Gradient Descent Feedback Correction for Robust Deadbeat Predictive Control in Permanent Magnet Synchronous Motor System

摘要

In order to optimize the control performance of the permanent magnet synchronous motor (PMSM) system, the deadbeat predictive control (DPC) is introduced. But due to the lack of rolling optimization and feedback correction, the control performance suffers from the mismatch of model parameters. To deal with the problem, the gradient descent feedback correction (GDFC) method is proposed, which automatically corrects the resistance and inductance in the operation process without the need of additional observer or identification model, enhancing the robustness of the DPC greatly. The proposed method also has the advantages of simple operation and low time-consuming. Then the effectiveness and applicability of the proposed method are verified by the simulation results which show the convergence values of resistance and inductance under different working conditions.