佐藤 之彦

Multi-level converters can essentially reduce the harmonics and Electro Magnetic Interference (EMI). The number of output levels should be higher drastically to realize extremely high quality output and low EMI. However, as the number of levels increases, the number of circuit components also increases and implementation will be difficult significantly. Thus, flexible building platform of the multi-level converters suitable for the higher number of the output levels should be developed. As a solution to this problem, a concept of Multi-Level converter Building Modules (MLBM) to realize the multi-level converter with a higher number of output levels systematically has been proposed in ECCE 2013. A prototype of the MLBM has been constructed as a 5-level flying capacitor converter with extendable configuration. However, in the previous version of the MLBMs (MLBM Type-I) without any heat sinks, there is a restriction to increase the rated power of the converter when plural MLBMs are combined. In this paper, the improved thermal management of the MLBM is investigated to realize higher power capacity as well as higher number of levels. The module configuration to improve the heat dissipation of devices by using aluminum-based printed circuit board (MLBM Type-II) is developed. The proposed configuration enables to obtain the higher power density in the multi-level converter with higher number of output levels.

In the field of motion control systems, voltage ripple, harmonics, and electromagnetic interferences (EMI) in an output of converters negatively affect the control performance. However, so far, there has been practically few works that explicitly consider the power converters in motion control systems. Currently, general 2-level converters are widely used. However, it is concerned that the 2-level converter becomes one of the obstacles to realize high-performance control because the output voltage has high harmonics and EMI. Those possibly degrade the position or force control performance. As the solution to this problem, linear amplifiers may be useful because they outputs continuous voltage without pulsed waveform. However, low efficiency of the linear amplifiers becomes critical issue as the power converter. In this paper, performance improvement of the motion control systems by applying multilevel converters is investigated as a solution to realize both high control performance and high efficiency at a higher level. From the results of some experiments, it is seen that current waveform is distorted in the case of using the 2-level converter due to the output voltage ripple of the converter. In addition, the operation of disturbance observer (DOB) causes current ripple with lower frequency. On the other hand, it is confirmed that 9-level converter reduces the ripples of the current effectively. Moreover, EMI which is major concern in high-performance control system such as extremely small motor applications for high-precision control, can be improved by using the multi-level converter even without any filters.

A method to determine minimum DC-link capacitance in PWM rectifier-inverter systems is proposed. First, a method to calculate the required DC-link capacitance at the steady state is investigated. Next, a method to calculate the DC-link voltage variation of PWM rectifier-inverter systems at a stepwise change in the output power is proposed. The results of the calculation are confirmed by experimental investigation. In addition, when the capacitance of the DC-link is reduced, the parameters that have a significant influence on the DC-link voltage variation are extracted. Then, based on the proposed calculation method for the DC link voltage variation and the constrained conditions determined by the parameters of the PWM rectifier, a calculation method for the minimum required DC-link capacitance at a stepwise change in output power is proposed. © 2013 The Institute of Electrical Engineers of Japan.

Multi-level converters can essentially reduce the harmonics and EMI. To realize extremely high quality output and low EMI, the number of output levels should be high. However, as the number of levels increases, the number of components also increases and implementation will be difficult significantly. Thus, flexible building platform of the multi-level converters suitable for the higher number of the output levels should be developed. In this paper, a concept of Multi-Level converter Building Modules (MLBM) is proposed to realize the multi-level converters with a higher number of output levels systematically. A prototype of the MLBM is constructed as a 5-level flying capacitor converter with extendable configuration. It is verified experimentally that single-phase and three-phase converters with 5-17 output levels are realized by various combinations of the six MLBMs. From these results, possibility of implementation of multi-level converters with further number of output levels with proposed MLBM is shown.

