φ+ / 3, the above equation is further simplified to,
{(vsb − vsc ) + β (vsb + vsc − 2 vsa )} isa
+ {(vsc − vsa ) + β (vsc + vsa − 2 vsb )} isb
+ {(vsa − vsb ) + β (vsa + vsb − 2 vsc )} isc = 0. (4.123)
Adding and subtracting vsa , vsb and vsc in β terms and expressing vs0 = (vsa + vsb + vsc )/3, in
above equation, we get the following.
{(vsb − vsc ) − 3β (vsa − vs0 )} isa
+ {(vsc − vsa ) − 3β (vsb − vs0 )} isb
+ {(vsa − vsb ) − 3β (vsc + vs0 )} isc = 0. (4.124)
The third objective of compensation is that the power supplied from the source (ps ) must be equal
to average load power (Plavg ). Therefore the following holds true.
ps = vsa isa + vsb isb + vsc isc = Plavg (4.125)
The above equation has important implications. For example when supply voltage are balanced,
the above equation is satisfied for balanced source currents. However if supply voltage are unbal-
anced and distorted, above equation gives set of currents which are also not balanced and sinusoidal
in order to supply constant power.

Equations (4.119), (4.125) and (4.124), can be written in matrix form as given below.
 
1 1 1
(vsb − vsc ) + β(vsb + vsc − 2vsa ) (vsc − vsa ) + β(vsc + vsa − 2vsb ) (vsa − vsb ) + β(vsa + vsb − 2vsc )
vsa vsb vsc
   
isa 0
 isb  =  0 
isc Plavg
Which can be further written as,
    
A isabc = Plavg (4.126)
Therefore,
 T  
a ac12 ac13 0
1  c11
isabc = A−1 Plavg =
    
ac21 ac22 ac23   0 
∆A a ac32 ac33 Plavg
c31
  
a ac21 ac31 0
1  c11
= ac12 ac22 ac32   0 
∆A a ac23 ac33 Plavg
c13

Where acij is the cofactor of ith row and j th column of A matrix in (4.126) and ∆A is the deter-
minant of matrix A. Due to the presence of zero elements in first two rows of column vector with

159
power elements, the cofactors in first two columns need not to be computed. These are indicated
by dots in the following matrix.
      
i sa · · a c31 0 a c31
  1  1 
isabc =  isb  = · · ac32   0  = ac32  Plavg
isc ∆ A · · ac33 Plavg ∆ A ac33
The determinant of matrix A is computed as below.
∆A = [(vsc − vsa ) + β(vsc + vsa − 2vsb )]vsc − [(vsa − vsb ) + β(vsa + vsb − 2vsc )]vsb
−[(vsb − vsc ) + β(vsb + vsc − 2vsa )]vsc + [(vsa − vsb ) + β(vsa + vsb − 2vsc )]vsa
+[(vsb − vsc ) + β(vsb + vsc − 2vsa )]vsb − [(vsc − vsa ) + β(vsc + vsa − 2vsb )]vsa

2 2 2
= β[vsc + vsa vsc − 2vsb vsc − vsa vsb − vsb + 2vsb vsc − vsb vsc − vsc + 2vsa vsc
2 2 2
+vsa + vsb vsa − 2vsc vsa + vsb + vsc vsb − 2vsa vsb − vsc vsa − vsa + 2vsb vsa ]
+(vsc − vsa )vsc − (vsa − vsb )vsb − (vsb − vsc )vsc
+(vsa − vsb )vsa + (vsb − vsc )vsb − (vsc − vsa )vsa

The above equation can be further simplified to,

2 2 2 2
∆A = β · 0 + vsc − vsa vsc − vsb vsa + vsb − vsb vsc + vsc + vsa
2 2
−vsa vsb + vsb − vsc vsb − vsa vsc + vsa
2 2 2
= 2vsa + 2vsb + 2vsc − 2vsa vsb − 2vsb vsc − 2vsc vsa
2 2 2 2 2 2
= vsa + vsb − 2vsa vsb + vsb + vsc − 2vsb vsc + vsc + vsa − 2vsc vsa
2 2 2
= (vsa − vsb ) + (vsb − vsc ) + (vsc − vsa )
2 2 2
= (vsab + vsbc + vsca )
2 2 2
Further adding and subtracting, vsa + vsb + vsc in above equation, we get the following.

