Three-vector model prediction of torque control with adaptive super twisting sliding mode observer with predictor

Abstract

In Model Predictive Torque Control (MPTC), To address the issues of high model complexity, significant torque/flux linkage ripple, and insufficient robustness in the speed loop existing in the MPTC for Interior Permanent Magnet Synchronous Motors (IPMSM), this study proposes an optimized control strategy incorporating an enhanced Adaptive Super-Twisting Sliding Mode Observer (ASSMO). Its innovations are reflected in: 1) Restructuring the reference flux linkage derivation process to simplify the model structure and enhance computational accuracy. 2)Replacing the conventional speed loop PI controller with a Nonlinear Active Disturbance Rejection Controller (NADRC), utilizing a novel smooth nonlinear function to strengthen disturbance rejection capability. 3) Designing an improved ASSMO, achieving synergistic optimization of disturbance observation accuracy and system stability through the introduction of a predictive factor via a variable exponential function, replacement with a hyperbolic tangent switching function, and incorporation of adaptive factors. Experimental verification demonstrates that, compared to traditional methods, this strategy reduces model complexity, decreases torque ripple, and significantly enhances system robustness and control accuracy.