Model-Free Adaptive Learning Control Scheme for Wind Turbines with Doubly Fed Induction Generators

Abstract

The classical control mechanisms of the wind turbines are generally based on precise modeling approaches to ensure robust and effective interplay between the wind turbines and the main power grids in both autonomous and grid-connected modes. The paper presents an innovative intelligent control system for the doubly fed induction generator wind turbines. The proposed system uses model-free control polices. The online controller is based on a policy iteration reinforcement learning paradigm along with an adaptive actor-critic technique. It is shown to be robust against the turbine's high nonlinearities and stochastic variations in the input-output conditions. These are associated with single and double rotor doubly fed large scale induction generators driven by wind turbines in the range of 5-7 MW. The performance of the controller is validated against challenging scenarios of coexisting undesired situations like severe wind changes with load excursions and abrupt shifts in the loads.