2 2 2 2 2 2
∆A = 3(vsa + vsb + vsc ) − (vsa + vsb + vsc + 2vsa vsb + 2vsb vsc + 2vsc vsa ) (4.127)
Further,
(vsa + vsb + vsc )2 = vsa
2 2
+ vsb 2
+ vsc + 2vsa vsb + 2vsb vsc + 2vsc vsa (4.128)
Using equations (4.127) and (4.128), we obtain the following.
2
∆A = 3(vsa + v 2 + vsc
2
) − (vsa + vsb + vsc )2
X sb
2 2
= 3 vsj − 9vso
j=a,b,c
" #
X
2 2
= 3 vsj − 3vso (4.129)
j=a,b,c

160
In above equation, the term vso is the instantaneous zero sequence component of the source voltage
and it is given as following.
(vsa + vsb + vsc )
vso = (4.130)
3
The determinant of matrix A, using (4.129), can also be expressed as,
 
2 2 2 (vsa + vsb + vsc )
∆A = 3 vsa + vsb + vsc − 3 vso
3
 2 2 2 2 2

= 3 vsa + vsb + vsc + 3vso − 2 × 3vso
2 2 2 2 2 2
= 3[vsa + vsb + vsc + vso + vso + vso − 2vso (vsa + vsb + vsc )]
2 2 2 2 2 2
= 3[(vsa + vso − 2vsa vso ) + (vsb + vso − 2vsb vso ) + (vsc + vso − 2vsc vso )]
2 2 2 2 2 2
= 3[(vsa − vso ) + (vsb − vso ) + (vsc − vso )]

The cofactors ac31 , ac32 , and ac33 are computed as below.

ac31 = [−(vsc − vsa ) − β(vsc + vsa − 2vsb ) + (vsa − vsb ) + β(vsa + vsb − 2vsc )]
= vsa − vsb − vsc + vsa + β(−vsc − vsa + 2vsb + vsa + vsb − 2vsc )
= (2vsa − vsb − vsc ) + 3β(vsb − vsc )
= (2vsa + vsa − vsa − vsb − vsc ) + 3β(vsb − vsc )
= 3 (vsa − vs0 ) + 3 β(vsb − vsc )

Similarly,

ac32 = [−(vsa − vsb ) − β(vsa + vsb − 2vsc ) + (vsb − vsc ) + β(vsb + vsc − 2vsa )]
= (−vsa + vsb + vsb − vsc ) + β(−vsa − vsb + 2vsc + vsb + vsc − 2vsa )
= (2vsb − vsc − vsa ) + 3β(vsc − vsa )
= 3(vsb − vso ) + 3β(vsc − vsa )

And,

ac33 = [(vsc − vsa ) + β(vsc + vsa − 2vsb ) − (vsb − vsc ) − β(vsb + vsc − 2vsa )]
= (2vsc − vsa − vsb ) + 3β(vsa − vsb )
= 3(vsc − vso ) + 3β(vsa − vsb )

Knowing the value of cofactors, we now have,

   
isa 3(vsa − vs0 ) + 3β(vsb − vsc ) 
1 
 isb  = hP i 3(vsb − vs0 ) + 3β(vsc − vsa ) Plavg
2 2
isc 3 j=a,b,c vsj − 3vso 3(vsc − vs0 ) + 3β(vsa − vsb )
 
(vsa − vso ) + β(vsb − vsc ) 
1 
= hP i (vsb − vso ) + β(vsc − vsa ) Plavg (4.131)
2 2
j=a,b,c vsj − 3vso (vsc − vso ) + β(vsa − vsb )

161
From the above equation, the desired source currents can be written as following.
(vsa − vso ) + β(vsb − vsc )
isa = P 2 2
Plavg (4.132)
j=a,b,c vsj − 3 vso
(vsb − vso ) + β(vsc − vsa )
isb = P 2 2
Plavg (4.133)
j=a,b,c vsj − 3 vso
(vsc − vso ) + β(vsa − vsb )
isc = P 2 2
Plavg (4.134)
j=a,b,c vsj − 3 vso

