PI and fuzzy logic controllers for shunt active power filter — A report

ISA Transactions 51 (2012) 163–169

Contents lists available at SciVerse ScienceDirect

ISA Transactions journal homepage: www.elsevier.com/locate/isatrans

PI and fuzzy logic controllers for shunt active power filter — A report Karuppanan P ∗ , Kamala Kanta Mahapatra Department of Electronics and Communication, National Institute of Technology, Rourkela-769 008, India

article

info

Article history: Received 8 March 2011 Received in revised form 11 September 2011 Accepted 12 September 2011 Available online 5 October 2011 Keywords: Active power filter PI-controller Fuzzy logic controller Adaptive-fuzzy hysteresis current controller Harmonic currents

abstract This paper presents a shunt Active Power Filter (APF) for power quality improvements in terms of harmonics and reactive power compensation in the distribution network. The compensation process is based only on source current extraction that reduces the number of sensors as well as its complexity. A Proportional Integral (PI) or Fuzzy Logic Controller (FLC) is used to extract the required reference current from the distorted line-current, and this controls the DC-side capacitor voltage of the inverter. The shunt APF is implemented with PWM-current controlled Voltage Source Inverter (VSI) and the switching patterns are generated through a novel Adaptive-Fuzzy Hysteresis Current Controller (A-F-HCC). The proposed adaptive-fuzzy-HCC is compared with fixed-HCC and adaptive-HCC techniques and the superior features of this novel approach are established. The FLC based shunt APF system is validated through extensive simulation for diode-rectifier/R–L loads. © 2011 ISA. Published by Elsevier Ltd. All rights reserved.

1. Introduction Recently a lot of research is being encouraged for power quality and custom power problems in the distribution system due to non-linear loads [1,2]. In practical application most of the loads are non-linear, such as power converters, SMPS, arc furnaces, UPS and ASDs [3]. These non-linear loads are introducing harmonic distortion and reactive power problems [4]. The harmonics in the system induce several undesirable issues, such as increased heating losses in transformers, low power factor, torque pulsation in motors, poor utilization of distribution plant and also affects other loads connected at the same Point of Common Coupling (PCC) [5,6]. In recent times Active Power Filters (APFs) or Active Power-Line Conditioners (APLCs) are developed for compensating the harmonics and reactive power simultaneously [7,8]. The APF topology can be connected in series or shunt and combinations of both (unified power quality conditioners) as well as hybrid configurations [9,10]. The shunt active filter is most popular than the series active filter, because most of the industrial applications required the current harmonics compensation [11,12]. The controller is the most significant part of the active power filter and currently various control strategies are proposed by many researchers [13–15]. There are two major parts of the controller,

∗

Corresponding author. Tel.: +91 9442704585; fax: +91 661 2462999. E-mail addresses: [email protected] (K. P), [email protected] (K.K. Mahapatra).

one is reference current extraction from the distorted line-current and another is the PWM-current controller to generate switching patterns for inverter [16,17]. Many control strategies are proposed in the literature to extract the harmonic components. However, the conventional PI controller requires precise linear mathematical model of the system, which is difficult to obtain under parameter variations and non-linear load disturbances. Another drawback of the system is that the proportional and integral gains are chosen heuristically [18,19]. Recently, Fuzzy Logic Controllers (FLCs) have been used in various power electronic applications and also in active power filters [20,21]. The advantage of FLCs over the conventional controllers is that it does not need an accurate mathematical model. It can handle nonlinearity and is more robust than conventional PI or PID controllers [22,23]. Similarly, various current control techniques are proposed for APF inner current control loop, such as triangular current controller, sinusoidal-PWM, periodical-sampling controller and hysteresis current controller [24,25]. The Hysteresis Current Controller (HCC) method attracts researchers’ attention due to unconditional stability and simple implementation [26]. However, this control scheme exhibits several unsatisfactory features such as uneven switching frequency where the switching frequency varies within a particular band limit [27]. Adaptive-HCC (A-HCC) overcomes these fixed-HCC demerits. The adaptive-HCC changes the bandwidth according to instantaneous compensation current variation [28–31]. However, the adaptive-HCC is having more switching losses due to high frequency switching. These problems can be overcome by the proposed Adaptive-Fuzzy-HCC (A-F-HCC) in this paper. We focus on an adaptive-fuzzy-hysteresis current

0019-0578/$ – see front matter © 2011 ISA. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.isatra.2011.09.004

