shunt connected hybrid active power filters by using …

80

Int. J. Elec&Electr.Eng&Telecoms. 2016 Vaisali and P D V S K Kishore, 2016

SHUNT CONNECTED HYBRID ACTIVE POWERFILTERS BY USING DC-LINK VOLTAGE

CONTROLLERVaisali1* and P D V S K Kishore2

*Corresponding Author: Vaisali,[email protected]

Active filters are widely used in power system for reactive power compensation and voltage/current harmonic elimination. Harmonic contents of the source current has been calculated andcompared for the different cases to demonstrate the influence of harmonic extraction circuit onthe harmonic compensation characteristic of the shunt active power filter. In this paper, a fuzzylogic controlled, three-phase shunt active filter is used to improve power quality by compensatingcurrent harmonics which is required by a nonlinear load. However, when the LC-HAPF isperforming dynamic reactive power compensation, this dc-link voltage control scheme willinfluence the reactive power compensation, and thus, makes the LC-HAPF lack of success tocarry out dynamic reactive power compensation. In this paper, a novel dc-link voltage controlscheme for LC-HAPF is proposed so that the dc-link voltage control with start-up self-chargingprocess can be obtained as well as providing dynamic reactive power compensation.Representative simulation and experimental results of the three-phase four-wire center-spiltLC-HAPF are presented to verify all deductions, and also show the effectiveness of the proposeddc-link voltage control scheme in dynamic reactive power compensation.

Keywords: Active Power Filters (APFs), dc-link voltage control, Hybrid Active Power Filters(HAPFs), Passive Power Filters (PPFs), Reactive power control

INTRODUCTIONIn this modern society, domestic customers’appliances nor-mally draw large harmonic andreactive current from the sys-tem. Highharmonic current causes various problems inpower systems and consumer products, suchas overheating in equip-ment and transformer,

ISSN 2319 – 2518 www.ijeetc.comVol. 5, No. 3, July 2016

© 2016 IJEETC. All Rights Reserved

Int. J. Elec&Electr.Eng&Telecoms. 2016

1 Student, EEE Department, Narasaraopeta Engineering College, Narasaraopeta, India.2 Assistant Professor, EEE Department, Narasaraopeta Engineering College, Narasaraopeta, India.

blown capacitor fuses, excessive neutralcurrent, low power factor, etc. (Subjek andMcquilkin, 1990; and Duarte and Alves, 2002).On the other hand, load-ings with low powerfactor draw more reactive current than thosewith high power factor. The larger the reactivecurrent/power, the larger the system currentlosses and lower the network stability.

Research Paper

81


Thus, electrical utilities usually charge theindustrial and com-mercial customers a higherelectricity cost with a low power factor situation.

To eliminate the harmonic and reactivecurrent problems, application of power filtersis one of the most suitable solutions. Since thefirst installation of Passive Power Filters(PPFs) in the mid 1940s, PPFs have beenwidely used to suppress harmonic current andcompensate reactive power in distributionpower systems (Senini and Wolfs, 2002) dueto their low-cost, simplicity, and high-efficiencycharacteristics. However, they have manydisadvantages such as low dynamicperformance, filtering characteristics easily beaffected by small variations of the systemparameter values and resonance problems.

Since the concept “active ac power filter”was first devel-oped by Gyugyi in 1976 theresearch studies of the Active Power Filters(APFs) for current quality compensation havebeen prospered since then. Although APFsovercome the disadvantages inherent in PPFs,the initial and operational costs are relativelyhigh due to its high dc-link operating voltageduring inductive loading. This results in slowingdown their large-scale applications indistribution networks.