400V DC distribution systems have attracted much attention recently as the next generation power distribution systems for data centers. For the reliability of these networks, the development of high speed DC circuit breaker for overcurrent protection is essential. In existing research, semiconductor DC circuit breakers using silicon carbide static induction transistors (SiC-SITs) show to be a promising candidate as these high speed DC circuit breakers. In this case, the overvoltage generated across the drain-source terminals of the SiC-SIT device during the turn-off process is highly dependent on the applied turn-off gate-source voltage and the variation in characteristic of the devices. In this paper, the active voltage control method is proposed to suppress the overvoltage to a predetermined range even when variation in characteristic of the devices exists. The effectiveness of the proposed method is demonstrated by experiments.

In recent years, versatile distribution systems for next generation to realize highly efficient and reliable operation are investigated actively. In such distribution systems, high-speed overcurrent protection is one of the important issues. In addition, dynamic routing of power flow to improve system reliability and efficiency is desired. To meet these requirements, several types of semiconductor power switches have been investigated. In this context, intelligent switches using SiC-SIT for 400V DC distribution systems have been investigated. In this paper, based on the analysis of turn-off process, the required features for the intelligent switch are investigated theoretically. Then, gate voltage waveform effective on overvoltage suppression during the interruption process is presented. Furthermore, a feedback control method to realize the optimum turn-off operation regardless of the variation in the device characteristics is discussed.

In this paper, we introduce silicon carbide (SiC) buried gate static induction transistors (BGSlTs) which can be applied for circuit breakers in DC distribution systems. SiC-BGSlTs have excellent electrical properties such as low on-resistance, first switching speed and robustness. These properties are indispensable for the realization of semiconductor circuit breakers in DC distribution systems and cannot be achieved by any other semiconductor materials.

This paper describes a damping control method of power converters for suppression of resonance in DC power network. The resonance occurs when a resonant frequency of the DC distribution line coincides with the frequency of the harmonic or interharmonic components generated by the power converters. For detailed investigation, a combined system which consists of a pulse-width modulated (PWM) rectifier and a PWM inverter is treated as the simplest example. To suppress the resonance, a DC-side damping control method is proposed and its implementation and design method are discussed in detail. Then, the proposed damping control method is applied to the combined system of a PWM rectifier and a PWM inverter. Experimental results verify the validity and practicability of the proposed damping control method. (c) 2012 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

A calculation method of DC-link voltage variation of PWM rectifier-inverter system at a stepwise change in output power is proposed. The results of the calculation are confirmed by experiment. In addition, when the capacitance of DC-link is reduced, the parameters that have significant influence on the DC-link voltage variation are extracted. © 2012 The Institute of Electrical Engineers of Japan.

Among the various multi-level topologies, flying capacitor converters are promising from the view point of the high power density because the size of the capacitors, which is a major concern, is expected to be reduced. Although the voltages of the capacitors theoretically can stay in the balanced condition, they tend to be unbalanced under some practical conditions. So far, it has been clarified quantitatively that the capacitor voltage becomes unbalance due to inequality in switching delay of power devices. In this paper, some countermeasures to reduce the unbalance in the capacitor voltages are investigated to realize simple multi-level converters. The effectiveness of the initial charging resistors of the capacitors on the voltage balance is clarified analytically. Based on the investigations, a prototype is constructed to demonstrate attainable power density of the converters with balanced capacitor voltages. Some experimental investigations confirm that high efficiency and high power density can be realized.

In this paper, minimum DC-link capacitance in PWM rectifier-inverter systems is investigated. Reduction of DC-link capacitor is one of the most effective issues to realize high power-density PWM rectifier-inverter systems. For the extreme reduction of the DC-link capacitor, the minimum required capacitance should be clarified theoretically, and the improved control method should be developed. Firstly, the calculation method of DC-link voltage variation of the PWM rectifier-inverter systems at a stepwise change in the output power is proposed. Next, based on the proposed calculation method of the DC-link voltage variation and the constraint conditions determined by parameters of the PWM rectifier, the calculation method of the minimum required DC link capacitance at a stepwise change in the output power is proposed. Then, a feed-forward controller based on the output power of the inverter is proposed and the least necessary capacitance is clarified.