164

K. P, K.K. Mahapatra / ISA Transactions 51 (2012) 163–169

Fig. 2. Block diagram of the control strategy. Fig. 1. Shunt active power filter system.

control scheme, where the hysteresis-bandwidth is calculated with the help of fuzzy logic [32–34]. This controller maintains the modulation frequency constant that facilitates reduction of switching loss. Furthermore, uneven switching does not occur that facilitates reduction in the high frequency ripples in the compensating current as well as load current. Therefore this control strategy for inner current loop improves the PWM–VSI performance and makes APF substantially. This paper presents a shunt APF that uses PI or fuzzy logic controller for reference current extraction. The active filter is implemented with VSI and switching patterns are generated from the proposed novel adaptive-fuzzy-HCC. The shunt APF system is validated through extensive simulation under nonlinear load conditions. Comparative assessments of the different PWM-current controller are presented. In Section 2, the design of shunt APF system architecture is presented. Section 3 describes about control strategies for reference current extraction and PWMcurrent controller proficiencies. Section 4, system performance has been modeled and analyzed from the obtained results under nonlinear load conditions. Finally, Section 5 describes the conclusions of this work. 2. Design of shunt APF system Shunt APF is connected in the distribution-grid at PCC through filter inductance that is shown in Fig. 1. The filter inductance suppresses the harmonics caused by the switching operation of the power inverter. The current harmonic compensation is achieved by injecting equal but opposite current harmonic components at PCC, there by canceling the original distortion and improving the power quality on the connected power distribution system [8]. The instantaneous source current is represented as is (t ) = iL (t ) − ic (t ).

(1)

The instantaneous source voltage is

vs (t ) = Vm sin ωt .

(2)

The nonlinear load current contains the fundamental component and harmonic current components, which is represented as [20] iL (t ) =

∞ −

In sin(nωt + Φn )

n=1

 = I1 sin(ωt + Φ1 ) +

∞ − n =2

 In sin(nωt + Φn ) .

(3)

The instantaneous load power can be computed from the source voltage and load current and the calculation is given as pL (t ) = is (t )∗ vs (t )

= Vm sin2 ωt ∗ cos ϕ1 + Vm I1 sin ωt ∗ cos ωt ∗ sin ϕ1   ∞ − ∗ + Vm sin ωt In sin(nωt + Φn ) n =2

= pf (t ) + pr (t ) + ph (t ).

(4)

This load power contains fundamental (active power), reactive power and harmonic power. From this Eq. (4), the real (fundamental) power drawn from the load is pf (t ) = Vm I1 sin2 ωt ∗ cos ϕ1 = vs (t )∗ is (t ).

(5)

If the active power filter provides the total reactive and harmonic power, the source current is (t ) will be in phase with the utility voltage and sinusoidal. The three-phase source currents after compensation can be expressed as i∗sa (t ) = pf (t )/vs (t ) = I1 cos ϕ1 sin ωt

= Imax sin ωt .

(6)

Similarly, i∗sb = Imax sin(ωt − 120°)

(7)

i∗sc = Imax sin(ωt + 120°).

(8)

This peak value of the reference current Imax is estimated by regulating the DC-bus capacitor voltage of the inverter using PI or fuzzy logic controller. 3. Control strategies The block diagram of the proposed control system is shown in Fig. 2 that consists of two parts. One is reference current extraction controller using unit current vector along with PI or fuzzy logic controller. Another is PWM-current controlled voltage source inverter switching control method using fixed-HCC or adaptiveHCC or adaptive-fuzzy-HCC. 3.1. Reference current extraction controller The reference current is extracted from distorted line-current using unit current vector with PI or FLC. Conventional PI and PID controllers have been used to estimate the required reference current, and to control the dc-bus capacitor voltage of the inverter [18,19]. When load is non-linear or reactive, it is

K. P, K.K. Mahapatra / ISA Transactions 51 (2012) 163–169

expected that the supply will supply only the active fundamental component of the load current and the compensator would supply the harmonic/reactive component. The outer voltage loop will try to maintain the capacitor voltage nearly constant which is also a mandatory condition for the successful operation of the APF. The system losses are provided by the source in steady state. The compensator supplies the harmonic power, which manifests itself only on the reactive component of power. In the transient conditions the load changes are reflected in the dc capacitor voltage as an increase (or decrease) as capacitor absorbs (or delivers) the excess (or deficit) power. This conservation of energy philosophy is used to obtain the reference current for compensator in this method. The perturbations in the capacitor voltage are related to the perturbations in the average power drawn by the non-linear load. This property is utilized which facilitates extracting compensator reference and maintains capacitor voltage nearly constant. This control algorithm is based on sensing source current only (no need to sense load currents and compensation currents) that reduces sensors and complexity. 3.1.1. Unit current vector The source currents are sensed and converted into the unit sine current(s) while corresponding phase angles are maintained. The unit current vectors templates are represented as [21] ia = sin ωt , ib = sin(ωt − 120°) and

