This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2015.2429593, IEEE Transactions on Power Electronics 1 A Novel Quasi-Z-Source Inverter Topology With Special Coupled Inductors For Input Current Ripples Cancellation Alexandre Battiston, Member, IEEE, El-Hadj Miliani, Member, IEEE, Serge Pierfederici, Farid Meibody-Tabar Abstract—This paper proposes an inductors coupling technique to cancel input current ripples of a quasi-Z-source inverter without adding any additional components. This means the suppression of input filters that are used to protect voltage sources, and this whatever the voltage boost factor value. This cancellation property is made possible by a suitable coupling of the two existing quasi-Z-source inductors according to mathematical condition. The considered system in this paper is an electric drive system composed of a motor fed by the proposed coupled quasi-Z-source inverter. An experimental prototype to validate both the theoretical and simulation analyses has been developed. The results have validated the proposed coupling strategy and show that it does not degrade the global efficiency of the system. K,D i L1 vs DC-source voltage C2 L1 Iinv C1 vC2 vC1 vDC iL2 L2 Z-source inverter Motor Fig. 1. Bidirectional Z-source inverter in an electric drive system. Index Terms—high-frequency ripples, quasi-Z-source inverter, coupled inductors, battery current, electric drive system I. I NTRODUCTION I N the last decade, the number of electric drives systems for hybrid/electric vehicles as well as renewable-energy applications have increased significantly and involved the development of cost-effective, high-efficient converters. The power is thus modulated and converters are controlled by means of power electronics switches. Generally, a boost-type architecture is required to convert energy and link the source voltage to the DC-AC inverter. This allows working at high range of voltage by stepping up the low source voltage (for instance that of a battery). Furthermore, continuous currents with minimum high-frequency ripples are generally required in some applications such as embedded or renewable-energy applications. Indeed, the lifetime of storage sources (battery, fuel cells, ...) is influenced by current ripples. DC-DC boost converter is traditionally adopted to step up source voltage when isolation is not required [1]–[4]. However, some isolated versions can be found with input current ripples cancellation techniques [5]. In 2002, a promising DC-AC converter, also known as Zsource inverter (ZSI), has been proposed by Prof. Fang Zheng Peng [6], [7]. In Fig. 1, a bidirectional version (the diode D is replaced by a bidirectional switch K,D) of the original Zsource inverter is presented in an electric drive system feeding an electrical machine. This inverter possesses particular impedance-source input that allows extra shoot-through states of the inverter legs (the upper and lower switches of a same inverter’s leg are turned-ON) to step up the source voltage vs . Such an impedance-source offers advantages for the inverter in terms of both robustness and reliability. Indeed, incorrect IGBT turned-ON switching no longer destroy the inverter. However, one drawback that can be pointed out concerns the discontinuous input current flowing through the bidirectional K, D switch (see Fig. 1). To overcome this problem, bulky capacitor and inductor are generally used as LC passive filter to protect the voltage source against current ripples. As a consequence, this filter increases both the volume and cost of the system, which are often limited in embedded applications. This represents a great disadvantage compared with classical DC-DC boost converter, which possesses input inductive current. However, the minimization of the input current ripple is combined with the use of large inductor or ripple cancellation techniques. In all cases, this leads to complexity of the architecture and additional passive elements [8]–[14]. In [15], authors show that boost converter with ripple cancellation network (RCN) takes advantage on classical boost converter from a weight point of view. Interesting results are also obtained in [16] where the