EPE Association: EPE-PEMC 2002: Special Session: Matrix Converters

EPE-PEMC 2002: Special Session: Matrix Converters
You are here: EPE Documents > 04 - EPE-PEMC Conference Proceedings > EPE-PEMC 2002 - Conference > EPE-PEMC 2002: Special Session: Matrix Converters

   [return to parent folder]

   A Direct Power Electronic Conversion Topology for Multi-Drive Applications
By F. Blaabjerg; C. Klumpner [View]
[Download]

Abstract: The matrix converter is known as a single stage AC/AC converter, which is able to provide sinusoidal PWM output voltages and input currents, inherent bi-directional power flow and no reactive elements (bulky DC-capacitors and large line inductors). However, the increased number of active switches (18), as well as the need of a protection circuit involving twelve diodes with rated voltage/current ratings used only during faulty situation, makes this topology not so attractive. In this paper, a two-stage converter topology consisting of a three-phase to two-phase matrix converter directly linked to a Voltage Source Inverter (VSI) that provides Direct Power Electronic Conversion (DPEC) is analyzed. While the functionality is similar to a matrix converter, the analysis of the conduction losses shows a different shape of the efficiency curve, which in HVAC application may be an advantage. In addition, a topology is proposed where several VSIs are connected to the same DC-link and their switching patterns are synchronized, which is especially useful for multi-drive application. The functionality of the proposed two-stage multi-drive direct power electronic conversion topology is analyzed by simulations.

   Analytically Closed Calculation of the Conduction and Switching Losses of Three-Phase AC-AC Sparse Matrix Converters
By J. W. Kolar; M. Baumann; F. Schafmeister [View]
[Download]

Abstract: For three-phase AC-AC power conversion a conventional matrix converter (CMC) or a DC side connection of a current DC link rectifier and a voltage DC link inverter comprising no energy storage components in the DC link could be employed. The combination of DC converters does show a lower number of turn-off power semiconductors and, therefore, has been denoted as Sparse Matrix Converter (SMC) or Very Sparse Matrix Converter (VSMC). A limitation of the phase displacement of the current and voltage fundamentals at the input and at the output to ±ð/6 does allow a further reduction of the system complexity, the respective circuit topology has been introduced as Ultra Sparse Matrix Converter (USMC) in the literature. In this paper a novel concept for the analytical calculation of the current stresses on the power semiconductors of the Sparse Matrix Converter Topologies (SMC, VSMC, and USMC) is proposed. Furthermore, the switching losses of the output stage which shows identical structure for the SMC, VSMC and USMC are calculated analytically based on an experimentally determined dependency of the switching loss energy on the switching voltage and current. As a comparison to a digital simulation shows, the analytical results do show a very good accuracy in a wide modulation range and for widely varying load current phase angle and widely varying ratio of output and mains frequency and therefore do provide an excellent basis for the dimensioning of the SMC, VSMC or USMC and/or for the determination of the rated output power and efficiency which could be achieved by employing given power transistors and diodes.

   Commutation Strategies for PWM Rectifier of Converter without DC Link Components for Induction Motor Drive
By K. Iimori; K. Shinohara; Y. Matsusita; M. Muroya [View]
[Download]

Abstract: This paper presents two different commutation methods for PWM rectifier section of the converter without dc link components. The first method is so commutated that each switching device of the bi-directional switches are controlled independently with sensing dc current direction. The second method is so commutated that the commutation of the PWM rectifier is forced to occur only in the zero voltage vector(V0, V7) period of the inverter. The main advantage of these methods is that the snubber circuits of the PWM rectifier section can be removed. Adopting the second method, not only the loss of the snubber circuit but also the switching loss in PWM rectifier section can be reduced. These methods have been verified by computer simulation on an induction motor drive.

   Comparison of Losses in Voltage Source Inverters and Direct AC-AC Converters
By P. Wheeler; L. Empringham; J. Clare; M. Bland [View]
[Download]

Abstract: This paper is concerned with modelling, predicting and experimental measurement of semiconductor losses in direct AC-AC (matrix) converters. All of the commutation scenarios possible in a matrix converter topology using two or four step commutation are identified and studied. A complete loss model for both switching loss and conduction loss for the purposes of comparison with other converter technologies is developed and the matrix converter compared with an equivalent back-to-back inverter employing the same devices.