(9)

ic = sin(ωt + 120°). The amplitude of the sine current is unity in steady state and in the transient condition it may increase or decrease according to the loads. The unit currents should be in phase with the source voltages. These unit currents are multiplied with peak-amplitude of the estimated reference current (using PI or FLC) which is used to generate the desired reference currents. 3.1.2. PI-controller The DC-side capacitor voltage is sensed and compared with a reference voltage. The comparison result of the voltage error e = Vdc ,ref − Vdc at the nth sampling instant is used as an input for Proportional Integral (PI)-controller. The error signal passes through Low Pass Filter (LPF) to filter the higher order components and passes only the fundamental. It transfer function is defined as [18], H (s) = KP + KI /s.

(10)

The proportional gain is derived using KP = 2ζ ωnv C [KP = 0.7] that determines the dynamic response of the DC-side voltage. The √ damping factor ζ = 2 / 2 and natural frequency ωnv should be chosen as the supply fundamental frequency. Similarly, the integral gain is derived using KI = C ωn2v [KI = 23] that determines settling time and eliminates steady state error in the DC-side capacitor voltage. This controller estimates the magnitude of peak reference current Imax and controls the dc-side capacitor. This estimated magnitude of peak current multiplies with output of unit current vector, which generates the required reference currents to compensate the harmonic components. 3.1.3. Fuzzy logic controller Fuzzy logic control is deduced from fuzzy set theory; it was introduced by Zadeh in 1965 [22]. In fuzzy set theory develops the transitions between membership and non-membership function. Therefore, limitation or boundaries of fuzzy sets can be undefined and ambiguous and useful for approximate systems design. In order to implement the fuzzy logic control algorithm of an APF in

165

Fig. 3. Fuzzy logic controller. Table 1 Rule base table. e(n)

NB NM NS ZE PS PM PB

ce(n) NB

NM

NS

ZE

PS

PM

PB

NB NB NB NB NM NS ZE

NB NB NB NM NS ZE PS

NB NB NM NS ZE PS PM

NB NM NS ZE PS PM PB

NM NS ZE PS PM PB PB

NS ZE PS PM PB PB PB

ZE PS PM PB PB PB PB

a closed loop, the dc-bus capacitor voltage is sensed and compared with the desired reference value. This computed error signal (e = VDC −ref − VDC ) is passed through a LPF. The error signal e(n) and integration of error signal or change of error signal ce(n) are used as inputs for fuzzy processing as shown in Fig. 3. Fuzzy logic is characterized by (1) Seven fuzzy sets (NB, NM, NS, ZE, PS, PM, PB) for each input and output variables. (2) Triangular membership function is used for the simplicity. (3) Implication using Mamdani-type min-operator. (4) Defuzzification using the height method. Fuzzification: Fuzzy logic uses linguistic variables instead of numerical variables. In a control system, the error between reference signal and output signal can be assigned as Negative Big (NB), Negative Medium (NM), Negative Small (NS), Zero (ZE), Positive Small (PS), Positive Medium (PM), Positive Big (PB). The processes of fuzzification convert numerical variables (real number) convert to linguistic variables (fuzzy number). Rule elevator: Conventional controllers like PI and PID have control gains which are combination of numerical values. In case of fuzzy logic controller uses linguistic variables instead of the numerical. The basic fuzzy logic operations required for evaluation of fuzzy set rules are AND (∩), OR (∪) and NOT (−) for intersection, union and complement functions respectively, it is define as AND—Intersection µA∩B = min[µA (X ), µB (x)] OR—Union µA∪B = max[µA (X ), µB (x)] NOT —Complement: µA = 1 − µA (x). Defuzzification: The rules of fuzzy logic generate required output in linguistic variables, according to real time requirements. The linguistic variables have to be transformed to crisp number. This selection of strategy is compromise between accuracy and computational intensity. Data and rule base: The database stores the definition of the triangular membership function, which is required by fuzzifier and defuzzifier. The rule-base stores the linguistic control rules which are used by rule evaluator (decision making logic). The 49-rules are used in this proposed controller as given in Table 1. The output of the fuzzy controller estimates the magnitude of peak reference current Imax . This current Imax takes response of the

