POWER electronic devices contribute an important part of

Transcription

1 130 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 1, JANUARY 2009 Static Var Compensator and Active Power Filter With Power Injection Capability, Using 27-Level Inverters and Photovoltaic Cells Patricio Flores, Juan Dixon, Senior Member, IEEE, Micah Ortúzar, Rodrigo Carmi, Pablo Barriuso, and Luis Morán, Fellow, IEEE Abstract An active power filter and static var compensator with active power generation capability has been implemented using a 27-level inverter. Each phase of this inverter is composed of three H bridges, all of them connected to the same dc link and their outputs connected through output transformers scaled in the power of three. The filter can compensate load currents with a high harmonic content and a low power factor, resulting in sinusoidal currents from the source. To take advantage of this compensator, the dc link, instead of a capacitor, uses a battery pack, which is charged from a photovoltaic array connected to the batteries through a maximum power point tracker. This combined topology make it possible to produce active power and even to feed the loads during prolonged voltage outages. Simulation results for this application are shown, and some experiments with a 3-kVA device are displayed. Index Terms Active filters, multilevel systems, solar power generation, static var compensators (SVCs). I. INTRODUCTION POWER electronic devices contribute an important part of harmonics in all kinds of applications such as power rectifiers, thyristor converters, and static var compensators (SVCs). On the other hand, the pulsewidth modulation (PWM) techniques used today to control modern static converters such as machine drives, power factor compensators, and active power filters do not give perfect waveforms, which strongly depend on the switching frequency of the power semiconductors. Normally, voltage (or current in dual devices) moves to discrete values, forcing the design of machines with good isolation and, sometimes, loads with inductances in excess of the required value. In other words, neither voltage nor current are as expected. This also means that harmonic contamination, Manuscript received January 11, 2007; revised May 28, First published November 18, 2008; current version published December 30, This work was supported by the Comisión Nacional de Investigación Cientifica y Tecnológica (CONICYT) through Project Fondecyt and Millenium Project P F. P. Flores and J. Dixon are with the Department of Electrical Engineering, Pontificia Universidad Católica de Chile, Santiago, Chile ( paflorel@ing.puc; jdixon@ing.puc.cl). M. Ortúzar is with the Compania Americana de Multiservicios Ltda. (CAM), Endesa, Santiago, Chile ( mortuzar@gmail.com). R. Carmi is with the National Transmission Electrical Utility Company, Santiago, Chile ( rcarmi@gmail.com). P. Barriuso is with CDEC-SIC, Santiago, Chile ( pbarriuso@ cdec-sic.cl). L. Morán is with the Department of Electrical Engineering, Universidad de Concepción, 160 Concepción, Chile ( luis.moran@udec.cl). Digital Object Identifier /TIE additional power losses, and high-frequency noise can affect the controllers. All these reasons have generated, for many years, a large amount of research works on the topic of PWM [1] [3]. Today, research is focused on multilevel converters, because they are able to generate voltage waveforms with less distortion than conventional inverters based on two-level topologies [4] [6]. However, one step ahead has been the new multistage converter technology, which allows us to generate [7], [9] many more levels of voltage with fewer power semiconductors. When the number of levels is high enough (more than 20), multilevel inverters are able to produce current waveforms with negligible total harmonic distortion. Furthermore, they can work using both amplitude modulation and PWM strategies. One of the multistage technologies that allows producing many levels of voltage with a few number of transistors is the one based on H bridges scaled in the power of three [10] [12]. This topology uses relatively few power devices, and each one of the H bridges work at a very low switching frequency, which gives the possibility of working in high power with low-speed semiconductors and generating low-switchingfrequency losses, which makes them very suitable for machine drive [13] or power grid