Performance Improvement of Two-Vectors-Based Model Predictive Control of PWM Rectifier

Finite-control-set model predictive control (FCSMPC) has been proposed as a powerful control strategy for the power control of the pulse-width modulation rectifier. However, conventional FCSMPC applies only one voltage vector during one control period, which still presents relatively high power ripples. Introducing a zero vector in combination with a nonzero vector during one control period can improve the steady-state performance of the conventional FCSMPC, but the fixed vector combination is not an optimal solution in minimizing power errors. This paper proposes an improved two-vectors-based MPC, which relaxes the vector combination to two arbitrary voltage vectors. Compared to prior improved MPC with fixed vector combinations, the proposed method achieves better steady-state performance without affecting the dynamic response. Furthermore, the average switching frequency is reduced by up to 29.5% in average. The principles of vector selection and optimal vector duration are introduced in detail. The results obtained from two kinds of three-vectors-based predictive control methods are also presented for the aim of comparison. Both simulation and experimental results confirm the theoretical study and the effectiveness of the proposed method.