Photovoltaic power generation prediction method based on dynamic time warping and transformer

Abstract To address the issue of low prediction accuracy caused by the strong randomness of photovoltaic (PV) power generation, this paper proposes a PV power forecasting method based on Dynamic Time Warping (DTW) and the Transformer model. Firstly, preprocess the data and use pearson correlation coefficient to select several meteorological factors that have a significant impact on PV power generation. Secondly, the training data is divided into three weather types: sunny, cloudy, and rainy using the DTW algorithm. Finally, based on the Transformer model, a PV power generation prediction model was established to predict the PV power generation under three different weather conditions. The verification results of the examples show that the method proposed in this paper achieves higher predictive accuracy compared to Support Vector Regression (SVR), Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRU) prediction methods.