IN THE LAST few decades, the field of high-power drives

Transcription

1 450 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 60, NO. 2, FEBRUARY 2013 A Novel Noninvasive Failure-Detection System for High-Power Converters Based on SCRs Victor Guerrero, Student Member, IEEE, Jorge Pontt, Senior Member, IEEE, Juan Dixon, Senior Member, IEEE, and Jaime Rebolledo, Student Member, IEEE Abstract In modern high-power industrial processes, multipulse cycloconverters are fundamental to feed synchronous motors. Given the importance of these critical processes, particularly for mining industry, the reliability of the high-power drive is critical. Having a noninvasive system capable of detecting and isolating faults in the thyristors is quite valuable, and it will reduce the economic losses caused by the long repair times of the converter. Currently, it is difficult and, in some cases, almost impossible to determine which of the SCRs experience an overcurrent or those which do not trigger properly, causing a malfunction. In this paper, we present a new method of online detection which is capable of delivering information about the SCR involved in a failure and the nature of it. This new method takes only a few input and output measurements (noninvasive feature) already available in the most of industrial cycloconverters and performs a timing analysis of these measurements. The algorithms of the method estimate and analyze the switching states of all thyristors, providing an impressive way of detecting the cause and consequences of the failure in the cycloconverter. Index Terms Diagnostics, drive, fault tolerance, industrial application, reliability. I. INTRODUCTION IN THE LAST few decades, the field of high-power drives has shown a great development mainly because all industrial processes have been increasing their power demand to achieve a large-scale economy [1], [2]. The high-power drive applications include pumps and compressors (petrochemical industry), laminators (metal industry), trucks and ships (transportation), active filters, high-voltage direct current and wind power conversion (energy industry), etc. [19]. In the mining industry, the use of multipulse cycloconverters to feed semiautogenous grinding mills is quite common [3]. Because of this, it is necessary to have monitoring and faultdetection systems capable of delivering an accurate and fast diagnostic of any failure that occurred in the converter, reducing the economic losses due to a long repair process. Manuscript received December 2, 2010; revised May 14, 2011 and December 21, 2011; accepted January 1, Date of publication February 16, 2012; date of current version September 13, This work was supported in part by the CONICYT-Chile Project , by the Chilean Millennium Scientific Initiative under Grant P F, and by the Universidad Técnica Federico Santa María, Valparaíso, Chile. V. Guerrero, J. Pontt, and J. Rebolledo are with the Universidad Técnica Federico Santa María, Valparaíso , Chile ( victor.guerrerob@ alumnos.usm.cl; jorge.pontt@usm.cl; jaime.rebolledo@usm.cl). J. Dixon is with the Department of Electrical Engineering, Pontificia Universidad Católica de Chile, Santiago , Chile ( jdixon@ing.puc.cl). Color versions of one or more of the figures in this paper are available online at Digital Object Identifier /TIE The new method presented in this paper consists of a series of algorithms which is capable of giving a detailed information about the failure that occurred in the cycloconverter, processing only the values of input and output currents of the converter (the noninvasive feature of the method). Although, in this paper, this method is presented for the 12-pulse cycloconverter without circulating current, it is important to say that this detection system works for any multipulse cycloconverter. This paper shows, in detail, the operation of the 12-pulse cycloconverter based on SCRs (Section II), showing the power circuit, firing-pulse-generation method, and the free cirulating current configuration of the system. Section III shows all the theoretical bases of two very important algorithms in the fault-detection system: switchingstate-estimation and fault-detection algorithms. Finally, in Section IV, it presents the simulation and experimental test under fault conditions which verify the diagnosis capability of the fault-detection system. II. THE 12-PULSE CYCLOCONVERTER WITHOUT CIRCULATING CURRENT Synchronous motors are the preferred solution for highpower mills [4], and cycloconverters are one of the most common ways to feed them. That is why a lot of research has been made to propose