多特征输入下电动汽车充电站充电负荷智能预测方法研究

打开文本图片集

中图分类号：U469.72 文献标识码：A 文章编号：1003-8639（2026）01-0009-03

Research on Intelligent Prediction Method of Charging Load forElectric Vehicle Charging Stations under Multi-feature Input

QiJunfei （China EnergyNational Electric Power Co.，Ltd.，Hami CoalPowerCompanyHuayuanPowerPlant，Hami 8390o0，China）

【Abstract】Electric vehicle charging load forecasting is crucialfor ensuring the stableoperation of the power grid. In viewof the problems thattheexisting methods ignoretheinfluence ofmultiplefactorsand have dificultytaking into account both long-termand short-term dependencies，this paper proposesa multi-feature input charging load prediction model combining Temporal Convolutional Networks （TCN）and Atention-enhanced Long Short-Term Memory （AttSTM） neural network.Theexperimental resultsof three public datasets show thatthe predictionaccuracyof this model is superior to that of the comparison model.The Mean Absolute Error（MAE）and Mean Squared Error （MSE）are reduced by （204 4.7% and 12.6% respectively compared with the optimal baseline model. This model can provide support for the stable operation of the power grid and the optimization of dispatching.

【Key words】 electric vehicle；charging load forecast;temporal convolutional network；attention long shorttermmemory

0 引言

随着电动汽车渗透率的提高，作为基础设施的充电站数量也在不断增加[]。（剩余3171字）