This paper demonstrates the effectiveness of a series active filter (SAF) for suppression of harmonic currents generated by multiple uncertain harmonic sources. The topology is based on a voltage source inverter (VSI). The SAF is controlled by employing digital loss-less resonator in order to realize a high gain to suppress the harmonic components in the source current. The PWM technique is used to generate the required gate drive signals to the VSI. Both of the extraction of harmonics and control are carried out using a Digital Signal Processor (DSP). The effectiveness of the proposed method is clarified analytically and verified by experiments.

Output power density of power converters has been increased. In power converters with lower harmonic distortion and lower electromagnetic interference (EMI) which are required in many applications, bulky harmonic and EMI filters are usually connected. The volume of the harmonic filters can be reduced by increasing the switching frequency. However, higher switching frequency will cause larger switching loss and severe EMI. As another approach, there is possibility of intrinsic elimination of harmonics and EMI by introducing multilevel converters. However, as the number of levels increases, the number of circuit components will exceed the practical limit of the implementation. Thus, from a practical viewpoint, the main circuit and related circuits should be integrated. However, it has not been clarified which multilevel topology is suitable for circuit integration. In addition, the effectiveness in reducing total device loss considering chip size has not been evaluated. In this paper, the advantage of a flying capacitor multilevel topology for circuit integration is clarified by theoretical and experimental investigations. Also, the total device loss and chip size of the flying capacitor converter are estimated. (C) 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

Multi-level converters can essentially reduce harmonics even when their switching frequency is low. Among the various topologies of the multi-level converters, flying capacitor converters are considered to be promising converters for realizing high power density. However, the main circuit of the flying capacitor converters has many capacitors. Therefore, in this study, the volume of the capacitors in the flying capacitor converters is determined by taking into consideration the allowable ripple voltage and temperature rise in the capacitor. © 2011 The Institute of Electrical Engineers of Japan.

Recently, power converters with lower power loss and lower electromagnetic interference (EMI) are required in many applications. To satisfy these requirements, diode-clamped multilevel inverters are investigated. In this case, the voltage balancing circuit must be connected to the DC capacitors because the DC voltages tend to unbalance when the number of the output level exceeds three. For this purpose, a voltage balancing circuit based on a RSCC is expected to be a practical solution. So far, the experimental investigations of the loss reduction in diode-clamped multilevel inverters with a voltage balancing circuit have hardly been carried out. In this paper, an experimental investigation employing prototype diode-clamped multilevel inverters of several different numbers of levels is presented. From these results, the effectiveness in the loss reduction by the increase of the number of level in diode-clamped multilevel inverters and the quality improvements of the output waveforms without deteriorating the efficiency are demonstrated.

Among the various multi-level topologies, flying capacitor converters will be promising multi-level converters from the view point of the high power density because the size of the capacitors is expected to be minimized in small converters with higher switching frequency. Although the voltages of the flying capacitors theoretically can stay in the balanced condition, they tend to unbalance under some particular conditions in the practical converters. So far, the mechanism of the voltage balance of the flying capacitors associated with the circuit operation has not been clarified fully. In this paper, the self-balancing function of the capacitor voltage is analyzed quantitatively. Based on the analysis, the operating condition under which the capacitor voltages are balanced is clarified. The analytical results are useful to realize the simple flying capacitor converters without additional forced balancing control.

This paper proposes a novel "Direct Space Vector PWM (Direct SVPWM)" strategy based on the direct AC/AC conversion approach for three-phase to three-phase matrix converters. This method allows the sinusoidal input and output waveforms as a major premise, and gives top priority to the improvement of output control performance in motor drive applications, such as providing maximum riding comfort in elevators, etc. Output voltage harmonics, switching losses, and common-mode voltages can be reduced across the entire voltage region. In addition, the switching count can be reduced even further by fully utilizing the output current detection value. Direct space vectors are first defined, and the method of selecting space vectors is described. Next, the PWM duty calculation technique is explained. Finally, the validity of the proposed method is demonstrated by comparison with the conventional virtual indirect method based on experimental and analytical results. (C) 2010 Wiley Periodicals, Inc. Electr Eng Jpn, 172(3): 52-63, 2010; Published online in Wiley Inter Science (www.interscience.wiley.com). DOI 10.1002/eej.20953

Recently, higher voltage direct current (HVDC) distribution systems are paid attention as highly efficient and reliable distribution systems. The over-current protection is one of the important issues in the realization of practical HVDC distribution systems. In this paper, as a solution to this problem, a semiconductor DC circuit breaker is investigated. A novel control method of the gate voltage waveform to reduce the overvoltage and resonance during the interruption process is proposed. The effectiveness of the proposed method is confirmed by actual operation tests employing a prototype of the DC distribution system.