Applying Kirchoff’s current law at the point of common coupling (PCC), we have,
i∗f a = ila − isa (4.135)
i∗f b = ilb − isb (4.136)
i∗f c = ilc − isc (4.137)
Replacing isa , isb and isc from equations (4.132)-(4.134), we obtain the reference filter currents as
given in the following.
(vsa − vso ) + β(vsb − vsc )
i∗f a = ila − isa = ila − P 2 2
Plavg (4.138)
j=a,b,c vsj − 3 vso
(vsb − vso ) + β(vsc − vsa )
i∗f b = ilb − isb = ilb − P 2 2
Plavg (4.139)
j=a,b,c vsj − 3 vso
(vsc − vso ) + β(vsa − vsb )
i∗f c = ilc − isc = ilc − P 2 2
Plavg (4.140)
j=a,b,c vsj − 3 vso

4.3.2 Compensating Delta Connected Load

The balancing of an unbalanced ∆-connected load is a generic problem and the theory of instan-
taneous symmetrical components can be used to balance the load. The schematic diagram of this
compensated scheme is shown in Fig. 4.12. This compensator is connected between the phases of

vsa isa
i*fab i*fca
vsb isb ilab
Lo
da

ad
Lo

vsc isc ilbc ilca

Load

i*fbc

Fig. 4.12 A compensation for delta connected load

the load. The aim is to generate the three reference current waveforms denoted by i∗f ab , i∗f bc , i∗f ca

162
respectively based on the measurement of system voltages and load currents such that the supply
sees balanced load. As was in previous case, the requirements for compensating source currents
are same. Therefore the following equations are valid.
isa + isb + isc = 0 (4.141)
Applying Kirchoff’s current law at nodes, we can express isa , isb and isc respectively as following.
isa = (ilab − i∗f ab ) − (ilca − i∗f ca )
isb = (ilbc − i∗f bc ) − (ilab − i∗f cb ) (4.142)
isc = (ilca − i∗f ca ) − (ilbc − i∗f bc )
As can be seen from above equations, equation (4.141) is satisfied. Since in ∆ connected load,
zero sequence current cannot flow, therefore
(ilab − i∗f ab ) + (ilbc − i∗f bc ) + (ilca − i∗f ca ) = 0 (4.143)
The source supplies the average load power, Plavg and following equation is satisfied.
vsa isa + vsb isb + vsc isc = Plavg (4.144)
From equation (4.124), the power factor between the source voltages and currents should be met.
Thus we have,
{(vsb − vsc ) − 3β (vsa − vs0 )} isa
+ {(vsc − vsa ) − 3β (vsb − vs0 )} isb
+ {(vsa − vsb ) − 3β (vsc + vs0 )} isc = 0. (4.145)
Replacing isa , isb and isc from (4.142), above equation can be simplified to the following.

(vsb − vsc ) − 3β(vsa − vs0 ) [(ilab − i∗f ab ) − (ilca − i∗f ca )]

+(vsc − vsa ) − 3β(vsb − vsa ) [(ilbc − i∗f bc ) − (ilab − i∗f ab )]
+(vsa − vsb ) − 3β(vsb − vsa ) [(ilca − i∗f ca ) − (ilbc − i∗f bc )] = 0 (4.146)
Simplifying above expression we get
[{(vsb − vsc ) + β(vsb + vsc − 2vsa )} − {(vsc − vsa ) + β(vsc + vsa − 2vsb )}] (ilab − i∗f ab )
| {z }
I
+ [{(vsc − vsa ) + β(vsc + vsa − 2vsb )} − {(vsa − vsb ) + β(vsa − vsb − 2vsc )}] (ilbc − i∗f bc )
| {z }
II
+ [{(vsa − vsb ) + β(vsa − vsb − 2vsc )} − {(vsb − vsc ) + β(vsb + vsc − 2vsa )}] (ilca − i∗f ca )
| {z }
III
(4.147)
The first term I is as follows:
I = {(vsb − vsc − vsc + vsa ) + β(vsb + vsc − 2vsa − vsc − vsa + 2vsb )} (ilab − i∗f ab )
= {(vsa + vsb − 2vsc ) − 3β(vsa − vsb )} (ilab − i∗f ab )
= −3 {(vsc − vs0 ) + β (vsa − vsb )} (ilab − i∗f ab ) (4.148)