166

K. P, K.K. Mahapatra / ISA Transactions 51 (2012) 163–169

Fig. 4. Diagram of hysteresis current control.

active power demand for harmonics and reactive power compensation. This estimated magnitude of peak-current multiplies with output of unit current vector and generates the required reference currents. These reference currents compared with actual currents to generate the inverter switching patterns using PWM-current controller. 3.2. PWM–VSI current controller The effectiveness of active power filter depends on the design and characteristics of the current controller. There is various PWMcurrent control strategies proposed for APF applications. But in terms of quick current controllability and easy implementation, HCC has highest rating among other methods such as sinusoidalPWM and triangular current controller [24,25]. This section describes the fixed-HCC, adaptive-HCC and adaptive-fuzzy-HCC for active power filter. 3.2.1. Fixed-hysteresis current controller The fixed-hysteresis current controller is utilized independently for each phase and directly generates the switching patterns for the PWM-voltage source inverter [26]. It imposes a bang–bang instantaneous control method that draws the APF compensation current to follow its reference signal within a certain band limits. The actual current iactual (t ) is compared with iref (t ) and the resulting error is subjected to a hysteresis controller to determine the gating signals of the inverter as shown in Fig. 4. If the error current exceeds the upper limit of the hysteresis band, the upper switch of the inverter arm is turned OFF and the lower switch is turned ON. As a result, the current starts to decay. If the error current crosses the lower limit of the band, the lower switch is turned OFF and the upper switch is turned ON. As a result, the current gets back into the hysteresis band. This switching performance is defined as

 S=

OFF if iactual (t ) > iref (t ) + H . ON if iactual (t ) < iref (t ) − H

Fig. 5. Adaptive hysteresis current controller.

following equations can be written in the switching intervals t1 and t2 [27,28] di+ a dt di− a dt

(11)

The advantages of fixed-HCC are simple design and unconditioned stability. However, this control scheme exhibits several unsatisfactory features such as uneven switching frequency, difficult to design the passive filters system to suppress ripples in compensating currents. This unpredictable switching function affects the APF efficiency and reliability. Adaptive-hysteresis current controller overcomes the fixed-HCC demerits. This adaptive-HCC changes the bandwidth according to instantaneous compensation current variation. 3.2.2. Adaptive-hysteresis current controller Fig. 5 shows current and voltage waves of the PWM-voltage source inverter for phase a. The current ia tends to cross the lower hysteresis band at point 1, where the transistor Q 1 is switched ON. The linearly rising current (ia +) then touches the upper band at point 2, where the transistor Q 4 is switched ON. The linearly falling current (ia −) then touches the lower band at point 3. The

1 L

(0.5Vdc − Vs )

(12)

1

= − (0.5Vdc + Vs )

(13)

L

− where L = phase inductance; (i+ a ) and (ia ) are the respective rising and falling current segments. From the geometry of Fig. 5, we can write

di+ a dt di− a

t1 −

di∗a dt

t1 = 2HB′

(14)

di∗a

t2 = −2HB′ dt t1 + t2 = Tc = 1/fc dt

t2 −

(15) (16)

where t1 and t2 are the respective switching intervals, and fc is the modulation frequency. Adding (14) and (15) and substituting (16) we can write di+ a dt

t1 −

di− a dt

t2 −

1 di∗a fc dt

= 0.

(17)

Subtracting (15) from (14), we get di+ a dt



=

t1 −

di− a dt

t2 − (t1 − t2 )

di∗a dt

= 4HB′ .

(18)

Substituting (17) in (18), we get di+ a dt

(t1 + t2 ) − (t1 − t2 )

di∗a dt

= 4HB′ .

(19)

Substituting (13) in (17) and simplifying,

(t1 − t2 ) =

di∗a /dt fc (di+ a /dt )

.

(20)

Substituting (20) in (19), it gives,

 ′

HB =

0.125Vdc fc L

 1−

4L2 2 Vdc



Vs L

2  +m

.

