International Journal of Electronics ISSN: 0020-7217 (Print) 1362-3060 (Online) Journal homepage: http://www.tandfonline.com/loi/tetn20 Shunt Hybrid Active Power Filter under Non ideal Voltage Based on Fuzzy Logic Controller Papan Dey & Saad Mekhilef To cite this article: Papan Dey & Saad Mekhilef (2016): Shunt Hybrid Active Power Filter under Non ideal Voltage Based on Fuzzy Logic Controller, International Journal of Electronics, DOI: 10.1080/00207217.2016.1138515 To link to this article: http://dx.doi.org/10.1080/00207217.2016.1138515 Accepted author version posted online: 06 Jan 2016. Submit your article to this journal Article views: 6 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tetn20 Download by: [University of Malaya] Date: 21 January 2016, At: 19:50 Publisher: Taylor & Francis Journal: International Journal of Electronics DOI: 10.1080/00207217.2016.1138515 Shunt Hybrid Active Power Filter under Non ideal Voltage Based on Fuzzy Logic Controller Corresponding Author: Downloaded by [University of Malaya] at 19:50 21 January 2016 Papan Dey Power Electronics &Renewable Energy Research Laboratory (PEARL), Dept. of Electrical Engineering, University of Malaya, Kuala Lumpur 50603, Malaysia; E-mail: papan@siswa.um.edu.my Contact no. 0060143247702 Co-author: Saad Mekhilef Senior member, IEEE, Power Electronics &Renewable Energy Research Laboratory (PEARL), Dept. of Electrical Engineering, University of Malaya, Kuala Lumpur 50603, Malaysia; 1 Shunt Hybrid Active Power Filter under Non ideal Voltage Based on Fuzzy Logic Controller In this paper a synchronous reference frame (SRF) method based on a modified phase lock loop (PLL) circuit is developed for a three phase four wire shunt Downloaded by [University of Malaya] at 19:50 21 January 2016 hybrid active power filter (APF). Its performance is analyzed under unbalanced grid conditions. The dominant lower order harmonics as well as reactive power can be compensated by the passive elements whereas the active part mitigates the remaining distortions and improves the power quality. As different control methods show contradictory performance, Fuzzy logic controller is considered here for DC-link voltage regulation of the inverter. Extensive simulations of the proposed technique are carried out in a MATLAB-SIMULINK environment. A laboratory prototype has been built on dSPACE1104 platform to verify the feasibility of the suggested SHAPF controller. The simulation and experimental results validate the effectiveness of the proposed technique. Keywords: Phase lock loop; Fuzzy logic controller; Harmonic compensation; Synchronous Reference frame (SRF) method; Total harmonic distortion (THD) 1. Introduction The presence of current harmonics, unbalanced loading, reactive power and excessive neutral current play an important role in deciding power quality in a three-phase fourwire distribution system. These harmonics interfere with sensitive electronic equipment and cause undesired power losses in electrical equipment. In order to solve the power 2 quality problem and to deliver clean power several methods have been proposed and developed by the researchers (El-Habrouk, Darwish, & Mehta, 2000). Passive filters that work as the least impedance path to tune harmonic frequencies, are simple and less expensive but have several drawbacks including fixed compensation, they are bulky devices and the resonance problem of the L-C filters; however, APF has been developed for complete compensation of distortions (Das, 2004). Owing to the rapid improvement in power semiconductor device technology, active Due to the Downloaded by [University of Malaya] at 19:50 21 January 2016 absence of any resonance problem in the system, APF has attracted favorable feedback compared to tuned passive filters. The APFs use power electronic converters that insert harmonic components into the electrical network that cancel out nonlinear load harmonic currents. With the harmonic compensating ability, APFs are also capable of reactive power reduction and balancing the unbalanced load (Patnaik & Panda, 2013). Although different types of active filters have been proposed to increase the electric system quality, pure APF is limited due to its high cost and rating constraints by power devices. To