methods to improve the power quality of these systems [5], [6], mitigating harmonics and interharmonics present on these drives [7] [9]. This section discusses the simulated and built cycloconverter that was used for the testing process of the detection system proposed in this paper. A. Power Circuit The electrical diagram of the 12-pulse cycloconverter without circulating current is shown in Fig. 1, where the presence of four rectifier bridges in every phase (each one with six SCRs) can be seen. There are two groups of bridges connected in parallel, the positive bridges and the negative bridges, each one is composed of two bridges in a series. The positive bridges carry the positive cycle of the load current, and the other group is active only when the load current is negative. Table I shows all the construction and simulation parameters of the converter, which were used for the diagnostic tests of the failure detection system. These parameters are based on typical requirements for these types of drives [10] /$ IEEE

2 GUERRERO et al.: NOVEL NONINVASIVE FAILURE-DETECTION SYSTEM FOR HIGH-POWER CONVERTERS 451 Fig. 1. Twelve-pulse cycloconverter. TABLE I DESCRIPTION OF THE SIMULATED AND BUILT CYCLOCONVERTER B. Firing Pulses Considering the one-phase equivalent shown in Fig. 2, four different firing signals can be identified. The α 1 and α 2 signals are related to the positive bridges, and the α 3 and α 4 signals are the ones which carry the firing angle of the negative bridges. This differentiation in the signals given to the cycloconverter is the basis of the free circulating current mode, where the two groups of bridges never work at the same time. The relationship between the firing angle α and the average output voltage (see Fig. 2) of each bridge of the system is given by V out = v m cos(α i ) (1) where V out represents the average output voltage of the bridge associated with the firing-pulse signal α i and v m is a proportional value to the voltage amplitude of the input in the bridge, which can be the line line voltage of the secondary delta or star depending on which secondary the bridge is connected to. III. SWITCHING-STATE-ESTIMATION AND FAILURE-DETECTION SYSTEM High-power cycloconverter-fed gearless motor drives experience a series of current and failure issues [11]; that is why it is very important to have an intelligent monitoring and an Fig. 2. One-phase equivalent.

3 452 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 60, NO. 2, FEBRUARY 2013 (see Fig. 2) with one of the 12 switching cases of the normal operation; see Table II. It is also possible to establish the currents in each primary of the power transformers with the currents through the secondaries of the transformers using the following equations: Fig. 3. Switching sequence. accurate diagnosis system to overcome these difficulties [12]. This section shows the way the fault-detection system works and gives the theoretical bases of its functioning. The detection system consists of a series of algorithms that runs in parallel for the three phases of the cycloconverter. They measure the i a, i b, and i load of the phase being studied; these currents are defined in Fig. 2. The system delivers information about the switching state of every SCR and the nature of any fault that occurred in the analyzed phase. If this method is implemented for the three phases of the cycloconverter, it is possible to establish the switching state of any of the 72 SCRs of the cycloconverter (CCV), as well as the possibility of detecting a failure in the CCV and the nature of it. The failure detection system consists of four stages. The first stage is to take the correct measurements. For proper operation, the system must take the measurements of the currents in the primary and the load of the phase under study, as shown in Fig. 2. These measurements are taken by the switching-stateestimation algorithm, the second stage. This stage provides information about the switching state of every SCR of the CCV. In the third stage of the detection system, there is an algorithm which takes the switching state estimation of the previous stage and generates a report of the fault and the nature of it, giving way to the last stage where all the information is delivered. This information includes the type of the fault that occurred and the malfunctioned SCR. A. Switching-State-Estimation Algorithm Analyzing the normal operation of the converter, it is possible to establish that the natural commutation of the SCRs is the pattern shown in Fig. 3, where any of the two pairs of rectifier bridges operates (either the