Modeling' And Analyzing Of Charging Behavior Of Electric Vehicle Usage For City Distribution Based On Data Driven

Under the dual challenges of the environment pollution and energy shortage that the global automotive industry in the face of,the electric vehicle has become the focus of people's attention by the advantages of zero pollution and low noise.Besides,as an important service industry in China,the logistics industry has become an important field for the promotion of electric cars.But there are many new problems to be solved in the process of development of electric vehicles and ancillary charging facilities industry.For example,How to reasonably plan the construction of charging facilities in the center of logistics distribution?How to optimize the distribution of vehicle charging in order to avoid the congestion for Logistics enterprises?How to plan the optimal distribution path of the logistic electric vehicle with the charging constraint?The truths behind these problems are closely related to the charging behavior of the logistic electric vehicle for urban distribution.Therefore,it is of great significance and practical value to analyze the charging behavior of the logistics electric vehicles for urban distribution and establish the model of charging behavior.At present,the research of charging behavior is mainly from the perspective of the operation of the power grid system to study the effect of the charging behavior of electric vehicles to the power grid.However,the research achievement of the charging behavior of the electric vehicle users is relatively little from the point of view of the electric vehicle travel.In addition,most of the studies are based on simulation or small sample data.In this paper,we will explore the charging behavior of logistic electric vehicle for city distribution and establish the prediction model of the charging behavior based on the mass data of the actual operation of logistic electric vehicles,which has more practicability and universality.In order to explore the charging behavior roles of the electric logistic vehicle users in the city distribution,this paper adopts the analysis methods in data mining to research on the state of charge(SOC)and charging time based on the data of charge and discharge of 70 electric logistic vehicles operating during 2014,which have been handled by the data processing method of deleting and interpolation.Moreover,a charging behavior model is established.The results show that:most drivers tend to charge the electric logistic vehicle when the SOC ranges from 30%to 50%;the remaining electricity before charging follows the normal distribution with the parameters are ?=0.48??=0.22;the charging time are mainly ranges from 14:00 to 16:00.The experimental results show that this charging behavior model has a high accuracy.In order to accurately predict the charging length,this paper first deeply analyses the relationships' among the collected data.Analysis shows there is obvious positive correlation between SOCc and the charging length.Thus the paper first develops a linear model combining SOCc with the charging length and model results show that the error sequence does not conform to the normal distribution.Next,the concept of time series is introduced to adjust the regression model and a combination forecasting model based on regression and ARMA model is established.The model validation shows that the combination model has higher accuracy and precision than the single regression model,and the residual error of the model is in a horizontal band,which is in the normal distribution.The charging behavior model considering the SOC and the prediction model of charging length have high accuracy and good practicability,which provide scientific making decision support about charging dispatch with the enterprises as well as the charging schedule with the users.