make the active filter smaller, cheaper and more practicable in industrial applications, a hybrid configuration of the APF appears very attractive in power distribution networks. However, series hybrid APF is complex, unreliable and expensive. Therefore, a shunt hybrid filter topology is characterized by a combination of a parallel-connected passive filter and a small rated active filter. This has been considered in this paper for its capability of attenuation and effective operation. Many harmonic revealing control concepts and approaches for active filters such as fast Fourier transform (FFT); recursive discrete Fourier transform (RDFT); instantaneous active and reactive power (p–q); direct testing and calculating (DTC) method; instantaneous active and reactive current component (id–iq); notch filter, sine multiplication, generalized integral, synchronous detection algorithm; adaptive filter, 3 flux based controller; artificial neural network (ANN); adaptive linear neuron (ADALINE); wavelet transform; particle swarm optimization (PSO); and genetic algorithm (GA) techniques have been proposed in have been suggested in the literature (Corasaniti, Barbieri, Arnera, & Valla, 2009). Among these control schemes, SRF is one of the most conventional and practically applicable method (Mikkili & Panda, 2012). Although it performs an excellent job, it requires a PLL circuit for synchronization. Different types of PLL have been introduced in the literature; however, the conventional Downloaded by [University of Malaya] at 19:50 21 January 2016 types of PLL have low performance in extremely unbalanced and distorted networks. So this paper is aimed at the competency of the SRF method with the modified PLL. An additional essential task in active filters development is to maintain a constant DC link voltage across the capacitor which is attached to the inverter. The energy loss due to the switching power and conduction losses associated with the IGBTs and diodes of the inverter in APF must be compensated that reduce the DC capacitor voltage. Normally, to control the DC link voltage a PI controller is used. However, precise linear mathematical models are required for the tuning of the PI controller. It is difficult to obtain and cannot give satisfactory performance under load disturbances and parameter variations (Ponpandi & Durairaj, 2011).In addition, in this conventional method, the shaping of source current is realized with a good amount of spikes. Recently, a great deal of interest has focused on Fuzzy logic controllers regarding their application to APFs. Fuzzy logic is considered more advantageous than conventional controllers as there is no requirement of an accurate mathematical model, they are capable of working with imprecise inputs, handling nonlinearity and are more robust (Suresh, Panda, & Suresh, 2011). 4 In this paper, a new technique has been developed on SRF using the PLL algorithm and its performance is analyzed and demonstrated for unbalanced source conditions. Here, a Hysteresis current regulator is used to generate switching signals of APF, and for maintaining the DC link capacitor a voltage fuzzy logic controller is introduced for robustness and superior performance compared to the conventional PI controller. The structure of this paper is as followed: The hybrid APF is discussed in section 2. Section 3 describes the control algorithm of APF. The simulation results and a brief Downloaded by [University of Malaya] at 19:50 21 January 2016 analysis are presented in section 4. In section 5 experimental results are shown. Finally, conclusions are drawn in section 6. 2. Hybrid APF topology The proposed hybrid filter structure shown in Figure1 comprises a shunt passive filter and shunt active filter. A 3-leg inverter with split capacitor works as the APF. It is designed to be associated in parallel with the single phase and three phase loads that are considered as a non-linear and unbalanced load for a 3-phase 4-wire distribution system while its complex control circuitry and massive DC link capacitors are necessary for perfect operation. The inner point of every branch is attached to the power network through an inductor, which is used to filter the ripples of inverter current. The considered LC passive filter at the 5th harmonic tuned frequency is connected in shunt to the power line before the load. This provides a low impedance trap for harmonics to which the filter is tuned, and, correspondingly, aids in reduction of the active filter power rating. 