163
Similarly, the second and third terms are given as below.
II = {(vsc − vsa − vsa + vsb ) + β(vsc + vsa − 2vsb − vsa − vsb + 2vsc )} (ilbc − i∗f bc )
= {(vsb + vsc − 2vsa ) − 3β(vsb − vsc )} (ilbc − i∗f bc )
= −3 {(vsa − vs0 ) + β (vsb − vsc )} (ilbc − i∗f bc ) (4.149)

III = {(vsa − vsb − vsb + vsc ) + β(vsa + vsb − 2vsc − vsb − vsc + 2vsa )} (ilca − i∗f ca )
= {(vsc + vsa − 2vsb ) − 3β(vsc − vsa )} (ilca − i∗f ca )
= −3 {(vsb − vs0 ) + β (vsc − vsa )} (ilca − i∗f ca ) (4.150)
Summing above three terms and simplifying we get,
{(vsc − vs0 ) + β(vsa − vsb )} (ilab − i∗f ab )
+ {(vsa − vs0 ) + β(vsb − vsc )} (ilbc − i∗f bc )
+ {(vsb − vs0 ) + β(vsc − vsa )} (ilca − i∗f ca ) = 0 (4.151)
The third condition for load compensation ensures that the average load power should be supplied
from the sources. Therefore,
vsa isa + +vsb isb + vsc isc = Plavg (4.152)
The terms isa , isb and isc are substituted from (4.142) in the above equation and the modified
equation is given below.
vsa (ilab − i∗f ab ) − (ilca − i∗f ca ) + vsb (ilbc − i∗f bc ) − (ilab − i∗f ab )
 

+vsc (ilca − i∗f ca ) − (ilbc − i∗f bc ) = 0


(4.153)
(4.154)
The above is simplified to,
(vsa − vsb )(ilab − i∗f ab ) + (vsb − vsc )(ilbc − i∗f bc ) + (vsc − vsa )(ilca − i∗f ca ) = 0 (4.155)
Equations (4.143), (4.151), (4.155) can be written in the matrix form as given below.

ilab − i∗f ab
  
1 1 1
(vsc − vs0 ) + β(vsa − vsb ) (vsa − vs0 ) + β(vsb − vsc ) (vsb − vs0 ) + β(vsc − vsa )  ilbc − i∗f bc 
vsa − vsb vsb − vsc vsc − vsa ilca − i∗f ca
 
0
= 0  (4.156)
Plavg
The above equation can be written in the following form.
ilab − i∗f ab
   
0
[A∆ ]  ilbc − i∗f bc  =  0  (4.157)
ilca − i∗f ca Plavg

164
Therefore,
ilab − i∗f ab
   
0
 ilbc − i∗f bc  = [A∆ ]−1  0  (4.158)
ilca − i∗f ca Plavg
The above equation is solved by finding the determinate of A∆ and the cofactors transpose as given
below.
T 
ilab − i∗f ab
   
ac11 ac12 ac13 0
 ilbc − i∗f bc  = 1 
ac21 ac22 ac23   0 
∗
ilca − if ca |A ∆| a
c31 ac32 ac33 Plavg
  
