基于改进LSTM神经网络的电动汽车充电负荷预测OACSTPCD

As the current research on charging load prediction of electric vehicles(EVs)lacks support from real data,the model considers scenarios that are too simple,and the influencing factors are not fully considered,the prediction results lack persuasiveness.On this basis,a charging load prediction method for electric vehicles that considers various factors affecting the charging load of electric vehicles is proposed.Considering of the impact factors such as weather,season,temperature,working days,and holidays on the charging load of electric vehicles,a three scale analytic hierarchy process is used to analyze the weights of each influencing factor.An LSTM neural network prediction model is established,the LSTM neural network model for prediction is obtained by training the real data,and the prediction model is modified based on the weight analysis results of impact factors to obtain the final improved LSTM neural network load prediction model.Real data from a residential area in Changzhou was used for experimental verification,and the results show that the proposed method can achieve accurate prediction of electric vehicle charging load,and the load prediction results can provide a foundation for the study of orderly charging strategies.