improvement applications. The objective of this paper is to show the advantages of a 27-level active power filter and var compensator, which is also used as an active power generator. Compared to conventional PWM techniques, this converter, in any of its functionalities, is able to produce current waveforms with negligible harmonic content. To produce active power, the system uses a battery pack kept charged using photovoltaic arrays. Furthermore, the batteries can also be charged from the mains when required. The paper describes the topology of the system and shows simulations and experiments with a small 3-kVA prototype. In summary, the final objective of this line of research is to take advantage of active filters and var compensators, which can also be used for power generation, and achieve a reliable and economic device for distributed generation, using renewable energy such as solar sources to support the grid in high-demand hours. II. OPERATION CHARACTERISTICS A. Basic Principle The circuit in Fig. 1 shows the basic topology of one converter used for the implementation of the multistage 27-level /$ IEEE

2 FLORES et al.: STATIC VAR COMPENSATOR AND ACTIVE POWER FILTER WITH POWER INJECTION CAPABILITY 131 Fig. 1. Three-level module for building multistage converters. inverter. It is based on the simple four-switch device ( H converter) used for single-phase inverters. These converters are able to produce three levels of voltage at the ac side: 1) +Vdc; 2) Vdc; and 3) zero. Reference [14] has proposed a per-phase power conversion scheme for synthesizing multilevel waveforms, linking many converters in a series connection, like the system shown in Fig. 1, but with all the dc voltages equal to Vdc. Such a multilevel inverter with n equal dc voltage levels can offer only 2n +1different voltage levels at the phase output. Reference [15] goes one step ahead with dc voltages varying in a binary fashion, which gives an exponential increase in the number of levels. For n such cascaded inverters, 2 n+1 1 distinct voltage levels may be achieved. In this paper, the outputs of the modules are connected thorough transformers whose voltage ratios are scaled in the power of three, allowing 3 n levels of voltage [16]. Then, with only three converters (n =3), 27 different levels of voltage are obtained: 13 levels of positive values, 13 levels of negative values, and zero. As a comparison, the first topology with equal dc voltages only achieves seven levels with three converters, and the second topology, scaled in power of two, achieves just 15 levels. This strategy represents an optimization of the number of levels, and their drawback is that there are no redundant levels, which are necessary to avoid negative power back to the dc link. However, in this particular application, redundant levels are unnecessary because the topology uses output transformers and a battery pack at the dc link that allows bidirectional power flow. B. System Components Fig. 2 displays the main components of the three-stage 27-level converter that is being used in this work. The figure only shows one of the three phases of the complete system. As can be seen, a battery pack, which is being charged through a maximum power point tracker (MPPT), is used in the dc link. MPPT [17] is a special dc dc converter, able to maximize the level of power transfer from the photovoltaic array to the battery pack, with a very high efficiency (normally around 99% at 80 Vdc). The solar array consists of two parallel groups of three serial connected panels, each one composed of 50 photovoltaic cells, producing, at peak sunlight, 370 W of power. During 8 h of charging, the photovoltaic array is able to deliver 2 kwh of energy to the battery pack, accounting for approximately 43% of the battery s energy. The solar panels were handmade at the Pontificia Universidad Católica de Chile, Santiago, Chile, using 300 cells, type MAIN-1530, which are produced by Schott Solar. They have an efficiency higher than 15%, and their size is 15 cm 10 cm. Fig. 2. Main components of the system (one phase). The transformer located at the bottom of the figure has the highest voltage ratio and will be called the main converter. The rest of the modules will be the auxiliary converters (Aux). The main converter manages most of the power (80%) [18], [19] but works at the lowest switching frequency, which is an additional advantage