390 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 49, NO. 2, APRIL 2002 Continuously Motor-Synchronized Ride-Through Capability for Matrix-Converter Adjustable-Speed Drives Eduardo P. Wiechmann, Senior Member, IEEE, Rolando P. Burgos, Student Member, IEEE, and José Rodríguez, Senior Member, IEEE Abstract—The ride-through capability of adjustable-speed drives has become an important issue due to its direct impact on production and revenue losses. Moreover, different industrial surveys have shown that voltage sags are the main cause of converter tripping. Disturbances such as swells, distortion, and impulses were found far less common and did not cause any tripping nor production losses. Matrix-converter (MC) drives are also prone to voltage sags, furthermore the lack of the dc-link capacitor renders them somehow more vulnerable. This paper presents a ride-through strategy for MC adjustable-speed drives. The strategy is based on the reduced speed/load approach for con- ventional drives and is capable of enforcing constant volts/hertz operation regardless of the supply voltage conditions by first regulating the modulation index of the matrix converter, which counteracts the supply voltage drop, and second by reducing the speed reference if required. This reduction seeks to maintain the maximum torque capability of the drive and not to reduce the motor load as in conventional drives. Hence, the proposed strategy is suitable for both variable and constant torque loads. Moreover, the converter never loses synchronization with the motor, so it is capable of immediate acceleration to its former speed after the disturbance disappears. The proposed strategy was experimentally verified under typical industry disturbances using a TMS320C32 DSP based system. Particularly, three-phase and single-phase sags varying from 10% to 60% were tested. Results obtained showed the effectiveness of the proposed strategy for MC adjustable-speed drives. Index Terms—Adjustable-speed drive, decision-making space- vector modulation, matrix converter, ride-through capability. I. INTRODUCTION T HE matrix converter (MC) was first introduced in 1979 by Peter Wood [1]. Based on Pelly and Gyugyi’s work in the naturally commutated cycloconverter [2], he derived the switching matrix as the simplest conceivable topology that could perform polyphase-to-polyphase power conversion. Later, Alesina and Venturini employed this topology to develop the first ac-to-ac forced commutated converter [3]. Results obtained with this converter were so promising that they named it the generalized transformer. Particularly, it achieved Manuscript received April 24, 2001; revised August 16, 2001. Abstract published on the Internet January 9, 2002. This work was supported by FONDECYT (Chilean Fund for Science and Technology Development) under Project 198-0463 and its international addendum 798-0005. The authors are with the Department of Electrical Engineering, University of Concepción, Concepción, Chile (e-mail: ep.wiechmann@ieee.org). Publisher Item Identifier S 0278-0046(02)02883-6. sinusoidal input and output waveforms, bidirectional power flow, controllable input power factor, did not require any energy storage reactive elements, and had a compact size. These appealing characteristics remain as the main advantages of MCs and throughout the years have driven a continuous research effort to further improve its performance. Besides all of the above characteristics, MCs have failed to reach industrial production level, having been mainly restrained to laboratory work. This has primarily been due to three rea- sons, namely: the inexistence of four-quadrant semiconductors, which has required MCs to use 18 switches to realize the 9 input–output line connections; the reduced maximum voltage gain of 0.866 (without overmodulation); and the lack of energy storage reactive elements. The latter has been considered a dis- advantage rather than an advantage. In fact, it is usually said that the MC has no ride-through capability and presents neither buffering nor protection from the input phases, as these are di- rectly connected to the output terminals of the converter. To account for these drawbacks, several authors have devel- oped and presented different solutions. The increased number of switches has been dealt with by initiating the development of a power electronics building block, seeking to integrate the design and reduce the overall cost of the converter [4], [5]. The transient and overvoltage issue has long been solved by using clamp cir- cuits [6]. Further, the feasibility of achieving ride-through capa- bility with the MC has already been shown in [7]. Finally, the reduced voltage gain hindrance could be solved by designing the motors to reach maximum flux at the MC rated voltage, or equivalently by using motors rated at the MC output voltage. In all, the progress attained with the MC has significantly improved its performance, rendering it a definite choice for compact and integrated converter-motor regenerative drives. Perhaps the most important and known limitation of the ones previously discussed is the ride-through capability, as it is di- rectly associated with production and revenue losses [8]. This liability, however, also affects conventional drives, as the en- ergy stored in the dc-link capacitor suffices only for a couple of milliseconds. Consequently, this impairment has obliged to de- velop different ride-through strategies for conventional drives [9], [10]. These can be classified into strategies employing or not employing external energy sources, where the latter defines the scope of this paper. Within the strategies not employing external energy sources, one trend is to ride through the voltage disturbance by simply drawing more energy from the mains [11], [12]. Naturally, this 0278-0046/02$17.00 © 2002 IEEE