A Systematic Consideration of Voltage Multiplier Topology for Power Factor Correction Using Voltage

Transcription

1 Volume2, Issue4, April 215 A Systematic Consideration of Voltage Multiplier Topology for Power Factor Correction Using Voltage Gopichand B1, Sreenivasulu P2 M.TECH (EEE),student, Dept of EEE, Siddarth Institute Of Engineering And Technology 2 Assoc professor (EEE), Dept of EEE, Siddarth Institute Of Engineering And Technology, Puttur, Chittor(Dist), India ABSTRACT This paper proposes a single-stage three-phase-to single-phase current-fed high step-up ac dc matrix converter. The proposed converter inserts a boost-type matrix converter, which is formed by three boost inductors and six bidirectional switches, between a three-phase ac source and a Cockcroft Walton voltage multiplier (CWVM). By using this topology associated with power factor correction technique, the proposed converter not only achieves almost unity power factor and sinusoidal input currents with low distortion but also obtains high voltage gain at the output end. Moreover, the matrix converter generates an adjustable-frequency and adjustable-amplitude current, which injects into the CWVM to regulate the dc output voltage and smooth its ripple. With this flexible injection current, the performance of the proposed converter is superior to the conventional CWVM, which is usually energized by a single-phase ac source. The operation principle, control strategy, and design considerations of the proposed converter are detailed in this paper. Finally, simulation results demonstrate the claims and validity of the proposed converter. Index Terms AC DC converter, Cockcroft Walton voltage multiplier (CWVM), matrix converter, power factor correction (PFC). I. INTRODUCTION Conventionally, most of the dc sources were usually generated from ac dc converters, except some dc-powered systems [1]. Depending on the application requirements, many circuit topologies have been well developed for different voltage levels, different power rating, and different ac sources. When high dc voltage was required, high voltage gain was an important issue to determine the suitable topologies to achieve the high voltage output. Boost-type nonisolated dc-dc converters, which not only provide high voltage gain but also promise high efficiency and compactness, were popular for high-voltage applications [2] [4]. When these kinds of converters were supplied by ac source, an ac dc stage had to deploy at the front end to form a two-stage topology. Thus, the advantages of these high step-up converters will be deteriorated. When necessary, these two-stage topologies could employ a high-frequency step-up transformer to obtain more voltage gain [5], [6]. However, this led to increasing cost and bulk, and decreasing efficiency further. When high power was required, three-phase ac sources were superior to single-phase ac sources in providing energy to the aforementioned ac dc converters. In [7], three-phase single-stage ac dc power converters, including buck type, boost type, buck boost type, multilevel type, and multipulse type, were summarized. In this review literature, most of these ac dc converters used power factor correction (PFC) techniques to improve the ac line conditions in terms of power factor (PF) and total harmonic distortion of input currents (THDi) and offered adjustable or regulated output for the variable loads. Theoretically, both boost-type and boost buck- type ac dc converters can provide high voltage gain with extremely high duty cycle. 1 84

