An Improved Deadbeat Predictive Current Control for Five-Phase PMSM

Deadbeat Predictive Current Control (DPCC) has many advantages, but its control effect is highly dependent on the mathematical model and parameter matching degree. As the speed increases, the error between the discrete model obtained by the traditional Euler discrete method and the continuous model will also increase. Besides, the one-step delay control caused by the system delay will also affect the DPCC control effect. In order to improve the control performance of DPCC and the robustness of parameter disturbance, this paper optimizes the traditional DPCC. Firstly, based on the Euler discrete method, the improved prediction model is given to reduce the error caused by discretization, and the influence of parameter disturbance on the improved prediction model is analyzed. Secondly, the system's delay time is obtained according to the current prediction value, reference value, and output voltage vector. And then, the next beat optimal voltage vector is compensated according to the delay time. Finally, a second-order sliding mode observer based on the variable-gain super-twisting algorithm(VA-STA) is proposed, which realizes the observation of parameter errors and compensates the DPCC output results in the form of feed-forward observations.