positive or negative). This figure shows that there are 12 switching cases in the normal operation of the converter. With these cases, it is possible to build a table that associates the switching state of each of the thyristors i a = N 2Y (i ay i by )+ N 2Δ i aδ (2) N 1 N 1 i b = N 2Y (i by i cy )+ N 2Δ i bδ (3) N 1 N 1 i c = N 2Y (i cy i ay )+ N 2Δ i cδ (4) N 1 N 1 where i a, i b, and i c represent the currents through the primary of the transformer; i ay, i by, and i cy are the currents through the secondary star; i aδ, i bδ, and i cδ are the currents through the secondary delta; N 2Y /N 1 represents the ratio of the number of turns between the primary and the star secondary; and N 2Δ /N 1 represents the ratio of the number of turns between the primary and the delta secondary. If the equations that relate the currents in the primary and secondary of the transformers are available and the currents in the secondary in the normal operation of the CCV can only take values as i load, i load, or zero depending on the switching state of each SCR (where i load represents the current through the connected load in the analyzed phase), it is possible to construct a table that relates a normal switching case with a particular relationship between primary and load currents (see Table III). For calculation reasons, it is considered that the relations of turns in the windings of the transformers allow an identical line line magnitude voltage in the primary delta and secondary delta. The main function of the switching-state-estimation algorithm is to take the numerical values of the currents in the primary and the load of each phase, compare the relationship between them using Table III, and delivering information about the switching case presented in the CCV. Once the switching state case is known, Table II is used to deliver the information about the switching state of every SCR ( on or off state). In normal conditions, the CCV should go from case to case, ascending in repetitive order (from case 1 to 12), and with a little transient which is defined as case 13. This case 13 represents the small moments in which the SCRs pass from the states on to off or vice versa (or where the relationship between the input and load currents is not defined as a normal conducting case). The normal pattern for the switching cases is shown in Fig. 4. This pattern is the output of the switchingestimation algorithm; this type of results is taken by the faultdetection algorithm in the next stage. B. Fault-Detection Algorithm This fault-detection system is focused on two common types of faults: firing-pulse failures, either by errors in the firing circuit or by faults on the thyristor itself, and loss of blocking capability [13]. The operation of the fault-detection algorithm for those types of faults is explained in the next subsections. 1) Firing-Pulse-Failure Detection: As shown in Fig. 3, there is a well-defined pattern for the switching state of every

4 GUERRERO et al.: NOVEL NONINVASIVE FAILURE-DETECTION SYSTEM FOR HIGH-POWER CONVERTERS 453 TABLE II SWITCHING CASES TABLE III PRIMARY CURRENTS ACCORDING TO SWITCHING STATE Fig. 5. Switching estimation under T4S firing-pulse-failure operation. Fig. 4. Normal switching pattern. thyristor of the converter. Therefore, it is necessary to study how a fault affects this switching pattern, depending on which SCR incurs in a fault and the nature of it. When a single SCR suffers a firing-pulse failure, the rest of the thyristors of the converter do not change their switching states [14]. In fact, the switching pattern for all the rest of the thyristors of the converter (which are not affected for any type of failure) remains identical. For example, if, for some reason, the thyristor T4 of the bridge, connected to the star secondary of the transformer (T4S in Fig. 2), suffers a firing-pulse failure (cannot pass to the on state), the output of the switching-estimation algorithm will be as seen in Fig. 5. As it can be observed in the resulting switching estimation due to a T4S firing-pulse failure, there are several signs that may indicate which was the failed SCR. As it can be seen, the switch pattern is lost after the case 6. In Table II of the switching cases, between the cases 6 and 7, the thyristor T4S located in the star connection should pass to the on state, which immediately delivers information about the failed SCR. Also, by analyzing Fig. 5, it can be established that the missing cases in the commutation pattern are the cases from 7 to 10. In Table II, it can be observed that the only thyristor that should be on the on state in