2 Volume2, Issue4, April 215 These topologies rarely appeared in high-voltage dc applications, such as X-ray systems, dust filtering, insulating test, and electrostatic coating, because some nonideal characteristics of power components seriously limit the theoretical gain. Providing the advantages of high voltage gain, low-voltage stress, compactness, and cost efficiency, the well-known Cockcroft Walton voltage multiplier (CWVM), which is constructed by cascading capacitor-and-diode pairs, as shown in Fig. 1(a), is very popular among highvoltage dc applications [8] [18]. Traditionally, some applications combined high step up ratio transformers with CWVM to generate high-voltage dc output [19]. However, these high step-up line-frequency transformers led to inefficiency of bulk and cost, and large amount of voltage drop and ripple appeared at the dc output. With the development of solid-state self-commutating devices, some ac dc converters based on CWVM, including both single-stage [2] [24] and two-stage configurations [25], [26] with PFC, were presented to obtain high voltage gain and high quality of line conditions. A previous work of this paper was presented by [27], in which the CWVM was fed by a single-phase-to-singlephase matrix converter, as shown in Fig. 1(b). This converter provided variable-frequency and variable-amplitude current to regulate the dc output voltage and smooth the ripple voltage. Moreover, the line condition was improved through PFC technology by using a commercial control integrated circuit. However, almost all these kinds of ac dc converters were sourced by single-phase utility, and only few of them were powered by three-phase source. In [23] and [24], two single-stage three-phase ac dc converters based on asymmetrical and symmetrical CWVMs were proposed, respectively. Both the two converters provided high voltage gain, high PF, low THDi, and regulated dc output; but in [23], it required an additional low-pass filter (LPF) to eliminate discontinuous input currents, and the voltage gain could only reach the value of conventional CWVM; and in [24], the symmetrical topology used more components in each stage, and the three center-tap transformers led the topology to inefficiency of bulk and cost. Recently, some matrix converters have been proposed to transfer three-phase source to single-phase source [31]. In this paper, a novel three-phase-to-single-phase current-fed high step-up ac dc matrix converter is proposed to energize the CWVM. The proposed converter provides higher voltage gain than conventional CWVM without using line- or highfrequency step-up transformer. Moreover, with PFC technology, the proposed converter draws power from three-phase utility with nearly unity PF and sinusoidal input currents [28] [3]. Furthermore, the matrix converter generates an adjustable frequency and adjustable-amplitude current, which injects into the CWVM to regulate the dc output voltage and smooth its ripple. With this flexible injection current, the performance of the proposed converter is superior to the conventional CWVM, which is usually energized by a singlephase ac source. In Section II, the mathematical model of the proposed converter is derived, and the operation principle of the proposed converter will be presented as well. The control strategy will be described in Section III. In Section IV, the theoretical analysis of ripple voltage at the output will be derived, and the design considerations will be given as well. In Section V, both simulation and experiment of the proposed converter will be conducted. A 5-W prototype is built for evaluation and measurement. Some selected operation cases with different parameters are investigated, and the results show the good performance of the proposed converter. Finally, some conclusions are given in Section VI. II. MATHEMATICAL MODEL AND OPERATION PRINCIPLE OF THE PROPOSED CONVERTER Here, the mathematical model of the proposed converter will be derived first, and then, the operation principle will be described as well. Fig. 2 shows the proposed converter, which is supplied by a three-phase ac source, and the phase voltages are denoted by van, vbn, and 85

3 Volume2, Issue4, April 215 vcn, respectively. A three-phaseto- single-phase matrix converter, which is composed of three boost inductors (La, Lb, and Lc) and six bidirectional switches (Sa1, Sb1, Sc1, Sa2, Sb2, and Sc2), is inserted between the ac source and a conventional N/2-stage CWVM circuit, where N is an even integer. Two antiseries insulated-gate bipolar transistors form one bidirectional switch, as shown in the bottom left part in Fig. 2. Fig. 1. (a) Conventional N/2-stage CWVM. (b) CWVM fed by a single phase to-single-phase matrix converter The six bidirectional switches are divided into two parts, namely, the upper part (Sa1, Sb1, and Sc1) and the bottom part (Sa2, Sb2, and Sc2). According to the first subscripts of the switches, the left terminals of the switches connect to the relative ac sources through three boost inductors. The three right terminals of the upper-part switches connect to the upper terminal D of the CWVM, whereas the three right terminals of the bottom-part switches connect to the bottom terminal E of the CWVM. For simplicity, the derivation of the mathematical model is divided into two parts, namely, the ac-side part and the CWVM, which are separated by terminals D and E. the six bidirectional switches, which are summarized in Table I,in which three auxiliary variables γ1, γ2, and γ3 are used to determine the connection between the three input lines and the voltage multiplier. Two equivalent circuits corresponding to two of the eight switching states are shown in Fig. 3, and the other six can be easily obtained. It has to be noted that, in Fig. 3, XY represents the terminal voltage of the CWVM, whose value is 86