Output power density of power converters has been increasing. At the same time, the power converters with lower harmonics and lower electromagnetic interference are required in many applications. To satisfy these requirements, the volume of filters and heat sinks should be reduced. As a solution to realize higher output power density and lower harmonics, diode-clamped multilevel inverters are investigated. In this case, the voltage balancing circuit must be connected to the DC capacitors because the DC voltages tend to unbalance when the number of the output level exceeds three. Therefore, the total volume of the converters including the voltage balancing circuit should be investigated. In this paper, resonant switched capacitor converter is treated as a promising solution to realize the voltage balancing circuit with smaller volume. The operating characteristics and estimated volume are clarified analytically. Finally, the applicability of the voltage balancing circuit is confirmed by the experimental results.

There is a possibility of the realization of high output power density converters by introducing flying capacitor multilevel topologies. However, as the number of the levels increases, the number of circuit components will exceed the practical limit of the implementation. Thus, from the practical viewpoint, the main circuit and related gate drive circuits including floating gate power supplies should be integrated. In this letter, charge pump circuits, already proposed for conventional 2-level converters, are extended to the floating gate power supplies for the flying capacitor multilevel converters. © 2010 The Institute of Electrical Engineers of Japan.

In this paper, we study a control method based on the virtual indirect circuit method for a single phase to three phase matrix converter (MC). The MC has a capacitor at the input side to compensate for die single phase input power fluctuation. We propose a method to determine the amplitude of the compensation capacitor voltage on the basis of the load power. The method is validated using a 1350 W class single phase to three phase matrix converter system. In addition, restrictions on the modulation ratio of the input current and the compensation capacitor current are revealed, showing the limit of the output voltage according to the load power. © 2010 The Institute of Electrical Engineers of Japan.

There is the possibility of the intrinsic elimination of harmonics and electromagnetic interference (EMI) by introducing multilevel converters with a large number of levels. As the number of levels increases, the number of the main switching devices on their higher side increases, and their floating gate power supplies become larger scale circuits. Because it is necessary to integrate the floating gate power supplies from the practical viewpoint, it is one of the important issues to realize the floating gate power supplies, which consist of the components suitable for the integration.In this paper, bootstrap circuits, charge pump circuits, and self power supply circuits, already proposed for the conventional 2-level converters, are extended to the floating gate power supplies for the multilevel converters: diode-clamped converters, flying capacitor converters, and series-connected H-bridge converters. The theoretical and experimental investigations of the floating gate power supplies for the multilevel converters considering the main circuit topology and the main circuit control are shown. From these results, the applicability and the required modification of the floating gate power supply methods are clarified.

This paper deals with a control method and a designing method of the capacitance of the compensation capacitor for a single phase to three phase matrix converters (MCs) with the variable speed drive capability of induction motors (IMs). Assuming the application of railway tractions, automobiles and elevators, motors drive inertial load. In such cases, both the load power and the load voltage increase in proportion to the rotor speed. The amplitude of the compensation capacitor voltage is controlled to absorb the single phase power fluctuation, along With the load power. We propose a method to decide the input side parameters such as the capacitance of the compensation capacitor, considering the input voltage and the power of the IM. We experimentally verify that a single phase to three phase MC can drive IMs with the proposed method.

