The conventional model predictive current control (MPCC) method is highly sensitive to motor parameters, resulting in decreased control performance when the model parameters are mismatched with the motor parameters. To address parameter sensitivity and improve the robustness of the control system, this paper proposes a data-driven model predictive current control method (BLR-MPCC), which utilizes the machine learning technique Bayesian linear regression. This method constructs a current prediction model based on the voltage difference and the current difference, considers the parameters of the linear model as random variables, and solves the model parameters online using the Markov chain Monte Carlo (MCMC) numerical method. Simulation results validate the effectiveness of the proposed method.

PMSM,AC motor drive,model predic- tive current control (MPCC),Machine learning techniques