Compared with AC-DC-AC converter, matrix converter (MC) has several advantages for its bidirectional power flow, controllable power factor, and the absence of large energy storage in dc-link. The topology of MC includes direct matrix converter (DMC) and indirect matrix converter (IMC). IMC has received great attention worldwide because of its easy implementation and safe commutation. Space vector PWM (SVM) algorithm for indirect matrix converter is realized on DSP and CPLD platform in this paper. The control of the rectifier and inverter in IMC can be decoupled because of the intermediate dc-link. The space vector modulation scheme for IMC is discussed and the PWM sequences for the rectifier and inverter are generated. And a two-step commutation of zero current switching (ZCS) in the rectifier is achieved. Input power factor of IMC can be changed by adjusting the angle of the reference current vector. Experimental tests have been conducted on a RB-IGBT based indirect matrix converter prototype. The results verify the performance of the SVM algorithm and the ability of power factor correction.
1. Introduction
AC/AC conversion has two main types: AC-DC-AC converters and direct AC/AC converters. AC-DC-AC converters include the current and voltage source topologies with a capacitor or inductor in the dc-link. Direct AC/AC converters include the cycloconverter of high-power applications and matrix converters in low power range [1]. Matrix converter has a compact structure without capacitor or inductor, which has received much attention in the academic and industrial fields worldwide in recent years [2–6]. Since Alesina and Venturini proposed a control method for matrix converter using a mathematical approach to generate the desired output voltage [7], the PWM algorithms for matrix converters become a research focus. New PWM strategies for matrix converters have been proposed in recent years [8–13]. Some researcher concentrated on the realization of the PWM schemes, such as dipolar modulation technique [8], SHE-PWM [9], and beatless synchronous PWM [10]. The PWM strategies for matrix converter are also proposed to reduce the ripple in the input and output waveforms [11, 12]. In the high-power applications, indirect SVPWM for multimodular matrix converters is realized [13].
The topologies of matrix converter, including direct matrix converter and indirect matrix converter [14–16], are another research focus in AC/AC conversion. Compared with the former one, IMC has several advantages. (1) With a real dc-link, the converter consists of two stages, a rectifier and an inverter. The control of the two stages can be decoupled and the control difficulty is lowered down. (2) With appropriate modulation method, zero-current switching commutation in the rectifier can be simply accomplished. The system efficiency is improved. (3) The number of power switches can be further reduced. (4) With the dc-link, multioutput IMC can be realized [17, 18]. The topology of IMC is shown in Figure 1. The bidirectional switch in the rectifier stage is composed by two RB-IGBTs.
Topology of indirect matrix converter using RB-IGBT.
The PWM strategies in [7–13] are proposed for DMC. In [19–22], the conventional PWM algorithm, based on two larger source line voltages, is proposed. The PWM period of the rectifier stage is divided in two segments to reduce the switching times and the harmonics of the input current. But these PWM strategies have no zero vectors. In [23–26], carrier-PWM realization and overmodulation method for IMC are discussed.
In the applications of indirect matrix converter, [27] proposes IMC to control doubly fed induction generation for wind power system. IMC is used to control a microturbine generator [28]. A multimodular indirect matrix converter topology is proposed to be a frequency converter used in the medium-voltage motion drive systems [29]. And three level neutral-point-clamped indirect matrix converters are proposed in [30].
In this paper, the rectifier is controlled as a current-source converter and inverter is controlled as a voltage-source converter. SVM is applied for the rectifier and SVPWM is applied for the inverter. The PWM sequence for indirect matrix converter is achieved. Two-step commutation of zero-current switching for the rectifier is presented. The principle of the input power factor correction for IMC is discussed. The design and implementation of indirect matrix converter prototype using RB-IGBT are introduced. The experimental results are shown to verify the performance of the proposed PWM algorithm and the ability of PFC.
2. Space Vector Modulation of IMC2.1. PWM Algorithm of IMC
The control of the rectifier and inverter for IMC can be decoupled because of the dc-link. The rectifier is controlled as a current-source converter, and the inverter can be seen as an inductive load in dc-link, shown in Figure 2.
