Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com

Energy
Procedia

Energy Procedia
00 (2011)
Energy Procedia
16000000
(2012) 1753 1759

www.elsevier.com/locate/procedia

2012 International Conference on Future Energy, Environment, and Materials

Recursive Integral with Fuzzy Control Method used in Shunt

Hybrid Active Power Filter
Wei Zhao1, Ruimin Chen1, An Luo2
1

Electric Power Research Institute of Guangdong Power Grid Corporation

2

Guangzhou, China

College of Electrical and Information Engineering,

Hunan University

Changsha, China

Abstract
A current control method for a shunt hybrid active power filter (HAPF) using recursive integral algorithm is
presented in this paper. The method improves performance of the HAPF system by reducing the influence of
detection accuracy, time delay of instruction current calculation and phase displacement of output filter. Fuzzy logic
based set-point weighing algorithm is combined in the control scheme to enhance its robustness and anti-interference
ability. The proposed algorithm is easy for engineering application and has small computation capacity. Experiment
results have verified the validity of the proposed controller. Furthermore the proposed recursive integral PI algorithm
can also be applied in the control of periodic current as in AC drivers.

by Elsevier
B.V. Selection
and/or peer-review
under responsibility
of InternationalofMaterials
Science Society.
2011
2011Published
Published
by Elsevier
Ltd. Selection
and/or peer-review
under responsibility
[name organizer]
Keywords: harmonic suppression; reactive power compensation; recursive integral PI; fuzzy inference

1. Introduction
With the wide use of power electronics-related appliances in industry, harmonics problem is more and
more serious. Harmonics not only influence the normal work state of electric equipments, but also
interfere communication equipments. At present, passive power filter (PPF), active power filter (APF)
and hybrid active power filter (HAPF) are mostly used to eliminate the harmonics [1][3].
In order to control the inverter to produce a current which is equal on the amplitude and opposite in
direction with the load harmonic current, a current control method is required. But due to the influence of
detection accuracy, time delay of instruction current calculation and phase displacement of output filter,
the harmonic currents are difficult to be acquired accurately, therefore the performance of the HAPF

1876-6102 2011 Published by Elsevier B.V. Selection and/or peer-review under responsibility of International Materials Science Society.
doi:10.1016/j.egypro.2012.01.271

1754

Wei Zhao et al. / Energy Procedia 16 (2012) 1753 1759

Author name / Energy Procedia 00 (2011) 000000

system is not so good [4]. Many papers which focus on obtaining the harmonic currents for APF have
been published to resolve the problems. Asynchronous frame proportion integral (SFPI) [5] and selfadapting filtering technique [6] are reported, but they both have their respective disadvantages. SFPI will
rotate every order harmonic, many frames will be created making the calculation intricacy, the control
period will be lengthened and the switch frequency will be limited too. Self-adapting filtering technique
relies on the accuracy of the reference current. Furthermore most proposed methods used to obtain the
control reference signal require several high precision analog multipliers or a high speed digital signal
processing chip with fast A/D chip, it will result in a complex circuitry, so they are mostly difficult to be
realized.
This paper proposed a new method using recursive integral algorithm for a shunt hybrid active power
filter to improve its steady-state and dynamic performance. The proposed algorithm is easy for
engineering application and has small computation capacity. Fuzzy logic based set-point weighing
algorithm is combined in the control scheme to enhance its robustness and anti-interference ability. A
prototype was developed to demonstrate the performance of the proposed HAPF and the control scheme.
Experiment results have verified the validity of the proposed control method.
2. Description of the system
HAPF system is composed of PPF and APF, as shown in Fig. 1, T stands for coupling transformer, L1
and C form into output filter. APF is used to improve the effect of PPF and restrain the resonance made
by PPF. The compensation characteristics of PPF are selected according to the characteristic harmonics
needing to be compensated. This paper chooses 5 and 7 single-tuned filters and a second order high pass
filter in allusion to the harmonic source of three-phase diode full-bridge rectifier. Coupling transformer is
used to realize electric isolation and make the voltage of inverter match up to the current. Output filter
eliminates switch harmonics made by switching power equipments.

