Analysis And Evaluation Method Of Influencing Factors Of Vehicle Fuel Consumption Based On Python

Increasing car ownership every year,with the decrease of non-renewable energy and increasingly serious environmental pollution,People have to pay attention to the problems that come with progress.But you can’t give up eating for choking,in an age of development,he who does not advance loses ground,with the development of the means of governance development brought about by the negative problems.Without affecting the dynamic performance of vehicles,improving the fuel economy of vehicles is the direction of people’s continuous efforts.Based on the Python programming language platform,using the real-time vehicle operation data collected by OBD detector,it takes the parameters such as idle time,acceleration,load rate,coolant temperature,speed,engine speed,throttle relative position and absolute pressure of intake pipe as independent variables and the average fuel consumption as the dependent variable,the characteristic parameters affecting fuel consumption were extracted to construct the multi-parameter regression prediction model of multi-linear,neural network and integrated instantaneous fuel consumption;Based on the original data collected by the OBD detector,it is processed for a second time,and a new speed ratio(AR)、reduction ratio(RR)、idle speed ratio(IR)、constant speed ratio(CR)、Average acceleration during acceleration(a+m)、Average acceleration during deceleration(a-m)、maximum acceleration(a max)、maximum deceleration(amin)、average speed(vm)、average speed(nm)、mean value of throttle position change rate(Tmc)、Hot time(TH)、mean value of throttle position change(T mp)、Standard deviation of added velocity during acceleration(a+s)、including standard deviation of speed(a-s)、standard deviation of speed(vs)etc 18 parameters related to fuel consumption are added.Perform cluster analysis on the outgoing travel segments of the operation,the effects of engine parameters and operating conditions on fuel consumption were found out.Due to the excessive dimension of parameters affecting fuel consumption,it is impossible to evaluate the level of fuel consumption based on a single parameter,so the principal component analysis method is used to extract 4 principal component factors through parameter dimensionality reduction,and according to the weight of principal component factors(eigenvalues).Finally,the comprehensive evaluation index of fuel consumption is calculated.The results of the above data analysis are as follows:the accuracy of the constructed multivariate linear,neural network and integrated instantaneous fuel consumption regression prediction models is 0.66,0.84 and 0.86,respectively,the root-mean-square errors were 4.40,0.43 and 0.39,and the mean absolute errors were 2.70,0.27 and 0.25,respectively;Based on the principal component factor extracted from principal component analysis and its corresponding weight,the comprehensive evaluation index of fuel consumption was constructed,Pearson correlation coefficient analysis shows that the coefficient between them is 0.89,which is a strong correlation.