Q-learning Method for Managing Wind Farm Uncertainties through Energy Storage System Control

In this paper, We are committed to improving the revenue of wind farm when wind farm and energy storage system (ESS) cooperate and interact with the main grid. The main challenge is the uncertainty of wind power generation (WPG). Based on WPG forecasting, the reinforcement learning (RL) method is used to overcome the impact of WPG uncertainty. The (RL) method used is classic Q-learning. Compared with other (RL) methods, Q-learning is widely applied and easy to converge. Especially, (RL) methods can realize online decision-making, and the decision-making will tend to be optimal. The simulation results show that the method in this paper can effectively reduce the uncertainty of WPG and increase the revenue of wind farms.