Reinforcement learning in wind energy - a review

Today's environmental concerns, particularly those related to global warming, have sparked a drive for the usage of renewable energy sources. One of the most significant sources of renewable energy is wind energy and wind energy conversion system are preferred to harvest wind energy. Due to the growing sophistication of wind energy conversion systems, new strategies based on advanced analytics are needed. In this study, reinforcement learning implemented in wind energy has been reviewed, the most popular approaches for various applications are identified, and it has been shown that reinforcement learning may be utilized in place of traditional approaches. According to the application, the techniques are examined and divided into four groups: optimal control, prediction and forecasting, optimization, and other techniques. Consequently, many literature has reported that, on an average, reinforcement learning has improved performance by 5% to 20% than existing methods. Moreover, around 85% of the 153 references included in this article were published after 2018. The purpose of the work is to provide a basis for future research on reinforcement learning applied in wind energy that may be crucial to energy sustainability. The report also addresses the discussion on the reinforcement learning current state, limitations, and future scope.