authors manage to cancel the input current ripples of a boost-type converter architecture. Some disadvantages of this proposal can be pointed out. For instance, it uses additional active and passive components and is slightly dependent on the operating point of the system. Other papers focus on coupled impedance-source inverter topologies but they do not aim at canceling the input current ripples [17], [18]. Most of the time, they present a reduction of the high-frequency ripples. In this paper, one focuses on DC-AC quasi-Z-source inverter [19]–[22]. This topology is an improvement of the original Z-source inverter [7]. Using quasi-Z-source inverter allows working with continuous input current. Thus, there is no need to use additional passive filter if the inductors are well- 0885-8993 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2015.2429593, IEEE Transactions on Power Electronics 2 designed. This topology is interesting as it naturally contains two inductors, which can be coupled or not. The coupled quasi-Z-source inverter (CQZSI) is presented in Fig. 2 in an electric traction system application that will be used in both simulation and experiment. One takes advantage of this property to propose in this paper a suitable coupling of the inductors that leads to the cancellation of the input current ripples. The advantages of the presented method in comparison with the other ones in the literature can be enumerated as follows: 1) There is no need to add other passive or active components. Only the coupling techniques of the inductors are modified according to the geometry used. 2) The proposed technique is not dependent on the operating point. Theoretically, it is valid for the entire range of the duty cycle. 3) The current flowing through the battery is theoretically perfectly continuous. In the next part, the modeling of the studied system is detailed before giving a mathematical point of view of the proposed input current ripple cancellation technique in part III. The realization of the coupling is studied in part IV. based on commercialized ferrite cores. Simulation and experimental validations are given in parts V. and VI. respectively before giving a conclusion in part VII. II. P RESENTATION OF THE STUDIED SYSTEM In Fig. 2, the Coupled quasi-Z-source inverter (CQZSI) is presented in a global electric traction system with an input source voltage and a motor. The source voltage vs can thus be stepped up by means of inserting extra shoot-through zero states in the inverter PWM scheme [23]–[25]. By considering iL1 , iL2 , vC1 and vC2 as state variables, the CQZSI can be modeled as follows using the logical command u ∈ {0, 1}. This latter indicates the state of the inverter (traditional or shoot-through state). When u = 1, inverter is shorten (shootthrough state) whereas u = 0 means it operates in its classical vC2 C2 iL1 L1 vs DC-source voltage 1 vDC = vs 1 − 2d (2) where d ∈ [0, 0.5] represents the duty cycle of the short-circuit states during a switching period T . It is thus the mean value of the logical variable u. III. M ATHEMATICAL CONDITION OF INPUT CURRENT RIPPLES CANCELLATION It is assumed the capacitors C1 and C2 are well-sized to consider the voltages vC1 and vC2 close to their mean values v̄C1 and v̄C2 respectively. These quantities are given by averaging (1): 1−d vs (3) vC1 ' v̄C1 = 1 − 2d d vC2 ' v̄C2 = vs (4) 1 − 2d Let vL1 and vL2 be the voltages across the two inductors. One has: vL1 = vs + vC2 u − vC1 (1 − u) (5) vL2 = vC1 u(t) − vC2 (1 − u) (6) Thus, for the sequence corresponding to u = 1, one has: vL1 = vs + vC2 (7) vL2 = vC1 And with the assumption that vC1 ' v̄C1 and vC2 ' v̄C2 , (7) is given by (8) according to (3) and (4). 