In this paper, a constrained optimization based approach is proposed for the control of inverter-fed induction motors. This approach has many distinguished features: the measurements are motor velocity and currents, the tracking of torque command is optimized, the switching signal of inverter is determined directly from the feedback signals. This approach may be viewed as an improved SVM (space vector modulation) method, but instead of the tracking of a voltage command we track the current and flux commands directly which is equivalent to the tracking of torque and flux commands. This method greatly simplifies the control structure and computation complexity, and the optimality is guaranteed even in the overmodulation region.

Recently, importance of the effective use of energy and demand for the introduction of renewable energy have risen more than before. Therefore, it is necessary to construct power network systems that consist of semiconductor power converters with various functions such as generation of the electric power, storing and discharge of the electric power. This paper describes highspeed coordination control methods for the multiple power converter systems. First, as a simplest example, a combined system of a PWM rectifier and a PWM inverter is investigated. Several coordination control methods are applied to minimize the required capacitance of the DC smoothing capacitor. The effectiveness of the coordination control methods are evaluated from the viewpoint of the over-voltage reduction and transient response of the DC link voltage. As a next step, to verify the effectiveness of the investigated coordination control methods in general multiple power converter systems, they are applied to a combined system that consists of 1 PWM rectifier and 2 PWM inverters. Furthermore, a general approach of the coordination control of multiple converter systems is discussed.

This paper proposes a designing method of a restart procedure speed-sensorless vector controlled induction generators for wind power generation systems. A designing method for a restart procedure is mainly discussed in this paper. Transfer function over the input and the output is linearized by using feedback for simple designing control system. This method is experimentally verified by a test set.

This paper describes an investigation of gate drive circuits of multilevel converters for circuit integration to realize high output power density (OPD). As key issues to implement the integration, methods for the isolation of the gate signal and the isolated power supplies suitable for the integrated diode-clamped multilevel converters are discussed. The usefulness of these methods has been confirmed by simulation and experimental results. Effectiveness of the multilevel converters on the reduction of the losses in the power devices and the total volume is evaluated for both silicon (Si) and gallium nitride (GaN) lateral devices.

This Paper proposes a novel conversion scheme of switching patterns for three-phase to three-phase matrix converters. The conventional virtual indirect conversion method is equivalent to PWM technique via an ordinary PWM rectifier/inverter, offering the advantage that no complicated specialized control is required. On the other hand, 6 of 27 switching patterns cannot be used in this method, because the inputs and outputs are always connected by way of the virtual DC link that is composed of 2 lines. This paper therefore defines the direct space vectors that can express the all 27 switching patterns and utilizes the geometric relationship of direct space vectors so that all switching patterns can be converted from other arbitrary vectors. This conversion scheme also allows the duty factor conversion with simple calculation by utilizing the duties of the virtual indirect conversion approach. In particular, the above-mentioned 6 switching patterns that have been restricted can be fully utilized for reducing output voltage harmonics, switching losses and common-mode voltages. The validity of the proposed conversion method is proven from the experimental results compared with a conventional virtual indirect method. © 2008 The Institute of Electrical Engineers of Japan.

This paper proposes a novel "Direct Space Vector PWM (Direct SVPWM)" strategy based on the direct AC/AC conversion approach for three-phase to three-phase matrix converters. This method allows the sine input and sine output waveforms as a major premise, and gives top priority to the improvement of the output control performance in motor drive applications, for instance, provides maximum riding comfort for an elevator, etc. Output voltage harmonics, switching losses, and common-mode voltage can be reduced across the entire voltage region. In addition, the switching count can be reduced even further by fully utilizing the output current detection value. Direct space vectors are first defined, and the selection method of space vectors is described. Next, the PWM duty calculation technique is explained. Finally, the validity of the proposed method is proven from the comparison with the conventional virtual indirect method based on the experimental and analysis results. © 2008 The Institute of Electrical Engineers of Japan.

