Single-Loop Model-Free Predictive Control for Maximum Power Point Tracking of Permanent Magnet Direct Drive Wind Power System

In the implementation of the maximum power point tracking (MPPT) strategy for wind energy conversion systems, model predictive control (MPC) is a common control method. However, MPC is sensitive to parameters, and the long-term operation of wind energy conversion systems in harsh environments will inevitably lead to parameter changes. The paper first designs an extended state observer (ESO) to obtain the lumped disturbance required by the ultralocal model, and then designs a single-loop model-free predictive controller that is different from the conventional dual-loop control. This algorithm does not require precise parameters of the generator and mainly relies on the system's inputs and outputs to obtain its ultralocal model. Simulation results show that, compared with the MPC using the traditional mathematical model, the model-free predictive control proposed demonstrates better dynamic response characteristics and steady-state performance. after the system parameters change, and exhibits strong parameter robustness.