A FCM-XGBoost-GRU Model for Short-Term Photovoltaic Power Forecasting Based on Weather Classification

Abstract

Aiming at the problem of low photovoltaic prediction accuracy, a short-term photovoltaic power prediction method based on fuzzy C-Means(FCM)- extreme gradient boosting (XGBoost)- gate recurrent unit (GRU) based on weather classification is proposed. First select the key meteorological factors as the clustering features, then use the FCM clustering method for cluster analysis, divide the historical data into sunny, cloudy, rainy and extreme weather, and then construct XGBoost-GRU combined forecasts for the four weather types The model predicts photovoltaic output power. Finally, the model proposed in this paper is compared with the prediction results of traditional XGBoost and GRU models. The results show that the proposed FCM-XGBoost-GRU short-term photovoltaic power prediction method can significantly reduce the error of photovoltaic prediction and improve the accuracy of short-term photovoltaic prediction. It is effective and scientific in practical application scenarios.