Later on, different Hybrid APF (HAPF)topologies composed of active and passiveparts in series and/or parallel have been pro-posed, in which the active part is a controllablepower electronic converter, and the passivepart is formed by RLC component. Thiscombination aims to improve thecompensation character-istics of PPFs andreduces the voltage and/or current ratings(costs) of the APFs, thus providing a cost-effective solution for compensating reactive

and harmonic current problems (Peng et al.,1990; Fujita and Akagi, 1991; Akagi, 1996;Park et al., 1999; Fujita et al., 2000; Rivaset al., 2003; Srianthumrong and Akagi, 2003;Inzunza and Akagi, 2005; Tangtheerajaroonwonget al., 2007; Jou et al., 2008; Luo et al., 2009a,2009b and 2009c; Rahmani et al., 2009; andSalmerón and Litrán, 2010). Among HAPFtopologies in Peng et al. (1990), Fujita andAkagi (1991), Akagi (1996), Park et al. (1999),Fujita et al. (2000), Rivas et al. (2003), Luoet al. (2009a, 2009b and 2009c) and Salmerónand Litrán (2010), a transformerless LC cou-pling HAPF (LC-HAPF) has been proposedand applied recently for current qualitycompensation and damping of harmonicpropagation in distribution power system, inwhich it has less passive components andlower dc-link operating volt-age comparingwith an APF.

In addition, LC-HAPF is normally designedto deal with harmonic current rather thanreactive power compensation, the inverter partis responsible to compensate har-moniccurrents only and the passive part provides afixed amount of reactive power. In practicalcase, the load-side reactive powerconsumption usually varies from time to time,and if the load-ing mainly consists of inductionmotors such as centralized an air-conditioningsystem, its reactive power consumption will bemuch higher than the harmonic powerconsumption (Tai et al., 2007). As a result, it isnecessary for the LC-HAPF to performdynamic reactive power compensationtogether with harmonic current compensation.

All LC-HAPF and other HAPFs discussedin Peng et al. (1990), Fujita and Akagi (1991),Akagi (1996), Park et al. (1999), Fujita et al.

82


(2000), Rivas et al. (2003), Srianthumrong andAkagi (2003), Inzunza and Akagi (2005),Tangtheerajaroonwong et al. (2007), Jou et al.(2008), Luo et al. (2009a, 2009b and 2009c),Rahmani et al. (2009) and Salmerón andLitrán (2010) are based on fixed dc-link voltagereference. For the purpose of reducingswitching loss and switching noise, an adaptivedc-link voltage reference control is proposedin Lam et al. (2012). However, the au-thor didnot discuss the dc-link voltage control in detailsand also the inherent influence between thereactive power compen-sation and dc-linkvoltage controls. Moreover, this influence hasnot been discussed.

Due to the limitations among the existingliteratures, this pa-per aims to investigate andexplore different dc-link voltage con-troltechniques for the three-phase four-wire LC-HAPF while performing dynamic reactivepower compensation:

1) By using the traditional dc-link voltagecontrol scheme as an active currentcomponent, an extra start-up dc-linkprecharging control process is necessary.

2) To achieve the start-up dc-link voltage self-charging func-tion, a dc-link voltage controlas a reactive current com-ponent for the LC-HAPF is proposed (Cui et al., 2011);however, the LC-HAPF with this dc-linkvoltage control scheme fails to providedynamic reactive power compensation.

3) A novel dc-link voltage control method isproposed for achieving start-up dc-link self-charging function, dc-link voltage control,and dynamic reactive power compensation.Moreover, the proposed method can beapplied for the adaptive dc-link voltage

reference control as discussed in Lamet al. (2012).

In this paper, the designed coupling LC isbased on the aver-age value of the loadingreactive power consumption, while thedesigned dc-link voltage level is based on theLC-HAPF maximum reactive powercompensation range specification. Therefore,even though the reactive power compensatingrange is small with a low dc-link voltage, theLC-HAPF can still provide dynamic reactivepower compensation. Thus, its reactive powercompensation ability (within its specification)is still effective. Given that most of the loads inthe distribution power systems are inductive,the following analysis and discussion will onlyfocus on inductive loads (Lam et al., 2008).