of this topology. P and Q are independently controlled. The reference for P depends on the state of charge (SOC) of the battery pack and thus can be controlled using the SOC information. On the other hand, Q can be controlled by keeping a unity power factor (for example) at the converter connection point. The load harmonics are filtered by forcing the current from the mains (I MAINS ) to be sinusoidal. The utilization of transformers make the system less efficient, but most of the time, active filters and var compensators use them when the voltage between the source and the line does not match. Furthermore, these transformers are fed with square voltages, which reduces even more the efficiency of the system. However, as the winding currents are sinusoidal, the extra loss of efficiency is only because of core losses, which do not drastically increase. For example, for a 100-kVA system (33.3 kva/phase), each main transformer has to be 27 kva. If this is a normal transformer, with 96% efficiency, this efficiency will be reduced, due to extra core losses, to 95.7%. If a transformerless topology is an option, then the topology shown in Fig. 3 is feasible, with the advantage of being much more efficient and cheaper. However, because of matching voltages and isolation problems, this is not always a viable solution. Moreover, with transformerless topologies, nine independent and isolated solar panels are required for the three phases, and additional control strategies to separately manage each of them must be implemented. For these reasons and mainly for the aforementioned matching problems, this paper has been focused on a system with output transformers, which is typical in most of the active filters and var compensators connected to the grid today. It should be noted that the main idea here

3 132 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 1, JANUARY 2009 Fig. 5. Main components of the control scheme. Fig. 3. Transformerless option (one phase). 0, or 1 (positive, zero, or negative output at the corresponding H bridge). Each one of the three digits of the trinary number is applied to each one of the three H bridges of the converter, which defines the output of the 27-level converter. For example, level 8 is obtained with the three-digit number {1, 0, 1} or =8. The maximum positive level is {1, 1, 1} or 13, and the minimum level (or maximum negative level) is { 1, 1, 1} or 13. It is important to mention that more complicated PWM strategies can also be applied in this kind of converters [20] [22]. III. CONTROL SCHEME Fig. 5 shows the block diagram of the control scheme, where P and Q are independently controlled through I P and I Q, respectively. Overall, the control system has been programmed using the DSP TMS-320 F241. The utilization of a DSP gives flexibility when changes in parameters, modifications, or improvements need to be included in the control scheme [23]. Fig. 4. (a) Switching frequency and voltage amplitude in each H converter. (b) Voltage waveforms for 27 levels. is to make active filters and var compensators more useful by adding active power sources at the dc link instead of a typical capacitor. C. Multiconverter Operation Fig. 4(a) shows the switching frequency in each one of the three H converters of the multistage converter implemented. It can be noted that the switching frequency of the main converter, which manages more than 80% of the total power, is the same frequency of the system, in this case, only 50 Hz. The frequency of the auxiliary converters is also low but increases as the voltage level of the converter becomes lower in the chain, as seen in Fig. 4(a). The modulation algorithm to synthesize the waveforms shown in Fig. 4 is described as follows: There is a three-digit trinary number associated with the instantaneous amplitude of the voltage. The three-digit number could be +1, A. Active Power Control The control of the active power generated (P ) is through the SOC of the battery pack and is shown in Fig. 6. If SOC > 70%, then the battery pack can inject power to the mains (I REF > 0 in Fig. 6). When SOC is between 70% and 60%, then no power is injected to the mains, and I REF will be set to zero. Now, when SOC is lower than 60% because solar panels are not giving enough energy to the battery pack, then I REF < 0, which means that the mains can charge and recover the battery pack. These SOC limits have been arbitrarily selected and are restrained by the physical limitations of the battery pack. For power injection, it could be the maximum current that the photovoltaic array can provide or the maximum discharge current of the pack. For a battery charge, the current is limited by the pack maximum charging current. The charging of the battery through the solar array is independent of the operation of the multilevel system and gives energy to the batteries whenever the sun is radiating. MPPT always delivers the optimum amount of energy from the solar panel, and the multilevel system will only take

