Short-term photovoltaic power generation prediction model based on fuzzy clustering-Elman neural network

Abstract

Photovoltaic power generation forecasting is the basis of safe and stable operation in power grids. This paper proposes a short-term photovoltaic power generation prediction model based on Isolation Forest, Fuzzy C Means and Elman. Firstly, similar daily datas are selected according to the forecast date and classified according to the weather. Secondly, the abnormal parts in the Isolation Forest cleaning training samples are adopted. Thirdly, Fuzzy C Means clustering method is used to cluster the meteorological data of similar and the forecasting days. Finally, combined with the Elman neural network algorithm, a fuzzy clustering-Elman neural network prediction model with isolated forest data cleaning is formed. The experimental simulation is carried out according to the actual measured data of a certain city in Anhui Province. The prediction results are respectively compared with the traditional Elman and BP model. It is demonstrated that higher prediction accuracy can be obtained.