A TRANSFORMERLESS TWO-LEVEL THREE-PHASEFOUR-WIRE CENTER-SPILTLC-HAPFCircuit Structure of LC-HAPFThe circuit structure of a transformerless three-phase four-wire center-spilt LC-HAPF is shownin Figure 1, where subscript “x” denotesphases a, b, c, and n. vsx is the source voltage,vx is the load voltage, Ls is the source inductorand normally neglected due to its low valuerelatively; thus, vsx”—,vx· isx, iLx, and icx are thesource, load, and compensating current foreach phase. Cc and Lc are the coupling partcapacitor and inductor for each leg of theconverter. CdcU, CdcL, VdcU, and VdcL are theupper and lower dc-link capacitor and dc-linkcapacitor voltages, and the dc-link voltage Vdc

= VdcU + VdcL. Since the transformerless two-level three-phase four-wire center-spilt LC-HAPF can be treated as three independent

83


single-phase circuits, its single-phaseequivalent circuit model is shown in Figure 2.

Modeling of the DC-Link Voltage ina LC-HAPF Single-Phase EquivalentCircuitFrom Figure 2, the compensating current icxcan flow either CdcU orCdcL and returns throughthe neutral wire in both directions of Insulated-Gate Bipolar Transistors (IGBT) switches. Thedc-link capacitor voltages can be expressedas

...(1)

where idcU and idcL are the dc currents of upperand lower dc-link capacitors, respectively, and

idcU = sxicx, idcL = (1 – sx)icx ...(2)

Substituting (2) into (1), the completedupper and lower dc capacitor voltages VdcU andVdcL become

...(3)

...(4)

In (3) and (4), sx represents the switchingfunction of one inverter leg in x phase basedon the hysteresis current Pulse WidthModulation (PWM) method, and that is thebinary state of the upper and lower switches(Tx and Tx). When the positive di-rection of icx

is assumed as Figure 2, the switching logicfor each phase is formulated as Singh andVerma (2005): if icx > (i”cx + hb), Tx is ON andTx is OFF, then sx = 1; if icx < (i”cx – hb), Tx isOFF and Tx is ON, then sx = 0, where i”cx is thereference compensating current and hb is thewidth of the hysteresis band.

According to the mathematical model in (3),if compensat-ing current icx > 0, the upper dc-link voltage VdcU is increased during switchingfunction sx = 1, while the lower dc-link voltageVdcL is decreased for sx = 0. The inverseresults can be obtained if icx < 0. Figure 3shows the operation of a single-phase voltagesource inverter under different switchingmodes by using hysteresis current PWMmethod (Aredes et al., 1997). The changes ofthe upper and lower dc-link voltages VdcU and

Figure 1: Circuit Structure of aTransformerless Three-Phase Four-Wire

Center-Spilt LC-HAPF

Figure 2: Single-Phase Equivalent CircuitModel of a Three-Phase Four-Wire

Center-Spilt LC-HAPF

84


Figure 3: Operation of a Single-Phase Voltage Source Inverter Under Different SwitchingModes by Using Hysteresis Current PWM Method

Table 1: The Change of the Capacitor Voltages (VdcU, VdcL) Under Different Modes

VdcL under different modes are summarized inTable 1.

INFLUENCE ON DC-LINKVOLTAGE DURING LC-HAPFPERFORMS REACTIVEPOWER COMPENSATION

This section aims to present and analyzethe influence on the dc-link voltage when LC-HAPF performs reactive power compensation.Through this analysis, the dc-link capacitorvoltage will either be increased or decreasedduring fundamental reactive powercompensation under insufficient dc-link

85


voltage. Moreover, the influence is proportionalto the difference between the LC-HAPFcompensating current icx and its pure reactivereference i”cxf q , where the subscript “f,” “p,”and “q” denote fundamental, active, andreactive components.