4 FLORES et al.: STATIC VAR COMPENSATOR AND ACTIVE POWER FILTER WITH POWER INJECTION CAPABILITY 133 R LIN and L LIN are the parameters of the line that connects the mains to the charge, and cos(ϕ) is the power factor of the inverter. Finally, KP MAINS and KI MAINS are the parameters of the PI block that controls the mains current. Obviously, it is necessary to configure the PI block for the mains current control before configuring the PI block for battery current control. The Appendix contains the set of equations used to obtain the transfer functions. Fig. 6. Active component generation. power from the batteries when SOC information allows this operation. It is important to mention that I P (ref) will be set automatically to a given value from the proportional integral (PI) block, even when I REF is set to zero. This is because I P (ref) has to satisfy the active power requirements of the contaminating loads, which will come from the mains supply. Looking at the connection between the V MAINS and the multilevel converter in Fig. 5, it can be seen that under a no-load scenario, I MAINS = I FILTER.IfI REF is set to zero (Fig. 6), then I BAT will be zero, and hence, automatically I P (ref) will be set to zero (no active power to or from the converter because I BAT =0), and there will be no current I LOAD because this is a no-load condition. When a load is connected to the system and I REF is still set to zero, then I BAT =0,butI P (ref) is automatically set by the PI control in Fig. 6 at a value required for the active load that has been connected. When I REF is set positive because the battery has SOC > 70%, then I BAT will follow I REF to give active power to the system, helping the mains to feed the load. Under a no-load condition, the battery energy will go directly to the mains supply. The opposite situation will be observed when I REF is set to a negative value (SOC < 60%). In both cases, the value of I REF will depend on the size of the battery pack and the power capability of the multilevel converter. Fig. 7 shows the complete control topology for the system, including the two PI blocks. Both PI blocks are configured using MATLAB s sisotool, which is a way to graphically configure a compensator for a transfer function based on the root locus method. The transfer functions, one for each block, are given in (1) and (2), shown at the bottom of the page. In both transfer functions, K INV represents the inverter gain, B. Reactive Power Control The injection of reactive power (Q) is only necessary when for some reason the system requires reactive power assistance. When Q is set to zero (I Q(ref) =0), the compensating system automatically generates the Q required for all loads connected after the multilevel converter. In other words, when Q is set to zero, the mains always see a unity power factor, independent of the type of contaminating load being compensated. If Q is set to a positive value, then the multilevel converter can assist the mains, injecting additional reactive power to the grid. Q can also be set to negative values, but this operation is normally not required. According to this explanation, the system works like a synchronous machine, where the reactive power is controlled through the excitation coil. In this case, it is controlled through I Q(ref). IV. SIMULATIONS Simulations were performed using the software PSIM [24]. Fig. 8 shows a typical battery charge discharge operation, when I LOAD =0and I Q(ref) =0(see Fig. 7). As can be observed, the quality of currents is quite good. As I Q(ref) has been set to zero, no reactive power comes from the mains, which means that I MAINS is in phase with the voltage V MAINS,asshownin the figure. Fig. 9 shows a reactive power reference change from zero to a positive and then to a negative number