5 Downloaded by [University of Malaya] at 19:50 21 January 2016 Figure 1. Schematic diagram of 3-phase 4-wire SHAPF 3. Proposed Control algorithm 3.1 Modified PLL Operation In order to extract the voltage peak and instantaneous phase information the best way is to convert the three phase unbalanced voltage in a two 2-D stationary frame. If voltage is perfectly balanced and this transformed voltage vector is drawn on a q-d frame, a circular vector rotating at speed equal to the frequency is seen. However, for the case of an unbalanced vector, there are two vectors, one corresponding to positive sequence rotating in one direction and the other because of the negative vector rotating in the opposite direction. Thus, if a transformed unbalanced voltage vector in the d-q frame is drawn, an eclipse is formed as a resultant of both positive and negative sequence components. Now, suppose the frequency and phase information are already known and taking an arbitrary frame that is rotating in the same direction as that of a positive sequence component and at a speed equal to the frequency, then a positive sequence component appears as a stationary and negative sequence component that seems to be rotating at a speed twice that of the frequency. For safe and consistent operation of an APF in an unbalanced and distorted grid voltage situation, the frequency and phase 6 extraction of the positive sequence component of the voltage should be obtained quickly and accurately. Because of the dynamic behavior, conventional SRF-PLL performance is unsatisfactory under non-ideal voltage mains (Kesler & Ozdemir, 2011). The improved PLL developed in this study is shown in Figure 2, which is put forward for the determination of positive sequence components with stability and rapidity. For nonideal mains voltage: Downloaded by [University of Malaya] at 19:50 21 January 2016 + V a V a + Vsabc = V b = V b + V c V c + V a − − V b + _ V c V a 0 V b0 V c 0 (1) Sinθ Sinθ − = V Sin(θ − 2π / 3) + V Sin(θ + 2π / 3) + Vo Sin(θ − 2π / 3 Sin(θ + 2π / 3) + Using (2) transform, the voltage vectors are: Vα Vαβ = = [Tαβ ]V abc V β Where, [ T αβ −1 2 3 2 1 ]= 2 3 0 (3) −1 2 3 2 Vα V + Sinθ + + V + Sinθ − Vαβ = = + + − − Vβ − V Cosθ + V Cosθ So: (4) Performing d-q transform: V d V dq = = [T dq ]V αβ V q ^ = Cos θ − Sin θ ^ Sin θ ^ V + Sin θ + + V + Sin θ − Cos θ ^ − V + Cos θ + + V − Cos θ − + + ^ − − ^ = V Sin(θ − θ ) + V Sin(θ + θ ) + + ^ − − ^ − V Cos(θ − θ ) + V Cos(θ + θ ) The estimated phase angle= θ ^ ; assuming θ ^ θ ^ =θ + =θ − ≈ ωt . (5) PLL successfully tracks the phase at . 7 Vd V − Sin ( 2θ ^ ) V ≈ + − ^ q − V + V Cos ( 2θ ) So, Here 2θ ^ (6) is the double frequency to be eliminated. It is the basic concept of the modified PLL structure. It can provide the positive sequence component by cancellation of 2ω oscillations (Venkatachalam, Chandrasekaran, & Rama Rao, 2009).The modified PLL circuit uses Clarke and Park transform for identifying the peak of positive sequence voltage. The three phase unbalanced voltages are converted to the stationary Downloaded by [University of Malaya] at 19:50 21 January 2016 coordinate system. Through the measured phase angle, the voltages in stationary coordination are transformed to the d-q frame. The PI controller forced the d-axis component Vd to zero so as to align the mains voltage space vector with the q-axis. The estimated phase angle θ, which is attained by integrating the proportion-integral (PI) controller output, is, in turn, used for the coordinate transformation process (Sinha & Sensarma, 2011). Here, a second order resonant filter is used instead of the conventional low pass filter for suppressing the double fundamental frequency, which is created for unbalancing and the gain adaptation block shows the convergence of any value of system voltages and creates normalized templates of fundamental voltage (De & Ramanarayanan, 2010).The rate limiter block works as an attenuator against the ripple. The modified PLL can perform satisfactorily as long as the gains of the PI are adjusted accordingly under highly inaccurate and disturbed system voltages. Figure 2. PLL circuit block diagram 8 The transfer function of the resonant filter is given in: H (s) = s 2 + 4ω 2 s + 4 ω cut −off + 4 ω 2 2 (7) and the open loop PLL gain is written as: 1 1 G ( S ) = V . .k p .