Reactive Power CompensationUnder Sufficient DC-Link VoltageUnder sufficient dc-link voltage, thecompensating current generated by the LC-HAPF icx can track its pure reactive refer-encei*cxf q, as shown in Figure 4, thus the amplitudeI*

cxfq ≈ Icx. Provided that the hysteresis band issmall enough for the LC-HAPF to be operatedat linear region, the PWM switching functionin (4) will be evenly distributed, as shown inFigure 4. According to Table 1, the LC-HAPFwill be changed between the operating modesof rectifier and inverter (modes a, b, c, and d),and keeping the average dc-link voltage as aconstant. In ideal lossless case, the dc-link

voltage will not be affected when LC-HAPFperforms reactive power compensation during

Icxfq ≈ Icx case.

Reactive Power CompensationUnder Insufficient DC-Link VoltageUnder insufficient dc-link voltage, thecompensating reactive current generated bythe LC-HAPF icx cannot track its pure reactivereference i*cxfq, in which there are two possiblesituations: 1) If the amplitude Icx > I*

cxfq; thus,the instantaneous relationship gives icx > i*cxfq

during icx > 0, and icx < i*cxfq during icx < 0, asshown in Figure 5; 2) If the amplitude Icx < I*

cxfq,the opposite instantaneous relationship isgiven as shown in Figure 6.

Hence, to force the compensating currenticx to track its reference icxfq correspondingly:

• For Icx > I*cxfq case, as shown in Figure 5,

when the hystere sis band is relatively small

Figure 4: icx and i*cxfq during icx ≈ i*cxfqCase and the Corresponding Switching

Function

Figure 5: icx and i*cxfq During icx > i*cxfqCase and the Corresponding Switching

Function

86


compared with the difference between icx

and i*cxfq, their instantaneous relationshipbetween icx and i*cxfq can be expressed asfollows:

...(5)

...(6)

Based on the hysteresis PWM technique,the switching functions of (5) and (6) are

...(7)

...(8)

According to Table 1, the PWM switchingsequences in (7) and (8) drive the LC-HAPFto be operated in rectifier mode (modes b andd), thus increasing the average dc-linkcapacitor voltage. When the dc-link voltage isincreased to sufficient high level, the LC-HAPFwill change the operating mode from Icx > I*

cxfq

into Icx ≈ I*cxfq, and keeping the dc-link voltage

value. Therefore, for a given fundamentalreactive current ref-erence i*cxfq (Icx > I*

cxfq), thedc-link voltage of LC-HAPF will be self-increased to a voltage level that lets thereference compensating current can betracked.

For Icx < I*cxfq case, as shown in Figure 6,

the instantaneous relationship between icx andi*cxfq is

...(9)...(10)

The switching functions of (9) and (10) are

...(11)

...(12)

According to Table 1, the PWM switchingsequences in (11) and (12) drive the LC-HAPF to be operated in inverter mode (modesa and c), thus decreasing the average dc-linkcapacitor voltage.

LC-HAPF OPERATION BYCONVENTIONAL DC-LINKVOLTAGE CONTROLMETHODSDC-Link Voltage Control Method asActive Current ComponentTraditionally, if the indirect current (voltagereference) PWM control method is applied, thedc-link voltage of the inverter is controlled bythe reactive current component feedbacksignal. However, when the direct current PWMcontrol method is applied in the dc-link voltageshould be controlled by the active currentcomponent feedback signal. Both dc-linkvoltage control methods are equivalent to eachother.

