This paper proposes an improved finite control set model predictive current control (FCS-MPCC) for variable-speed wind-energy conversion system (WECS) based on the permanent magnet synchronous generator (PMSG) connected to the grid via an LCL filter. For the generator side, the control is achieved through an adaptive FCS-MPCC based on an unscented Kalman filter (UKF) and a reinforcement learning (RL) MPPT based on Q-learning algorithm for PMSG speed regulation and maximum power extraction (MPE). A UKF is used to perform a real-time parameter estimation for a PMSG prediction model, while a Q-learning algorithm dynamically optimizes the control policy to ensure MPE under wind speed varying conditions. For a grid-side, an adaptive model-free FCS-MPCC is proposed, which combines a dc-link voltage outer loop controller for stabilizing the dc-bus voltage and a model-free FCS-MPCC inner loop current controller based on a piecewise ARX model and a standard RLS algorithm. Findings demonstrate that the RL-MPPT outperforms other popular MPPT controllers, such as Adaptive P&O (AP&O) and TSR over a wide range of operating conditions. Further, the improved FCS-MPCC achieves the lowest computation time and THD, while enhancing reliability and operational stability for both the MSC and GSC compared to the traditional FCS-MPC.
Hamid et al. (Fri,) studied this question.