An Improved Model Predictive Control Method for Parallel Three-Level Converters System with Reduced AC Current Harmonics

Parallel operation of three-level converters is a popular solution to increase the system capacity. However, the parallel system faces the coupled problems of high-quality AC current output, neutral point (NP) voltage balance and circulating current suppression. To solve these coupled problems, an improved model predictive control (MPC) method for the parallel three-level converters (P3LCs) system is proposed. Firstly, the equivalent mathematical model of the P3LCs is developed. Then, a novel three-layer optimization strategy of the P3LCs is designed. The first layer selects the optimal five-level (5L) basic vector to track the AC current. Remarkably, to reduce the AC current harmonics, the P3LCs are controlled as a whole to provide five-level characteristic. The second layer selects proper redundant 5L voltage vector to balance NP voltage. The third layer selects the optimal redundant three-level (3L) switching states to suppress the circulating current. Besides, there are only three candidates for current tracking and four candidates for circulating current suppression, and they are directly obtained from the coordinates of the reference voltage to avoid look-up tables, which is computationally efficient. As a result, the proposed MPC method can significantly reduce the AC current harmonics under low switching frequency, while the NP voltage is balanced well and the circulating current is suppressed well. Finally, the simulation and experimental results verified the feasibility and correctness of the proposed MPC method.