When LC-HAPF performs reactive power

Figure 6: icx and i*cxfq During icx < i*cxfqCase and the Corresponding Switching

Function

87


compensating, the reference compensatingcurrent i*cx is composed by

...(13)

However, to perform the reactive power anddc-link voltage control action in (13), a sufficientdc-link voltage should be provided to let thecompensating current icx track with itsreference i*cx. As a result, this conventional dc-link voltage control method fails to control thedc-link voltage during insufficient dc-linkvoltage, such as during start-up process. Dueto this reason, when this dc-link voltage controlmethod is applied in LC-HAPF, an extra start-up precharging control process is nec-essary.Usually, a three-phase uncontrollable rectifieris used to supply the initial dc-link voltagebefore operation. Moreover, when theadaptive dc-link voltage control idea is ap-plied, the reference dc-link voltage may bechanged from a low level to a high level; at thatoccasion, the dc-link voltage may beinsufficient to track the new reference value.Therefore, the conventional dc-link voltagecontrol with precharging method may not workproperly in the adaptive dc-link voltagecontrolled system.

The LC-HAPF in showed that this dc-linkvoltage control can achieve start-up self-charging function without any external supply.That is actually due to the LC-HAPF in whichis initially operating at Icx > I*cxf q condition.According to the analysis in Section III, duringIcx > I*

cxfq condition, the dc-link voltage will beself-charging to a sufficient voltage level thatlets the LC-HAPF’s compensating currenttrack with its reference i*cxfq ≈ icx. Thus, the dc-link voltage can be maintained as its referencevalue in steady state. However, if the LC-HAPFis initially operating at Icx < I*

xfq condition, this

dc-link voltage control method fails c to carryout this function.

Figures 7 and 8 show the simulation resultsof LC-HAPF start-up process by using theconventional dc-link voltage control methodunder different cases. From Figure 7, whenLC-HAPF starts operation during Icx > I*

cxfq

condition, the dc-link voltage Vdc can carry outthe start-up self-charging function and i*cx ≈ icx

in steady state. On the contrary, from Figure 8,neither the dc-link voltage nor the reactivepower compensa-tion can be controlled whenLC-HAPF is operating during Icx < I*cxfq

condition, in which the simulation resultsverified the previous analysis.

DC-Link Voltage Control Method asReactive Current ComponentAccording to the analysis in Section III, the dc-link voltage can be influenced by varying thereference compensating fundamental reactivecurrent i*cxfq under insufficient dc-link voltage;thus, the dc-link voltage can also be controlled

Figure 7: Vdc, i*cx, and icx During Icx > I*cxfqCondition with Conventional dc-Link

Voltage Control Method as Active CurrentComponent

88


I*cxfq. Moreover, it has been concluded that for

a given reference i*cxfq(Icx > I*cxfq), the dc-link

voltage will be self-charging to a voltage level,then icx can trend its reference i*cxfq gradually,and maintaining this voltage level in steadystate. Therefore, to control the dc-link voltageVdc as its reference V*dc, it will have acorresponding fixed reference value, that is,I*

cxfq = I*cxfq fixed.

Therefore

...(15)

Equation (15) implies that to control the dc-link voltage, the LC-HAPF must restrict toprovide a fixed amount of reactive power.Therefore, by using the dc-link voltage controlmethod as reactive current component, theLC-HAPF fails to perform dynamic reactivepower compensation, in which thecorresponding simulation and experimentalresults will be given in Section VI.

PROPOSED DC-LINKVOLTAGE CONTROL METHODFrom the previous analysis, the dc-link voltagecontrol method as active current componentis effective only under sufficient dc-link voltage.In an insufficient dc-link voltage case, such asthe LC-HAPF start-up process, both dc-linkvoltage control and reactive powercompensation may fail. On the contrary, whenthe dc-link voltage control method as reactivecurrent component is applied, the dc-linkvoltage control can be effectively controlledwith start-up self-charging function. However,by using this control method, the LC-HAPF failsto perform dynamic reactive powercompensation. As a result, a novel dc-linkvoltage control method by both reactive and

by utilizing this influence. Hence, this dc-linkvoltage control method can achieve the start-up self-charging function. Moreover, it has beenshown that this method is more effective thanthat conventional method as active currentcomponent. However, when it is being appliedin LC-HAPF, although the dc-link voltage canbe controlled as its reference value, the LC-HAPF cannot perform dynamic reactive powercompensation. The corresponding reason canbe derived as follows.

