Research On Planning Method Of Electric Vehicle Charging Infrastructure Based On Data Driven

With the gradual emergence of energy crisis and environmental issues around the world,electric vehicles have attracted more and more attention from the government and the public for their environmentally friendly and low-carbon features,which can bring many opportunities and challenges at the same time.Reasonable layout planning of charging network for EV is an important basis to promote the rapid development of EV.However,the charging stations that have been built currently have problems such as low utilization rate of charging facilities,uneven number of service vehicles between charging stations,and mismatch between construction scale of charging stations and charging demand.In response to the above problems,this article is based on data-driven research on the planning methods of basic charging facilities for electric vehicles.First,this article briefly introduces the data background around the data set needed by the research,then preprocesses the data to prepare for subsequent analysis.Data preprocessing specifically includes data normalization,deletion of duplicates,processing of missing and outliers.Using the region type as the classification standard,the probability descriptions of each characteristic variable in the related region are carried out by using the Gaussian Mixed Distribution Model and the Negative Exponential Distribution Model,in which the Expected Maximum algorithm is used to solve the parameters of the Gaussian Mixed Distribution Model.Secondly,a two-stage composite sampling method is proposed as the sampling method of the Gaussian mixture distribution model to sample and simulate the arrival of vehicles in each area;then a hierarchical charging decision model is proposed to describe the charging decision behavior of electric vehicle users,and the users are classified as emergency types.Users and random users;emergency users will definitely choose to charge,while random users use fuzzy inference algorithms to solve their charging probability,and adjust the fuzzy rule base according to the real charging data to make the fuzzy inference algorithm more implementable;finally aggregated the average daily charging demand curve for working days and rest days in each region.Finally,after obtaining the regional charging demand distribution,the queuing theory model is used to plan the charging station capacity allocation problem;by analyzing the charging behavior of electric vehicles in the charging station,it is found that the charging queuing behavior in the charging station conforms to the M/G/s model in the queuing theory,Where the time interval of electric vehicles arriving at the station conforms to the negative exponential distribution law,and the charging service time conforms to the normal distribution law.With the minimum total social cost in the charging station as the objective function,the cost is divided into three parts,including the specified cost in the charging station,variable cost in the charging station and the time cost of stay in the charging station for the users of electric vehicles,taking into account the investment operator of the charging station and the users of electric vehicles.The average travel time cost of Beijing residents is calculated by the production method.Considering the constraints such as queue waiting time of electric vehicle users and utilization rate of charging facilities,the optimal solution for the number of fast charging facilities in charging stations in each typical functional area is found.