4 Volume2, Issue4, April 215 determined by the combination of the capacitor voltage(s) of the CWVM according to diodeconducting states. Consequently, the ac-side mathematical model of the proposed converter is given as. Fig. 2. Configuration of the proposed converter with an N/2-stage CWVM. A. Mathematical Model of the Circuit in the AC Side Assuming that the PFC technique is properly applied to the proposed converter and that the circuit operates under continuous conduction mode (CCM), there are eight switching states of TABLE I SUMMARY OF THE SWITCHING STATES IN THE AC SIDE,,, , 87

5 Volume2, Issue4, April 215 Fig. 3. Two circuit states of the proposed converter in ac source side. (a) Sa1,Sb2, and Sc1 are turned on. (b) Sa2, Sb1, and Sc2 are turned on. (1) Where the CWVM; and the inductance of the boost inductor;,, is the terminal voltage of are defined as (2) Moreover, the current that flows into the CWVM circuit can be expressed as (3) where the current is the current feeding into the CWVM and can be deemed a pulse-form current source. B. Steady-State Operation of the Proposed Converter In this paper, the CWVM is operated and analyzed by current-fed method, in which the matrix converter generates the adjustable-frequency and adjustable-amplitude current, i.e.,, and feeds it into the CWVM. Before analyzing, some assumptions are made, as follows [27]. 1) All of the circuit elements are ideal, and there is no power loss in the system. 2) All of the capacitors in the N/2-stage CWVM are large enough; thus, the voltages across them are equal and ripple free, except the first capacitor, whose voltage is one half of the others. 3) The boost-type matrix converter operates in CCM and under steady-state condition. 88

6 Volume2, Issue4, April 215 4) When one of the boost inductors transfers the storage energy to the CWVM, i.e.,, only one of the diodes in the CWVM will conduct. 5) The three-phase ac source is balanced and undistorted. 6) To avoid the open circuit of the inductors, a safe commutation technology provides enough overlap of the trig signals between the relative switches. Analysis of circuit will ignores the safe-commutation states for simplicity. According to these assumptions, each capacitor voltage in an N/2-stage CWVM can be expressed as. (4) where is the steady-state voltage of the kth capacitor, Vo is the steady-state output voltage. The phase voltages of the ac source are given as (5) where Vm and ω are the amplitude of the phase voltages and the angular frequency of the ac source, respectively. If the matrix converter operates with PFC control, the ac line currents will be nearly sinusoidal and in phase with their corresponding phase voltages. Thus, the three ac line currents can be given as. (6) where Im is the amplitude of the line currents. For systematic analysis, one cycle of the ac source is divided into six sectors, which are denoted as sectors I VI, according to the polarities of the three-phase voltages. Fig. 4(a) shows the relationship of the phase voltages and the six sectors, and the corresponding line currents are shown in Fig. 4(b). Moreover, according to the polarity of the current fed into the CWVM, there are two operation modes in each sector, which are denoted as mode 1 for < and mode 2 for >. To achieve the goal of PFC and satisfy (6), this paper proposes a strategy to determine the switching patterns of the six bidirectional switches. Considering the first half of sector I in Fig. 4(a), which is denoted as mode 1, Sb1(Sb2) always turns on (off), and the four combinations of the switching states of Sa1(Sa2) and Sc1(Sc2) can be used to determine the slopes of and, respectively. Thus, there are four circuit states in this period, as shown in Fig. 5(a) (d). The operation of these four circuit states is detailed in the following. 1) State 1 [Fig. 5(a)]: In this state, Sa1(Sa2) and Sc1(Sc2) turn off (on). The two boost inductors La and Lc and the two input phase voltages van and vcn transfer energy to the CWVM through one of the odd diodes in the CWVM. Thus, iγ = ib = (ia + ic) < and vγ = Vo/N, and the equations relative to ia and ic are given as 89

