Research On Energy Consumption Prediction Of Pure Electric Vehicle Based On Path Planning And Driving Style

In recent years,pure electric vehicles have entered the fast lane of development due to their advantages of green,pollution-free,low noise,and stable starting.However,the issue of "mileage anxiety" in pure electric vehicles remains an important factor affecting their further development.Research on energy consumption prediction of pure electric vehicles has become a hot topic,The differences in road conditions and driving styles are the main external factors affecting their energy consumption.Based on this,this article conducts a comprehensive evaluation of the proposed energy consumption prediction method by studying and analyzing the trend of energy consumption changes in pure electric vehicles under different driving styles and path planning conditions.This article will conduct research from four aspects,and the main research content is as follows:(1)For the modeling of pure electric vehicles: Develop a complete vehicle model based on MATLAB/Simulink,forming a closed-loop control system for the entire vehicle,including driver model,vehicle model,motor model,and battery model.For the establishment of energy consumption prediction model: firstly,the influence factors of pure electric vehicle energy consumption are analyzed from both internal and external factors.Based on the analysis results,the kinematics segments of typical working conditions are divided,and a variety of characteristic parameter data that can characterize the characteristics of working conditions are extracted from the segments.Secondly,in order to reduce computational complexity,dimensionality reduction based on Principal Component Analysis(PCA)was performed on the extracted multiple feature parameters.In order to further analyze the feature parameter category attributes after dimensionality reduction,the K-means clustering algorithm(K-means)is used to classify the extracted multiple feature parameters.Define the classification results as low speed,medium speed,high speed,and ultra high speed operating conditions.Then,a mathematical model for energy consumption prediction was established with unit mileage energy consumption as the core.Based on the characteristic parameter characteristics of each of the four types of operating conditions,the design and simulation of simulated operating conditions within different speed ranges were completed,resulting in the unit mileage energy consumption of different speed ranges.Finally,simulation verification was completed on the feasibility of the energy consumption prediction method proposed in this article.(2)Pure electric vehicles have significant differences in energy consumption under different road conditions.Based on this,the path planning studied in this section incorporates actual road conditions as traffic information.In the case of completely unobstructed roads,the concept of Ideal Path Planning(IPP)is proposed,and ant colony algorithm is used to iteratively optimize the established road network model,ultimately obtaining an optimal travel route.Considering the situation of road congestion,the concepts of shortest distance path planning(SDPP)and comprehensive path planning(CPP)were proposed,and a combination model based on road resistance model and road network model was established.The A *algorithm was used to iteratively optimize them to obtain the optimal travel route under SDPP and CPP.(3)Cluster analysis and identification of driving style: The construction of a virtual driving scene based on MATLAB/Simulink,Pre Scan,and driving simulator,as well as the collection of driving data,were completed.Three feature parameters that best reflect the characteristics of driving style were extracted from the collected samples.In order to obtain different types of driving styles,the K-means algorithm is used to classify the feature parameter data,ultimately obtaining three types of driving styles: radical,general,and mild.In order to identify different driving styles,an LSTM neural network model is used to train the feature parameter data corresponding to different driving styles.From the final identification results,it can be seen that the LSTM neural network model developed in this paper has a high accuracy in identifying driving style.(4)In order to achieve the final energy consumption simulation,an ideal path planning virtual driving cycle is obtained based on MATLAB/Simulink,Pre Scan,and a joint simulation of MATLAB/Simulink,Pre Scan,and driving simulator is used to obtain virtual driving cycles with different driving styles under SDPP and CPP.Simulate and analyze the energy consumption of different driving styles under the same driving style,different path planning under the same driving style,energy consumption simulation analysis of different driving styles under the same path planning,and simulation analysis of the optimal travel plan with integrated path planning and driving style based on the obtained virtual driving conditions under SDPP and CPP.Finally,integrating the above simulation analysis results,a comprehensive evaluation is made for the energy consumption prediction method of pure electric vehicles proposed in this article.This study comprehensively evaluates the energy consumption prediction model from two perspectives: practical significance and accuracy of energy consumption prediction,and ultimately obtains the best applicable scope.From a practical perspective,the "CPP general" combination type is one of the optimal types considering energy conservation and travel efficiency comprehensively.From the perspective of energy consumption prediction accuracy:The energy consumption prediction model has the lowest percentage of energy consumption prediction error for the "CPP general" combination type.Finally,it is concluded that the "CPP general" combination type is the best applicable category for the energy consumption prediction method of pure electric vehicles developed in this article.