An optimization-based Space Vector Modulation method for the three-phase to three-phase Matrix Converter.

摘要

The Matrix Converter (MC) provides an all-silicon solution for AC-to-AC conversion, which directly connects the power supply to the load without the intermediate DC stage. The MC has many advantages over the conventional DC-Link converters. The characteristics of MC, such as compact size, long life span, inherent regenerative capability, are attractive for many industrial and military applications.; The Pulse Width Modulation (PWM) method is a key enabling technology for MC and lays the basis for the control of a MC. The objective of the PWM is to synthesize the desirable voltage (or current) waveforms from a set of available voltage (or current) waveforms. Based on the modulation methodology, PWM methods for matrix converters can be roughly categorized into two major groups, namely the Venturini Modulation (VM) methods and the Space Vector Modulation (SVM) methods. The SVM methods are particularly suitable for the vector control of MC-based AC drives. Therefore, our focus is on the SVM methods.; In the thesis, based on our SVM mathematical model, we introduced an innovative Optimization-Base Space Vector Modulation (OSVM) method for 3o-3o MC. The OSVM is a direct SVM method with many desirable properties. The OSVM maximizes the modulability of the reference vectors. The solution of OSVM can immediately tell the modulability of the reference vectors in a switching cycle. When performing the modulation computations, the OSVM doesn't require information of the input voltages and the output currents in advance. Therefore, theoretically, the OSVM can work with any voltage and current sources to produce the desirable voltage/current waveforms, as long as the reference vectors are modulable. The OSVM paves the way for the design of advanced control systems for MC.; An efficient algorithm for OSVM is essential for the purpose of real-time control. We designed an innovative Modified Quadratic Programming (MQP) method to solve the minimum-1-norm optimization challenge in OSVM. The MQP algorithm was successfully implemented and tested through computer simulations.; Simulation results were presented to show the performance of OSVM in various system operation conditions. Comparisons between the OSVM and the conventional SVM (CSVM) method show that the OSVM has a similar harmonic performance to the CSVM, while the OSVM can achieve higher voltage ratios than the CSVM. The OSVM method and the MQP algorithm were experimentally verified on the hardware matrix converter testbed.