cases 7 10 is precisely the SCR T4S of the star connection; therefore, with this information, it is possible to establish which thyristor was affected by the fault. Performing the same procedure explained earlier for all the thyristors in the converter, it can be established that if a firingpulse failure occurs, there will be four missing cases in the switching pattern. These missing cases are precisely the ones where the failed SCR should be on the on state. So, the failure-detection algorithm takes the switching state estimation of the previous stage (output of the switching-state-estimation algorithm) and analyzes the missing cases. Depending on these cases, the algorithm can inform about the failed SCR and the characteristic of the failure. 2) Loss-of-Blocking-Capability Detection: It is known that the SCR can pass to the on state only when the voltage across its terminals is positive and an adequate current is injected by

5 454 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 60, NO. 2, FEBRUARY 2013 Fig. 7. Load current and voltage for phase A: Normal operation. Fig. 6. Switching estimation with a loss of blocking capacity of T5S. the gate. The operation of passing to the off state is different because it only depends on the external conditions of the SCR. For this reason, we can say that the commutation of the SCR is semicontrolled. In some occasions, the SCR cannot leave the on state, generating a short circuit with the parallel leg of the rectifier bridge where the failure is presented [15], [17]. If the SCR T5S of the bridge connected to the secondary star has a loss-of-blocking-capacity failure (so it cannot leave the on state), there will appear a short circuit between the thyristors T1S and T5S, both located at the star connection (see Fig. 2). This happens because the thyristor T1S is the one that must turn off the failed SCR T5S. This short circuit generates an overcurrent in the primary of the transformer and an output for the switching-state-estimation algorithm as it is seen in Fig. 6. The switching estimation shows that eight cases are missing, remaining only the cases between 9 and 12. The Table II of switching cases shows that the only thyristor that should be conducting in these four cases is precisely the T5S, which was assumed in failure at the beginning. In this way, the detection algorithm identifies the presence of any overcurrent in the converter and analyzes the missing cases. With this data, it is possible to establish the failed SCR and differentiate the type of failure. IV. RESULTS OF SIMULATIONS AND EXPERIMENTAL TESTS This section presents the simulations and experimental tests realized in the laboratory. The results of these tests indicate a correct diagnostic capacity of the detection system. A. Simulation Results The entire drive was simulated with the computational tool PSCAD/EMTDC (converter plus synchronous machine) with the field-oriented control technique. In the first stage, the results in the normal operation of the CCV are shown, and then, the veracity of the diagnosis given by the detection system is tested for the different scenarios of failures that occurred in the drive. 1) Normal-Operation Simulation: In order to observe if the switching estimation algorithm works properly, the complete system, the converter plus synchronous machine, is simulated without any kind of failure. Fig. 8. Fig. 9. Input current for phase A: Normal operation. Switching pattern for phase A: Normal operation. In Fig. 7, the current and the voltage in the load for phase A can be observed. The current is quite normal and smooth because of the inductive characteristic of the synchronous machine. The input currents for the phase A can be observed in Fig. 8. In this oscillogram, the characteristic waveforms of a 12-pulse configuration are displayed. The currents previously mentioned were taken from PSCAD and processed in MATLAB, where the algorithms of the switching estimation and failure detection were programmed. In Fig. 9, the output for the switching-estimation stage for the phase A can be observed. As it is shown, the cases evolve from 1 to 12, through all the normal states, without any presence of overcurrents or other irregular signs in the system. That is why the failure-detection algorithm, noting a normal evolution of states, does not detect any failure. 2) Firing-Pulse-Failure Simulation: Now, it is necessary to simulate a failure in the CCV and observe if the switching estimation will behave in such a way that the failure-detection algorithm is able to establish the nature of the failure and what SCR incurred on it.