when I LOAD =0.It can be appreciated that the current I MAINS changes from zero to leading and then to lagging. In this simulation, I REF has been set to zero. Moreover, the current waveform is quite good. Fig. 9 shows the filter operation when a power rectifier is connected as contaminated load. The current of the rectifier is shown as I LOAD and changes from zero (no load) to 20 A dc and then to almost 40 A. In this case, I Q(ref) has been set to zero, and the battery current is also zero. As a result, the source current I MAINS remains in phase with the voltage V MAINS, and I FILTER (from the multilevel converter), generates all the harmonics required by the load. In this case, only the 27-level operation is displayed. I MAINS V MOD = K INV L LIN s + R LIN (1) I BAT IREF MAINS = 3 K INV cos(φ) (KP MAINS s + KI MAINS ) V MAINS V BAT s (L INV s + R INV ) (2)

5 134 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 1, JANUARY 2009 Fig. 7. Complete control topology. is connected. This inductive load is again disconnected at t = 30 ms. In the simulation, the current I REF is changing from positive to negative values every 10 ms. As can be seen in Fig. 11(a), the filter current is compensating the harmonics of the rectifier to keep the source current I MAINS sinusoidal [Fig. 11(c)]. On the other hand, Fig. 11(b) is showing I REF, I BAT, and I P (ref). It can be noted that the battery current I BAT oscillates when the rectifier is connected to the system. This is due to the harmonics generated by the multilevel converter to compensate the load. Fig. 8. Fig. 9. Step battery charge discharge. Step reactive power change. In Fig. 10, the oscillogram located at the bottom shows the small phase displacement between the mains voltage V MAINS and the inverter voltage V MULTILEVEL. It can be noted that they are in phase when I LOAD =0, and then, they begin to shift, as expected, when the active load appears. Fig. 11 shows a combined operation of the multilevel converter. In the first 5 ms, the converter is injecting Q to the power systems (I MAINS leads V MAINS, as shown in the last oscillogram). Then, the battery begins to be charged from the mains, and I MAINS becomes more in-phase with V MAINS. Around t =8ms, the power rectifier is connected. Later on, at t =10ms, an inductive load (in parallel with the rectifier) V. E XPERIMENTAL RESULTS The prototype used for the experiments is shown in Fig. 12. It is a four-stage 81-level converter, able to operate as a threestage 27-level converter as well as a two-stage nine-level converter. The power connections are like the drawing showed in Fig. 2. All the control tasks are executed by a DSP, which allows the programming of the algorithms related with this system. Fig. 13 clearly shows the 27 levels of the power converter. On other hand, Fig. 14 shows the phase voltage and currents in one of the three phases of the multiconverter when it feeds a Graetz bridge rectifier. Finally, in Fig. 15, a transient response under rectifier connection is showed. In this case, the control keeps active power from the mains constant, and reactive power Q =0(changes in reference Q were not experimentally implemented). Finally, the last oscillogram shows the battery current, which is being charged from the mains, from large to small current. This oscillogram also shows perturbations produced by harmonic voltage in the waveform obtained with the 3-kVA prototype in Fig. 12. VI. CONCLUSION A 27-level active power filter and SVC, with active power generation capability, has been implemented simulated and tested. Each phase of the filter is composed of three H converters per phase, all of them connected to a battery pack. The filter can compensate load currents with a high harmonic