7 Volume2, Issue4, April 215 Fig. 4. Some selected simulation waveforms of the proposed converter during one line period in CCM operation. (a) van, vbn, and vcn. (b) ia, ib, and ic. (c) iγ. (d) vγ. (e) Sa1, Sb1, and Sc1. (7) (8) To ensure boost operation, the dc output voltage has to meet the following condition (9) 2) State 2 [Fig. 5(b)]: In this state, Sa1(Sa2) turns on (off),and Sc1(Sc2) turns off (on). The boost inductor La is charged by van, and the boost inductor Lc and the phase voltage vcn transfer energy to the CWVM through one of the odd diodes.thus, iγ = ic, vγ = Vo/N, and the equations relative to ia and ic are given as. (1) (11) 9

8 Volume2, Issue4, April 215 3) State 3 [Fig. 5(c)]: In this state, Sa1(Sa2) turns off (on), and Sc1(Sc2) turns on (off). The boost inductor La and the phase voltage van transfer energy to the CWVM through one 91

9 Volume2, Issue4, April 215 Fig. 5. Four circuit states of the proposed converter in mode 1 of sector I. (a) State 1. (b) State 2. (c) State 3. (d) State 4. of the odd diodes, and the boost inductor Lc is charged by vcn.thus, iγ = ia, vγ = Vo/N, and the equations relative to ia and ic are given as (12) (13) TABLE II SWITCHING STATES OF THE PROPOSED CONVERTER FOR SECTOR I 4) State 4 [Fig. 5(d)]: In this state, Sa1(Sa2) and Sc1(Sc2) turn on (off). The boost inductors La and Lc are charged byvan and vcn, respectively. Due to no current path, iγ =, and all diodes in the CWVM are off; thus, all even capacitors C6,C4, and C2 are in series and supply power to load RL, whereas the odd capacitors C5, C3, and C1 are floating. The equationsrelative to ia and ic are given as (14) (15) 92

10 Volume2, Issue4, April 215 Fig. 6. Control block diagram. (a) Control block of the proposed converter. (b) Block diagram of the sector/mode generator. For the second half of sector I, which is also denoted as mode 2,Sb2(Sb1) always turns on (off), and there are four circuit states in this mode as well. The operation of these four states is similar to that of mode 1, except that iγ and vγ have opposite polarities. Table II summarizes the switching pattern of the six bidirectional switches and the relative circuit characteristics in sector I. From this table, it can be seen that the switching states of the switch pairs (Sa1, Sa2), (Sb1, Sb2), and (Sc1, Sc2) are complementary in sector I, and this rule is also available for the operation in other five sectors. Consequently, through a similar process, the circuit operation and switching patterns can be obtained. The simulated waveforms of and vγ are shown in Fig. 4(a) and (d), respectively; and the corresponding switching signals of the three upper-part switches are shown in Fig. 4(e). It can be seen that changes polarity once in each sector, i.e., the frequency of is six times as high as the line frequency in this case; and in this paper, the frequency of i is defined as the alternation frequency. The alternation frequency is determined by the number of mode swaps in one sector, which can be implemented by changing the conduction state of Sb1 in 93

