A CONTROL STRATEGY FOR SHUNT ACTIVE POWER FILTER UNDER

THE DISTORTED SOURCE VOLTAGES CONDITION

S. H. Hosseini1, Sh. Sarlak2, M. Sarhagzadeh3, R. Yadipour1
1, 2
Center of Excellence for Mechatronics, University of Tabriz, 51666, Tabriz, Iran
e-mail:[email protected], shirinsarlak@ yahoo.com
3
Tabriz Electric Power Distribution Company, Tabriz, Iran

Key words: APF, p-q theory, positive sequence, compensation

ABSTRACT to determine the APF reference compensation currents in

The shunt active power filter (APF) is a useful device to the three-phase three-wire systems. p-q theory is not
eliminate the harmonic currents and to compensate reactive suitable for three-phase four-wire systems in its primary
power. This paper proposes a useful approach to determine form. Peng extended the p-q theory to apply for three-
the reference compensation currents when the source phase four-wire systems by handling the zero-sequence
voltages are severely distorted and imbalanced. The
approach is simulated by PSCAD/EMTDC .The simulation
power compensation [4]. p-q theory has caused many
results show that this strategy can be applied for such works dealing with active filter compensation strategies.
situations with acceptable results. One of them is that APF can be designed without active
energy source units, such as batteries or in other forms in
I. INTRODUCTION its compensation machines. In other words, an ideal APF
Due to the recent advancements device technology that does not consume any average real power supplied by the
makes high-speed high-power switching devices such as source.
power MOSFETs, MCTs, IGBTs, IEGTs and etc. the use Another one is the synchronous reference frame (SRF)
of shunt active power filters (APF) to eliminate harmonic method [5]. SRF con not compensate for reactive power.
currents and to compensate reactive power has attracted Reference [6] proposed an algorithm in the a-b-c
much attention, in the last years [1]. APFs basically work reference frame for maintaining ideal three-phase source
by detecting the harmonic components from the distorted currents when the source voltages are amplitude-
signals and injecting the right amount of counter harmonic imbalanced. This method works very well on harmonic
compensating currents to the coupling point of the load. and/or reactive power compensation for nonlinear loads
Fig.1 shows the schematic diagram of a three-phase four- under ideal source voltages. This theory did not work
wire shunt APF, where APF senses the source voltages under the imbalanced and/or distorted source voltages
and load currents to determine the reference compensation conditions.
currents. The methods mentioned above, do not use the strategies to
apply under the imbalanced and/or distorted source
Va
Ila voltages conditions. To achieve full compensation of both
Ilb Linear/Nonlinear reactive power and harmonic/neutral currents of the load,
Vb
Ilc this paper presents an approach to determine the shunt
Vc
load
APF reference compensation currents, even if the source
Null
