Dual-Layer Modulated Model Predictive Control Scheme for the Cascaded H-Bridge Converter

Abstract

To simultaneously achieve fast dynamics and fixed switching frequency, a dual-layer modulated model predictive control scheme is proposed for the cascaded H-bridge (CHB) converter in this article. For the output current control layer, firstly, each phase cluster is considered as a whole, and only six nonzero vectors are evaluated for output current prediction. Then, two optimal vectors are selected, and their dwell times are calculated based on the current prediction results. As a result, fast output current dynamics can be obtained with a greatly reduced number of evaluated control options. For the cluster voltage balancing control layer, instead of fundamental-frequency zero-sequence voltage (ZSV), the adaptive ZSV is injected according to the predicted cluster voltage differences. As a result, the active power flows are adjusted in each control period, and the cluster voltage balancing speed can be significantly improved. By the proposed scheme, complicated weight factor design is avoided, and fixed switching frequency is obtained through the integration of pulsewidth modulation. Experimental results verify that the proposed scheme can achieve fast output current and cluster voltage balancing dynamics with fixed switching frequency.