11 Volume2, Issue4, April 215 sector I and Sa1 in sector II, etc. Obviously, higher alternation frequency leads to smoother ripple, and the ripple factor of the dc output will be significantly improved. III. DESIGN CONSIDERATIONS Here, the voltage and current stresses on key components will be considered. The ripple components in boost inductors and the capacitors in CWVM are investigated as well. Thus, the values of these two energy-storage elements can be determined by the specified ripple components posed on them. A. Capacitor Voltage Stress According to the assumptions in Section II, the maximum voltage stress across all capacitors in CWVM is 2Vo,max/N, where Vo,max is the maximum value of output voltage, except the first capacitor with half of that value. It has to be noted that, in real situation, the capacitor located farther from the source has lower voltage. This is a natural phenomenon of CWVM and should be taken into consideration when choosing the voltage specification for these capacitors. B. Voltage and Current Stresses on Switches and Diodes From the circuit states in Fig. 5, the voltage stress of the bidirectional switches is Vo,max/N, and their current stress can be calculated by (16) where Po is the rated output power, η is the efficiency at rated power, and Kover is the proportion factor of the overload. The current stress on the diodes is identical to that of the switches, i.e., IL,max, whereas the voltage stress of the diodes is twice as large as that of the switches, i.e., 2Vo,max/N. C. Boost Inductor Design Even the topology of the proposed converter is different from traditional boost-type switchmode rectifiers (SMRs); the switching strategies, such as current hysteresis method or space-vector modulation, used in SMRs can easily adapt to the proposed converter with only few modifications. Consequently, the design criteria of the boost inductors proposed by SMRs are also available for the proposed converter. According to [33], which proposed a simplified method to calculate the duty ratio of the two switching switches in each sector, the duty ratios of Sa1 and Sc1 are given as (17) (18) The minimum duty cycle happens when ωt = for Da1 and ωt = 6 for Dc1. Thus, the minimum duty cycle Dmin can be calculated as 94

12 Volume2, Issue4, April 215 (19) (2) where ton,min and fc are the minimum turn-on time and the switching frequency, respectively. The value of the inductance can be determined by (21) where KI is the expected percentage of the maximum peak-to peak ripple of the line current. D. Determination of the Capacitance Corresponding to the output and input frequencies of the matrix converter, there are two major components contained in the output ripple of the proposed converter. The first one is corresponding to the alternation frequency, and the second one is relative to the ac source frequency. Moreover, similar to conventional capacitor-filtered ac dc converters, the peakto- peak ripple is a function of output current. For the proposed converter, the load is connected to the even-side capacitors; thus, only the ripple in even capacitors is under concerned. The ripple component corresponding to the alternation frequency was analyzed by current-fed mode analytical principle [34], from which the peak-to-peak ripple corresponding to the alternation frequency can be derived as (22) where Io is the average load current, and C is the value of the capacitance, assuming that all capacitors are identical. Because PFC is a control goal of the proposed converter, it is assumed that unity PF is obtained at the ac side. Consequently, for a balanced and undistorted three-phase system with unity PF, the instantaneous power is constant, i.e., no power pulsation at the ac side. Therefore, unlike in a single-phase system [27], the line frequency causes no ripple effect in the dc output. The peak-to-peak ripple given in (24) can be used as a criterion to design the value of the capacitance depending on a given ripple specification. However, when the proposed converter was sourced by unbalanced ac source, the instantaneous power will no longer be constant; thus, the ripple will contain line frequency components, particularly of the lower order harmonics, such as second, third, fifth, etc. IV. SIMULATION RESULTS Extending from previous work simulations are conducted to demonstrate the performance of the proposed converter. Based on the proposed scheme, operation principle, and design considerations described in previous sections, two simulation cases with different alternation frequency were conducted by the software tool MATLAB/Simulink/SimPower to evaluate the performance of the proposed converter. The system specification for the proposed converter is summarized as follows: 95

