Research On Construction Of Electric Vehicle Urban Driving Cycle And Driving Style Recognition

With the increase in car ownership in China year by year,the road safety and fuel consumption problems are becoming increasingly serious.Among them,the driving cycle play an important role in the performance evaluation of ordinary fuel vehicles such as emissions and economy,but for the performance evaluation of electric vehicle battery energy consumption and mileage,the driving cycle of fuel vehicles are not applicable;in addition,changes in driver driving style also have an important impact on the traffic environment and tram energy consumption.Therefore,the study of driving cycle and the identification of driver driving style have important guiding significance and practical application value for the in-depth study of electric vehicle technology.Taking a certain type of electric vehicle as the research object,this paper collects the actual driving data of an electric vehicle in a certain city,constructs its driving cycle,and completes the matching of vehicle parameters based on the constructed electric vehicle driving cycle,establishes a driving cycle recognition model,and realizes a driving style recognition algorithm that considers driving cycle.The specific work completed is as follows:(1)The construction of the driving cycle of an electric vehicle in a certain city.The driving data of pure electric vehicles in a certain city are collected,and the pre-treatment analysis is carried out,the kinematic fragments are divided,the characteristic parameters are reduced by principal component analysis,the number of clusters is determined by the contour coefficient and the elbow method,and the cluster center of the K means clustering algorithm is optimized by the particle swarm algorithm,and the driving cycle of electric vehicles in a certain city are constructed;the comprehensive characteristic parameters of the original test data and the domestic and foreign driving cycle are compared with the traditional K-means clustering algorithm to verify the effectiveness of the construction method.(2)Vehicle parameter matching of electric vehicles based on driving cycle.According to the driving cycle and design performance objectives of the pure electric vehicles constructed,the speed,power and speed distribution of electric vehicles under the driving cycle are analyzed;the battery,motor and transmission system parameters are matched in combination with the dynamic theory of pure electric vehicles;and the vehicle model of pure electric vehicles is established for dynamic and economic simulation analysis.(3)Determination and accuracy analysis of driving cycle identification model.Based on the driving cycle database,the driving cycle is divided into sample segments and the driving cycle characteristic parameters are calculated by the recognition cycle,the driving cycle are classified and analyzed by the K means cluster analysis and its characteristics are analyzed,various driving cycle samples are sampled by the random number method,and the driving cycle identification parameters are selected according to the sample recognition speed and accuracy rate through the correlation coefficient and probability neural network algorithm,and the influence of the identification period and the prediction period on the driving cycle accuracy is analyzed to obtain the optimal driving cycle identification model.(4)Driving style recognition algorithm considering driving cycle.The driving style database is divided by a fixed length of time,and the driving style is classified and analyzed by K means cluster analysis,and the driving style identification and analysis are carried out based on the analysis of random vehicle speed and the CLTC-P driving cycle of Chinese passenger cars and the selection of characteristic parameters to characterize the driving style;since the driving cycle will have an impact on the driving style,the driving style recognition factor is introduced in the driving style recognition coefficient,and the driving style identification algorithm is combined with the driving cycle recognition algorithm.