6 FLORES et al.: STATIC VAR COMPENSATOR AND ACTIVE POWER FILTER WITH POWER INJECTION CAPABILITY 135 Fig. 10. Step rectifier load. Fig. 11. Combined operation of the multilevel converter (phase a). (a) I FILTER, I REACTOR, I LOAD,andI RECT.(b)I BAT, I P (ref),andi REF.(c)I MAINS, V MAINS,andI LOAD. possible to deliver active power to the net when the SOC of the batteries is higher than 70%, allowing feeding the contaminating load during prolonged voltage outages. Simulation results for this application are shown, and some experiments with a 3-kVA device allow us to verify the operation of the proposed system. Future work will focus on the efficiency and losses of transformers. APPENDIX The basic equations for the converter control are the voltage drop between the mains (V MAINS ) and the inverter (V MOD ) V MAINS V MOD = Z LIN I MAINS Z INV I FILTER (3) Fig. 12. Twenty-seven-level cascade H-bridge prototype. distortion and a low power factor, resulting in sinusoidal currents from the source. The battery pack can be charged from either the ac mains supply during nights or a photovoltaic array during sunny days. This solar panel is connected to the batteries through MPPT, which is able to maximize the power conversion from the sun. These characteristics makes it where Z LIN and Z INV are the line impedance and the transformer series impedance, respectively. The inverter is assumed to be an amplifier with a gain K INV (inverter s simplified model) V MOD = K INV V REF. (4) The mains line impedance is given as follows: Z LIN = R LIN + L LIN d/dt. (5)

7 136 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 1, JANUARY 2009 Fig. 13. (Bottom) Voltage step waveforms in the 27-level converter (scale: 100 V/div). Fig. 14. Unity power factor operation with contaminating load (rectifier). (Top) Rectifier current I RECT. (Middle) Filter current I FILTER. (Bottom) Mains current I MAINS and mains voltage V MAINS (in phase). (scale: 10 A/div and 100 V/div). Fig. 15. Transient response under rectifier connection. (Top to bottom) Rectifier current I RECT, filter current I FILTER, mains current I MAINS, and battery current I BAT, which changes from heavy to light load (scale: 10 A/div). The inverter s connection line impedance is given as follows: Z INV = R INV + L INV d/dt. (6) Replacing (4), (5), and (6) into (3), we obtain V MAINS K INV V REF =R LIN I MAINS +L LIN di MAINS /dt R INV I FILTER L INV di FILTER /dt. (7) Linearization is possible where the inverter s output current is considered to be constant, given that it derives from the reference signal imposed to the inverter. The mains voltage is assumed to be constant. The linearization around an operation point for V MOD, I MAINS, and di MAINS /dt is K INV ΔV REF =R LIN ΔI MAINS +L LIN ΔdI MAINS /dt. (8)

8 FLORES et al.: STATIC VAR COMPENSATOR AND ACTIVE POWER FILTER WITH POWER INJECTION CAPABILITY 137 V MAINS K INV ( KPMAINS (I ) MAINS IREF MAINS ) +KI MAINS ( ) I MAINS IREF MAINS dt V BAT (R INV I BAT + L INV di BAT /dt) = R LIN I MAINS ( KPMAINS (I ) MAINS I MAINS ) REF 3 K INV +KI MAINS ( ) I MAINS IREF MAINS cos(φ) dt (14) In the Laplace domain, we have K INV V REF (s)=r LIN I MAINS (s)+l LIN s I MAINS (s). (9) The transfer function is [the same as (1)] I MAINS (s) V REF (s) = K INV L LIN s + R LIN (10) Now, the transfer function between I BAT and IREF MAINS (see Fig. 7) will be obtained. Starting from (6), the transfer function for the battery current has to include the already configured PI block for the mains current control. The input signal of the inverter is the output of the previous PI block: ( ) V REF = KP MAINS + KI MAINS dt (I MAINS IREF MAINS ). (11) The inverter current can be deducted from the active power flow equation of the inverter (balance of the active power between the ac and the dc sides plus the losses) V BAT I BAT =3 V MOD I FILTER cos(φ)+p L I FILTER =(V BAT I BAT P L )/ (3 V MOD cos(φ)). (12) Replacing (12) into (7), we obtain V MAINS K INV V MOD = R LIN I MAINS R INV (V BAT I BAT P L )/ (3 V MOD cos(φ)) L INV d ( VBAT I BAT P L 3 V MOD cos(φ) )/ dt V MAINS K INV V REF = R LIN I MAINS R INV VBAT I BAT P L 3 V MOD cos(φ) V BAT L INV 3 V MOD cos(φ) di BAT. (13) dt It can be seen that V MOD has been left out of the differentiation caused by L INV. The reason for this is that V MOD is the result of the control block for the ac sinusoidal mains current, which has a considerably faster time constant than the control block for the dc battery current in the tuning process. Hence, for differentiation purposes, the sinusoidal parameters can be considered as constants, which means that their differentiations are zero (the same applies for the differentiation of I MAINS caused by L LIN ). Replacing (4) and (11) into (13), we obtain (14), shown at the top of the page. The expansion of the equation will introduce quadratic expressions for the state variables and again linearization comes in handy. For space considerations, the steps of linearization and replacement of I REF0 = IREF0 MAINS will not be displayed. In the Laplace domain, we