   Design of Pulse Patterns for Matrix Converters
By J. Mahlein; H. Schierling; M. Bruckmann; O. Simon [View]
[Download]

Abstract: To enhance the behavior of voltage dc-link converters several methods to optimize the pulse patterns are known, for example the Flat Top Modulation ([ 1 ], Fig. 15) to minimize switching losses or the symmetrical use of zero switching states ([ 1 ], Fig. 12) to overcome control problems with minimum turn-on or turn-off times. It is quite obvious to ask if these methods can be adopted to the control of matrix converters. This article will give appropriate solutions and further optimizations suitable for matrix converters only. The influence of the pulse patterns on switching losses and current ripple is discussed also.

   Implementation of a Matrix Converter Space Vector Control in Programmable Logic
By J. Igney; J. Mahlein; O. Simon; J. Weigold [View]
[Download]

Abstract: An implementation of a matrix converter vector modulation and commutation control within a PLD is presented. The rectifying and inverting vector modulation (RIVM) is used. This was possible after the optimization and the reduction of the necessary calculation steps. This solution uses the desired output voltage space vector, calculates pulse patterns and executes them. This new solution works without a micro-controller or a digital signal processor (DSP) in closed loop configuration. It allows great reduction in costs of matrix converter control’s implementation and increases the performance.

   Minimization of Matrix Converter Commutation Times
By P. Wheeler; L. Empringham; J. Clare [View]
[Download]

Abstract: This paper examines a new technique that minimizes the commutation time in Matrix Converters. The technique avoids unnecessary waveform distortion, which is particularly important at low frequencies. This minimization gives the Matrix Converter superior waveform quality in comparison to a conventional inverter. This paper describes the design, construction and testing of a 10kVA threephase to three-phase Matrix Converter induction motor drive. The converter has been built using discrete 65Amp MOS Controlled Thyristors (MCTs), although the techniques applied could be used with most power semiconductor switching device. The commutation time minimization has been demonstrated using the MCT Matrix Converter with a 12HP induction motor load; practical results from these tests are presented.

   Rectifier based robust control of bidirectional switches in AC-AC matrix converters
By W. Hofmann; M. Ziegler [View]
[Download]

Abstract: Due to the non-existence of natural freewheeling paths the commutation in matrix converters is more complicated than in voltage source converters. The proposed interlock-free control for the bidirectional switches where both current directions can be controlled independently is based on the knowledge of the signs of the line to line input voltages and the load current directions. The suggested policy is compared with multi-step commutation methods. Finally some measurement results of the forward recovery effect of new reverse and non reverse blocking IGBTs are presented.

EPE-PEMC 2002: Special Session: Matrix Converters
You are here: EPE Documents > 04 - EPE-PEMC Conference Proceedings > EPE-PEMC 2002 - Conference > EPE-PEMC 2002: Special Session: Matrix Converters
	[return to parent folder]

	A Direct Power Electronic Conversion Topology for Multi-Drive Applications By F. Blaabjerg; C. Klumpner	[View] [Download]
Abstract: The matrix converter is known as a single stage AC/AC converter, which is able to provide sinusoidal PWM output voltages and input currents, inherent bi-directional power flow and no reactive elements (bulky DC-capacitors and large line inductors). However, the increased number of active switches (18), as well as the need of a protection circuit involving twelve diodes with rated voltage/current ratings used only during faulty situation, makes this topology not so attractive. In this paper, a two-stage converter topology consisting of a three-phase to two-phase matrix converter directly linked to a Voltage Source Inverter (VSI) that provides Direct Power Electronic Conversion (DPEC) is analyzed. While the functionality is similar to a matrix converter, the analysis of the conduction losses shows a different shape of the efficiency curve, which in HVAC application may be an advantage. In addition, a topology is proposed where several VSIs are connected to the same DC-link and their switching patterns are synchronized, which is especially useful for multi-drive application. The functionality of the proposed two-stage multi-drive direct power electronic conversion topology is analyzed by simulations.

	Analytically Closed Calculation of the Conduction and Switching Losses of Three-Phase AC-AC Sparse Matrix Converters By J. W. Kolar; M. Baumann; F. Schafmeister	[View] [Download]
Abstract: For three-phase AC-AC power conversion a conventional matrix converter (CMC) or a DC side connection of a current DC link rectifier and a voltage DC link inverter comprising no energy storage components in the DC link could be employed. The combination of DC converters does show a lower number of turn-off power semiconductors and, therefore, has been denoted as Sparse Matrix Converter (SMC) or Very Sparse Matrix Converter (VSMC). A limitation of the phase displacement of the current and voltage fundamentals at the input and at the output to ±ð/6 does allow a further reduction of the system complexity, the respective circuit topology has been introduced as Ultra Sparse Matrix Converter (USMC) in the literature. In this paper a novel concept for the analytical calculation of the current stresses on the power semiconductors of the Sparse Matrix Converter Topologies (SMC, VSMC, and USMC) is proposed. Furthermore, the switching losses of the output stage which shows identical structure for the SMC, VSMC and USMC are calculated analytically based on an experimentally determined dependency of the switching loss energy on the switching voltage and current. As a comparison to a digital simulation shows, the analytical results do show a very good accuracy in a wide modulation range and for widely varying load current phase angle and widely varying ratio of output and mains frequency and therefore do provide an excellent basis for the dimensioning of the SMC, VSMC or USMC and/or for the determination of the rated output power and efficiency which could be achieved by employing given power transistors and diodes.

