Non-cascaded Robust Model Predictive Direct Speed Control Based on the Reconstruction Model and Third-Order ESO for SPMSM

This paper presents a more robust model predictive direct speed control based on the third-order extended state observer (RMPDSC-TESO) method for surface-mounted permanent magnet synchronous motor (SPMSM) servo systems to attain a superior disturbance rejection performance and simplify parameter adjustment. First, by taking advantage of the fast current response to reconstruct the load disturbance, the disturbances in the motion and voltage equations are transformed into a unified equivalent current disturbance, thereby constructing a second-order speed ultra-local model. Subsequently, a TESO that only requires consideration of a single control parameter is adopted to estimate the speed control model's lumped disturbance, improving the system's robustness. Moreover, the deadbeat method is adopted to obtain the optimal voltage control solution. Then, by conducting derivations on the root locus and disturbance sensitive characteristics, the system's stability, anti-interference performance, and the influence of the input gain on the system's dynamic performance are demonstrated. Finally, experimental verification was carried out on a SPMSM test platform based on a DSP F28379D chip, and the results confirm the effectiveness of this control strategy in terms of dynamic performance and unmodeled disturbance suppression.