LAUSR

The model predictive control (MPC) has its superiorities in simple algorithm and exceptional dynamic performance. However, the mismatch of the model parameters will strongly affect the control performance on MPC and will also cause current error in prediction. Aiming to enhance the robustness of MPC while eliminating the predicted current error under parameters mismatch, in this article, a simple robust MPC method is put forward. In the proposed control strategy, accurate inductance information can be extracted simply by designing a discrete integral controller based on the predicted current error of the d-axis. Moreover, two methods are proposed to deduce the gain coefficient of the integration controller theoretically. Furthermore, in the new current prediction model, the flux-linkage parameter has been replaced by the extracted accurate inductance information. This indicates that the flux-linkage parameter has been removed from the model of current prediction. Since the resistance does not have much influence on the accuracy of the current prediction model, the inductance parameter becomes the only factor that predominantly affects the accuracy of the current prediction model. Finally, the experimental results prove the effectiveness of the proposed method.