(21)

Here, m = di∗a /dt is the slope of reference current signals. The hysteresis band HB′ can be modulated at different points of fundamental frequency to control the switching pattern of the inverter. The bandwidth HB′ is applied to the variable HCC and it created by s-functions in Matlab to produce switching patterns. This switching frequency of the inverter depends on the dc-side capacitor

K. P, K.K. Mahapatra / ISA Transactions 51 (2012) 163–169 Table 2 System parameters. Parameters

Values

Line to line source voltage (Vm ) System frequency (f ) Source impedance: Source resistor (RS ) Source inductor (LS ) Non-linear load: Diode rectifier Load resistor (RL ) Load inductor (LL ) Filter:Inductor (LF ) Resistor (RF ) DC-side capacitance (CDC ) Reference voltage (VDC , ref ) Power converter

440 V 50 Hz 1 0.1 mH 6-diode 20  100 mH 1 mH 1 2100 µF 400 V MOSFETs/diodes

voltage and line inductance of the APF configuration. This adaptiveHCC is having more switching power losses due to high frequency and this problem is mitigated by the proposed adaptive-fuzzy-HCC. The adaptive-fuzzy-HCC calculates the bandwidth effectively with the help of FLC and reduces the switching power losses by optimizing switching frequency. 3.2.3. Adaptive-fuzzy hysteresis current controller The proposed adaptive-fuzzy-HCC calculates the hysteresis bandwidth very effectively with the help of fuzzy logic controller. The adaptive-HCC based band-width is calculated and is given in the Eq. (21). From this equation, the hysteresis bandwidth HB′ is derived from the modulation frequency fc , supply voltage Vs , dc-side capacitor voltage Vdc , and slope of the reference current signals di∗a /dt and decoupling inductance L of the active power filter. The adaptive hysteresis band-width can be modulated as a function of di∗a /dt, Vs and Vdc . This adaptive hysteresis bandwidth HB′ as an error signal E (HB′ ) and change of error signal CE (HB′ ) are used as inputs for fuzzy logic processing. The adaptive-fuzzyhysteresis bandwidth (HB) is the output of the fuzzy controller. The fuzzy logic rules are stored as linguistic variables, which are required by the rule-evaluator. The 49-rules are used in this adaptive-fuzzy-HCC to calculate the required band-width (HB) and the rules are given in Table 1. The output of FLC of HB (hysteresis bandwidth) can be modulated at different points of the fundamental frequency cycle to control the switching pattern of the inverter. For symmetrical operation of all three-phases is denoted as HBa , HBb and HBc of same value, but having 120° phase difference. The adaptivefuzzy-HCC based hysteresis-bandwidth HB should maintain the modulation frequency fc constant. This controller reduces the switching power losses and improves the PWM–VSI performances for APF substantially. 4. Result and analysis The performance of the proposed control strategy is evaluated through Matlab/Simpower tools. The three-phase APF system comprises six-IGBTs with diodes, a dc-bus capacitor, RL-filter, compensation controller and switching pattern generator. The compensation control process is developed from unit current vector along with PI or FLC for extract the reference currents. The switching patterns are generated from fixed-HCC, adaptive-HCC and adaptive-fuzzy-HCC. These three-different types of current control techniques are simulated and investigated. The designs of the shunt APF system parameters value are given in Appendix (see in Table 2). Case 1 PWM-current controller. The effectiveness of an active power filter basically depends on the design and characteristics of the current controller. The switching losses are increased in fixed-HCC method and currents

167

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0

0.002 0.004 0.006 0.008

0.01

0.012 0.014 0.016 0.018

0.02

0

0.002 0.004 0.006 0.008

0.01

0.012 0.014 0.016 0.018

0.02

0

0.002 0.004 0.006 0.008

0.01

0.012 0.014 0.016 0.018

0.02

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Fig. 6. Switching patterns (a) HCC (b) Adaptive-HCC and (c) Adaptive-fuzzy-HCC.

