Adaptive Model Predictive Current Control for PMSM Drives Based on Bayesian Inference
Xiaoguang Zhang, Xiang Yu, Guofu Zhang
Abstract
Fully utilizing artificial intelligence (AI) algorithms to develop more flexible and robust intelligent control methods is a current hot spot in research. To solve the parameter robustness problem in the conventional finite control set model predictive current control (FCS-MPCC), an adaptive model predictive current control (BI-AMPCC) method for permanent magnet synchronous motor drives based on Bayesian inference (BI) is proposed in this article. The method uses Bayesian inference to achieve a beneficial combination of AI and model predictive control, which can achieve accurate control without knowing the motor parameters. In addition, as for the proposed simplified prediction model, the algorithm only needs to identify the inductance parameter online in real time to achieve adaptive control, which enhances the robustness of the FCS-MPCC method and reduces the complexity of the algorithm. Simulations and experiments demonstrate the effectiveness of the proposed method.