Fig. 1. System configuration of the HAPF

*

Control system for the HAFP system is shown in Fig2. In Fig2 icx and icx respectively stand for the x
*
phases compensate current and its reference value, while isx for source reference current; vdc and
*
vdc respectively denote the dc side voltage of PWM inverter and its reference value while iLx and iLf
respectively the load current and load fundamental current gained by i p , iq theory; iREF represents the
reference signal needed buy the inverter. It is made up of current control inner loop and voltage control
outer loop, current loop is used to control current, while voltage loop maintain the dc-link voltage to
enhance the inner loops performance. The reference current of inverter can be obtained from

icx* = i*sx - iLx

(1)

1755

Wei Zhao et al. / Energy Procedia 16 (2012) 1753 1759

Author name / Energy Procedia 00 (2011) 000000

i*sx

i REF

i Lf

Fig. 2. A controller for the HAPF system

It is obvious that harmonics will be eliminated when the compensate current and the harmonic
components of load current are equal in amplitude and opposite in phase.
3. Control algorithm
3.1 Structure of the Controller
Recursive integral PI controller is composed of recursive integral PI and fuzzy logic based set-point
weighing control as shown in Fig.3, e and f respectively stand for the input and output of the controller,
while +e for variation rate of e; b and f respectively denote the weighing coefficient and output of fuzzy
inference; w is a constant between 0 and 1.

e,e
e

f+ b

w+

i*CX

iCX
+
+
+

Fig. 3. The structure of the proposed controller

PI controller is widely used in industrial process control for its simple arithmetic and high reliability,
but conventional PI is difficult to achieve the satisfied results in controlling sinusoidal current as in AC
driver [7]. The current controlled in this paper is periodic, so recursive integral PI is adopted to integrate e
in every period, and it is equivalent to have N PI controllers to work realizing fast zero steady-state error
adjusting. Fuzzy logic based set-point weighing algorithm is combined in the control scheme to enhance its
robustness and anti-jamming ability.
3.2 Recursive integral pi algorithm
The input of the controller is periodic and its period is 20ms. It can be seen as a composite of sinusoid
whose periods are multiple of 20ms. To explain the calculation, the input of the controller is shown in
Fig.4.

1756

Wei Zhao et al. / Energy Procedia 16 (2012) 1753 1759

Author name / Energy Procedia 00 (2011) 000000

Fig. 4. The input of the controller

Traditional PI algorithm will integrate every sampling point. The output of the controller should be
K

u(K)=K i e(K) + K1 e(i)

(2)

i =0

In formula (2) u and e respectively denote the output and input of the controller, while N the sample
amount in every period; K P and K i respectively stand for process gain and integral constant, while k
for sample time.
Recursive integral PI algorithm is proposed based on the traditional PI algorithm, it means that the
variable at the time of K, K-N, K-2N, , K-CN will be integrated. According to the algorithm the output
should be
C

u(K)=K P e(K) + K1 e(K-iN)

In formula (3) C=int

(3)

i=0

K
.
N

At the time of K-N, the above formula can be predigested as

C-1

u(K-N)=K P e(K-N) + K1 e[K-(i+1)N]

(4)

i =0

Subtract4from3the output of the controller should be

u(K-N)=u(K-N) + K P e(K)-K P e(K-N) + K1 .e(K)

(5)

3.3 The control algorithm combined with fuzzy logic based set-point weighing algorithm
To enhance robustness and anti-interference ability of the proposed recursive integral PI algorithm,
fuzzy logic based set-point weighing algorithm [11] is combined in the control scheme. After it has been
adopted, the control formula can be changed as

u(K)=u(K-N) + K P [b(K) i*c (K)-ic (K)

b(K-N) i*c (K-N)+ic (K-N)] + K1e(K) e (6)

b(K)=w+f (K)

Wei Zhao et al. / Energy Procedia 16 (2012) 1753 1759

Author name / Energy Procedia 00 (2011) 000000

This paper supposes that the fuzzy sets of e and e are {NBNSZOPSPB} and fuzzy set of
f is{NVBBNMNSZOPSPMPBPVB}. Membership degrees of e , e and f submit to
normal distribution model. Fuzzy rule is founded based on Macvicar-Whelan matrix and engineering
experience as shown in Tab.1 [8]. The method of weighted means is used to unwrap ambiguity.
Table 1 F rule base.

