Application and Prospect of Reinforcement Learning in Power Prediction on Source and Load Sides

Abstract

With the large-scale access of a high proportion of new energy sources, there is a high degree of uncertainty on both sides of the source and load, which brings huge challenges to the optimal dispatch of the power system. Therefore, accurate power prediction information of the source and load can provide important decision support for the dispatch of the new power system. In recent years, with the development of artificial intelligence technology, reinforcement learning (RL) has been gradually used in uncertain power supply power prediction, load prediction, etc., so as to support the stable and safe operation of power grid under the uncertainty of source and load. Therefore, reinforcement learning has a great application prospect in the new power system dominated by new energy. Based on this, this paper will conduct a research review on the application of reinforcement learning technology to wind power forecasting, photovoltaic power forecasting, load power forecasting, and source-load power forecasting under extreme weather. Besides, the reinforcement learning algorithm is used to predict the distributed power supply of a wind farm in Jilin Province, China and its region, which is used as the support of the calculation example. Finally, the development direction of reinforcement learning applied to source-load power prediction is prospected and analyzed.