1−d vs ' vC1 vL1 ' (8) 1 − 2d v =v L2 C1 Thus, one obtains that for u = 1, vL1 ' vL2 . The same mathematical description is made by considering the second sequence for which u = 0. One has: vL1 = vs − vC1 (9) vL2 = −vC2 Iinv K,D iL2 L2 vC1 C vDC 1 Coupled Quasi Z-source inverter active or zero-sequence states. diL2 diL1 +M = vs + vC2 u − vC1 (1 − u) L1 dt dt di di L2 L1 L2 +M = vC1 u − vC2 (1 − u) dt dt (1) dv C1 C = −i u + i (1 − u) − I (1 − u) 1 L2 L1 inv dt C dvC2 = −i u + i (1 − u) − I (1 − u) 2 L1 L2 inv dt M represents the mutual inductance. By averaging the two first equations in (1), the elevating ratio of CQZSI is given by (2) noting vDC = vC1 + vC2 : Motor And, according to (3) and (4), (9) is given by: −d vL1 ' vs ' −vC2 1 − 2d v = −v L2 Fig. 2. Bi-directional coupled quasi-Z-source inverter in an electrical traction system. (10) C2 As a result, vL1 ' vL2 on this second sequence. The voltages across the two inductors can be thus considered equal for all 0885-8993 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2015.2429593, IEEE Transactions on Power Electronics 3 time. From (1), by considering current ripples on an operating sequence, the equality (11) can be expressed. ∆iL2 ∆iL2 ∆iL1 ∆iL1 +M = L2 +M (11) ∆T ∆T ∆T ∆T with ∆T the duration of a sequence when u = 1 or u = 0. This latter is theoretically known. (11) leads to the expression of the input current ripple: L1 ∆iL1 = L2 − M ∆iL2 L1 − M (12) One finally obtains the mathematical cancellation condition given by: L2 = M (13) IV. M AGNETIC CONDITION OF INPUT CURRENT RIPPLES CANCELLATION This part now focuses on the magnetic realization of the coupled inductors with respect on the above mathematical condition. An E magnetic core is for instance considered with two sets as illustrated in Fig. 3. In order to present results in general case, three different lengths l1 , l2 and l3 as well as three different air gap e1 , e2 and e3 are taken into account. The third dimension length is noted lz . A permeance network iL1 vL1 air lz iron vL2 l1 e1 e2 l2 e3 l3 iL2 Fig. 3. Geometry of the considered magnetic core. modelling the magnetic geometry in Fig. 3 is presented in Fig. 4. This representation allows pre-dimensioning the system and will be verified by means of finite elements methods in the following lines. The two windings are replaced by Ampereturns n1 iL1 and n2 iL2 with n1 and n2 the turns number on the primary and secondary side of the coupled inductors. The magnetic fluxes in the different legs are noted ϕ1 and ϕ2 . Magnetic reluctances < are given by: <material = l 1 µmaterial A (14) with l the length of the circuit in meters, µmaterial the permeability of the material (air or iron) and A the cross-sectional area of the circuit in square meters. From this concept and the analogy with electrical circuit, one has on the assumption that <iron ' 0: n1 iL1 = <air1 ϕ1 + <air2 (ϕ1 − ϕ2 ) (15) n2 iL2 = <air3 ϕ2 + <air2 (ϕ2 − ϕ1 ) (16) Fig. 4. Permeance network model of the magnetic core. From (15) and (16), the expression of the fluxes ϕ1 and ϕ2 are given by: n2 <air2 ϕ1 = iL2 (<air2 + <air3 ) (<air1 + <air2 ) − <2air2 n1 (<air2 + <air3 ) + iL1 (<air2 + <air3 ) (<air1 + <air2 ) − <2air2 n1 <air2 ϕ2 = iL1 2 (< + < ) air2 air3 (<air1 + <air2 ) − <air2 n2 (<air1 + <air2 ) iL2 + (<air2 + <air3 ) (<air1 + <air2 ) − <2air2 (17) From an electrical circuit point of view, the total fluxes φ1 and φ2 through the primary and secondary sides of the inductors are given by: φ1 = n1 ϕ1 = L1 iL1 + M iL2 (18) φ2 = n2 ϕ2 = L2 iL2 + M iL1 From (17) and (18), one finally obtains the expressions of the inductances L1 , L2 and the mutual inductance M : n21 (<air2 + <air3 ) L1 = (19) (<air2 + <air3 ) (<air1 + <air2 ) − <2air2 n22 (<air1 + <air2 ) L2 = (20) (<air2 + <air3 ) (<air1 + <air2 ) − <2air2 n1 n2 <air2 (21) M= (<air2 + <air3 ) (<air1 + <air2 ) − <2air2 The mathematical equality (13) is finally geometrically given by: <air2 n2 = n1 (22) <air1 + <air2 From the definition (14) of magnetic reluctance, (22) gives a condition on the turns numbers n1 and n2 : A2 e 1 n1 = 1 + · n2 (23) A1 e 2 with A1 = l1 · lz and A2 = l2 · lz . By considering symmetrical commercialized ferrite cores, with l1 = l3 = l22 and e1 = e2 = e3 the above condition becomes: n1 = 3 · n2 0885-8993 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. (24) This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2015.2429593, IEEE Transactions on Power Electronics 4 Designing the experimental coupled inductors in compliance with condition (24), input current ripples are theoretically canceled. The values of inductances and mutual (19), (20) and (21) becomes with this technique and for the chosen