voltages and load currents are both imbalanced and
Ifc Ifb Ifa distorted. The proposed method is somehow similar to
that presented in [6].
Active Power This paper, reviews the p-q theory as an example of the
Filter methods that don’t work under the unbalanced and/or
distorted source voltages conditions. Next the proposed
Fig.1.Schematic diagram of a three-phase four-wire shunt active filter.
method is explained analytically and the PSCAD/EMTDC
Many approaches to determine the reference simulation results are followed to compare the proposed
compensation currents are being developed. One of them method with p-q theory.
is the instantaneous reactive power theory (i.e. p-q theory)
introduced by Akagi [2]. This method requires II. p-q THEORY
transmission of both source voltages and load currents This theory introduced by Akagi. In this theory both
from the a-b-c reference frame to the α-β reference frame source voltages and load currents transfer from the a-b-c
reference frame to α-β reference frame. Then the load source currents and the positive-sequence source voltages
voltages and load currents are: at the fundamental frequency must be:
⎡vla ⎤
⎡vlα ⎤ 2 ⎡1 − 1 / 2 − 1 / 2 ⎤ ⎢ ⎥ (1) isa=Is sin(ωt+θp) (11)
⎢v ⎥ = ⎢ ⎥ ⎢vlb ⎥
⎣ lβ ⎦ 3⎣ 0 3 2 − 3 2⎦⎢ ⎥ isa=Is sin(ωt-120+θp) (12)
⎣vlc ⎦ isa=Is sin(ωt+120+θp) (13)
⎡ila ⎤
⎡ilα ⎤ 2 ⎡1 − 1 / 2 − 1 / 2 ⎤ ⎢ ⎥ and
⎢i ⎥ = ilb (2)
⎣ lβ ⎦ 3 ⎢⎣0 3 2 − 3 2⎥⎦ ⎢ ⎥
⎢⎣ilc ⎥⎦
vpfa=Vmp sin(ωt+θp) (14)
vpfb= Vmp sin(ωt-120+θp) (15)
Where vla, vlb, vlc, ila, ilb, ilc load voltages and currents in vpfc= Vmp sin(ωt+120+θp) (16)
the a-b-c reference frame and ilα, ilβ are the load voltages
and currents in α-β reference frame, respectively. Where Vmp and θp are voltage magnitude and phase angle
The active power and reactive power are as follow: of the positive-sequence components at the fundamental
frequency, respectively. The load average real power
p=vlα ilα + vlβ ilβ (3) should be supplied by the source and the APF dos not
provide or consume any average real power. This requires
q=vlα ilβ _ vlβ ilα (4) that Is-current amplitude- expressed as a function of the
sequential instantaneous voltage and real power
Therefore ilα and ilβ are components. According to the symmetrical components
−1 transformation of (17) for the three-phase root-mean-
⎡ilα ⎤ ⎡ vlα v lβ ⎤ ⎡ p ⎤ (5)
⎢i ⎥ = ⎢− v vlα ⎥⎦ ⎢⎣q ⎥⎦
square (rms) voltages at each harmonic order h=1,2,3,…
⎣ lβ ⎦ ⎣ lβ the three phase instantaneous voltages can be expressed
by equations (18) to (20).
we can also write
⎡V zh ⎤ ⎡1 1 1 ⎤ ⎡Vah ⎤
p= p+ ~p (6) ⎢ ⎥ 1⎢
α 2 ⎥⎥ ⎢⎢Vbh ⎥⎥
~ (7) ⎢V ph ⎥ = 3 ⎢1 α (17)
q =q +q ⎢V ⎥ ⎢⎣1 α 2 α ⎥⎦ ⎢⎣Vch ⎥⎦
⎣ nh ⎦
Where p , q and ~ p , q~ are the DC and AC components ∞ ∞ ∞
v a = ∑ v pha + ∑ v nha + ∑ v zha (18)
of instantaneous active power and reactive power, h =1 h =1 h =1
respectively. ∞ ∞ ∞
Similarly, for APF
vb = ∑ v phb + ∑ v nhb + ∑ v zhb (19)
h =1 h =1 h =1
~
⎡i fα ⎤ ⎡ vlα v lβ ⎤ ⎡ p + p c ⎤ ∞ ∞ ∞
⎢ ⎥=⎢ ⎢
vlα ⎥⎦ ⎢ ql ⎥
⎥ (10) v c = ∑ v phc + ∑ v nhc + ∑ v zhc (20)
⎣⎢i fβ ⎦⎥ ⎣− vlβ ⎣ ⎦ h =1 h =1 h =1