have 3 s IREF MAINS (s) K INV cos(φ) V MAINS KP MAINS +3 s IREF MAINS (s) R LIN I MAINS0 K INV cos(φ) KP MAINS 3 IREF MAINS (s) K INV cos(φ) V MAINS KI MAINS +3 IREF MAINS (s) R LIN I MAINS0 K INV cos(φ) KI MAINS +3 s I MAINS (s) K INV cos(φ) V MAINS KP MAINS 3 s I MAINS (s) R LIN I MAINS0 K INV cos(φ) KP MAINS +3 I MAINS (s) K INV cos(φ) V MAINS KI MAINS 3 I MAINS (s) R LIN I MAINS0 K INV cos(φ) KI MAINS + V BAT L INV s 2 I BAT (s)+v BAT R INV s I BAT (s) =0. (15) Given that the only interest is on the transfer function between I BAT (s) and IREF MAINS (s), I MAINS (s) is set to zero, and the transfer function (2) is obtained I BAT (s) (s) I MAINS REF = 3 K INV cos(φ) V MAINS (KP MAINS s+ki MAINS ). (16) V BAT s (L INV s+r INV ) Since the resistive component of the mains line impedance is expected to be low, the mains current equilibrium value can be neglected from the final transfer function, which is also evaluated for a unity power factor. REFERENCES [1] H. Akagi, The state-of-the-art power electronics in Japan, IEEE Trans. Power Electron., vol. 13, no. 2, pp , Feb [2] B. Bose, Power electronics and motion control Technology status and recent trends, IEEE Trans. Ind. Appl., vol. 29, no. 5, pp , Sep./Oct [3] Ó. Lopez, J. Alvarez, J. Doval-Gandoy, and F. D. Freijedo, Multilevel multiphase space vector PWM algorithm, IEEE Trans. Ind. Electron., vol. 55, no. 5, pp , May [4] A. Draou, M. Benghanen, and A. Tahri, Multilevel converters and VAR compensation, in Power Electronics Handbook, M. H. Rashid, Ed. New York: Academic, 2001, ch. 25, pp [5] J. Rodríguez, J.-S. Lai, and F. Zheng Peng, Multilevel inverters: A survey of topologies, controls, and applications, IEEE Trans. Power Electron., vol. 49, no. 4, pp , Aug [6] F. Zheng Peng, A generalized multilevel inverter topology with self voltage balancing, IEEE Trans. Ind. Appl., vol. 37, no. 2, pp , Mar./Apr

9 138 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 1, JANUARY 2009 [7] M. D. Manjrekar, P. K. Steimer, and T. A. Lipo, Hybrid multilevel power conversion system: A competitive solution for high power applications, IEEE Trans. Ind. Appl., vol. 36, no. 3, pp , May/Jun [8] J. Dixon and L. Morán, High-level multistep inverter optimization using a minimum number of power transistors, IEEE Trans. Power Electron., vol. 21, no. 2, pp , Mar [9] A. Chen and X. He, Research on hybrid-clamped multilevel-inverter topologies, IEEE Trans. Ind. Electron., vol. 53, no. 6, pp , Dec [10] M. Ortúzar, R. Carmi, J. Dixon, and L. Morán, Voltage-source active power filter based on multilevel converter and ultracapacitor DC link, IEEE Trans. Ind. Electron., vol. 53, no. 2, pp , Apr [11] J. Dixon and L. Morán, A clean four-quadrant sinusoidal power rectifier, using multistage converters for subway applications, IEEE Trans. Ind. Electron., vol. 52, no. 5, pp , May/Jun [12] F.-S. Kang, S.-J. Park, M. H. Lee, and C.-U. Kim, An efficient multilevelsynthesis approach and its application to a 27-level inverter, IEEE Trans. Ind. Electron., vol. 52, no. 6, pp , Dec [13] S. Lakshminarayanan, G. Mondal, P. N. Tekwani, K. K. Mohapatra, and K. Gopakumar, Twelve-sided polygonal voltage space vector based multilevel inverter for an induction motor drive with common-mode voltage elimination, IEEE Trans. Ind. Electron., vol. 54, no. 5, pp , Oct [14] F. Z. Peng, J. S. Lai, J. McKeever, and J. Van Coevering, A multilevel voltage source inverter with separate dc sources for static VAr generation, in Conf. Rec. IEEE IAS Annu. Meeting, 1995, pp [15] N. Mohan and G. R. Kamath, A novel, per-phase interface of power electronic apparatus for power system applications, in Proc. NAPS,1995, pp [16] F.-S. Kang, S.-J. Park, S. Eog Cho, C.-U. Kim, and T. Ise, Multilevel PWM inverters suitable for the use of stand-alone photovoltaic power systems, IEEE Trans. Energy Convers., vol. 20, no. 4, pp , Dec [17] BRUSA: Maximum Power Tracking Model MPT-N15, Brusa Elektronik AG, Sennwald, Switzerland. [Online]. Available: brusa.li/products/g_mpt_n15207.htm [18] S. Kouro, J. Rebolledo, and J. Rodríguez, Reduced switching-frequencymodulation algorithm for high-power multilevel inverters, IEEE Trans. Ind. Electron., vol. 54, no. 5, pp , Oct [19] M. Rotella and J. Dixon, Simplified 27-level traction drive system with low part count and single battery pack, in Proc. 22th EVS, Yokohama, Japan, Oct. 7 10, CD-ROM. [20] J. Rodríguez, B. Wu, S. Bernet, J. Pontt, and