core geometry: 3 n22 (25) L2 = 4 <air1 (26) M = L2 L1 = 9 L2 (27) (n1 = 3 n2 ), the ratio 3/5 is validated according to the inductions values for the considered point in the air gaps (|B1 | = 0.061 T and |B2 | = 0.037 T ). V. S IMULATION RESULTS Simulation parameters as well as experimental ones are given in TABLE I. A simulation of electrical system in Fig. TABLE I S IMULATION AND EXPERIMENTAL PARAMETERS Theoretical design has been obtained according to the perSymbol Description Value vs source voltage DC-bus voltage inductors capacitors PM-motor power 65 V 100 V 230 µH 680 µF < 500 W vDC L1 = L2 C1 = C2 P0max 2 is conducted. The CQZSI is controlled by means of Sliding Mode Control and the source voltage vs = 65 V is stepped up to vDC = 100 V . The control diagram will not be detailed in this paper. The mechanical speed of the machine is controlled by means of PI regulators. In Fig. 6, the DC-bus voltage vDC (a) with n1 = n2 . DC−bus voltage 150 vDC (V) 100 50 0 0.07 0.0701 0.0702 0.0703 iL2 (A) 4 Inductive currents 0.0704 iL1 − 1A (A) 3 2 1 0.07 (b) with n1 = 3 n2 . Fig. 5. Finite elements results: flux lines in the magnetic circuit plotted for the two considered configurations. meance network in Fig. 4 under hypothesis. The mean value ¯ = iL2 ¯ ), of the two inductive currents being the same (iL1 ϕ1 = ϕ2 with n1 = n2 . This comes from (15) and (16) with the considered geometry. By contrast, with n1 = 3 n2 , (15) and (16) lead to ϕ2 = 3/5 ϕ1 . This property is thus verified using a finite elements software which takes into account iron reluctance or boundaries effects. The results obtained in Fig. 5 for the two configurations are comply with theoretical expectation. For the first configuration (n1 = n2 ), the results show that the flux lines share between the two sides of the magnetic material (inductions are equal in the air gaps |B1 | = |B2 | = 0.025 T ). For the second configuration 0.0701 0.0702 simulation time (s) 0.0703 0.0704 Fig. 6. Simulation results: inductive currents waveforms in steady state with L1 = L2 = 230 µH. and the inductive currents iL1 and iL2 are presented. The DCmax bus voltage evolves between two values, vDC , which represents the DC-bus voltage reference (100 V ) and zero when a shoot-through zero state is added in the inverter PWM scheme. In simulation, four shoot-through states are added during the switching period T = 10−4 s (see [7] for more details). In this figure, the currents waveforms are obtained with simple coupling n1 = n2 . Thus, the inductive currents are equal so as the high-frequency ripples (about 1 A according to the figure). In Fig. 7, the proposed coupling strategy is investigated with respect to condition (24). As expected, this result validates the input current ripples cancellation (iL1 ). The input current is thus perfectly continuous and does not contain high-frequency ripples. A third test in Fig. 8 is conducted and focuses on 0885-8993 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2015.2429593, IEEE Transactions on Power Electronics 5 DC−bus voltage 150 vDC (V) 100 50 0 0.09 0.09 0.0901 0.0901 0.0902 0.0902 0.0903 0.0904 0.0904 Inductive currents 6 iL2 (A) 5 i L1 4 (A) 3 2 1 0.09 0.09 0.0901 0.0901 0.0902 0.0902 0.0903 0.0904 0.0904 (a) d = 10%. Speed DC−bus voltage Fig. 7. Simulation results: inductive currents waveforms in steady state with L1 = 9 · L2 (n1 = 3 n2 ) and L2 = M = 230 µH. (b) d = 30%. Fig. 9. DC-bus voltage and currents for two operating points depending on the duty cycle d value. 150 C1 = C2 inverter 100 vDC (V) 50 0 0.18 0.2 0.22 0.24 0.26 0.28 0.3 0.32 0.34 Alternator 0.36 1000 Ω* (rpm) 500 Inductive currents 0.18 10 0.2 0.22 8 iL1−4A (A) 6 iL2 (A) 0.24 0.26 0.28 0.3 0.32 0.34 Motor 0.36 4 L1= 9 L2 2 0 −2 0.18 0.2 0.22 0.24 0.26 0.28 0.3 simulation time (s) 0.32 0.34 0.36 Fig. 10. Test bench for validation. Fig. 8. Simulation results: study of transient state after a speed step. transient state after a speed step from 500 rpm to 1000 rpm. The two currents have been shifted for convenience. The result allows validating the proposed coupling technique even in transient state. The input current does not have any highfrequency ripples. One advantage of the