Analysis And Prediction Method Of Commercial Vehicle Fuel Consumption Based On Vehicle Operating Status

Road transportation has always been the main mode of transportation in the transportation industry.Commercial vehicles are the main means of road transportation and their sales have been rising steadily.The increase of commercial vehicles will inevitably increase fuel consumption and emit a lot of polluting gases.Emissions will cause greater pressure on the environment and resources.Fuel consumption is also one of the major costs of the transportation industry.If the fuel efficiency of commercial vehicles has not been improved,it will increase the expenditure of transportation operators,reduce profits and inhibit the development of the industry.Reducing fuel consumption can not only relieve the pressure on the environment and resources,but also reduce the cost of transportation and increase the vitality of the whole transportation industry.Therefore,consumption research has attracted extensive attention in related fields.The research on fuel consumption is mainly divided into two categories,namely physical research and statistical research.Since physical research is difficult to reduce fuel consumption after the vehicle is put into use,statistical fuel consumption research has gradually become a hot spot.In the fuel consumption research based on statistics,some methods are to consider the influence of external factors and vehicle driving conditions on fuel consumption.It analyzes the correlation among them and fuel consumption to find out which factors have a greater impact on fuel consumption,and uses them to predict fuel consumption.Suggestions to reduce fuel consumption are given according to the results.Other methods analyze and predict fuel consumption based on driving behavior,which is the main factor that affects fuel consumption.By summarizing and evaluating driving behavior,predicting low fuel consumption driving behavior and high fuel consumption driving behavior.Although the above fuel consumption analysis and prediction methods have made some achievements,they still have some shortcomings.First of all,the existing fuel consumption analysis and prediction methods consider the factors that affect fuel consumption,and rarely consider the influence of vehicle internal state on fuel consumption.Secondly,when some methods consider the influence of various factors on fuel consumption,most of them individually analyze the influence of factors on fuel consumption,but they fail to comprehensively consider the mutual influence relationship among these factors.This will affect the effect of fuel consumption prediction to some extent.Based on the above problems,since the impact of various factors on the vehicle will be directly reflected in the operating state of the vehicle,this paper combines a large amount of data provided by the Internet of Vehicles to propose a commercial vehicle fuel consumption analysis and prediction method based on the operating state of the vehicle.First,relevant factors that can describe the operating status of vehicles are mined from the data,and the interaction relationship among factors and fuel consumption is analyzed to build a correlation diagram that can describe the interaction relationship among them.Then,a fuel consumption prediction model is established.Through in-depth study of the temporal and spatial dependence of various factors,the vehicle operating status data in the historical time period is used to predict the average fuel consumption in the next time period,and a good prediction effect has been achieved in practical applications.The analysis and prediction method of commercial vehicle fuel consumption based on vehicle operating status is mainly divided into three steps:The first step is to analyze the interaction among various factors of vehicle operating status and fuel consumption,then build its correlation diagram.This paper firstly uses the method of gray correlation analysis to obtain a gray correlation matrix composed of gray correlation degrees between various factors and fuel consumption,then establishes a correlation diagram describing the relationship between them based on the matrix.The second step is to build a model to learn and integrate the spatial and temporal characteristics of the influencing factors of fuel consumption.In the process of vehicle operating,there is interdependence among factors and fuel consumption in space and time.In this paper,the graph convolutional neural network and LSTM network are used to learn the spatial and temporal characteristics among factors and fuel consumption respectively.The results of spatial and temporal feature learning are further enhanced by attention mechanism,so we can measure the importance of factors and integrate the spatial and temporal characteristics information in this period.The third step is to predict the average fuel consumption of the vehicle in the future period.In this paper,the LSTM unit and the attention unit are connected alternately in chronological order,and the output result of the last time feature module is decoded and output through a full connection layer as the final prediction result.The main works of this paper are summarized as follows:1.A new method of fuel consumption analysis based on grey correlation analysis was proposed.In order to comprehensively consider the influence of various factors on fuel consumption,this paper focuses on the influence of vehicle operating status related factors on fuel consumption.This paper analyzes the relationship among the factors related to vehicle operating status and fuel consumption through the grey correlation analysis,and describes the complex mutual influence relationship among factors and fuel consumption intuitively with the correlation diagram.2.A new fuel consumption prediction model based on vehicle operating status and spatiotemporal feature learning was constructed.This model can predict the average fuel consumption of vehicles in the next period by using the data related to fuel consumption in the historical period.In this paper,the study of spatiotemporal characteristics is applied to the fuel consumption prediction model.By fully analyzing the spatiotemporal dependence between the factors related to vehicle operating status and fuel consumption,the fuel consumption prediction model has achieved a better prediction effect.3.In this paper,the accuracy of fuel consumption analysis and prediction method is verified by the relevant data of running vehicle provided by a large commercial vehicle manufacturing enterprise.The experimental results show that our method can accurately predict the fuel consumption value in the next period of time according to the vehicle operating status in historical time.In addition,the ablation experiment in this paper verifies the effectiveness of the spatiotemporal feature learning method used in fuel consumption prediction.Finally,the method proposed in this paper is compared with other methods,and the results show that our method is more suitable for fuel consumption prediction than other forecasting methods.