A Data-Driven Short-Term Prediction Method for Photovoltaic Power Generation

This paper describes a data-driven photovoltaic power generation prediction strategy based on grey correlation analysis and long and short term memory (LSTM) model. The strategy utilizes grey correlation analysis to select main influencing factors of photovoltaic power generation, and utilizes LSTM regression network to predict intraday power sequences. The proposed method is only based on data, avoiding complex modeling processes. And it is validated with actual photovoltaic power data.