Control principle of the rectifier.
Three phase input currents can be defined as follows:(1)ij=Sj·idcj=a,b,c.
In (1), Sj is the switch function. Sj can be 1, 0, and −1, which are shown in Figure 3. The switching functions of the rectifier are shown in Table 1.
Switching functions of the rectifier.
Number
Sa
Sb
Sc
1
1
0
−1
2
0
1
−1
3
−1
1
0
4
−1
0
1
5
0
−1
1
6
1
−1
0
7
0
0
0
The statues of Sa.
Sa=1
Sa=-1
Sa=0
Current space vector can be defined as follows:(2)I=ia+ib·ej2π/3+ic·e-j2π/3.
From (1) and (2), (3)I=idcSa+Sb·ej2π/3+Sc·e-j2π/3.
The principle of space vector modulation (SVM) is shown in Figure 4. Iref is the reference current space vector. θin is the angle of Iref.
Current space vectors of the rectifier.
According to the principle of space vector PWM, (4)t1=mrecTssinπ/2-θinsinπ/3,t2=mrecTssinθin-π/6sinπ/3,t0=Ts-t1-t2,(5)d1=t1Ts,d2=t2Ts,d0=t0Ts.
In (4), mrec is modulation index for the rectifier, which can be defined as (6)mrec=IrefIdc.
In order to ensure the sinusoidal input current, the modulation index mrec cannot be larger than 1,(7)mrec⩽1.
In order to maintain the input power factor as unity, phase angle of Iref, θin should follow the phase angle of input space vector voltage Uin, so the voltage of dc-link is as follows:(8)Udc=mrec·Uin,(9)Uin=ua+ub·ej2π/3+uc·e-j2π/3.
In (9), ua, ub, and uc are the three phase voltages of input sides. In order to obtain the maximum voltage ratio of the rectifier, mrec should be 1.
The inverter is controlled as voltage source converter, and the rectifier can be seen as a voltage source in dc-link. The conventional SVPWM algorithm can be used in the inverter, shown in Figure 5:(10)T1=minvTssinπ/3-θsinπ/3,T2=minvTssinθsinπ/3,T0=Ts-T1-T2,D1=T1Ts,D2=T2Ts,D0=T0Ts,minv=UrefUdc.
SVPWM for inverter when 0<θ≤π/3.
In (10), Ts is the period of a switching cycle. minv is the modulation index of inverter. D0, D1, and D2 are the duty cycles of the zero vectors, vectors V1 and V2, respectively. Uref is the reference space voltage vector and Udc is the voltage of dc-link.
From (8) and (9), when the input voltages are symmetrical, the voltage of dc-link is constant. The modulation index of inverter is determined by the reference output space voltage vector. The calculation of duty cycles D0, D1, and D2 will be simple.
2.2. Commutation and PWM Sequence
Generally, a bidirectional switch in matrix converters is constructed by two IGBTs and two diodes or two RB-IGBTs. Commutations between two bidirectional switches need four steps with voltage/current signal. Figure 6 is the process of four-step commutation with voltage signal. tc0, tc1, tc2, and tc3 are the times for four-step commutation.
Four-step commutation for bidirectional switches of RB-IGBT with voltage signal.
In IMC, the inverter is a conventional voltage source converter and its commutation is traditional. When the inverter is on zero-vector state, the current of dc-link is zero. The commutation of the bidirectional switches in the rectifier can be achieved during this period, and zero-current switching commutation can be realized. The PWM sequence of IMC is shown in Figure 7 and ZCS commutation is shown in Figure 8. tc0 and tc1 are the switching times for two-step commutation.
PWM sequence for IMC.
Two-step commutation for bidirectional switches of ZCS.
Compared with the conventional PWM algorithm [19–22], the proposed PWM algorithm increases one switching action in rectifier, which is a ZCS process. ZCS commutation is much simpler than four-step commutation and it needs no voltage or current signals. So voltage or current sensor in dc-link can be omitted.