(1 + ). H ( s ) s sT (8) Where T and Kp are the time constant and gain of the PI controller correspondingly. Now a critically damped system is taken into consideration and the value of the cut-off Downloaded by [University of Malaya] at 19:50 21 January 2016 frequency is chosen as it is equal to double the fundamental frequency of the system (Indu Rani, Aravind, Saravana Ilango, & Nagamani, 2012). For selection of the PI controller gains, the resonant filter is approximated as a low pass filter and the parameter design is obtained via the calculation through the optimum symmetrical method: T = 4τ (9) And: kP = 1 2V τ (10) Where τ is the time constant of the low pass filter. 3.2. Reference Current Generation The three phase load currents are measured using a hall-effect current sensor and converted into d-q-0 by means of a rotational frame that is synchronous with the system voltage positive sequence in equation 11. 2π 2π sin(ω s t ) sin(ω s t − 3 ) sin(ω s t + 3 ) i d i A i = 2 cos(ω t ) cos(ω t − 2π ) cos(ω t + 2π ) i q s s s B 3 3 3 i 0 i 1 1 1 C 2 2 2 (11) 9 ωst ω is considered here as the transformation-angle of the positive sequence source Downloaded by [University of Malaya] at 19:50 21 January 2016 voltage and is delivered by the phase lock loop (Jianben, Qiaofu,Jun,2012). Figure 3. SRF based control block diagram The system under observation is a three phase-four wire in which the active component id and oscillating part iq are reflected. After the load currents id and iq are found, they are allowed to pass over a low pass filter to separate the AC and DC part through which the active and reactive fundamental current components (id-DC, iq-AC) are obtained. Both the currents (id-DC, iq-AC) AC parts are related with the responsibility for active and non-active harmonic components. The filters used in the circuit are the 2nd order Butterworth type and their cut off frequency is identical to one half of the fundamental frequency (Mikkili & Panda, 2011). In consideration of the reactive current, the passive filter provides its DC value while the VSI delivers an AC voltage to sink the harmonics. The filtered active and non-active currents from equation 12 are used for generation of the accurate references to the modulator: i d i dDC i dAC i = − i qAC i q qDC (12) As well as providing harmonic currents, the DC voltage of the PWM VSI should be maintained for accurate operation (Saad , Tarek & Nasruddin, 2011). The voltage of the 10 capacitor is controlled by regulating the reactive current (Salmeron & Litrán, 2010) as shown in Figure 3. Then the a-b-c frame reference currents are: Downloaded by [University of Malaya] at 19:50 21 January 2016 cos(ω s t ) iCA sin(ω s t ) i = sin(ω t − 2π ) cos(ω t − 2π ) i d * s s CB 3 3 i q * i CC sin(ω t + 2π ) cos(ω t + 2π ) s s 3 3 (13) 3.3. DC Link Capacitor Voltage Control 3.3.1. Fuzzy controller Figure 4. DC voltage optimization with Fuzzy approach A Fuzzy logic controller is used to regulate the DC link voltage and added to the active part of the fundamental load current. The obtained voltage of the capacitor is matched with a preset reference value (Mikkili & Panda, 2011). The loss component is then handled by a Fuzzy logic controller, which contributes to the zero steady error in tracking the reference current signal, as shown in Figure 4. A low pass filter and a rate limiter block have been used here for optimum performance. The supply current peak value is calculated using the output component of the fuzzy logic controller (Park,Jeong,Kwon,2001),(Younus, Rahim& Mekhilef,2009). 