S. Kouro, Multilevel voltage-source-converter topologies for industrial medium-voltage drives, IEEE Trans. Ind. Electron. Special Section on High Power Drives, vol. 54, no. 6, pp , Dec [21] A. K. Gupta and A. M. Khambadkone, A space vector PWM scheme for multilevel inverters based on two level space vector PWM, IEEE Trans. Ind. Electron., vol. 53, no. 5, pp , Oct [22] C. Rech and J. R. Pinheiro, Hybrid multilevel converters: Unified analysis and design considerations, IEEE Trans. Ind. Electron., vol. 54, no. 2, pp , Apr [23] E. Roman, R. Alonso, P. Ibanez, S. Elorduizapatarietxe, and D. Goitia, Intelligent PV module for grid-connected PV systems, IEEE Trans. Ind. Electron., vol. 53, no. 4, pp , Jun [24] Power Electronics Simulations, User Manual, Powersim Technol., Vancouver, BC, Canada. PSIM Version 6.1. [Online]. Available: Patricio Flores was born in Santiago, Chile. He received the B.S. degree from the Pontificia Universidad Católica de Chile, Santiago, in He is currently working toward the M.S. degree in the Department of Electrical Engineering, Pontificia Universidad Católica de Chile. Since 2005, he has been working on active power filter projects. Juan Dixon (M 90 SM 95) was born in Santiago, Chile. He received the professional degree in electrical engineering from the Universidad de Chile, Santiago, in 1977, and the M.Eng. and Ph.D. degrees from McGill University, Montreal, QC, Canada, in 1986 and 1988, respectively. Since 1979, he has been with the Department of Electrical Engineering, Pontificia Universidad Católica de Chile, Santiago, where he is currently a Professor. He has presented more than 70 works at international conferences. He is the author of more than 30 papers related to power electronics in international journals. His main areas of interest are in electric traction, PWM rectifiers, active filters, power factor compensators, and multilevel converters. He has created an electric vehicle laboratory, where state-of-the-art vehicles are investigated. Micah Ortúzar received the professional degree in electrical engineering and the Ph.D. degree in 2005 from the Pontificia Universidad Católica de Chile, Santiago, Chile. He worked in the areas of power active filters, power electronics, and electric vehicles research projects while working toward his Ph.D. degree. He continued working on power-electronics-related research projects at the same university from 2005 to He is currently with the Compania Americana de Multiservicios Ltda. (CAM), Endesa, Santiago, in the energy distribution industry. Rodrigo Carmi was born in Santiago, Chile. He received the professional degree in electrical engineering from the Pontificia Universidad Católica de Chile, Santiago, in Between 2003 and 2005, he worked on active power filter projects. He is currently with the National Transmission Electrical Utility Company, Santiago. Mr. Carmi is a coauthor of two papers published in IEEE conference proceedings. Pablo Barriuso received the electrical engineering and M.Sc. degrees from the Pontificia Universidad Católica de Chile, Santiago, Chile, in Currently, he is an Engineer with CDEC-SIC, Santiago, Chile, the ISO of the Chilean electrical interconnected system. During his M.Sc. preparation, he was involved with the design and application of power electronics devices and studied the fields of photovoltaic generation, harmonic filtering, and fault-tolerant control. Luis Morán (S 79 M 81 SM 94 F 05) was born in Concepción, Chile. He received the B.Eng. degree in electrical engineering from the Universidad de Concepción, Concepción, in 1982, and the Ph.D. degree from Concordia University, Montreal, QC, Canada, in Since 1990, he has been with the Department of Electrical Engineering, Universidad de Concepción, where he is a Professor. His main areas of interests are in ac drives, power quality, active power filters, FACTS, and power protection systems. Dr. Morán is the author of more than 30 papers on active power filters and static var compensators published in IEEE TRANSACTIONS. He is the principal author of the paper that received the IEEE Outstanding Paper Award from the IEEE Industrial Electronics Society for the best paper published in the IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS in 1995 and is the coauthor of the paper that was awarded in 2002 by the IEEE Industry Applications Society Static Power Converter Committee.