Research On Electric Vehicle Short-Term Load Forecasting Of Smart Distribution Grid

According to the characteristics of electric vehicle charging of electric vehicles, from the scale of the network load and improve the temporal and spatial distribution of charging electric vehicles power prediction accuracy of two.Research on the large scale EVs’ load distribution in time and space is the foundation to realize the good interaction between the EVs and power grid. This paper studied from two aspects:on the one hand, considering the different types of EVs with different driving rules, different charging full, respectively discussed the factors such as user behavior, charging mode, battery characteristics and so on, also studied the factors’ influence on the EVs’ charging power characteristics, to ensure the input parameters’rationality of the power calculation model, analysed the EVs’ load distribution under the complex environment. On the other hand, charging price、permeability and the guide policy will affect the EV charging load, and different area has different charging load, these influences are all studied to calculating the load distribution in different regions.Based on the temporal and spatial distribution of the EVs’ load, this paper designed a multi-time scale load forecasting scheme, as recently predicted,before rolling prediction and forecasting. Recently predicted binding to EVs’ actual load calculated before the days, under the constraint of the prediction accuracy forecast power in a week of 24 hours a day. Before rolling prediction combined with the new data, through the electric vehicle power the latest actual value of recently predicted in the times of the result, to reduce the prediction error on the day before the power load forecast effect.To improve the prediction accuracy of electric vehicle power grid dispatch and operation optimization is the key of large-scale network of electric vehicle in electric vehicle load sequence, distribution in time and space is disordered and random, on the basis of the traditional electric power, the prediction consists of GA algorithm, GA-BP algorithm and the improved grey algorithm was proposed for the electric vehicle forecast by the improved multi time scale electric vehicle decision the prediction model, the improved GA algorithm, GA-BP algorithm and the improved grey algorithm of three kinds of algorithm in decision algorithm to determine the weight through the combination of AHP and expert experience, give full play to their respective advantages of three algorithms in most recently predicted, rolling forecast, real-time prediction. To a certain extent, the three algorithms are complement each other, improve the prediction accuracy, and improve the electric vehicle to interact with the grid level, reduce the influence of randomness and volatility of electric vehicle charging on the grid.