Low Complexity Model Predictive Control Method for Three-Level Converters with Fixed Switching Frequency Mode

The three-level neutral-point-clamped (NPC) converter is a promising topology for motor control. However, its large number of voltage vectors imposes a heavy computational burden on the conventional finite-control-set model predictive control (MPC). Furthermore, variable frequency is another barrier limiting the development of MPC, i.e., challenging to evaluate the power loss and EMI suppression. To address these concerns, this paper proposes a low complexity MPC method for three-level converters. While maintaining lower current distortion and fast dynamic response, the novel method has the desired characteristics of fixed switching frequency and no need for cumbersome weighting factor' tuning. First, the vector space of the three-level converter is divided into 6 small hexagonal sectors. Second, the sector judgment and execution time calculation are greatly simplified by using the similar two-level action law in each sector. More importantly, multiple-voltage-vector is adopted during each control period and the fixed switching frequency is realized through the proposed allocating action sequence. The validity of the proposed algorithms is evaluated by simulation and experiment results in contrast with other state-of-the-art MPC.