Model-free predictive control based on an extended state observer

Abstract To solve the problem that the traditional model predictive control of dual-three-phase permanent magnet synchronous motor is greatly affected by parameter changes and the model construction is inaccurate, this paper proposes to apply model-free predictive control to the dual-three-phase motor control system. By establishing the current hyperlocal model of the system, the design of a model-free controller for stator current based on the extended state observer is carried out, and the current required for the next moment is calculated. Then, the synthesized 12 virtual vectors were brought into the value function calculation, and the final vectors were determined by traversal optimization. The comparison of simulation results shows that compared with the traditional prediction method, the proposed strategy improves the parameter robustness of the system and effectively enhances the stability of the motor.