generated due to this approach contain high frequency ripples. The PWM-current controller performance can be improved by using the adaptive control technique. A novel A-F-HCC, based on the adaptive control concept is used in this paper where hysteresis band implementation is done using fuzzy logic; this approach facilitates to optimize the PWM performance. The A-F-HCC utilizes a look-up table method using the instantaneous supply voltage and mains current reference slope as input variables and the hysteresis band as an output variable to maintain the modulation frequency constant. Fig. 6 illustrates the switching patterns of fixed-HCC, A-HCC and A-F-HCC methods. The ripples in the currents are reduced even if the switching frequency is increased; the A-F-HCC facilitates increase of switching frequency as per the system requirements. This A-F-HCC reduces the arbitrary switching and current ripples, which improves the APF performance compared to fixed-HCC and A-HCC. Case 2 PI-controller. The ac-main grid feeds diode-rectifier load. Shunt APF is connected in parallel at PCC that injects the anti-harmonics. PI-controller is used to estimate the peak amplitude of the reference current and fixed-HCC or A-HCC and A-F-HCC controllers used to generate switching patterns of the inverter. These three-different types of current controllers are simulated and investigated separately, but simulation of the Fig. 7 is focused on A-F-HCC. The rectifier R–L load current or source current before compensation is shown in Fig. 7(a). The shunt active filter supplies the compensating current that is shown in Fig. 7(b). Consequently current harmonic compensation is achieved by injecting equal but opposite current harmonic components at PCC by canceling the original distortion. The simulation result of source current after compensation is presented in Fig. 7(c) that indicates the current is sinusoidal. These figures are presented for A-phase only; other phase waveforms are phase shifted by 120°. Case 3 Fuzzy logic controller. FLC-controller is used to estimate the magnitude of peak reference current by controlling dc-side capacitor voltage of the inverter. The simulation result of the 3-phase source voltage is shown in Fig. 8(a) that indicates the source voltage is balanced. The six-pulse diode-rectifier load current or source current before compensation is shown in Fig. 8(b). The peak reference current is multiplied with unit current vector templates and generates the

168

a

K. P, K.K. Mahapatra / ISA Transactions 51 (2012) 163–169

a

60 40 20 0 20 40

b

c

60 40 30 20 10 0 10

b

20 30 40 80 60 40 20 0 20 40 60 80

c 0.02

0.03

0.04

0.05

0.06

0.07

0.09

0.08

0.1

d Fig. 7. Simulation results (a) source current before compensation (b) compensation current (c) source current after active filter compensation. Table 3 VDC settling time under various techniques. Current controller

e

VDC -settling time

Fixed-HCC (s) Adaptive-HCC (s) Adaptive-fuzzy HCC (s)

PI

FLC

0.29 0.26 0.25

0.25 0.23 0.23

Table 4 Real and reactive power measured under various methods. Current controller

Without APF

Fixed-HCC Adaptive-HCC Adaptive-Fuzzy HCC

P = 10.17 kW Q = 269 VAR

f

With APF PI

FLC

P = 10.85 kW Q = 118 VAR

P = 10.93 kW Q = 108 VAR

P = 11.14 kW Q = 067 VAR

P = 11.42 kW Q = 053 VAR

P = 11.27 kW Q = 059 VAR

P = 11.81 kW Q = 048 VAR

required reference-currents as shown in Fig. 8(c). That reference currents compared with actual current and generates switching patterns using fixed-HCC or A-HCC or A-F-HCC current controllers. The controllers are tested individually; however, the simulation of the Fig. 8 is focused on A-F-HCC. The shunt active filter produces required compensation currents as shown in Fig. 8(d). The source current after compensation is presented in Fig. 8(e) that indicates the current is sinusoidal. The APF is suppressing reactive/harmonic power and simultaneously improves power factor as shown in Fig. 8(f); the phase-a voltage is in phase with current. The DC-side capacitance voltage is controlled by PI or FLC. This controller reduces the ripples in the capacitor voltage to certain level and makes settling time to a low value (t = 0.23 s) and it is plotted in Fig. 8(g). This figure is for FLC in the outer loop with A-F-HCC in the inner current loop. The performance of the PI or FLC with different PWM-current control based dc-side capacitor voltage settling time is measured and given in Table 3. The Real (P ) and Reactive (Q ) power are calculated for different cases and are given in Table 4. These results are measured under diode-rectifier load using PI and FLC with three-different PWMcurrent controller. These results indicate that reactive power is suppressed by APF and hence improves power quality in the distribution system. The Fast Fourier Transform (FFT) is used to measure the order of harmonics in the source current. The Total Harmonic Distortion

g

Fig. 8. Simulation results (a) source-voltage, (b) load current, (c) reference-current, (d) compensation current (e) source current after compensation and (f) unity power factor, (g) DC-side capacitor-voltages.

