A Robust Torque and Flux Prediction Model by a Modified Disturbance Rejection Method for Finite Set Model Predictive Control of Induction Motor

The extended state observer (ESO) has been found as an efficient solution for the model sensitivity in the model predictive control. Generally, ESO is used for disturbance rejection when the voltage reference is the output of the control law. The observed disturbance is subtracted from the control law. Thus, using disturbance rejection for finite set model predictive control (FS-MPC) is a challenge when it is utilized by a cost function consists of the errors of the torque and the flux. In this research, a modified disturbance rejection method is used in a feed-forward shape in order to improve the robustness of the finite set model predictive torque control. In this regard, the motor parameters have not appeared in the torque prediction algorithm. Only a rough approximation of the motor parameters is needed to design the ESO which means the proposed method has a very low dependency on the motor parameters. Besides, a thorough tuning guideline is proposed to tune the local parameters based o the convergence analysis of the ESO by using the self-stable region (SSR) approach and the Lyapunov function. The performance of the proposed MPC scheme is evaluated through the simulations and experimental tests. The results of the proposed method are compared with the classic MPC results.