PSO–LSTM–Markov Coupled Photovoltaic Power Prediction Based on Sunny, Cloudy and Rainy Weather

With the advancement of peak carbon and carbon neutrality targets, renewable energy power generation is gradually gaining popularity. As a representative of renewable energy applications, photovoltaic (PV) power generation, due to its uncertainty and volatility, can cause impacts on the power grid when it is connected to the grid on a large scale, leading to the phenomenon of light abandonment. To ensure the stable operation of the power system, this paper proposes a PSO–LSTM–Markov coupled model for PV power prediction. The model considers sunny, cloudy and rainy weather, and uses PSO–LSTM to establish a PV power prediction model with weighted Markov chain correction of the residuals; at the same time, the algorithm is optimised and corrected for the weather conditions that have a large prediction error for cloudy and rainy weather; and finally, the meteorological data provided by a region in Shenzhen and a historical power dataset are selected for experimental evaluation. By comparing the prediction results with the existing methods, it is verified that the method constructed in this paper has high prediction accuracy, and the MAPE is reduced by 116.8835 under cloudy weather conditions and 97 under cloudy and rainy weather conditions.