6 GUERRERO et al.: NOVEL NONINVASIVE FAILURE-DETECTION SYSTEM FOR HIGH-POWER CONVERTERS 455 Fig. 10. Current and voltage for phase A under T4S firing-pulse-failure operation. Fig. 13. Voltage and load current: The loss-blocking-capability failure of T5S. Fig. 11. Input currents for phase A under T4S firing-pulse-failure operation. Fig. 14. of T5S. Input currents of the converter: The loss-blocking-capability failure Fig. 12. Output for the switching-estimation algorithm under T4S firingpulse-failure operation. To simulate a firing-pulse failure in PSCAD, the control signal of the SCR T4S, located in the star connection of the positive cycle of the phase A (in t =3.19 s), was disabled. In other words, firing pulses will not reach the gate of SCR T4S. In Fig. 10, the current and the voltage in the load for the phase A can be observed. It is clear that from t =3.19 s, the current and the voltage of the positive cycle look quite affected. This occurs principally because from that time, the T4S remains on the off state. In the same way, Fig. 11 shows how the currents of the transformer of the phase A through the primary also experience an abnormal behavior, although there is no overcurrent in the converter or any transformer. The output for the switching-estimation stage can be seen in Fig. 12. It is clearly seen that the cases 7 10 disappear. Table II shows that the only SCR that works in those four cases is precisely the T4S of the star connection. Hence, the failuredetection algorithm, with the information about the absence of any overcurrent in the converter and the data of the four missing Fig. 15. Switching estimation: The loss-blocking-capability failure of T5S. cases, can deliver a correct diagnosis about the failed SCR and the nature of it. 3) Loss-of-Block-Capability Simulation: The detection system is able to detect a loss in the blocking capability of any of the thyristors of the converter. That is why an incapability of the T5S (located at the star connection of the positive cycle of the phase A) to leave the on state is simulated. In Fig. 13, the current and the voltage for the analyzed phase (phase A) can be seen. It can be observed that both cycles, the positive and the negative, are affected by the short circuit between the T1S and T5S in the star connection of the positive cycle. Also, it is possible to see overcurrents in the converter [16], particularly in the currents of the primary (see Fig. 14). The output of the switching-estimation algorithm is shown in Fig. 15. This oscillogram has quite relevant aspects that can be highlighted: 1) The overcurrent appears right after the failure (t =3.25 s), indicating the presence of a short circuit in the converter, and 2) the overcurrent occurs during the case 12. Looking at Table II, to pass from case 12 to case 1, it is necessary that the thyristor T1S located in the star connection

7 456 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 60, NO. 2, FEBRUARY 2013 Fig. 18. Switching estimation: Normal operation. Fig. 16. Input and load currents: Normal operation. Fig. 19. Load current of phase A under T1S firing-pulse operation. Fig. 17. Currents and switching estimation: Normal operation. of the positive cycle passes to the on state, turning off the thyristor T5S located in the same bridge. However, this situation does not happen; in this case, the estimation fails to pass from case 12 to case 1, which clearly indicates problems in the blocking capacity of the T5S. That is why the failuredetection algorithm is able to detect any failure of this kind by analyzing the last switching case before the presence of any overcurrent. B. Experimental Results In order to verify the proper functioning of the failuredetection system, it is proceeded to measure the currents in the primary and in the load of the cycloconverter built in the laboratory. These data were processed with the algorithms programmed in MATLAB. A series of forced failures in the converter was made, which was compared to the diagnosis given by the detection system. 1) Normal Operation: In order to establish whether the detection system is able to give an accurate diagnosis and a correct switching estimation, a series of measurements was made (primary and load currents) without forcing any type of failure. Fig. 16 shows a normal behavior of the load current and primary current of phase A. The switching case estimation can be seen in Figs. 17 and 18. These figures also show that the switching cases are ascending between 1 and 12 through the case 13, which represents the transient switching of each of the thyristors. Fig. 20. Input currents and switching estimation under T1S firing-pulse operation. 2) Firing-Pulse Failure: In order to force a failure in the cycloconverter, the firing-pulse signal of the thyristor T1S located at the star connection of the positive cycle of the phase A was disabled. In Fig. 19, it can be observed that the load current through this phase is clearly affected, particularly in the positive cycle of the current, which is quite logical because the SCRs of the negative bridges are working properly. Fig. 20 shows how the primary currents are distorted due to the failure forced in the converter and shows an abnormal behavior in the switching-estimation pattern. This anomalous behavior in the evolution of the switching cases can be seen in Fig. 21. There is an absence of four cases, from 1 to 4. Table II shows that the only thyristor which must be on the on state for these four missing cases is precisely the T1S. With this information, the failure-detection system is capable of delivering a correct diagnosis since it accords with the forced failure at the beginning (the firing-pulse failure of the T1S of the secondary star).