For the speed sensor-less control in the ultra lower speed range, the mechanical simulator method is applied to cope with the problems due to the extremely lower induced voltage. In the method, however, the mechanical parameters' error, such as the error of the inertia, affects the starting performance. To reveal the conditions for the stable starting, we carry out the phase plane analysis under the condition that the actual inertia J changes from the set inertia J(s). Consequently, we find out J(s) >= J is the condition for the stable acceleration. The derived condition is useful as a design index of the mechanical simulator. Additionally, to cope! with the large variation of the inertia, we carry out the phase plane analysis of the IM simulator method which compensates the estimation error of the rotor frequency. The analysis reveals that the method enables the stable acceleration while the estimation error of speed is remained in the case of inertia variation. Furthermore, we propose the modified IM simulator method which eliminates the deviation. By the phase plane analysis and experimental tests, we confirm the effectiveness of modified method.

This paper proposes 3 types of new direct space vector modulation methods for three-phase to three-phase matrix converters. The space vector modulation method allows a direct understanding of switching patterns and their characteristics from the viewpoint of analysis and control. Output voltage harmonics and switching losses can be reduced across the entire voltage region by making use of the direct AC/AC conversion approach. The switching count can be reduced even further by fully utilizing the output current detection value. The conventional virtual indirect conversion approach is first described and then the selection method of space vectors is explained along with the duty calculation technique in direct space vector modulation. Finally, the validity of the proposed method is proven from simulation and analysis results.

This paper proposed the simple method to develop observer for detects the speed of Induction motor.. Direct field-oriented induction motor drive system need rotor flux observer and rotor angular speed identifier. ANFIS is used for identifying parameter dynamics and system variable estimation, linear either non-linear. ANFIS with back propagation learning algorithm has applied to estimate flux rotor and identify rotor angular speed of three-phase induction motor. The simulation result is found good for the speed up to 200 rpm, and no good result for the speed less then 200 rpm.

In this paper, a control method to improve ac input current waveforms of current type PWM rectifiers which has been proposed by the present authors is investigated. In this control strategy, state feedback of the ac side LC filter is introduced to eliminate waveform distortion and transient oscillation of input current of the rectifier. Theoretical investigation is made on the basis of equivalent circuits of the ac side of the rectifier which are derived in this paper. The function of the proposed feedback control is analyzed from the view point of the power flow between the LC filter and the bridge circuit of the rectifier.<br>From the results of the analysis, it is shown that the proposed control strategy provides damping effect in the form of equivalent resistance connected in parallel with the filter capacitor. The waveform distortion and transient oscillation of the input current are suppressed by this damping effect. On the basis of these results, a brief comparison between the improvement mechanism of the input current waveform by the proposed feedback control and that by active power filters is made. The effects of the feedback gain, control delay, and the circuit parameters of the LC filter on the elimination of harmonic current and on the stability of operation are investigated in detail. Analytical results shown in this paper are confirmed by computer simulation and experimental investigation.

PWM rectifiers which can realize unity power factor and sinusoidal input current have been widely investigated to overcome problems of reactive power and harmonic current in electric power systems. However, in such PWM rectifiers, an LC filter which is necessary for absorption of carrier harmonics and for reduction of commutation spike voltage may cause waveform distortion and transient oscillation of the ac input current. As a solution to these problems, a new control method introducing state feedback control of the LC filter was proposed in our previous paper. To achieve stable operation in this method, the control frequency or carrier frequency of the PWM rectifier must be higher than 5 times the resonant frequency of the LC filter. So, a method to reduce the required carrier frequency is necessary to implement high power PWM rectifiers.<br>In this paper, a parallel connection of two bridge circuits is introduced to reduce the required carrier frequency. A new control method for the parallel connected PWM rectifiers is proposed. The main circuit configuration and inter-bridge current between the two component bridge circuits are investigated. To achieve the reduction of the required carrier frequency without affecting the controllability, a new method of generating PWM pattern is developed. The effectiveness of the proposed control method is confirmed by some experimental results employing a test system.