13 Volume2, Issue4, April 215 Fig:7 simulation model Fig:8 input wave forms 96

14 Volume2, Issue4, April 215 Fig:9 out put wave forms Fig:1 Inverter out put wave forms 97

15 Volume2, Issue4, April 215 V. CONCLUSION In this paper, a three-phase-to-single-phase current-fed high step-up ac dc matrix converter based on CWVMwithout a lineand high-frequency step-up transformer has been presented to obtain high voltage gain. By using current hysteresis control to implement PFC, the proposed converter offers almost unity PF with low THDi at the ac mains; meanwhile, the matrix converter provided an adjustable-frequency and adjustableamplitude current to feed the CWVM to regulate the dc output and smooth ripple voltage. The programmable alternation frequency can be used to improve the ripple effect according to the application requirement. However, too much higher alternation frequency deteriorates the performance of the system; thus, the highest alternation frequency should be limited. The circuit operation, control strategy, and design considerations of the proposed converter were detailed in this paper. simulation work were conducted, and their results demonstrated the validity and the good performance of the proposed converter. Compared with conventional CWVM, the proposed converter sourced by a three-phase source is quite suitable for highvoltage and high-power applications. VI.REFERENCES [1] Y. Xue, L. Chang, S. B. Kjaer, Jr., J. Bordonau, and T. Shimzu, Topologies of singlephase inverters for small distributed power generators: An overview, IEEE Trans. Power Electron., vol. 19, no. 5, pp , Sep. 24. [2] L. S. Yang, T. J. Liang, and J. F. Chen, Transformerless dc dc converters with high step-up voltage gain, IEEE Trans. Ind. Electron., vol. 56, no. 8, pp , Aug. 29. [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 , Apr [4] D. Zhou, A. Pietkiewicz, and S. Cuk, A three-switch high-voltage converter, IEEE Trans. Power Electron., vol. 14, no. 1, pp , Jan [5] J. Tanaka and I. Yuzurihara, The high frequency drive of a new multistage rectifier circuit, in Proc. IEEE Power Electron. Spec. Conf., Apr. 1988, pp [6] J. F. Chen, R. Y. Chen, and T. J. Liang, Study and implementation of a single-stage current-fed boost PFC converter with ZCS for high voltage applications, IEEE Trans. Power Electron., vol. 23, no. 1, pp , Jan. 28. [7] B. Singh, B. N. Singh, A. Chandra, K. Al-Haddad, A. Pandey, and D. P. Kothari, A review of three-phase improved power quality ac dc converters, IEEE Trans. Ind. Electron., vol. 51, no. 3, pp , Jun. 24. [8] M. D. Bellar, E. H. Watanabe, and A. C. Mesquita, Analysis of the dynamic and steadystate performance of Cockcroft Walton cascade rectifiers, IEEE Trans. Power Electron., vol. 7, no. 3, pp , Jul [9] F. Hwang, Y. Shen, and S. H. Jayaram, Low-ripple compact high-voltage dc power supply, IEEE Trans. Ind. Appl., vol. 42, no. 5, pp , Sep./Oct. 26. [1] I. C. Kobougias and E. C. Tatakis, Optimal design of a half-wave Cockcroft Walton voltage multiplier with minimum total capacitance, IEEE Trans. Power Electron., vol. 25, no. 9, pp , Sep. 21. [11] H. J. Chung, A CW CO2 laser using a high-voltage dc dc converter with resonant inverter and Cockroft Walton multiplier, Opt. Laser Technol., vol. 38, no. 8, pp , Nov

16 Volume2, Issue4, April 215 [12] M. M.Weiner, Analysis of Cockcroft Walton voltage multipliers with an arbitrary number of stages, Rev. Sci. Instrum., vol. 4, no. 2, pp , Feb [13] S. M. Sbenaty and C. A. Ventrice, High voltage dc shifted RF switchmode power supply system design for gas lasers excitation, in Proc. Appl. Power Electron. Conf. Expo., Mar. 1991, pp [14] P. G. Maranesi, F. Raina, M. Riva, and G. Volpi, Accurate and nimble forecast of the HV source dynamics, in Proc. IEEE Power Electron. Spec. Conf., Jun. 2, pp [15] F. Belloni, P. Maranesi, and M. Riva, Parameters optimization for improved dynamics of voltage multipliers for space, in Proc. IEEE Power Electron. Spec. Conf., Jun. 24, pp [16] E. Chu, L. Gamage, M. Ishitobi, E. Hiraki, and M. Nakaoka, Improved transient and steady-state performance of series resonant ZCS highfrequency inverter-coupled voltage multiplier converter with dual mode PFM control scheme, J. Electr. Eng. Jpn., vol. 149, no. 4, pp. 6 72, Dec