(THD) of the source current is measured without and with active filter using PI and FLC along with different current controllers. The THD measurement is compared for PI and FLC controller that is presented in Table 5. The order of the harmonics versus magnitude of the source current is plotted that is shown in Fig. 9(a). It is shown that the supply current is distorted due to the presence of fifth, seventh, eleventh and higher order of harmonic spectral components and the measurement of THD is 26.67%. The Fig. 9(b) is plotted using A-F-HCC current controller based active power filter and it ensures that THD is less than 5% (THD = 1.72%) in compliance with the harmonic standards. The PI or FLC based active filter system is simulated and validated under diode-rectifier load conditions. The adaptivefuzzy-HCC reduces the switching power losses and improves the active filter performance in comparison with the fixed-HCC and

K. P, K.K. Mahapatra / ISA Transactions 51 (2012) 163–169

a

b

Fig. 9. Order of Harmonics (a) without APF and (b) with APF. Table 5 THD is measured under various methods. Current controller

Fixed-HCC Adaptive-HCC Adaptive-fuzzy HCC

Without APF (%)

26.67

With APF PI (%)

FLC (%)

3.57 2.91 2.38

3.21 2.66 1.72

the adaptive-HCC. Similarly FLC is giving better performance than conventional PI-controller in terms of Vdc -settling time, reactive power compensation and THD. The FFT analysis of the active filter confirms that the THD of the source current is less than 5% that is in compliance with IEEE-519 and IEC 61000-3 harmonic standards. 5. Conclusions The fuzzy logic controller based three-phase shunt active power filter facilitates current harmonics and reactive power compensation in the distribution system. The compensation process uses a PI or a fuzzy logic controller to extract reference-current while controlling DC-side capacitor voltage of the inverter. The shunt APF is implemented with PWM current controlled voltage source inverter. This inverter switching patterns are generated from a novel adaptive-fuzzy-hysteresis current controller. The adaptivefuzzy-HCC calculation of the hysteresis bandwidth is done with the help of fuzzy logic. The shunt APF system is investigated under fixed-HCC, adaptive-HCC and adaptive-fuzzy-HCC current controller techniques. It is observed that adaptive-fuzzy-HCC reduces the switching power losses compared to fixed and adaptive-HCC. The measured total harmonic distortion of the source currents is in compliance with IEEE 519 and IEC 61000-3 harmonic standards. This proposed controller based APF system can be implemented in field programmable gate array (FPGA) that would be attempted as a future work. Appendix See Table 2. References [1] Gyugyi L, trycula ECS. Active AC power filters. In: Proc of IEEE/IAS annual meeting 1976. 19-C: p. 529–35. [2] Hosseini Mehdi, Ali Shayanfar Heidar, Fotuhi-Firuzabad Mahmoud. Reliability improvement of distribution systems using SSVR. ISA Transactions 2009;48: 98–106. [3] Duffey Christopher K, Stratford Ray P. Update of harmonic standard IEEE519: IEEE recommended practices and requirements for harmonic control in electric power systems. IEEE Transactions on Industry Application 1989;25(6): 1025–34.