In current type PWM rectifiers which can realize unity power factor and almost sinusoidal ac input current, an LC filter is required on the ac side for the absorption of the carrier harmonics and voltage spikes due to the PWM operation. However, the LC filter may cause waveform distortion and transient oscillation of the ac input current.<br>In this paper, the ac input current waveform is analyzed quantitatively. From the analysis, the amplification effect of the LC filter on the harmonic currents is clarified. An expression for the prediction of the transient oscillation is derived. It is shown by the expression that the transient oscillation can contain an unbalanced harmonic component whose frequency is equal to the resonant frequency of the LC filter regardless of the ratio of the resonant frequency to fundamental frequency of the power supply.<br>From the results of these analysis, a method of selecting the resonant frequency of the LC filter to avoid the steady state harmonic distortion of the supply current is proposed. It is also clarified that conventional methods of improving ac input current waveform such as optimization of the PWM pattern or introduction of the high carrier frequency operation do not have any effect on the reduction of the transient oscillation. In contrast to this, a control strategy introducing state feedback control of the LC filter proposed by the present authors is effective on the reduction of the transient oscillation.<br>The analytical results shown in this paper are confirmed by some results of computer simulation.

Resonant DC link inverters, which can realize high switching frequency power conversion without switching loss in principle, have been proposed and investigated. In such inverters, power loss is produced in the resonant circuit though power loss in the main circuit is reduced by the amount of the switching loss Therefore, an evaluation method of the total power loss should be established from the view point of efficiency.<br>In this paper, first we summarize the circuit operation and the control method of the resonant DC link inverter. Based on this, the expressions for various power losses produced both in the inverter main circuit and in the resonant circuit are derived. In this derivation, since the turn-on signal for an inverter arm is assumed to be determined only by comparison of the actual output current and its reference, a concept of probability is introduced for the turn-on signal of each inverter arm.<br>As an investigation of the resonant reactor, which affects the total power loss significantly, two types of reactor are constructed and the frequency characteristics of their resistances are clarified experimentally.<br>To confirm the validity of the expressions for losses derived in this paper, a comparison of the calculated and measured total losses is made for an experimental system.

In recent years, PWM rectifiers, which realize high power factor and almost sinusoidal AC input current, have been proposed and some of them have already been put into practice.<br>In many cases the PWM rectifiers are of current type, that is, they have a smoothing reactor on the DC side. In such rectifiers, an LC filter has to be inserted on the AC side to reduce current harmonics due to the PWM operation. However, the LC filter may cause waveform distortion and transient oscillation of the AC side current.<br>As a solution to this problem, we propose a new control method introducing AC side current control by means of state feedback of the LC filter. In this case, both the DC output current and the AC input current controls should be achieved by the PWM operation of the rectifier. Furthermore, to obtain sufficient controllability regardless of the operating condition, the DC output and AC input current controls should be done independently. To meet these requirements, existing control methods for PWM rectifiers are not applicable directly.<br>In this paper, a novel control circuit configuration and a method of generating PWM pulse pattern which are suitable for the proposed control method are described in detail. From the view point of stability, feedback coefficients and circuit parameters are investigated on the basis of root locus analysis.<br>Some experimental results are also shown to confirm the effectiveness of the proposed control method and the validity of the analytical results. Improvement of the AC side current waveform and stability of the transient response are demonstrated by these experimental results.

High performance power conversion requires high switching frequency power converters. Resonant DC link inverters, which are investigated in this paper, are suitable for this purpose because they cause no switching loss in principle. But in these inverters, loss of the resonant oscillation and occurrence of high peaks in the resonant capacitor voltage are serious problems. In this paper, we investigate a control method for resonant DC link inverters which can overcome these problems.<br>First of all, we summarize the basic control principle of an ideal resonant DC link inverter. Then we present an analyzing method for the practical resonant DC link inverter introducing two equivalent circuits in which the influence of the power losses in the resonant link and in the inverter load is considered. From these equivalent circuits, the problems of existing control methods are clarified. Based on the results of this analysis, we propose an optimum control strategy which can sustain the oscillation and keep the capacitor voltage at an allowable level. A simple compensation method for the influence of storage time of switching devices, which affects the implementation of the proposed control method, is described.<br>Some experimental results are included to confirm the validity of the analytical results and the effectiveness of the proposed control method.