2.3. Realization of SVM Algorithm and Commutation Method
In order to realize the SVM algorithm and commutation strategy, a control platform based on DSP and CPLD is implemented, shown in Figure 9. DSP TMS320F2812 of TI is to realize SVM algorithm and CPLD EPM9320LC84-15 of Altera is to realize the commutation strategy for bidirectional switches in rectifier.
Control platform based on DSP and CPLD.
The input sectors Sci1~Sci6 and output sectors Svo1~Svi6 are determined with the angle of input voltage and output voltage with DSP. And 8 PWM signals, which are shown in Figure 10, are generated by SVM and SVPWM algorithms with DSP. The 8 PWM signals are not the drive signals for the IGBTs. A decoder for drive signals is needed. CPLD is used to decode drive signals from PWM sequences and realize two-step ZCS commutation. With the PWM signals PWM1–8, input sectors, and output sectors, the drive signals for the bidirectional switches can be shown as expression (11), where Si+ is the upper bridge and Si- is the lower bridge, i=a,b,c.
PWM segments generated by DSP.
According to the SVM algorithm for rectifier, the commutation of bidirectional switches occurs between the upper bridges (Sa+,Sb+,Sc+) and lower bridges (Sa-,Sb-,Sc-). From Figure 4, the input current vector I1(1,0,-1) will change to I2(0,1,-1), and the commutation is Sa+ to Sb+. The commutation process can be described by the state of RB-IGBTs in the rectifier Sa+Sa-Sb+Sb-Sc+Sc-. The process of commutation is 100001-000001-001001. Because the commutation is zero-current switching, the commutation of rectifier is the same as inverter. Only the dead time is needed:(11)Sa+=Sci1+Sci6·1⊕PWM2+Sci2·PWM1+Sci4·PWM1⊕PWM2,Sa-=Sci4+Sci3·1⊕PWM2+Sci5·PWM1+Sci1·PWM1⊕PWM2,Sb+=Sci3+Sci2·1⊕PWM2+Sci4·PWM1+Sci6·PWM1⊕PWM2,Sb-=Sci6+Sci5·1⊕PWM2+Sci1·PWM1+Sci3·PWM1⊕PWM2,Sc+=Sci5+Sci4·1⊕PWM2+Sci6·PWM1+Sci2·PWM1⊕PWM2,Sc-=Sci2+Sci1·1⊕PWM2+Sci3·PWM1+Sci5·PWM1⊕PWM2.
3. Power Factor Control of IMC
According to the SVM algorithm, power factor of IMC can be controlled by changing the angle of the reference input current vector, shown in Figure 11. The phase difference of current vector and voltage vector is φ and power factor is cosφ.
The principle of power factor control.
It is assumed that, on the input side,(12)ua=Umcosωit,ub=Umcosωit-2π3,uc=Umcosωit+2π3.
And the reference input currents are(13)ia=Imcosωit-φ,ib=Imcosωit-2π3-φ,ic=Imcosωit+2π3-φ.
The active power on input side is(14)P=uaia+ubib+ucic=32UmImcosφ.
From (14), φ is power angle. When φ>0, IMC absorbs reactive power. When φ<0, IMC generates reactive power.
From Figure 1, the inverter of IMC is voltage source converter. The voltage of dc-link upn should be positive. Otherwise, the reverse recovery diode for IGBT will be short-circuited and the semiconductor may be damaged.
The sectors of input current space vector can be divided by the angle of input voltage space vector, which is shown in Table 2.
Sectors of input current space vector.
Input sector
Phase angle ωit
1
[-π/6,π/6)
2
[π/6,π/2)
3
[π/2,5π/6)
4
[5π/6,7π/6)
5
[7π/6,3π/2)
6
[3π/2,11π/6)
According to Table 2, the corresponding relationship of input sector and input voltage interval is shown in Figure 12. Input sector I–VI is corresponding with voltage interval 1–6. In order to maintain the DC-link voltage being positive, the relationship of line voltage and phase angle in each sector is shown in Table 3.