4. Experimental investigations
To verify the validity of the proposed control method discussed in this paper, a prototype of HAPF was
manufactured as in Fig. 1.
Multi-objective optimal optimization[9] is adopted to design parameters of passive filter. The major
parameters are as follows. 5th Passive Filter: L5=15.2mH, C5=29.6uF; 5th Passive Filter: L7=13.7mH,
C5=16.7uF, High pass filter : LH=13.7mH, CH=16.7uF, RH=1 , Vs= 380V, Ls =50uH, Load type : threephase diode full-bridge, Dc link inductor: L=4mH.
Power switching equipment is IPM modules PM300CLA120. Parameters of controller KP and KI are
confirmed by tries in experiment and they are all simulated in MATLAB. Control algorithm is realized
based on TMS320F2407, and photos of the control system are shown in fig.5.

Fig.5 Photo of the control system

The waveforms and spectrums are shown in Fig.6, measure results are obtained by intelligent
harmonic surveillance, analysis and protection incorporate equipment [10].

1757

1758

Wei Zhao et al. / Energy Procedia 16 (2012) 1753 1759

Author name / Energy Procedia 00 (2011) 000000

Fig. 6. The waveforms of load, inverter and utility current (a) Use P controller. (b) Use controller proposed.

It can be seen from fig.6 that the harmonics of source current are mostly compensated by HAPF which
adopts recursive integral PI controller, the amplitudes of harmonics are almost minished to zero while the
other HAPF system which only adopts integral PI controller are minished finitely. The above results show
that the proposed controller improves the performance of HAPF system distinctly.
5. Conclusions
In allusion to the solve problems of HAPF system, this paper presents recursive integral PI algorithm
combined with fuzzy logic based set-point weighing algorithm applied. After the algorithm has been
adopted, the performance of the system and its robustness, anti-interference ability are improved
distinctly. Experiment results show the feasibility of the proposed algorithm. The proposed recursive
integral PI algorithm is not only easy for engineering application, but also can be applied in the control of
periodic current to obtain a zero steady-state error regulation, as in AC drivers.
Acknowledgment
T The financial support of The National Natural Science Foundation of China (Project No. 60774043)
and The National High Technology Research and Development of China (863 Program) (Project No.
2008AA05Z211) .
References
[1]H. Fujita, H. Agaki, The Unified Power Quality Conditioner of Seriesand Shunt -Active Filters, IEEE Transactions on
Power Electronics, vol. 13, pp. 34 41, Mar. 1998.

Wei Zhao et al. / Energy Procedia 16 (2012) 1753 1759

Author name / Energy Procedia 00 (2011) 000000

[2]R. S. Herrera, P. Salmern, H. Kim, "Instantaneous Reactive Power Theory Applied to Active Power Filter Compensation:
Different Approaches, Assessment, and Experimental Results," IEEE Transactions on Industrial Electronics, vol. 55, pp. 184-196,
Feb. 2008.
[3]S. Buso, L. Malesani, and P. Mattavelli, Comparison of current control techniques for active filter applications, IEEE
Transactions on Industrial Electronics, vol. 45, pp. 722-729, Oct. 1998
[4] Teresa Esther Nez-Ziga, Jos Antenor Pomilio, Shunt Active Power Filter Synthesizing Resistive Loads, IEEE
Transactions on Power Electronics, 2002,17(2)273-278
[5] Oscar C. Montero-Hernndez, Prasad N. Enjeti, A Fast Detection Algorithm Suitable for Mitigation of Numerous Power
Quality Disturbances, IEEE Transactions on Industry Application., 2005,41(6),16841690.
[6] V. Soares, P. Verdelho, G. D. Marques, An instantaneous active and reactive current component method for active filters,
IEEE Transactions on Power Electronics, 2000,5(4):660669.
[7] Lorenz R D, Lipo T A, Novotny D W, Motion control with induction motors, Proceedings of the IEEE199482(8)
1115-1135
[8] Leonid Reznik, Omar Ghanayem, Anna Bourmistrov, PID plus fuzzy controller structures as a design base for industrial
applications, Engineering Applications of Artificial Intelligence, 2000, 13:419-430.
[9] Tu Chun Ming, Luo An, Liu Juan, Multi-objective optimal design of passive filters, Proceedings of the CSEE2002
22(3)17-21
[10] Lin Gan, He Zhihua, Li Zhengguo, An Intelligent Instrument for Integrated Harmonic Detection, Analysis and Protection,
Automation of Electric Power Systems,2000,24(1)64-66

1759