	Commutation Strategies for PWM Rectifier of Converter without DC Link Components for Induction Motor Drive By K. Iimori; K. Shinohara; Y. Matsusita; M. Muroya	[View] [Download]
Abstract: This paper presents two different commutation methods for PWM rectifier section of the converter without dc link components. The first method is so commutated that each switching device of the bi-directional switches are controlled independently with sensing dc current direction. The second method is so commutated that the commutation of the PWM rectifier is forced to occur only in the zero voltage vector(V0, V7) period of the inverter. The main advantage of these methods is that the snubber circuits of the PWM rectifier section can be removed. Adopting the second method, not only the loss of the snubber circuit but also the switching loss in PWM rectifier section can be reduced. These methods have been verified by computer simulation on an induction motor drive.

	Comparison of Losses in Voltage Source Inverters and Direct AC-AC Converters By P. Wheeler; L. Empringham; J. Clare; M. Bland	[View] [Download]
Abstract: This paper is concerned with modelling, predicting and experimental measurement of semiconductor losses in direct AC-AC (matrix) converters. All of the commutation scenarios possible in a matrix converter topology using two or four step commutation are identified and studied. A complete loss model for both switching loss and conduction loss for the purposes of comparison with other converter technologies is developed and the matrix converter compared with an equivalent back-to-back inverter employing the same devices.

	Design of Pulse Patterns for Matrix Converters By J. Mahlein; H. Schierling; M. Bruckmann; O. Simon	[View] [Download]
Abstract: To enhance the behavior of voltage dc-link converters several methods to optimize the pulse patterns are known, for example the Flat Top Modulation ([ 1 ], Fig. 15) to minimize switching losses or the symmetrical use of zero switching states ([ 1 ], Fig. 12) to overcome control problems with minimum turn-on or turn-off times. It is quite obvious to ask if these methods can be adopted to the control of matrix converters. This article will give appropriate solutions and further optimizations suitable for matrix converters only. The influence of the pulse patterns on switching losses and current ripple is discussed also.

	Implementation of a Matrix Converter Space Vector Control in Programmable Logic By J. Igney; J. Mahlein; O. Simon; J. Weigold	[View] [Download]
Abstract: An implementation of a matrix converter vector modulation and commutation control within a PLD is presented. The rectifying and inverting vector modulation (RIVM) is used. This was possible after the optimization and the reduction of the necessary calculation steps. This solution uses the desired output voltage space vector, calculates pulse patterns and executes them. This new solution works without a micro-controller or a digital signal processor (DSP) in closed loop configuration. It allows great reduction in costs of matrix converter control’s implementation and increases the performance.

	Minimization of Matrix Converter Commutation Times By P. Wheeler; L. Empringham; J. Clare	[View] [Download]
Abstract: This paper examines a new technique that minimizes the commutation time in Matrix Converters. The technique avoids unnecessary waveform distortion, which is particularly important at low frequencies. This minimization gives the Matrix Converter superior waveform quality in comparison to a conventional inverter. This paper describes the design, construction and testing of a 10kVA threephase to three-phase Matrix Converter induction motor drive. The converter has been built using discrete 65Amp MOS Controlled Thyristors (MCTs), although the techniques applied could be used with most power semiconductor switching device. The commutation time minimization has been demonstrated using the MCT Matrix Converter with a 12HP induction motor load; practical results from these tests are presented.

	Rectifier based robust control of bidirectional switches in AC-AC matrix converters By W. Hofmann; M. Ziegler	[View] [Download]
Abstract: Due to the non-existence of natural freewheeling paths the commutation in matrix converters is more complicated than in voltage source converters. The proposed interlock-free control for the bidirectional switches where both current directions can be controlled independently is based on the knowledge of the signs of the line to line input voltages and the load current directions. The suggested policy is compared with multi-step commutation methods. Finally some measurement results of the forward recovery effect of new reverse and non reverse blocking IGBTs are presented.