ac11 ac21 ac31 0
1 
= ac12 ac22 ac32   0  (4.159)
|A∆ | a a a P
c13 c23 c33 lavg

The determinant |A∆ | and cofactors in above equation are calculated below.
|A∆ | = [(vsa − vso ) + β(vsb − vsc )](vsc − vsa ) − [(vsb − vso ) + β(vsc − vsa )](vsb − vsc )
−[(vsc − vso ) + β(vsa − vsb )](vsc − vsa ) + [(vsb − vso ) + β(vsc − vsa )](vsa − vsb )
+[(vsc − vso ) + β(vsa − vsb )](vsb − vsc ) − [(vsa − vso ) + β(vsb − vsc )](vsa − vsb )
Separating all the terms containing β and rearranging the above equation, we get,
|A∆ | = (vsa − vso )(vsc − vsa ) − (vsb − vso )(vsb − vsc ) − (vsc − vso )(vsc − vsa )
+(vsb − vso )(vsa − vsb ) + (vsc − vso )(vsb − vsc ) − (vsa − vso )(vsa − vsb )
+β [(vsb − vsc )(vsc − vsa ) − (vsc − vsa )(vsb − vsc ) − (vsa − vsb )(vsc − vsa )
+(vsc − vsa )(vsa − vsb ) + (vsa − vsb )(vsb − vsc ) − (vsb − vsc )(vsa − vsb )]
It is seen that the terms containing β cancel each other and give zero. Thus, above equation
becomes,
|A∆ | = (vsa − vso )(vsc − vsa − vsa + vsb ) + (vsb − vso )(−vsb + vsc + vsa − vsb )(4.160)
+(vsc − vso )(−vsc + vsa + vsb − vsc ) + β × 0
In the above, using (vsa + vsb + vsc ) = 3vs0 ,
vsc − vsa − vsa + vsb = vsa + vsb + vsc − 3vsa = −3(vsa − vso )
−vsb + vsc + vsa − vsb = vsa + vsb + vsc − 3vsb = −3(vsb − vso )
−vsc + vsa + vsb − vsc = vsa + vsb + vsc − 3vsc = −3(vsc − vso ).
Replacing above term in (4.160), we get the following.
A∆ = −3 [(vsa − vs0 )(vsa − vs0 ) + (vsb − vs0 )(vsb − vs0 ) + (vsc − vs0 )(vsc − vs0 )]
= −3[(vsa − vs0 )2 + (vsb − vs0 )2 + (vsc − vs0 )2 ]
X
= −3 (vsj − vs0 )2 (4.161)
j=a,b,c

165
The above equation can also be written as,
2 2 2 2
|A∆ | = −3 [vsa + vsb + vsc + 3 vs0 − 2 vs0 (vsa + vsb + vsc )]
2 2 2 2 2
= −3 [vsa + vsb + vsc + 3 vs0 − 6 vs0 ]
2 2 2 2
= −3 [vsa + vsb + vsc − 3vs0 ]
!
X
2 2
= −3 vsj − 3vs0 (4.162)
j=a,b,c

Calculation of cofactors of A∆

We need to calculate ac31 , ac32 and ac33 . These are computed as following.

ac31 = [(vsb − vso ) + β(vsc − vsa ) − (vsa − vso ) + β(vsb − vsc )]

= [(vsb − vsa ) − β(vsa + vsb − 2 vsc )]
= [(vsb − vsa ) − β(vsa + vsb + vsc − 3 vsc )]
= [(vsb − vsa ) − β(3 vs0 − 3 vsc )]
= − [vsab − 3 β(vsc − vs0 )]

Similarly,

ac32 = −[(vsb − vso ) + β(vsc − vsa ) − (vsc − vso ) + β(vsa − vsb )]

= −[(vsb − vsc ) + β(vsc + vsb − 2 vsa )]
= −[(vsb − vsc ) + β(vsa + vsb + vsc − 3 vsa )]
= −[(vsb − vsc ) + β(3 vs0 − 3 vsa )]
= − [vsbc − 3β(vsa − vs0 )]

and

ac33 = [(vsa − vso ) + β(vsb − vsc ) − (vsc − vso ) + β(vsa − vsb )]

= [(vsa − vsc ) − β(vsa + vsc − 2 vsb )]
= [(vsa − vsc ) − β(vsa + vsb + vsc − 3 vsb )]
= [(vsa − vsc ) − β(3 vs0 − 3 vsb )]
= − [vsca − 3β(vsb − vs0 )]

Therefore, the solution of the equation is given by,

T 
ilab − i∗f ab
   
ac11 ac12 ac13 0
 ilbc − i∗f bc  = 1 ac21 ac22 ac23   0 
ilca − i∗f ca A∆ ac31 ac32 ac33 Plavg
   
a ac21 ac31 0
1  c11
= ac12 ac22 ac32   0  (4.163)
A∆ a ac23 ac33 Plavg
c13

166
From the above equation and substituting the values of cofactors obtained above, we get the fol-
lowing.
ac31
ilab − i∗f ab = Plavg
|A∆ |
− [vsab − 3β(vsc − vs0 )]
= hP i Plavg
2 2
−3 v
j=a,b,c sj − 3 vs0

