An EV load forecasting method for using GCN-LSTM

Traditional electric vehicle （EV） load forecasting methods often fail to fully utilize the spatial correlation among EV loads， resulting in low forecasting accuracy. To address this issue， a load forecasting method using graph convolution network-long short-term memory （GCN-LSTM） is proposed. Firstly， graph data is constructed to describe the distribution of charging stations in the region， and the spatial dependency information between the charging station under study and neighboring charging stations is extracted using a GCN. Secondly， the information extracted by the GCN at different time periods is formed into a time series and input into the LSTM to forecast the EV charging loads. Finally， the proposed algorithm is validated by using load data from charging stations in an urban area in China as an example. The results show that the proposed method can effectively improve forecasting accuracy.