Predictive Control of Modular Multilevel Converters Using Sphere Decoding Algorithm

Modular Multilevel Converters, by increasing the number of voltage levels the quality of output waveform is improved and the semiconductor switches tolerate low voltage values. However, design of switching strategy in these complicated circuits is challenging. In this paper, based on sphere decoding algorithm, a predictive controller with finite control set is proposed to regulate load current while minimizing both capacitor voltage variations and circulating currents. The suggested scheme decreases the computational burden of optimization stage which is considerable for long prediction horizons and modular converters with high number of voltage levels. The suggested scheme is simulated for a practical modular multilevel converter to demonstrate its performance compared to some rival methods.