proposed strategy using a quasi-Z-source inverter is that it is valid on the entire range of duty cycle. It does not depend on the boost ratio vDC /vs . The simulation results in Fig. 9 allow proving this for two tested duty cycles (d = 10% and d = 30%). n1=n2 n2 n1=3n2 n2 Fig. 11. Two coupling techniques for comparison and validation. VI. E XPERIMENTAL RESULTS A. Presentation of the test bench The experimental test bench is presented in Fig. 10. It is composed of a PM-motor fed by a CQZSI with two capacitors (C1 = C2 ) and two coupled inductors with the proposed coupling strategy (n1 = 3 n2 ). In order to present results for the two considered coupling techniques, two coupled inductors are built according to the turns numbers in primary and secondary sides (n1 = n2 or n1 = 3 n2 ). The two configurations are presented in Fig. 11. The typical waveforms of inductive currents with classical coupling n1 = n2 are given in Fig. 12. When a shoot-through zero state is added, the DC-bus voltage equals zero and the inductive currents iL1 and iL2 increase. These waveforms are obtained by inserting four short-circuits 0885-8993 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2015.2429593, IEEE Transactions on Power Electronics 6 vDC (100V/div) vDC (100V/div) Ω (500 rpm/div) iL1 (1A/div) iL2 (1A/div) iL2 (1A/div) iL1 (1A/div) Fig. 12. Experimental results: inductive currents iL1 and iL2 waveforms in steady state for L1 ' L2 (n1 = n2 ). Fig. 14. Experimental results: inductive currents waveforms after a speed step (from 500 to 1000 rpm) with the proposed couplig strategy. vDC (100V/div) Ω (500 rpm/div) iL2 (1A/div) vDC (100V/div) Ω (500 rpm/div) iL2 (1A/div) iL1 (1A/div) iL1 (1A/div) Fig. 13. Experimental results: inductive currents iL1 and iL2 waveforms in steady state for L1 = 9 L2 (n1 = 3 n2 ) and L2 = M . states during a switching period T [7]. The waveforms with the proposed coupling strategy are presented in Fig. 13 by keeping n2 constant and modifying n1 according to (24). These results are presented in steady state for Ω = 1000 rpm. As expected, the high-frequency ripples of input current iL1 have been canceled in comparison with previous results in Fig. 12. This confirms the simulation results obtained above. The Fig. 14 presents the inductive currents iL1 and iL2 with the proposed coupling strategy after a speed step from 500 to 1000 rpm. This validates the behavior of the two currents and confirms the ripples cancellation of iL1 . A zoom in Fig. 14 when Ω = 1000 rpm is given in Fig. 15 for several switching periods to point out the effectiveness of the coupling technique. In Fig. 16, the transient response of the inductive currents is given after a voltage step from 80 V to 100 V Fig. 15. Experimental results: inductive currents waveforms in steady state for Ω = 1000 rpm with the proposed coupling strategy. with constant speed Ω. This test represents the worst case of using quasi-Z-source inverter as it is preferable to adapt the DC-bus voltage to the mechanical speed of the machine so that the efficiency is better [26]. Indeed, the reference ∗ voltage vDC generally evolves with the same dynamic as that of the mechanical speed. Nevertheless, the experimental results remains interesting and show that the coupling strategy is still valid. VII. E FFICIENCY RESULTS ON THE TEST BENCH It is interesting in this part to study the effect of the coupling strategy (n1 = 3 n2 ) over the efficiency of the global system. Note that n2 is always constant in the two configurations and only n1 is modified (see Fig. 11). With vs = 100 V and vDC = 180 V , experimental efficiencies are plotted in Fig 17. 0885-8993 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2015.2429593, IEEE Transactions on Power Electronics 7 vDC (100V/div) vDC* (50 V/div) iL1 (5A/div) in low speed range. A remark can be expressed as regards the efficiency values that evolve between 45 to 80 %. These low values can be explained by the fact the source voltage vs is always stepped up to 180 V . Thus, shoot-through states are always added and increase losses in the inverter even when the source voltage is high enough to control the machine. Using a quasi-Z-source inverter, it is preferable to adapt the DC-bus voltage to the power demand of the machine so that losses are reduced [26]. Nevertheless, the worst case study has thus been considered to plot the efficiency. VIII. C ONCLUSION iL2 (5A/div) Fig. 16. Experimental results: inductive currents waveforms after a voltage step from 80 V to 100 V . 