By using the dc-link voltage control methodas reactive current component, the referencecompensating current i*cx is composed by

...(14)

where i*cxfq dc is the dc-link voltage controlsignal related to reactive current component.

Based on the analysis in Section III, underinsufficient dc-link voltage, the change of thedc-link voltage is directly proportional to thedifference between the amplitude of Icx and

Figure 8: Vdc, i*cx, and icx During Icx < I*cxfqCondition with Conventional dc-Link

Voltage Control Method as Active CurrentComponent

89


active current components as the feedbacksignals is proposed in this section, so as tocombine the advantages of both methods,which can achieve the start-up self-chargingfunction, dc-link voltage control, and dynamicreactive power compensation.

...(16)

...(17)

where Qdc andPdc are the dc control signalsrelated to the reactive and active currentcomponents and kq and kp are thecorresponding positive gains of the controller.By using the proposed control method, thereference compensating current i*cx iscalculated by

...(18)

where i* = i*Lxfq + i*

cxfqcxf q d c.

In (17), Pdc represents the active powerflow between the source and the LC-HAPFcompensator, Pdc > 0 means the LC-HAPFabsorbs active power from the source, andPdc < 0 means the LC-HAPF injects activepower to the source. According to the analysisin Section III, the dc-link voltage will be chargedfor Icx > I*cxfq, and discharged for Icx < I*cxfq duringperforming reactive power compensation.When V*

dc – Vdc > 0, in order to increase thedc-link voltage, I*

cxfq should be decreased byadding a negativeQdc. On the contrary, whenV*dc – Vdc < 0, in order to decrease the dc-linkvoltage, I*cxfq should be increased by adding apositiveQdc. Therefore, the “_” sign is addedin (16). In this paper, in order to simplify thecontrol process, Qdc and Pdc in (16) and(17) are calculated by the same controller, i.e.,kq = kp.

In (18), i*Lxfq, i*cxfq dc and i*cxfp dc are

calculated by using the three-phaseinstantaneous p – q theory

Control action in (18). During this period,the dc-link voltage will be self-charging underIcx > I*

cxfq condition. As the dc-link voltage isincreased and approaching the referencevalue, the compensating current trackingability of the LC-HAPF will be improvedgradually, and the control signals Qdc andPdc in (16) and (17) will also be decreasedgradually; thus, i*cx≈ icx eventually. Accordingto the analysis in Section III, the dc-linkvoltage will not be affected by the reactivecomponent when i”cx ≈ icx. Hence, during thisperiod, the dc-link voltage control signal asactive current component will dominate thecontrol action in (18). Therefore, theproposed dc-link voltage control method canbe realized as:

90


• Qdc control signal is used to step changethe dc-link voltage under insufficient dc-linkvoltage that can be effectively applied forstart-up process and the adaptive dc-linkvoltage control;

• Pdc control signal is used to maintain thedc-link voltage under sufficient dc-linkvoltage.

By using the proposed method, the LC-HAPF can compensate the reactive powerconsumed by the load, and keep the dc-linkvoltage as its reference one as shown inFigure 9 (Qsxf ≈ 0 var, VdcU = VdcL = V*

dc/2 = 30V). There-fore, the proposed dc-link voltagecontrol method can effectively control the dc-link voltage without any extra prechargingprocess and lets the LC-HAPF providedynamic reactive power compensation. Table2 shows the comparison among theconventional and proposed dc-link voltagecontrol methods error. Since the parameterdesign of the dc-link voltage controller is notthe main theme of this paper, a pure pro-portional controller with an appropriate valueis selected. A limiter is applied to avoid theoverflow problem. After that, the final dc-linkvoltage control reference currents i*cxfq dc and

Figure 9: Control Process of the Proposeddc-link Voltage Control Method

Table 2: Comparison Between Conventional and Proposed dc-Link Control Methods

i*cxfp dc can be calculated, and they will be sentto current PWM control block to perform thedc-link voltage control.