Relationship of line voltage and phase angle.
Sector
Line-line voltage
Phase angle ωit
1
uab≥0
uac≥0
[-π/3,π/3)
2
uac≥0
ubc≥0
[0,2π/3)
3
ubc≥0
uba≥0
[π/3,π)
4
uba≥0
uca≥0
[2π/3,4π/3)
5
uca≥0
ucb≥0
[π,5π/3)
6
ucb≥0
uab≥0
[4π/3,2π)
Input voltage interval and sector of input current space vector.
Input voltage intervals
Sector of input current space vector
In order to maintain the dc-link voltage being positive, the regulation range of φ is -π/6,π/6. In the applications of power factor control, the parameter of input LC filters has effect of power factor control. The simplified circuit of input LC filters is shown in Figure 13.
Simplified circuit of input LC filters.
In Figure 13, Us is input source voltage vector. Is is input source current vector. Uin and Iin are input voltage and current vector. Because Uin will be influenced by SVM algorithm, Us is used as the reference input voltage vector. Iin is the reference input current vector. The parameters of input LC filters are La=Lb=Lc=L=0.7 mH, Ca=Cb=Cc=C=20μF. Assume that input source voltage vector is Us=Um∠α and input current vector is Iin=Im∠α-φ. Im is decided by the load current and can be calculated with the modulation index of IMC and the amplitude of output currents.
From Figure 13, the differential equations for LC filters can be built as follows:(15)Us-Uin=LdIsdt,Is-Iin=CdUindt.
Then, the input source current Is is as follows:(16)LCd2Isdt2+Is-Iin+CdUsdt=0.
Considering the fundamental part of input current and ignoring harmonics current and dynamic process, input source current is expressed as Is=Ix∠β. From (16), (17)Ix1-LCωi2∠β=Im∠α-φ-CωiUm∠α-12π.
The amplitude and angle of input source current can be got as follows:(18)Ix=Im2+CωiUm2-2CωiUmImsinφ1-LCωi2,β=α+arctanCωiUm-ImsinφImcosφ.
From (18), power angle is related to the capacitor of LC filters, the amplitude of load current, and the angle of reference input current vector.
4. Experimental Results
The prototype of RB-IGBT based indirect matrix converter has been developed, shown in Figure 14. It mainly includes power circuit, RB-IGBT gate drive circuit, signal processing circuit, CPLD and DSP based control board, and heat sink. The rectifier is composed of the RB-IGBT module 18MBI100W-060A from Fuji Electric [31].
Prototype of RB-IGBT based indirect matrix converter.
In order to verify the proposed PWM algorithm and sequence, experimental tests on IMC prototype have been carried out. The parameters of the input filters are L=0.7 mH, C=20μF, and R=200 Ω. The sampling frequency is 10 kHz. Input voltage is 50 Hz and output frequency is 25 Hz. The load is an induction motor.
Figure 15 gives the experimental waveforms of SVM algorithm. The THD of output current is 1.06%. The THD of input current is 6.39%. Figure 16 shows the voltage of the bidirectional switch Sap-Usap and dc-link current idc when the commutation of bidirectional switches occurs. It can be seen that dc-link current is around zero at the commutation point, which means that zero-current switching is realized.
Experimental waveforms of proposed PWM algorithm.
Input phase voltage (25 V/div) and current (5 A/div)
Output line-to-line (25 V/div) and current (5 A/div)
The voltage (25 V/div) across RB-IGBT and the dc-link current (5 A/div) during zero-current switching.
The waveforms of input voltage and current when the power factor angle is setting as −2π/15, zero, and 2π/15 are shown in Figures 17–19, respectively.
The experiment waveforms of input (25 V/div) and current (5 A/div) when power factor angle is −2π/15.
The experiment waveforms of input (25 V/div) and current (5 A/div) when power factor angle is zero.