0.8 0.75 0.7 Efficiency 0.65 0.6 0.55 0.5 with classical coupling strategy (n1=n2) with proposed coupling strategy (n1=3 n2) 0.45 0.4 200 In this paper, an input current ripples cancellation technique is presented using a DC-AC coupled quasi-Z-source inverter. The mathematical derivation is established by considering commercialized E ferrite core. Others geometries can be used and may be optimized. The validation has been conducted on an electric traction prototype composed of a PM-motor fed by a quasi-Z-source inverter. Experimental results are in accordance with simulation ones. The input current is thus ”perfectly” continuous (without high-frequency ripples), hence interest as regards voltage sources (battery, fuel cell, ...) from a lifetime point of view. Other advantages can be pointed out in comparison with widely used boost-type architecture. For instance, there is no need to use additional components in the system as the two inductors exists in the quasi-Z-source topology. Furthermore, as the input current high-frequency ripples are canceled, no passive filter is necessary to protect the source, which is interesting in such applications like embedded or renewable-energy ones. Finally, in addition to be valid for all dudty cycle d, the coupling strategy does not impact the efficiency, which is slightly the same as the classical quasi-Zsource topology. R EFERENCES 400 600 800 1000 1200 Mechanical speed Ω (rpm) 1400 1600 1800 Fig. 17. Experimental results: Efficiency comparison between proposed and classical inductors coupling with vs = 100 V and vDC = 180 V . For this test, the efficiency is given for different mechanical speeds of the machine and for the two coupling techniques. The efficiency is calculated according to the ratio of the power absorbed by the machine Pm = Ts ·Ω over the power provided by the source Ps = iL1 ·vs . Ts represents the shaft torque given by an Luenberger estimator not detailed in this paper. It is a viscous-type load torque that is generated according to a threephase resistor fed by an alternator coupled in the same machine shaft. The calculated efficiency thus takes into account all the losses in the system (machine losses, switching and conduction losses in the inverter, resistive and iron losses in the coupled inductors, resistive losses in the capacitors ...). The results show that the proposed coupling strategy does not seem to have high influence on efficiency. However, it can be pointed out that a slight advantage is given to the proposed coupling [1] B. Williams, “Dc-to-dc converters with continuous input and output power,” IEEE Trans. Power Electron., vol. 28, no. 5, pp. 2307–2316, May 2013. [2] C. Leu, P.-Y. Huang, and M.-H. Li, “A novel dual-inductor boost converter with ripple cancellation for high-voltage-gain applications,” IEEE Trans. Ind. Electron., vol. 58, no. 4, pp. 1268–1273, April 2011. [3] W. Li and X. He, “Review of nonisolated high-step-up dc/dc converters in photovoltaic grid-connected applications,” IEEE Trans. Ind. Electron., vol. 58, no. 4, pp. 1239–1250, April 2011. [4] J. Balestero, F. Tofoli, R. Fernandes, G. Torrico-Bascope, and F. J. M. De Seixas, “Power factor correction boost converter based on the threestate switching cell,” IEEE Trans. Ind. Electron., vol. 59, no. 3, pp. 1565–1577, March 2012. [5] S.-J. Cheng, Y.-K. Lo, H.-J. Chiu, and S.-W. Kuo, “High-efficiency digital-controlled interleaved power converter for high-power pem fuelcell applications,” IEEE Trans. Ind. Electron., vol. 60, no. 2, pp. 773– 780, Feb 2013. [6] F. Peng, X. Yuan, X. Fang, and Z. Qian, “Z-source inverter for adjustable speed drives,” IEEE Power Electron. Letters, vol. 1, no. 2, pp. 33 –35, june 2003. [7] F. Z. Peng, “Z-source inverter,” IEEE Trans. Ind. Appl., vol. 39, no. 2, pp. 504–510, Mar 2003. [8] B.-R. Lin and C.