Indirect model predictive control strategies for a direct matrix converter with mitigation of input filter resonances

Abstract

The direct matrix converter has twenty-seven available switching states. This implies that the implementation of predictive control techniques in this converter requires high computational cost while an adequate selection of weighting factors in order to control both input and output sides. In addition, the technique presents a variable switching frequency which could produce resonances in the input filter damaging the performance of the system. In this paper, two indirect model predictive current control strategies are proposed in order to simplify the computational cost while avoiding the use of weighting factors. The first method consists in a predictive current control strategy with minimization of the reactive power minimization enhanced with an active damping implementation which allows mitigate resonances in the input side. The second proposal consist in an indirect model predictive current control with imposed sinusoidal source currents in the input side. Simulated results confirm the feasibility of the proposals demonstrating that with both alternatives it is possible to mitigate the resonances on the input filter.