11 A linguistic control approach is converted into an automatic control strategy, and with a knowledge database, Fuzzy rules are constructed by proficient experience. To begin with, ‘µ(E)’, the input error, and ‘µ(∆E)’, the change in error, have been used as the input variables of the fuzzy logic controller (Singh, Singh, & Mitra, 2007). Then the control current ‘Imax’ represents the output variable. Seven Fuzzy sets are chosen here to adapt these numerical variables into linguistic variables: PB (positive big), PM (positive medium), PS (positive small), ZE (zero), NS (negative small), NM (negative Downloaded by [University of Malaya] at 19:50 21 January 2016 medium) and NB (negative big), as can be seen in Figure 5. The intransient state requires coarse input/output variables for controlling based on large errors and fine input/output variables are essential for controlling the small errors properly with the development of Fuzzy logic rules in the steady state (Saad & Zellouma, 2009). With this perception, the rule table elements and view of the surface is obtained, as shown in Fig. 6, where ‘µ (E)’ and ‘µ (∆E)’ are used as inputs. Figure 5. Normalized membership functions for the input and output variables 12 ΔE NB NM NS Z PS PM PB E NB NB NB NB NB NM NS Z NM NB NB NB NM NS Z PS NS NB NB NM NS Z PS PM Z PS NB NM NM NS NS Z Z PS PS PM PM PB PB PB PM NS Z PS PM PB PB PB PB Z PS PM PB PB PB PB (a) (b) Downloaded by [University of Malaya] at 19:50 21 January 2016 Figure 6. (a) Fuzzy control rule (b) Surface Viewer (E vs ∆E) 3.4. Harmonic current generator The active filter switching patterns are decided by the current controller. A hysteresis comparator is used here for a quick response and appropriate sinusoidal current tracking capability (W.u,2011). The desired current, Ia (t), and the injected inverter current, Ia*(t) are compared with one another. The logic of switching is as follows: If Ia< (Ia*-hb) upper switch S1 is OFF & lower switch S4 is ON in leg of ‘a’ of APF. If Ia> (Ia*-hb) upper switch S1 is ON & lower switch S4 is OFF in leg of ‘a’ of APF. Correspondingly, in the legs ‘b’ & ‘c’ the switches are initiated. Here ‘hb’ is the hysteresis band and determines the variation of load current harmonics and switching frequency of the devices. 4. Simulation results and analysis The performance of a 3-legs 4-wire Hybrid APF for filtering current distortions, compensation of neutral current, load current balancing and reactive power mitigation has been examined under unbalanced source-balanced load. The goal of this case study is to demonstrate the validity and evaluation of the proposed approach, where the source is greatly unbalanced. 13 The results are specified in the following simulation studies for before compensation, compensation using only passive filter and after the operation of the hybrid filter. Three and single phase diode rectifier nonlinear loads are coupled in the power system for producing unbalanced, reactive current and harmonics in the load current as well as a neutral current. The widespread simulation outcomes are discussed below. 4.1 Unbalanced source condition Voltage unbalance occurs when the relative magnitudes and phase angle of the three Downloaded by [University of Malaya] at 19:50 21 January 2016 phases are not equal. The source voltage magnitude of phase-b is 20% smaller than the source voltage of phase-a, and the phase-c voltage is 50% lower than phase-a. This is done by using a three phase programmable voltage source block in Simulink. The three phase unbalanced voltages are considered as: Vsa= 100 sin ωt, Vsb= 80 sin (ωt-120) and Vsc= 50 sin (ωt+120). In unbalance source voltages, the fundamental components are first derived using PLL to make the unit templates for the balanced and sinusoidal voltages. Then the synchronizing phase angle is extracted and simulation is performed through it. The source voltage is shown in Figure 7 (a). The load current and its FFT graph without compensation are given in Figure 7 (b) and (c), which is highly non-linear in nature. After the application of the passive filter lower order harmonics are reduced but non-linearity and distortion still exists, which is presented in Figure 7 (d). The THD of the source current is reduced from 21.07% to 1.33% in the case of the fuzzy controller, which is below the IEEE standard as shown in Figure 7 (g). In Figure 7 (l) the extracted transformation angle from PLL is also shown. The values of source current and THD levels are listed in table 1.Although, the source currents are not purely balanced but the results are quite satisfactory. 