Research On Load Forecasting Of Electric Vehicle

As people’s concerns about environmental and energy issues increase,the development of electric vehicles is receiving more and more attention and the penetration rate is getting higher and higher.The large-scale commissioning of electric vehicles,while mitigating energy and environmental issues,is accompanied by the challenge of random and disorderly charging behavior to the stability of the grid.In order to cope with this challenge,the load forecasting problem of electric vehicles has become a hot research direction.Through research on the charging behavior of electric vehicles,it is found that the charging behavior of electric vehicles has strong regularity on both time and space,and the electric vehicles of different uses and the charging points of different areas both have fixed behavioral characteristics.Therefore,this paper divides the charging area into three representative areas: residential area,public area and commercial area,respectively,to study the charging load forecast of electric vehicles separately.By analyzing the different charging behaviors at different charging points,appropriate prediction methods are proposed for different regions for prediction.The number of electric vehicles in the charging area in residential areas is relatively fixed and the use habits are also fixed,Therefore,the charging load is not volatile,and the charging behavior is mainly affected by time,date and weather.And the data collection difficulty is also low,that is,there is stable and continuous data for analysis,which is not much different from the daily resident load.Therefore,the paper designs a BP neural network algorithm based on similar day clustering to predict the charging load of this area.The public areas mainly include industrial areas,large parking lots in office area,and some public parking lots for all social vehicles.The types of charging cars in this type of parking lot are complex,and the charging behavior is diverse and disordered.In order to fully simulate the random area of the public area,this paper uses Monte Carlo simulation algorithm to simulate the random charging behavior in the public area by extracting random numbers.Most of the vehicles in the parking lot of commercial area will not stay for a long time.Most of cars at the charging points in this area take fast charging as a temporary short-term fast power supplement,so the arrival time is the key factor affecting the choice of charging point for electric vehicles.This paper studies the commercial charging area of the vehicle traffic and the travel habits of the vehicle,and regards the commercial charging point in a region as a competitive relationship.Introducing the Bureau of Public Road path resistance function(BPR)to properly represent road weights.and then using the Dijkstra algorithm to calculate the time of each commercial charging point in the arrival area,each charging point takes time as the bid,and the charging rate of the charging point with the least time is higher,and the load is more.Using the selection probability of the electric vehicle to charge the different road conditions in the same area,the predicted value is divided into each charging place in detail to improve the accuracy of the prediction.