Analysis Of Driving Behavior Of Electric Vehicles Driven By Data

As one of the research focuses in the field of electric vehicles,driving behavior research has important theoretical and practical significance.Research shows that driving behavior will greatly affect the energy efficiency of electric vehicles.In order to estimate the potential of energy efficiency while considering the operating conditions,based on the actual collected driving data of electric vehicles,this theies uses data mining method to study the relationship between energy consumption and driving behavior of electric vehicles in depth,and to find the driving behavior parameters with the highest correlation with energy consumption in actual driving conditions;based on the analysis of the correlation between driving behavior and energy consumption,a new multi-classification model of energy consumption-related driving patterns is established by using K-means clustering algorithm and artificial neural network algorithm.Firstly,based on the actual driving data of electric vehicles,the driving data preprocessing model of electric vehicles is established.Based on this preprocessingt model,the following data preprocessing tasks are completed: driving data collection,driving data analysis and driving data cleaning.Among them,electric vehicle driving data cleaning includes the identification and filling of outliers and missing values,outliers recognition and filling are based on box diagram analysis,the missing value mechanism is completely random missing,and the missing mode is single variable missing mode.Secondly,based on the preprocessing data set,the driving behavior analysis is carried out: driving behavior parameters extraction,driving behavior factors extraction,driving behavior and energy consumption correlation analysis.Among them,there are55 driving behavior parameters,which can describe driving behavior in detail;nine potential driving behavior factors are extracted by exploratory factor analysis;correlation analysis between driving behavior parameters and energy consumption includes correlation analysis between driving behavior parameters and energy consumption,driving behavior factors.Based on the correlation analysis with energy consumption,the driving behavior characteristic parameters with the high correlation with energy consumption are found as input features of driving pattern classification model.Finally,based on energy consumption data,K-means clustering model is established to determine four categories of driving modes: low energy consumption driving mode,medium energy consumption driving mode,high energy consumption driving mode and abnormal driving mode.Based on the correlation analysis between driving behavior parameters and energy efficiency of EV,the driving behavior characteristic parameters with high correlation with energy consumption are selected:average speed,speed change factor,parking times per kilometer,kinetic energy change factor,and the multi-classification model of driving mode with high correlation with energy consumption of EV is established.Based on the driving data mining and driving behavior analysis of EV,this theies explores the relationship between driving behavior and energy consumption of EV,and provides a research idea and method for energy efficiency research of EV.Based on the characteristic parameters of driving behavior with high correlation with energy consumption,a new multi-classification model of driving pattern with high correlation with energy consumption is proposed by using K-means clustering algorithm and artificial neural network algorithm.Based on this model,the driving pattern can be classified and evaluated,which provides guidance for the establishment of high energy efficiency driving patterns for electric vehicles.