An Improved FCS-MPC Based on Novel Sector Optimization and Capacitor Charge Balance Algorithm for T-Type 3P-3L Converters

Conventional finite control set model predictive control (FCS-MPC) suffers from the high computational burden and cumbersome weight factor design, especially for three-phase three-level (3P-3L) converters. In this article, an improved FCS-MPC is proposed for T-type 3P-3L converters to reduce the computational burden, eliminate the weight factor, and improve system performance. Three improvements are as follows. First, a sector optimization method based on effective vector radiation range is proposed to determine the optimal vector for current tracking without vector enumeration, and a combined virtual vector is adopted to supersede the basic vector and avoid affecting neutral-point (NP) potential. Second, a capacitor charge balance algorithm is proposed to determine the optimal vector for NP potential balance without weight factor, which minimizes negative effects on current tracking. Third, a hybrid vector output mode is proposed to give full consideration to current tracking and NP potential balance performance. Finally, the advantages of the proposed MPC are verified in simulation and experiment, where code execution time and current THD are reduced by 16.1
μ
s and 1.27%, steady-state current error and capacitor voltage deviation are experimentally reduced by 0.8 A and 0.6 V, and transient dc voltage recovery time and capacitor voltage deviation are experimentally reduced by 2 ms and 0.4 V compared with conventional FCS-MPC.