In (17), the operator α equals ej2П/3 and p, n and z

where pc is switching losses and ql is the load required
represents the positive-, negative- and zero-sequence
reactive power. ifα and ifβ are the desired reference
components, respectively. In equations (18) to (20), the
compensation currents.
voltage components for h=1,2,3,… are defined as
III. PROPOSED APPROACH FOR
DETERMINING THE REFERENCE vpha=Vmph sin(h(ωt+θph) (21)
COMPENSATION CURRENTS vphb=Vmph sin(h(ωt-120)+θph) (22)
This method is based on the requirement of the source vphc=Vmph sin(h(ωt+120)+θph) (23)
currents need to be balanced, undistorted, and in phase vnha=Vmph sin(h(ωt+θnh)) (24)
with the positive-sequence source voltages. The goals of vnhb=Vmph sin(h(ωt+120)+θnh) (25)
this approach are: harmonic and neutral current vnhc=Vmph sin((ωt-120)+θnh) (26)
compensation, unity power factor at positive-sequence vzha=vzhb=vzhc=Vmzh sin(hωt+θzh) (27)
fundamental frequency and minimum average real power
consumed or supplied by the APF. The approach provides The average real power consumed by the load over the
a full compensation for the non-linear load. Therefore the one period of time T should be supplied by the source and
the APF should not consumes or supplies any average real (32) and (33) must satisfy:
power. Then: p s = p spf = pl (42)
p s = pl + p s
T (28)
1 Then according to (40) and (42), the desired source
T ∫0
ps = (v a i sa +vb i sb + vc i sc )dt current amplitude at each phase is determined as follow
(29)
T
1 2 pl
T ∫0
pl = (v a ila +vb ilb + vc ilc )dt (30)
Is =
3Vmp (43)
pf = 0 (31)
and the source currents are expressed by
p s = pl (32) v pfk 2 pl
i sk = I s = v pfk T K=a,b,c. (44)
Where pl includes the average fundamental and harmonic Vmp 3(Vmp ) 2
real power. Substituting (18 )-( 20) into (29 ) yields The required compensation current that should be injected
to network is obtained by subtracting the desired source
p s = p spf + p snf + p szf + p sph + p snh + p szh (33) current from the load current as given in (45)

Where the positive-, negative-, and zero-sequence average 2 pl

real power components at the fundamental frequency and
i * fk = ilk − v pfk
3(Vmp ) 2 K=a,b,c. (45)
at harmonic frequencies are given in (34) to (39),
respectively.

T The average real power consumed or supplied by the APF

1
pspf = ∫ (v pfaisa + v pfbisb + v pfcisc )dt (34) is expressed as
T0
T (46)
T 1
psnf
1
= ∫ (vnfaisa + vnfbisb + vnfcisc )dt (35) p f = ∫ (va i fa +vb i fb + vc i fc )dt
T0 T0
T
1 By substituting (45) into (46)
pszf = ∫ (v zfaisa + v zfbisb + v zfcisc )dt (36)
T0 T
1
T ∫0
∞
1
T
pf = (v a i fa +vb i fb + v c i fc )dt
psph = ∑ { ∫ (v phaisa + v phbisb + v phcisc )dt} (37)
h =2 T0 T
2 pl 1
3(Vmp ) 2 T ∫0
∞
1
T
− . (v a v pfa +vb v pfb + vc v pfc )dt
psnh = ∑ { ∫ (vnhaisa + vnhbisb + vnhcisc )dt} (38)
h =2 T0
∞
1
T =0 (47)
pszh = ∑ { ∫ (v zhaisa + v zhbisb + vzhcisc )dt} (39)
h=2 T0
Then, the APF does not consume or supply average real
power.
Substituting the desired source currents of (11) to (13) and Fig.2 shows the block diagram of the control circuit for
fundamental positive-sequence voltages of (14) to (16) proposed approach.
into (34), results
Ila +
Vfa Upfa - Ifar
3Vmp I s Va Positive- Ilb +
p spf = -
Low-Pass Vfb Sequence Upfb Ifbr
Vb
(40) Filter Component
2 Vc
Vfc
Calculator
Upfc Ilc
-+ Ifcr
(Vmf)^2

Similarly, it can be shown that all other average real Pa

Ila
power compensation satisfy (41) after substituting (11) to Low- Is
(13) and (21) to (27) into (35) to (39) Ilb
Pb
∑ Pass 2/3
Filter
Pc
Ilc

p snf = p szf = p sph = p snh = p szh = 0 (41)

Fig.2. Block diagram of the control circuit for proposed approach..
The SPWM current controller is used to control the APF
inverter switching.

