Photovoltaic Power Prediction Method Based on Fluctuating Weather Identification

Abstract

At present, photovoltaic power prediction has problems of low prediction accuracy and weak correlation between meteorological factors and power fluctuation process, so a photovoltaic power prediction method based on fluctuating weather identification is proposed in the paper. First, the weather process is initially divided into five types based on PV power fluctuation characteristics, and then the clarity index Kt is introduced to perform weather type cross-segmentation to decompose the full time PV power into smooth process and fluctuation process. Finally, a PV power prediction model is established. The model fully considers the specificity of the deep learning algorithm to classify the fluctuating process and the smooth process, and the simulation results show that the proposed method can effectively improve the prediction accuracy.