A novel method based on the encoder-decoder structure of attention mechanism was proposed to improve the accuracy of fine-time-granularity charging load prediction.Firstly,a multi head self-attention mechanism was used to encode the time dependence of the station's historical charging load sequence;then,a non-autoregressive spatiotemporal decoder was used to calculate the correlation between site spatial dependence and historical charging load using multi head self-attention mechanism and multi head mutual attention mechanism,respectively;finally,comparative experiments were conducted on actual charging load data to verify whether the proposed method has higher accuracy.The results indicate that the encoder-decoder method based on attention mechanism can effectively capture the time dependence of load sequence and the spatial dependence of charging station.
维普
万方数据