8 GUERRERO et al.: NOVEL NONINVASIVE FAILURE-DETECTION SYSTEM FOR HIGH-POWER CONVERTERS 457 Fig. 21. Switching estimation of phase A under T1S firing-pulse operation. V. C ONCLUSION This research has been based on explaining and demonstrating the operation of a failure-detection system for the 12-pulse cycloconverter previously described. One of the main advantages of this system is a noninvasive characteristic since it only considers measurements (input currents on the converter and the stator currents of the synchronous machine) that, in most of the industrial monitoring systems, are already present. So, it is only necessary to take these measurements which are already available, process them according to the algorithm proposed earlier for detecting a failure, and have an estimation of the switching state of the SCRs of the converter. The easy implementation of this method can provide a much more complete drive monitoring system with a better diagnostic capability. The experimental tests under a firing-pulse-failure operation show a remarkable diagnosis capability of the detection system. Further work includes the development of a fault-tolerant multipulse cycloconverter. Taking the fault report given by the detection system presented in this paper and designing a CCV capable of using this information, it is possible to build a converter which could operate under abnormal conditions. REFERENCES [1] J. Rodríguez, B. Wu, S. Bernet, and J. Pontt, Guest editorial high-power drives part I, IEEE Trans. Ind. Electron., vol. 54, no. 6, pp , Dec [2] J. Rodríguez, B. Wu, S. Bernet, and J. Pontt, Guest editorial special edition high-power drives Part II, IEEE Trans. Ind. Electron., vol. 55, no. 3, pp , Mar [3] ABB, ACS 6000c Cycloconverter. [Online]. Available: com/product/seitp322/8bfb9e28c05b62eac f3.aspx [4] J. Rodríguez, J. Pontt, P. Newman, R. Musalem, H. Miranda, L. Moran, and G. Alzamora, Technical evaluation and practical experience of high power grinding mill drives in mining applications, IEEE Trans. Ind. Appl., vol. 41, no. 3, pp , May/Jun [5] D. Basic, V. S. Ramsden, and P. K. Muttik, Performance of combined power filters in harmonic compensation of high-power cycloconverter drives, in Proc. 7th Int. Conf. Power Electron. Var. Speed Drives (Conf. Publ. No. 456), Sep , 1998, pp [6] P. Syam and A. Chattopadhyay, Power Quality Improvement in Cycloconverter-Fed Systems. Saarbrücken, Germany: VDM Verlag Dr. Müller, Jul. 26, [7] M. Loskarn, K. D. Tost, C. Unger, and R. Witzmann, Mitigation of interharmonics due to large cycloconverter-fed mill drives, in Proc. 8th Int. Conf. Harmonics Qual. Power, Erlangen, Oct , 1998, vol. 1, pp [8] D. Basic, V. S. Ramsden, and P. K. Muttik, Selective compensation of cycloconverter harmonics and interharmonics by using a hybrid power filter system, in Proc. 31st IEEE Power Electron. Spec. Conf., Sydney, Australia, Jun , 2000, vol. 3, pp [9] P. Syam, G. Bandyopadhyay, P. K. Nandi, and A. K. Chattopadhyay, Simulation and experimental study of interharmonic performance of a Cycloconverter-FedSynchronous motor drive, IEEE Trans. Energy Convers., vol. 19, no. 2, pp , Jun [10] J. Rodriguez, B. Wu, S. Bernet, N. Zargari, J. Rebolledo, J. Pontt, and P. Steimer, Design and evaluation criteria for high power drives, in Conf. Rec. IEEE IAS Annu. Meeting, Oct. 5 9, 2008, pp [11] P. J. Rodriguez, J. Valderrama, W. Sepulveda, G. Chavez, P. Cuitino, B. Gonzalez, and P. Alzamora, Current issues on high-power cycloconverter-fed gearless motor drives for grinding mills, in Proc. ISIE, Rio de Janeiro, Brazil, 2003, pp [12] P. Jorge, Current topics on reliability of high power electrical machines employed in mining applications, in Conf. Rec. IEEE IAS Annu. Meeting, Oct. 4 8, 2009, pp [13] J. Pontt, J. Rodríguez, E. Cáceres, and I. Illanes, Cycloconverter behavior for a grinding mill drive under firing pulses fault conditions, in Conf. Rec. IEEE IAS Annu. Meeting, Hong Kong, China, 2005, pp [14] I. Illanes and J. Rodriguez, Model-based fault diagnosis applied to 6-pulse cycloconverter, in Proc. IEEE Power Electron. Spec. Conf.,2007, pp [15] J. Pontt, J. Rodriguez, E. Cáceres, I. Illanes, and C. Silva, Cycloconverter drive system for fault diagnosis study: Real time model, simulation and construction, in Proc. IEEE Power Electron. Spec. Conf., Jeju, Korea, 2006, pp [16] J. Pontt, J. Rodríguez, E. Cáceres, and I. Illanes, Integrated monitoring and control of cycloconverter drive system for fault diagnosis and predictive maintenance, in Conf. Rec. IEEE IAS Annu. Meeting, Tampa, FL, 2006, pp [17] J. Pontt, J. Rodriguez, J. Rebolledo, and K. Tischler, Operation of highpower cycloconverter-fed gearless drives under abnormal conditions, IEEE Trans. Ind. Appl., vol. 43, no. 3, pp , May/Jun [18] V. Guerrero and J. Pontt, Oscillatory torque caused by dead time in the current control of high power gearless mills, in Proc. 37th IEEE IECON, Nov. 7 10, 2011, pp [19] M. I. Milanés-Montero, E. Romero-Cadaval, and F. Barrero-González, Hybrid multiconverter conditioner topology for high-power applications, IEEE Trans. Ind. Electron., vol. 58, no. 6, pp , Jun Victor Guerrero (S 10) was born in Puerto Montt, Chile, in He received the Engineer and M.Sc. degrees in electronics engineering from Universidad Técnica Federico Santa María, Valparaíso, Chile, in 2010, where he is currently working toward the Ph.D. degree. His main research interests include fault-detection systems for power converters, mining industry, fault analysis for high-power drives, electric protections, etc. Jorge Pontt (M 00 SM 04) received the Engineer and M.S. degrees in electrical engineering from the Universidad Técnica Federico Santa María (UTFSM), Valparaíso, Chile, in He was with the Technische Hochschule Darmstadt, Darmstadt, Germany ( ), the University of Wuppertal, Wuppertal, Geramany (1990), and the University of Karlsruhe ( ), Karlsruhe, Germany. Since 1977, he has been with UTFSM, where he is currently a Professor in the Department of Electronics Engineering and the Director of the Laboratory for Reliability and Power Quality and of the Nucleus for Industrial Electronics and Mechatronics. He is a Consultant to the mining industry, particularly, in the design and application of power electronics, drives, instrumentation systems, and power quality issues, with the management of more than 80 consulting and R&D projects. He is the coauthor of the software Harmonix used in harmonic studies in electrical systems and of the patent applications concerning innovative instrumentation systems employed in highpower converters and large grinding mill drives. He has authored more than 90 international refereed journals and conference papers.

9 458 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 60, NO. 2, FEBRUARY 2013 Juan Dixon (M 90 SM 95) was born in Santiago, Chile. He received the B.S. 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. In 1976, he was with the State Transportation Company, in charge of trolleybus operation. In 1977 and 1978, he was with the Chilean Railways Company. 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 and has published more than 30 papers related to power electronics in IEEE TRANSACTIONS and Institution of Electrical Engineers Proceedings. His main areas of interest are in electric traction, power converters, pulsewidth modulation rectifiers, active power filters, power factor compensators, and multilevel and multistage converters. He has performed consulting work related to trolleybuses, traction substations, machine drives, hybrid electric vehicles, and electric railways. He has created an electric vehicle laboratory, where he has built state-of-the-art vehicles using brushless dc machines with ultracapacitors and high-specificenergy batteries. Jaime Rebolledo (S 07) was born in Curicó, Chile, in He received the Engineer and M.Sc. degrees in electronics engineering from the Universidad Técnica Federico Santa María, Valparaíso, Chile, in 2006, where he is currently working toward the Ph.D. degree. He is the Cofounder of a spin-off company focused on fault-detection and monitoring systems for high-power drives used in mining applications. His main research interests include high-power converters, embedded systems, and instrumentation.