IV. SIMULATION RESULTS

To verify the performance of the proposed compensation
strategy for the APF, results obtained by using
PSCAD/EMTDC are given. First the results for p-q theory
are given, as an example of the strategies that do not apply
when the source voltages are imbalanced and distorted.
Next the results for the proposed approach are given. Fig.3)d. APF compensation reactive power , load reactive power, and source
supplied reactive power.
In both p-q theory and the proposed approach, the
magnitude imbalance for source voltages is ٪10 and the Fig.3. Simulation results for p-q theory.
distortion with 5th harmonic is ٪7. Also a single phase
breaker, breaks phase 'a' at t=0.8 sec to show that how the As shown in Fig3.)c, this theory can not compensate the
APF follows the reference currents to compensate the source currents when the source voltages are distorted and
harmonic currents and reactive power. imbalanced. A single phase breaker applies at t=0.8 sec.
As shown in Fig2)d,, this theory compensates the reactive
A)simulation results for p-q theory power, instantly.
Fig.3 shows the simulation results for p-q theory.
B)simulation results for proposed approach
Fig.4 shows the simulation results for p-q theory.

Fig.3)a. Imbalanced and distorted source voltages.

Fig.4)a. Imbalanced and distorted source voltages.

Fig.3)b. Distorted load currents.

Fig.4)b. Distorted load current

Fig.3)c. Source currents after compensation for p-q theory. Fig4)c. Source currents after compensation for proposed approach.
 requirement. Therefore the proposed approach yields a
simpler design of the shunt APF controller. The
PSCAD/EMTDC simulation results for both p-q theory
and proposed approach show that this strategy can
compensate harmonic components and reactive power
required by the non-linear load, when the source voltages
are severely distorted.

REFERENCES
[1] L.Gyugyi and E.C. Strycula, "Active AC power filters," in Proc.
IEEE/Ind. Applicant. Soc. Annu. Meeting, 1976, p.529.
[2] H. Akagi, Y.Kanazaw, and A.Nabae, "Generalized theory of the
Fig.4)d. APF compensation reactive power, load reactive power, and source instantaneous reactive power in three-phase circuits," in Proc. IEEJ
supplied reactive power. Int. power electronic conf., Tokyo, Japan, 1983, pp. 1375-1386
[3] F.Z. Peng, H.Akagi, and A.Nabae, "A novel harmonic power filter,"
Fig4. Simulation results for proposed approach. PESC '88, pp. 1151-1158, 1996
[4] F.Z. Peng, George W. Ott, and D.J. Adams, "Harmonic and reactive
As shown in Fig.4)c, this theory compensates the source power compensation based on Generalized theory of the
currents when the source voltages are distorted instantaneous reactive power in three-phase four-wire system," IEEE
Trans. On power electronic, vol. 13, no. 6, pp. 1174-1181, Nov.
imbalanced. A single phase breaker applies on phase 'a' at 1988.
t=.8 sec. As shown in Fig.4)c, this strategy responds [4] S. kim, N. Enjeti, "Control strategies for active power filter in three-
rapidly. Fig.4)d, shows that this theory compensates the phase four-wire systems," IEEE 2000
reactive power instantly. [5] S.Bhattacharya, D.M. Divan, and B. Banerjee, "Synchronous frame
harmonic isolator using active series filter," in Proc 4th Eur. Power
electronics Conf., vol. 3, 1991, pp. 030-035.
V. CONCLUSIONS [6] C.L. Chen, C.E. Lin, and C.L. Hung, "Reactive and harmonic
This paper proposed a simple, useful, and efficient current compensation for unbalanced three-phase systems using the
approach to determine the reference compensation synchronous detection method," Elect. Power Syst. Res., no. 26, pp.
163-170, 1993.
currents to control the shunt active power filter when the [8] Gray W.Chang, senior member IEEE, and Tai-Chang Shee, "A novel
source voltages are imbalanced and/or distorted. In this reference compensation current strategy for shunt active power filter
approach, there is no reference frame transformation control," IEEE Trans. On power delivery, vol. 19, no. 4, pp. 1751-1758,
Oct. 2004.