The experiment waveforms of input (25 V/div) and current (5 A/div) when power factor angle is 2π/15.
The power factor angle of experiment can be shown in Table 4. The theoretical value of power factor angle can be calculated by expression (18).
Comparison of theoretical value and experimental result of power factor angle.
Setting power factor angle
Calculated by (18)
Experimental result
φ=-24°
-26.3°
-28.8°
φ=0
-7.7°
-14.4°
φ=24°
14°
10.8°
5. Conclusions
In this paper, a SVM algorithm for indirect matrix converter with zero-current switching commutation for bidirectional switches is realized based on a control platform of DSP and CPLD. SVM for the rectifier and SVPWM for the inverter stage are described. PWM sequence of the SVM algorithm can achieve two-step ZCS commutation for the rectifier bidirectional switches. The realization of PWM sequence by DSP and decoding process for drive signals of RB-IGBTs by CPLD are explained in detail.
Based on the modulation strategy, input power factor correction of IMC is discussed. Because the inverter of IMC is VSI, the dc-link voltage should be positive and the regulation range of power angle is [-π/6,π/6]. The influence of the parameters in LC filters is analyzed. A prototype of RB-IGBT based indirect matrix converter is implemented. The experimental results show the performance of SVM algorithm and validate the analysis of power factor control of IMC.
Conflict of Interests
The author declares that there is no conflict of interests regarding the publication of this paper.
WheelerP. W.RodríguezJ.ClareJ. C.EmpringhamL.WeinsteinA.Matrix converters: a technology review200249227628810.1109/41.9932602-s2.0-0036543575SunK.ZhouD. N.HuangL. P.MatsuseK.Compensation control of matrix converter fed induction motor drive system under abnormal input voltage conditions1Proceedings of the IEEE Industry Applications Conference2004623630WeiL.LipoT. A.ChanH.Matrix converter topologies with reduced number of switches1Proceedings of the IEEE 33rd Annual Power Electronics Specialists Conference (PESC '02)June 200257632-s2.0-0036440967WangL. N.SunK.HuangL. P.A novel method to enhance the voltage transfer ratio of matrix converter1Proceedings of the 30th Annual Conference of IEEE Industrial Electronics Society (IECON '04)November 2004IEEE72372810.1109/IECON.2004.1433401KlumpnerC.BlaabjergF.Two stage direct power converters: an alternative to the matrix converterProceedings of the IEE Seminar on (Digest No. 2003/10100) Matrix ConvertersApril 2003IET7/17/9HelleL.LarsenK. B.JorgensenA. H.Munk-NielsenS.BlaabjergF.Evaluation of modulation schemes for three-phase to three-phase matrix converters200451115817110.1109/tie.2003.8219002-s2.0-1342286913AlesinaA.VenturiniM.Solid-state power conversion: a fourier analysis approach to generalized transformer synthesis1981284319330KiatsookkanatornP.SangwongwanichS.A unified PWM method for matrix converters and its carrier-based realization using dipolar modulation technique2012591809210.1109/tie.2011.21518232-s2.0-80053627616DahidahM. S. A.AgelidisV. G.SHE-PWM technique for single-phase AC-AC matrix convertersProceedings of the Australasian Universities Power Engineering Conference (AUPEC '08)December 2008Sydney, AustraliaIEEE18ItohJ.-I.ChiangG. T.MakiK.Beatless synchronous PWM control for high-frequency single-pulse operation in a matrix converter20132831338134710.1109/TPEL.2012.22079182-s2.0-84867892539TakeshitaT.FukagawaH.PWM strategy of voltage harmonics minimization for matrix convertersProceedings of the Annual Conference on IEEE Industrial Electronics SocietyNovember 2010668673KimS.YoonY.-D.SulS.