-L. Huang, “Interleaved zvs converter with ripplecurrent cancellation,” IEEE Trans. Ind. Electron., vol. 55, no. 4, pp. 1576–1585, April 2008. [9] Y. Gu, D. Zhang, and Z. Zhao, “Input current ripple cancellation technique for boost converter using tapped-inductor,” IEEE Trans. Ind. Electron., vol. PP, no. 99, pp. 1–1, 2014. 0885-8993 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TPEL.2015.2429593, IEEE Transactions on Power Electronics 8 [10] J. Rosas-Caro, F. Mancilla-David, J. Mayo-Maldonado, J. GonzalezLopez, H. Torres-Espinosa, and J. Valdez-Resendiz, “A transformerless high-gain boost converter with input current ripple cancelation at a selectable duty cycle,” IEEE Trans. Ind. Electron., vol. 60, no. 10, pp. 4492–4499, Oct 2013. [11] Y.-P. Hsieh, J.-F. Chen, T.-J. Liang, and L.-S. Yang, “Novel high step-up dc-dc converter with coupled-inductor and switched-capacitor techniques,” IEEE Trans. Ind. Electron., vol. 59, no. 2, pp. 998–1007, Feb 2012. [12] Y. Gu and D. Zhang, “Interleaved boost converter with ripple cancellation network,” IEEE Trans. Power Electron., vol. 28, no. 8, pp. 3860– 3869, Aug 2013. [13] Y. Gu, D. Zhang, and Z. Zhao, “Input/output current ripple cancellation and rhp zero elimination in a boost converter using an integrated magnetic technique,” IEEE Trans. Power Electron., vol. 30, no. 2, pp. 747–756, Feb 2015. [14] D. Diaz, O. Garcia, J. Oliver, P. Alou, Z. Pavlovic, and J. Cobos, “The ripple cancellation technique applied to a synchronous buck converter to achieve a very high bandwidth and very high efficiency envelope amplifier,” IEEE Trans. Power Electron., vol. 29, no. 6, pp. 2892–2902, June 2014. [15] D. Diaz, D. Meneses, J. Oliver, O. Garcia, P. Alou, and J. Cobos, “Dynamic analysis of a boost converter with ripple cancellation network by model-reduction techniques,” IEEE Trans. Power Electron., vol. 24, no. 12, pp. 2769–2775, Dec 2009. [16] C. Soriano-Rangel, J. Rosas-Caro, and F. Mancilla-David, “An optimized switching strategy for a ripple-cancelling boost converter,” IEEE Trans. Ind. Electron., vol. PP, no. 99, pp. 1–1, 2014. [17] M.-K. Nguyen, Y.-C. Lim, and S.-J. Park, “Improved trans-z-source inverter with continuous input current and boost inversion capability,” IEEE Trans. Power Electron., vol. 28, no. 10, pp. 4500–4510, Oct 2013. [18] M. Shen, J. Wang, A. Joseph, F. Z. Peng, L. Tolbert, and D. Adams, “Constant boost control of the z-source inverter to minimize current ripple and voltage stress,” IEEE Trans. Ind. Appl., vol. 42, no. 3, pp. 770–778, May 2006. [19] J. Anderson and F. Peng, “Four quasi-z-source inverters,” in Proc. IEEE Power Electron. Spec. Conf., June 2008, pp. 2743–2749. [20] Y. Li, S. Jiang, J. Cintron-Rivera, and F. Z. Peng, “Modeling and control of quasi-z-source inverter for distributed generation applications,” IEEE Transactions on Industrial Electronics, vol. 60, no. 4, pp. 1532–1541, 2013. [21] M.-K. Nguyen, Y.-C. Lim, and G.-B. Cho, “Switched-inductor quasi-zsource inverter,” IEEE Trans. Power Electron., vol. 26, no. 11, pp. 3183 –3191, nov. 2011. [22] F. Guo, L. Fu, C.-H. Lin, C. Li, W. Choi, and J. Wang, “Development of an 85-kw bidirectional quasi-z-source inverter with dc-link feedforward compensation for electric vehicle applications,” IEEE Trans. Power Electron., vol. 28, no. 12, pp. 5477–5488, 2013. [23] F. Peng, “Z-source inverter,” in 37th IAS Annual Meeting. Conference Record of the Industry Applications Conference, vol. 2, oct. 2002, pp. 775 –781 vol.2. [24] P. C. Loh, D. Vilathgamuwa, Y. Lai, G. T. Chua, and Y. Li, “Pulse-width modulation of z-source inverters,” IEEE Trans. Power Electron., vol. 20, no. 6, pp. 1346 – 1355, nov. 2005. [25] P. Loh, F. Blaabjerg, and C. P. Wong, “Comparative evaluation of pulsewidth modulation strategies for z-source neutral-point-clamped inverter,” IEEE Trans. Power Electron., vol. 22, no. 3, pp. 1005 –1013, may 2007. [26] A. Battiston, E.-H. Miliani, J.-P. Martin, B. Nahid-Mobarakeh, S. Pierfederici, and F. Meibody-Tabar, “A control strategy for electric traction systems using a pm-motor fed by a bidirectional z -source inverter,” IEEE Trans. Veh. Technol., vol. 63, no. 9, pp. 4178–4191, Nov 2014. 0885-8993 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.