Model Variation Observer Based Model Predictive Control for PMSM Drives

Due to the coupling of the discrete nature of voltage vector and the disturbances, it is challenging to achieve a high-performance finite control set model predictive control (FCS-MPC) algorithm under various disturbances. To overcome this issue and enhance the robustness of drive systems, this paper proposes a model variation observer (MVO) of predictive current control (PCC) algorithm for the permanent-magnet synchronous machine (PMSM). Based on the detailed analysis of prediction error, an improved stator current mathematical model considering all disturbances is modeled. The lumped current prediction error is divided into independent errors generated by the variation of each component in the improved model through an adaptive weighting technique. Then, these variations are estimated by a designed fast-Terminal integral sliding mode observer (FISMO) and compensated into the improved model. The proposed method overcomes the coupling of the voltage and disturbances, and avoids the shortcomings of the stator current disturbance observer. Finally, the comparative experiments demonstrate the high performance and better robustness of the proposed method against different disturbances.