169

[4] Subjak Jr Joseph S, Mcquilkin John S. Harmonics — causes, effects, measurements, and analysis: an update. IEEE Transactions on Industry Application 1990;26(6):1034–42. [5] Emanuel Alexander E, Orr John A, Cyganski David, Gulachenski Edward M. A survey of harmonic voltages and currents at the customer’s bus. IEEE Transactions on Power Delivery 1993;8(1):411–21. [6] Grady WM, Samotyj MJ, Noyola AH. Survey of active power line conditioning methodologies. IEEE Transactions on Power Delivery 1990;5(3):1536–42. [7] Singh Bhim, Al-Haddad Kamal, Chandra Ambrish. A review of active filter for power quality improvements. IEEE Transactions on Industrial Electronics 1999;46(5):960–70. [8] Akagi Hirofumi, hirokazu watanabe Edson, Aredes Mauricio. Instantaneius power theory and applications to power conditioning. IEEE-press; 2007. [chapter 3–4]. [9] Kale Murat, zdemir Engin O. Harmonic and reactive power compensation with shunt active power filter under non-ideal mains voltage. Electric Power Systems Research 2005;74:363–70. [10] Habrouk MEl, Darwish MK, Mehta P. Active power filters: a review. IEE Proceedings Electrical Power Applied 2000;147(5):403–13. [11] Kumar Jain Shailendra, Agarwal Pramod, Gupta HO. A control algorithm for compensation of customer-generated harmonics and reactive power. IEEE Trans Power Delivery 2004;19(1):357–66. [12] Rahmat Allah Hooshmand, Mahdi Torabian Esfahani. A new combined method in active filter design in power systems. ISA Transactions 2011;50:150–8. [13] Herrera RS, Salmeron P, Kim H. Instantaneous reactive power theory applied to active power filter compensation: different approaches, assessment, and experimental results. IEEE Transactions on Indstrial Electronics 2008;55(1): 184–96. [14] Rahmat Allah Hooshmand, Mahdi Torabian Esfahani. Adaptive filter design based on the LMS algorithm for delay elimination in TCR/FC compensators. ISA Transactions 2011;50:142–9. [15] Czarnecki Leszek S. Instantaneous reactive power p–q theory and power properties of three-phase systems. IEEE Transactions on Power 2006;21(1): 362–7. [16] Shatshat REl, Kazerani M, Salama MMA. Power quality improvement in 3phase 3-wire distribution systems using modular active power filter. Electric Power Systems Research 2001;61:185–94. [17] Massoud AM, Finney, Williams BW. Review of harmonic current extraction techniques for an active power filter. In: International conference on harmonies and quality of power. 2004. p. 145–59. [18] Chaoui Abdelmadjid, Gaubert Jean Paul, Krim Fateh, Champenois Gerard. PI controlled three-phase shunt active power filter for power quality improvement. Electric Power Components and Systems 2007;35:1331–44. [19] Karuppanan Pitchaivijaya, Kamala Kanta Mahapatra. A novel PLL with fuzzy logic controller based shunt active power line conditioners. Journal of Electrical and Electronics Engineering 2010;3(2):153–8. [20] Jain SK, Agrawal P, Gupta HO. Fuzzy logic controlled shunt active power filter for power quality improvement. IEE Proceedings Electrical Power Applied 2002;149(5):317–28. [21] Saad S, Zellouma L. Fuzzy logic controller for three-level shunt active filter compensating harmonics and reactive power. Electric Power Systems Research 2009;79:1337–41. [22] Raviraj VSC, Sen PC. Comparative study of proportional–integral, sliding mode, and fuzzy logic controllers for power converters. IEEE Transactions on Industry application 1997;33(2):518–24. [23] Singh GK, Singh AK, Mitra R. A simple fuzzy logic based robust active power filter for harmonics minimization under random load variation. Electric Power Systems Research 2007;77:1101–11. [24] Nabae Akira, Ogasawara Satoshi, Akagi Hirofumi. A novel control scheme for current-controlled PWM inverters. IEEE Transactions on Industry Application 1986;1A-22(4):562–70. [25] Brod DM, Novotny DM. Current control of VSI-PWM inverter. IEEE Transactions on Industry Application 1985;21:562–70. [26] Zeng Jiang, Yu Chang, Qi Qingru, Yan Zheng, Ni Yixin, Zhang BL, Chen Shousun, Wu Felix F. A novel hysteresis current control for active power filter with constant frequency. Electric Power Systems Research 2004;68:75–82. [27] Kale Murat, Ozdemir Engin. An adaptive hysteresis band current controller for shunt active power filter. Electric Power Systems Research 2005;73:113–9. [28] Bose Bimal K. An adaptive hysteresis-band current control technique of a voltage-fed PWM inverter for machine drive system. IEEE Transactions on Industrial Electronics 1990;31(5):402–8. [29] Ezoji E, Fazlali M, Ghatresamani A, Nopour M. A novel adaptive hysteresis band voltage controller for dynamic voltage restorer. European Journal of Scientific Research 2009;37(2):240–53. [30] Dalvand H, Moghani JS, Talebi N. An adaptive hysteresis band current controller for hybrid power filter. In: International conference on industrial and information systems. 2006. p. 613–8. [31] Belhaouchet N, Rahmani L. Development of adaptive hysteresis-band current control of PWM three-phase AC chopper with constant switching frequency. Electric Power Components and Systems 2009;33(10):583–98. [32] Kadjoudj M, Abdessemed R, Golea N, Benbouzid ME. Adaptive fuzzy logic control for high performance PM synchronous drives. Electric Power Components and Systems 2001;29:789–807. [33] Cecati Carlo, Corradi Stefano, Rotondale Nicola. A fuzzy logic-based adaptive PWM technique. conference on IECON 1997; 3: p. 1142 – 47. [34] Nounou Hazem N, Rehman Habib-ur. Application of adaptive fuzzy control to ac machines. Applied Soft Computing 2007;7:899–907.