Model Predictive Speed Control of Double-Sided Linear Switched Reluctance Motors Using Extended State Observer

In this paper, a model predictive speed control method based on an extended state observer (MPC-ESO) is proposed for the double-sided linear switched reluctance motor (DLSRM) to compensate for the discrepancies between the predictive model and the actual system behavior, aiming to reduce speed error and ripple. First, the discrete state-space model of speed and speed increments is derived from the dynamic model, leading to the development of the predictive model. A model predictive controller is designed using the predictive model and a defined cost function. However, the predictive model derived from the dynamic model neglects the nonlinear disturbance. To address this issue, an ESO is introduced to estimate and compensate the disturbance for speed control system. Experimental results demonstrate that the proposed method effectively reduces the speed error, and particularly during startup, while significantly mitigating speed ripple. The effectiveness of the proposed method is verified through experiments.