An Improved Model Predictive Control method with Luenberger Observer for Measurement Noise Suppression and Reduction of Current Ripples

Model predictive control (MPC) methods have gained popularity in the control of permanent magnet synchronous motors (PMSM) due to its simplicity, easy implementation, and ability to accommodate multiple constraints. However, the current ripple in traditional MPC scheme could be noticeable due to limited vector choices. Moreover, the introduction of sampling noise causes additional current ripple into the feedback loop. Nevertheless, the utilization of an analog filter introduces extra time delay that negatively affects the overall effectiveness of the MPC approach. To address these issues, this paper proposes a predictive control method that combines generalized two-vector-based model predictive current control with Luenberger-observer. The approach employs the Luenberger observer as an alternative for analog filters to suppress current sampling noise, besides the optimal vector combination is obtained by combining deadbeat predictive current control with duty cycle calculation, avoiding the need for exhaustive calculations. The simulation demonstrates that the proposed method can effectively reduce current ripple comparing to traditional MPC approaches, and the total harmonic distortion (THD) is significantly improved.