Then, the final reference compensatingcurrent i*cx can be obtained by summing up thei*Lxfq ,i*cxfq dc, and i*cxfp dc. Then i*cx together withcompensating current icx will be sent to thecurrent PWM control part for generating PWMtrigger signals to control the power electronicswitches of the inverter.

91


Figure 10: Circuit Diagram

Figure 11: Input Voltage and Current

SIMULATION RESULTS

92


CONCLUSIONThe shunt APF is implemented with three-phase supply controlled with voltage sourceinverter and is connected at the point ofcommon coupling for filtering the currentharmonics. The VSI switching signals arederived from hysteresis band current controller.The hysteresis controller changes thebandwidth based on instantaneouscompensation current variation. The proposedcontroller based shunt active power filterperforms perfectly for mitigate the harmonicsThrough the proposed dc-link voltage controlmethod:

1) The LC-HAPF can achieve start-up dc-linkself-charging function;

2) The dc-link voltage of the LC-HAPF can becontrolled as its reference level;

3) The LC-HAPF can provide dynamicreactive power compensation;

4) The adaptive dc-link voltage referencecontrol can be implemented.

Finally, simulation results of the three-phasefour-wire center-spilt LC-HAPF under dynamicreactive power compensation application arepresented to verify all discussions andanalysis, and also show the effectiveness ofthe proposed dc-link voltage control method.

REFERENCES1. Akagi H (1996), “New Trends in Active

Filters for Power Conditioning”, IEEETrans. Ind. Appl. , Vol. 32, No. 6,pp. 1312-1322.

2. Aredes M, Hafner J and Heumann K(1997), “Three-Phase Four-Wire ShuntActive Filter Control Strategies”, IEEETrans. Power Electron., Vol. 12, No. 2,pp. 311-318.

3. Cui X-X, Lam C-S and Dai N-Y (2011),

Figure 12: Load Currents

93


“Study on dc Voltage Control of HybridActive Power Filters”, in Proc. 6th IEEEConf. Ind. Electron. Appl. (ICIEA), June,pp. 856-861.

4. Duarte L H S and Alves M F (2002), “TheDegradation of Power Capacitors Underthe Influence of Harmonics”, in Proc. IEEE10th Int. Conf. Harmonics Quality Power,Vol. 1, October, pp. 334-339.

5. Fujita H and Akagi H (1991), “A PracticalApproach to Harmonic Compensation inPower Systems—Series Connection ofPassive and Active Filters”, IEEE Trans.Ind. Appl., Vol. 27, No. 6, pp. 1020-1025.

6. Fujita H, Yamasaki T and Akagi H (2000),“A Hybrid Active Filter for Damping ofHarmonic Resonance in Industrial PowerSystems”, IEEE Trans. Power Electron.,Vol. 15, No. 2, pp. 215-222.

7. Inzunza R and Akagi H (2005), “A 6.6-kVTransformerless Shunt Hybrid Active Filterfor Installation on a Power DistributionSystem”, IEEE Trans. Power Electron.,Vol. 20, No. 4, pp. 893-900.

8. Jou H-L, Wu K-D, Wu J-C, Li C-H andHuang M-S (2008), “Novel PowerConverter Topology for Three Phase Four-Wire Hybrid Power Filter”, IET PowerElectron., Vol. 1, pp. 164-173.

9. Lam C-S, Choi W-H, Wong M-C and HanY-D (2012), “Adaptive dc-link VoltageControlled Hybrid Active Power Filters forReactive Power Compensation”, IEEETrans. Power Electron., Vol. 27, No. 4,pp. 1758-1772.

10. Lam C-S, Wong M-C and Han Y-D (2008),“Voltage Swell and Overvoltage

Compensation with Unidirectional PowerFlow Controlled Dynamic VoltageRestorer”, IEEE Trans. Power Del.,Vol. 23, No. 4, pp. 2513-2521.