[vsab /3 − β(vsc − vs0 )]

= P 2 2
Plavg
j=a,b,c vsj − 3 vs0

From the above equation, the reference compensator current (i∗f ab ) can be given as follows.

vsab /3 − β(vsc − vs0 )

i∗f ab = ilab − P 2 2
Plavg (4.164)
j=a,b,c vsj − 3 vs0

Similarly,
vsbc /3 − β(vsa − vs0 )
i∗f bc = ilbc − P 2 2
Plavg (4.165)
j=a,b,c vsj − 3 vs0

and
vsca /3 − β(vsb − vs0 )
i∗f ca = ilca − P 2 2
Plavg (4.166)
j=a,b,c vsj − 3 vs0

When the source power factor is unity, β = 0, for balanced source voltages (fundamental) vs0 = 0.
Substituting these values in above equations, we get,
vsab
i∗f ab = ilab − P 2
Plavg
3 j=a,b,c vsj
vsbc
i∗f bc = ilbc − P 2
Plavg (4.167)
3 j=a,b,c vsj
vsca
i∗f ca = ilca − P 2
Plavg
3 j=a,b,c vsj
Further it can be seen that,
2 2 2
vsab + vsbc + vsca = (vsa − vsb )2 + (vsb − vsc )2 + (vsc − vsa )2
2 2 2 2 2 2
= vsa + vsb − 2vsa vsb + vsb + vsc − 2vsb vsc + vsc + vsa − 2vsc vsa
2 2 2
= 2(vsa + vsb + vsc ) − (2vsa vsb + 2vsb vsc + 2vsc vsa )
2 2 2 2 2 2
= 3(vsa + vsb + vsc ) − (vsa + vsb + vsc + 2vsa vsb + 2vsb vsc + 2vsc vsa )
= 3(vsa + vsb + vsc ) − (vsa + vsb + vsc )2
2 2 2

2 2 2 2
= 3[(vsa + vsb + vsc ) − 3 vs0 ]
" #
X
2 2
= 3 vsj − 3 vs0 (4.168)
j=a,b,c

167
 hP i
2 2
Replacing 3 v
j=a,b,c sj − 3 v s0 in equations (4.164), (4.165) and (4.166), we get the following.

vsab − 3β (vsc − vs0 )

i∗f ab = ilab − 2 2 2 )
Plavg
(vsab + vsbc + vsca
vsbc − 3β (vsa − vs0 )
i∗f bc = ilbc − 2 2 2 )
Plavg (4.169)
(vsab + vsbc + vsca
vsca − 3β (vsb − vs0 )
i∗f ca = ilca − 2 2 2 )
Plavg
(vsab + vsbc + vsca
For unity power factor and balanced source voltages (fundamental), the reference compensator
currents are given as following.
vsab
i∗f ab = ilab − Plavg
9V 2
vsbc
i∗f bc = ilbc − Plavg (4.170)
9V 2
vsca
i∗f ca = ilca − Plavg
9V 2
Where V is the rms value of phase voltages.

References
[1] H. Akagi, Y. Kanazawa, and A. Nabae, “Instantaneous reactive power compensators compris-
ing switching devices without energy storage components,” IEEE Transactions on Industry
Applications, no. 3, pp. 625–630, 1984.
[2] A. Ghosh and G. Ledwich, “Load compensating DSTATCOM in weak ac systems,” IEEE
Transactions on Power Delivery, vol. 18, no. 4, pp. 1302–1309, Oct. 2003.
[3] V. George and Mahesh K. Mishra, “DSTATCOM topologies for three-phase high power appli-
cations,” International Journal of Power Electronics, vol. 2, no. 2, pp. 107–124, 2010.
[4] A. Ghosh and A. Joshi, “A new approach to load balancing and power factor correction in
power distribution system,” IEEE Transactions on Power Delivery, vol. 15, no. 1, pp. 417–
422, 2000.
[5] L. S. Czarnecki, “Budeanu and Fryze: Two frameworks for interpreting power properties of
circuits with nonsinusoidal voltages and currents,” Electrical Engineering (Archiv fur Elek-
trotechnik), vol. 80, no. 6, pp. 359–367, 1997.

168