Robust Model Predictive Torque Control Strategy for Permanent Magnet Synchronous Motors Based on Improved Nonlinear Flux Linkage Observer

Non-uniform internal temperature distribution in permanent magnet synchronous motors (PMSMs) induces parameters perturbation. Such variations can significantly degrade control performance and potentially lead to propulsion system failure. To enhance system robustness against flux linkage parameter mismatches, a robust model predictive torque control (MPTC) strategy based on an improved nonlinear flux linkage observer is proposed. An iterative learning control (ILC) mechanism is integrated into the nonlinear flux observer to improve its dynamic tracking capability, thereby mitigating torque control deviations caused by flux estimation errors. Furthermore, to increase the accuracy of the prediction model and reduce truncation errors arising from discretization, a composite discretization method is developed by combining explicit and implicit Euler formulations. Simulation and experimental results demonstrate that the proposed control strategy substantially improves system robustness and dynamic response under conditions of permanent magnet flux linkage parameter perturbation.