Actor-critic continuous state reinforcement learning for wind-turbine control robust optimization

The control of Variable-Speed Wind-Turbines (VSWT) extracting electrical power from the wind kinetic energy are composed subsystems that need to be controlled jointly, namely blade pitch and generator torque controllers. Previous state art approaches decompose joint problem into independent subproblems, each with its own subgoal, carrying out separately design tuning a parameterized controller for subproblem. Such neglect interactions among which can introduce significant effects. This paper applies Actor-Critic Reinforcement Learning (ACRL) as whole, simultaneous parameter optimization both without neglecting their interactions, aiming globally optimal whole system. innovative architecture uses an augmented input space so parameters fine-tuned working condition. Validation results conducted on simulation experiments using state-of-the-art OpenFAST simulator show efficiency improvement relative best controllers used benchmarks, up 22% in average error performance after ACRL training.