Model-free predictive control based on ultra-local model for permanent magnet synchronous machines

To conquer the performance deterioration of the traditional model predictive control when parameter mismatch exists, an ultra-local predictive model of motor current is established, and a model-free predictive control method is proposed. The current variation during each control cycle was decomposed into natural response which is only affected by operating conditions, and forced response which is determined by selected voltage vector. The natural response of current was calculated alone to achieve fast tracking effect. A recursive least squares observer was designed to estimate the coefficient of voltage on current, and thus the forced response of current can be obtained. A favorable model-free predictive control method which requires not much computation was achieved. Simulation and experimental results affirm that the proposed method can achieve fine static and dynamic performance without knowing exact motor parameters, and is superior to traditional model predictive control with parameter mismatch. The proposed method can be applied in a wide range of motor control fields. © 2021, Harbin University of Science and Technology Publication. All right reserved.