-K.Pulsewidth modulation method of matrix converter for reducing output current ripple201025102620262910.1109/TPEL.2010.20421792-s2.0-77957724074WangJ.WuB.XuD.ZargariN. R.Indirect space-vector-based modulation techniques for high-power multimodular matrix converters20136083060307110.1109/tie.2012.22002152-s2.0-84876216478WeiL. X.LipoT. A.A novel matrix converter topology with simple commutation3Proceedings of the 36th IEEE IAS Annual Meeting on Industry Applications Conference200117491754KlumperC.BlaabjergF.ThoegersenP.Alternate ASDs: evaluation of the converter topologies suited for integrated motor drives2006122718310.1109/MIA.2006.1598029JussilaM.TuusaH.Comparison of simple control strategies of space-vector modulated indirect matrix converter under distorted supply voltage200722113914810.1109/TPEL.2006.8866542-s2.0-33846927040KlumpnerC.BlaabjergF.Modulation method for a multiple drive system based on a two-stage direct power conversion topology with reduced input current ripple200520492292910.1109/tpel.2005.8509652-s2.0-22944479929LiG.SunK.HuangL. P.IgarashiS.RB-IGBT gate drive circuit and its application in two-stage matrix converterProceedings of the 23rd Annual IEEE Applied Power Electronics Conference and Exposition (APEC '08)February 200824525110.1109/apec.2008.45227292-s2.0-49249095299XiaodongQ.YanchunX.YuyeW.Indirect matrix converter based on investigations of field-oriented control for induction motorProceedings of the 4th IEEE Conference on Industrial Electronics and Applications (ICIEA '09)May 2009Xi'an, China70971410.1109/iciea.2009.51382972-s2.0-70349336949FriedliT.HeldweinM. L.GiezendannerF.KolarJ. W.A high efficiency indirect matrix converter utilizing RB-IGBTsProceedings of the IEEE Power Electronics Specialists ConferenceJune 200617KolarJ. W.BaumannM.SchafmeisterF.ErtlH.Novel three-phase AC-DC-AC sparse matrix converterProceedings of the 17th Annual IEEE Applied Power Electronics Conference and ExpositionsMarch 20027777912-s2.0-0036075180SchafmeisterF.KolarJ. W.Novel modulation schemes for conventional and sparse matrix converters facilitating reactive power transfer independent of active power flow4Proceedings of the IEEE 35th Annual Power Electronics Specialists ConferenceJune 20042917292310.1109/pesc.2004.13552972-s2.0-8744255496QiC.ChenX.QiuY.Carrier-based randomized pulse position modulation of an indirect matrix converter for attenuating the harmonic peaks20132873539354810.1109/TPEL.2012.22241402-s2.0-84872032185NguyenT. D.LeeH.-H.Dual three-phase indirect matrix converter with carrier-based PWM method201429256958110.1109/TPEL.2013.22550672-s2.0-84883206662ChiangG. T.ItohJ.-I.Comparison of two overmodulation strategies in an indirect matrix converter2013601435310.1109/TIE.2012.21850212-s2.0-84866245674NguyenT. D.LeeH.-H.Modulation strategies to reduce common-mode voltage for indirect matrix converters201259112914010.1109/TIE.2011.21411012-s2.0-80053646491PeñaR.CárdenasR.ReyesE.ClareJ.WheelerP.Control of a doubly fed induction generator via an indirect matrix converter with changing dc voltage201158104664467410.1109/TIE.2011.21093342-s2.0-80052370625OrtegaM.JuradoF.CarpioJ.Control of indirect matrix converter with bidirectional output stage for micro-turbine20125665966810.1049/iet-pel.2011.02102-s2.0-84865848744WangJ.WuB.XuD.ZargariN. R.Phase-shifting-transformer-fed multimodular matrix converter operated by a new modulation strategy201360104329433810.1109/TIE.2012.22177142-s2.0-84878126803LohP. C.BlaabjergF.GaoF.BabyA.TanD. A. C.Pulsewidth modulation of neutral-point-clamped Indirect matrix converter20084461805181410.1109/tia.2008.20063212-s2.0-57049130939Fuji Electric Corporation2004Tokyo, JapanFuji Electric Corporation