14 Downloaded by [University of Malaya] at 19:50 21 January 2016 (a) (b) (c) (d) (e) (f) (g) (h) (i) (j) (k) (l) 15 Figure 7. Simulation results for operational performance of SHAPF a) Source voltage b) load current c) THD before compensation d) passive filtering source current e)THD after passive filtering f) source current g) THD after SHAPF h) injecting current i) DC link voltage J) Real power (p) k) Reactive power (Q) l)Transformation angle (wt) Table 1.System performance for unbalanced source condition Source current Before filter After Passive After Hybrid filter active filter Downloaded by [University of Malaya] at 19:50 21 January 2016 THD (%) A-phase 21.07 14.69 1.33 B-phase 48.59 27.11 1.52 C-phase 21.18 10.64 1.03 Neutral 52.5 - A-phase 1.58 2.32 5.84 B-phase 0.87 1.41 5.67 C-phase 1.54 2.23 5.39 Neutral 5.25 3.21 0.41 RMS(A) Table 2. Fuzzy logic controller performance evaluation Condition DC voltage settling time with Fuzzy Real & Reactive power without Active filter Real & reactive power with Active filter Unbalanced source 0.24s 910w, 370var 2.89kw,130var Table 3.Simulation parameters Parameters Source impedance Load impedance Values R = 0.01Ω, L = 2 mH 3-phase diode rectifier R = 80 Ω, L = 60 mH 1-phase diode rectifier R= 0.1Ω, L=1 mH Passive Filter Shunt APF 5th harmonic tuned, R=0.2Ω, L=5mH, C=60μF R=0.001Ω, L=2mH, Two DC link capacitor = 2100μF,Vdc=220 V 100 V (r.m.s), 50 HZ Source voltage 16 It has been observed that the shunt hybrid APF performance with fuzzy logic controller shows outstanding characteristics compared to the conventional controller under unbalance voltage condition. The calculation of real (p) and reactive (Q) power for unbalance case are arranged in Table 2. It also shows the effectiveness of the fuzzy logic controller. The parameters used in this study for SHAPF system are listed in Table 3. Downloaded by [University of Malaya] at 19:50 21 January 2016 Table 4. THD comparison between recommended topology and conventional controller Source current System voltage conditio -n Unbalance Phase A B C After SHAPF Before SHAPF THD (%) Q (VAR) 21.0 48.5 21.1 370 P-Q theory (Mikkili) Modified P-Q theory (Patnaik) PI controller with SRFPLL(corasanti) Proposed controller THD (%) Q (VAR) THD (%) Q (VAR) THD (%) Q (VAR) THD (%) Q (VA R) 6.32 5.23 6.49 305 5.21 5.03 5.43 280 3.41 3.32 2.89 150 1.33 1.52 1.03 130 A comparative analysis between the proposed control method, conventional method (Corasaniti et al., 2009), P-Q theory (Mikkili & Panda, 2012) and modified P-Q theory (Patnaik & Panda, 2013) is shown in Table 4 in terms of THD % and reactive power. The P-Q theory and modified P-Q theory are simulated with the used parameters. From the table it is observed that they are capable of mitigating harmonics but fail to reduce the THD % below 5% under a non-ideal voltage source. The proposed method allows 2% THD values of source current and compensates reactive power, which is satisfactory when compared with the conventional control technique. 5. Experimental Results The system parameters selected for the experimental verification: Vs=50V (peak), f=50HZ, Ls=5mH, Lc=5mH, RL=100Ω, Ll=20mH, Cdc=2000μF, Vdc =70V. Parameters 17 of passive filter, Lf =3mH, Cf =60μH. A prototype model of three-phase- four wire APF is developed in the laboratory for experimentation shown in Figure 8. For the generation of unbalancing in the voltage, two different values of resistance are connected in series with the two phases of supply, and, just after that, a constant 3-phase load is fixed. Another phase is kept as usual. The flow of current through the series impedance will cause a different voltage drop in each phase, so that unbalanced voltages occur in the system. In this case, the equivalent supply will be after the series Downloaded by [University of Malaya] at 19:50 21 January 2016 impedance and constant 3-phase load. There are two single-phase bridge inverters connected in series in each phase. Suitably designed snubber circuit is connected across each device. DC link capacitor is connected at DC side of the inverter. Three coupling inductors are connected in series with the three lines on AC-side of the inverter. Three-phase diode rectifier with R-L elements on DC side is used as a nonlinear load which draws nonlinear current. The AC