11. Luo A, Shuai Z K, Shen Z J, Zhu W J andXu X Y (2009b), “Design Considerationsfor Maintaining dc-side Voltage of HybridActive Power Filter with Injection Circuit”,IEEE Trans. Power Electron., Vol. 24,No. 1, pp. 75-84.

12. Luo A, Tang C, Shuai Z K, Zhao W,Rong F and Zhou K (2009c), “A NovelThree-Phase Hybrid Active PowerFilter with a Series Resonance CircuitTuned at the Fundamental Frequency”,IEEE Trans. Ind. Electron., Vol. 56,No. 7, pp. 2431-2440.

13. Luo A, Zhao W, Deng X, Shen Z JandPeng J-C (2009a), “DividingFrequency Control of Hybrid Active PowerFilter with Multi-injection Branches UsingImproved ip – iq Algorithm”, IEEE Trans.Power Electron. , Vol. 24, No. 10,pp. 2396-2405.

14. Park S, Sung J-H and Nam K (1999), “ANewparallel Hybrid Filter ConfigurationMinimizing Active Filter Size”, in Proc.IEEE 30th Annu. Power Electron. Spec.Conf. (PESC), Vol. 1, pp. 400-405.

15. Peng F Z, Akagi H and Nabae A (1990),“A New Approach to HarmonicCompensation in Power Systems—ACombined System of Shunt Passive andSeries Active Filters”, IEEE Trans. Ind.Appl., Vol. 26, No. 6, pp. 983-990.

16. Rahmani S, Hamadi A, Mendalek N andAl-Haddad K (2009), “A New Control

94


Technique for Three-Phase Shunt HybridPower Filter”, IEEE Trans. Ind. Electron.,Vol. 56, No. 8, pp. 2904-2915.

17. Rivas D, Moran L, Dixon J W andEspinoza J R (2003), “Improving PassiveFilter Compensation Performance withActive Techniques”, IEEE Trans. Ind.Electron., Vol. 50, No. 1, pp. 161-170.

18. Salmerón P and Litrán S P (2010), “AControl Strategy for Hybrid Power Filterto Compensate Four-Wires Three-PhaseSystems”, IEEE Trans. Power Electron.,Vol. 25, No. 7, pp. 1923-1931.

19. Senini S T and Wolfs P J (2002),“Systematic Identification and Review ofHybrid Active Filter Topologies”, in Proc.IEEE 33rd Annu. Power Electron. Spec.Conf. (PESC), Vol. 1, pp. 394-399.

20. Singh B and Verma V (2005), “An IndirectCurrent Control of Hybrid Power Filter forVarying Loads”, IEEE Trans. Power Del.,Vol. 21, No. 1, pp. 178-184.

21. Srianthumrong S and Akagi H (2003), “AMedium-Voltage Transformerless ac/dcPower Conversion System Consisting ofa Diode Rectifier and a Shunt HybridFilter”, IEEE Trans. Ind. Appl., Vol. 39,No. 3, pp. 874-882.

22. Subjek J S and Mcquilkin J S (1990),“Harmonics-Causes, Effects, Measurementsand Analysis”, IEEE Trans. Ind. Electron.,Vol. 26, No. 6, pp. 1034-1042.

23. Tai S-U, Wong M-C, Dong M-C and HanY-D (2007), “Some Findings on HarmonicMeasurement in Macao”, in Proc. 7th Int.Conf. Power Electron. Drive Syst.(PEDS), pp. 405-410.

24. Tangtheerajaroonwong W, Hatada T,Wada K and Akagi H (2007), “Designand Performance of a TransformerlessShunt Hybrid Filter Integrated into aThree-Phase Diode Rectifier”, IEEETrans. Power Electron., Vol. 22, No. 5,pp. 1882-1889.

shunt connected hybrid active power filters by using …

Documents