考虑关键气象因素的时间卷积网络充电桩负荷预测

Charging Pile Load Prediction Using Temporal Convolutional Network with Consideration of Key Meteorological Factors

为解决电动汽车EV(electric vehicle)充电负荷易受气象因素影响产生“时移”的问题,提出一种考虑关键气象因素的时间卷积网络充电桩负荷预测方法。首先分析不同气象因素即气温、降雨、风速、降雪、相对湿度、日照总强度对EV充电负荷的影响,利用最大相关最小冗余准则提取关键气象因素。其次,为简化相似日选取步骤并保证分类准确性,引入孪生网络选取不同天气类别下充电负荷相似日。最后以关键气象因素和相似日历史负荷作为时间卷积网络的输入向量进行EV充电负荷预测。对比实验表明,考虑气象影响可有效提高负荷预测精度。

To solve the problem that the charging load of electric vehicle(EV)is easily affected by meteorological fac⁃tors and thus produces“time shift”,a charging pile load prediction method using temporal convolutional network(TCN)is proposed with the consideration of key meteorological factors.First,the influences of different meteorological factors including temperature,rainfall,wind speed,snowfall,relative humidity and total sunshine intensity on the EV charging load are analyzed,and the maximum correlation minimum redundancy criterion is used to extract the key mete⁃orological factors.Second,to simplify the selection steps of similar days and ensure the accuracy of classification,the siamese network is introduced to select days with similar charging loads under different weather categories.Finally,the key meteorological factors and historical load on similar days are used as input vectors of TCN to predict the EV charg⁃ing load.Comparative experiments show that the consideration of meteorological influences can effectively improve the load prediction accuracy.