and DC voltages are sensed using the isolation amplifier (AD202JN). The AC source currents are sensed using the PCB-mounted Hall-effect current sensors (LEM25-NP). The IGBT driver circuits are used for pulse amplification and isolation purposes. The driver circuit has special features like the fault protection and protection against the under-voltage lockout. The load current, the source voltage and injected filter currents, ifabc, are measured. The sensed signals are used for processing in the designed control algorithm. Because real-time simulation is an important aspect for control engineering, the same is true for the automatic generation of real-time code, which can be implemented on the hardware. 18 Downloaded by [University of Malaya] at 19:50 21 January 2016 Figure 8. Experimental setup for shunt hybrid APF 5.1 Unbalance source condition For an unbalanced source, 14% amplitude unbalance is introduced in the ‘b’ and ‘c’ phases so that the phase voltages are 50V in phase ‘a’, 43 V in phase ‘b’, and 38 V in phase ‘c’. The unbalance is also reflected in the load currents in the three phases. However, the source currents are balanced with the proposed APF. Figure 9 shows the total response of the system for this condition. From the results it is observed that the proposed control method gives satisfactory results under unbalanced source. The load current waveform of phase ‘a’ is distorted before filtering shown in Figure 9(b). FFT result is also listed. After the operation of hybrid APF controlled by fuzzy controller sinusoidal source current waveform is obtained instantaneously after the insertion of active filter as illustrated in Figure 9 (f). In figure it is also shown that the experimental results of the dc-side capacitor voltage which is nearly constant with small ripple. The current THD is reduced from 28% to 4.8% as shown in the frequency current spectrum. 19 (a) Downloaded by [University of Malaya] at 19:50 21 January 2016 (b) (c) (d) (e) (f) (g) (h) Figure 9. Experimental results for unbalance source a) Supply voltage (25v/div) b) Load current before compensation (0.8A/div) c) %THD before compensation d) Source current after passive filtering e) %THD after passive filter f) Source current after compensation g) %THD after SHAPF compensation h) DC link voltage 20 6. Conclusion A modified PLL with SRF theory is developed for computation of the reference current and employed effectively for grid voltage synchronization under unbalanced mains conditions. It is seen from the simulated figures that the Fuzzy logic controller shows better dynamic performance to mitigating current harmonics and compensating reactive power. The figures illustrate that the feedback currents of the filter approximately agree with the generated reference currents. The proposed method has been evaluated with Downloaded by [University of Malaya] at 19:50 21 January 2016 simulations under non-ideal mains voltage conditions where it is shown that percentage of THD is reduced from 21.07% to 1.33% for unbalance source. Moreover, 53% amount of reactive power has been compensated after using the shunt hybrid APF. The impact of source current unbalance has also been minimized. A laboratory prototype is developed and it has been found that value of THD has been reduced from 28% to 4.8% for unbalance voltage which is in the mark of 5% specified in the IEEE-519 standard. The value of power factor has been increased to 0.632 to 0.963. Therefore, with the combination of fuzzy logic and modified SRF theory approach, SHAPF can be considered as a reliable harmonic isolator for its quick response and good quality of filtering. Acknowledgement The authors wish to acknowledge the financial support from university of Malaya HIRMOHE project UM.C/HIR/MOHE/ENG/24. References Corasaniti, Victor Fabián, Barbieri, Maria Beatriz, Arnera, Patricia Liliana, & Valla, María Inés. (2009). Hybrid active filter for reactive and harmonics compensation 21 in a distribution network. IEEE Transactions on Industrial Electronics, 56(3), 670-677. Das, JC. (2004). Passive filters-potentialities and limitations. IEEE Transactions on Industry Applications, 40(1), 232-241. De, D, & Ramanarayanan, V. (2010). An active shunt compensator for reactive, unbalanced and harmonic loads under balanced and unbalanced grid voltage conditions. Paper presented at the 2010 Joint International Conference on Downloaded by [University of Malaya] at 19:50 21 January 2016 Power Electronics, Drives and Energy Systems (PEDES) & 2010 Power India. El-Habrouk, M, Darwish, MK, & Mehta, P. (2000). Active power filters: A review. IEE Proceedings-Electric Power Applications, 147(5), 403-413. Indu Rani, B, Aravind, CK, Saravana Ilango, G, & Nagamani, C. (2012). A three phase PLL with a dynamic feed forward frequency estimator for synchronization of grid connected converters under wide frequency variations. International Journal of Electrical Power & Energy Systems. Jianben, L., Qiaofu, C., & Jun, T. (2012). A novel DC voltage control strategy of series hybrid active power filter. International Journal of Electronics, 100(10), 14141428. doi: 10.1080/00207217.2012.743091 Kesler, Metin, & Ozdemir, Engin. (2011). Synchronous-reference-frame-based control method for UPQC under unbalanced and distorted load conditions. IEEE Transactions on Industrial Electronics, 58(9), 3967-3975. Mikkili, Suresh, & Panda, AK. (2011). RTDS Hardware implementation and Simulation of 3-ph 4-wire SHAF for Mitigation of Current Harmonics with pq and Id-Iq Control strategies using Fuzzy Logic Controller. International Journal of Emerging Electric Power Systems, be press, 12(5), 1-24. 22 Mikkili, Suresh, & Panda, AK. (2012). Real-time implementation of PI and fuzzy logic controllers based shunt active filter control strategies for power quality improvement. International Journal of Electrical Power & Energy Systems, 43(1), 1114-1126. Mekhilef, Saad,Tarek, Messikh, & Rahim, N (2011). Single-phase hybrid active power filter with adaptive notch filter for harmonic current estimation. IETE Journal of Research,57,20-8 Downloaded by [University of Malaya] at 19:50 21 January 2016 Park, T.-J., Jeong, G.-Y., & Kwon, B.-H. (2001). Shunt active filter for reactive power compensation. International Journal of Electronics, 88(12), 1257-1269. doi: 10.1080/00207210110092735 Patnaik, Sushree Sangita, & Panda, Anup Kumar. (2013). Real-time performance analysis and comparison of various control schemes for particle swarm optimization-based shunt active power filters. International Journal of Electrical Power & Energy Systems, 52, 185-197. Ponpandi, Rathika, & Durairaj, Devaraj. (2011). A Novel Fuzzy—Adaptive Hysteresis Controller Based Three Phase Four Wire-Four Leg Shunt Active Filter for Harmonic and Reactive Power Compensation. Journal of Energy and Power Engineering, 3, 422-435. Saad, S, & Zellouma, L. (2009). Fuzzy logic controller for three-level shunt active filter compensating harmonics and reactive power. Electric power systems research, 79(10), 1337-1341. Wu, J.-C. (2011). Utility-current feedforward-based control for a single-phase active power filter. International Journal of Electronics, 98(2), 185-196. doi: 10.1080/00207217.2010.506840 23 Salmeron, Patricio, & Litrán, Salvador P. (2010). A control strategy for hybrid power filter to compensate four-wires three-phase systems.IEEE Transactions on Power Electronics, 25(7), 1923-1931. Singh, GK, Singh, AK, & Mitra, R. (2007). A simple fuzzy logic based robust active power filter for harmonics minimization under random load variation. Electric Power Systems Research, 77(8), 1101-1111. Sinha, Rakesh Kumar, & Sensarma, Parthasarathi. (2011). A pre-filter based PLL for Downloaded by [University of Malaya] at 19:50 21 January 2016 three-phase grid connected applications. Electric Power Systems Research, 81(1), 129-137. Suresh, Mikkili, Panda, Anup Kumar, & Suresh, Y. (2011). Fuzzy controller based 3phase 4wire shunt active Filter for mitigation of current harmonics with combined pq and Id-Iq control strategies. Journal of Energy and Power Engineering, 3(1), 43-52. Venkatachalam, PA, Chandrasekaran, K, & Rama Rao, KS. (2009). Simulation model of Power Quality provider. Proceedings of the International Conference on Man-Machine Systems (ICoMMS),11-13 oct.,2009, Batu Ferringhi, Penang, MALAYSIA Younis, MAA,Rahim, NA, & Mekhilef, S,(2009). Distributed generation with parallel connected inverter,4th IEEE Conference on Industrial Electronics and Applications, 2009. ICIEA 2009. 2935-2940 24