Development Of Simplified Vehicle Fuel Consumption Model Based On Vehicle Weight

In recent years,it has become a research hotspot to quantify traffic fuel consumption in order to support fuel-saving strategies in the transportation sector due to the severe global and domestic energy status.The fuel consumption models based on the vehicle specific power(VSP)have become the mainstream models.Such models calculate fuel consumption by four steps:1)categorize vehicles into different types;2)develop fuel consumption rate(FR)for each vehicle type,which is defined as the fuel consumption per unit time,based on fuel consumption tests;3)develop the VSP distribution,which is defined as the time fraction in each VSP bin,based on vehicle activity data;and 4)develop the fuel consumption factor(FF),which is defined as the fuel consumption per unit distance,being derived by multiplying the FR by the VSP distribution,and then the traffic fuel consumption in an area is calculated by multiplying the FF by the vehicle kilometers of travel of all vehicles.However,current research on fuel consumption estimations faces following problems and challenges.On the one hand,the combination of the vehicle type,the road type and the speed bin in traffic flows will increase the level of complexity of fuel consumption calculation and revisions geometrically;while on the other hand,the traditional model cannot quantify the internal relationships of the fuel consumption estimations for different vehicle types.Thus,there exists some room for the traditional fuel consumption model to be simplified in the model development,model application and theoretical study.In this context,the study in this these intends to develop a simplified vehicle fuel consumption model based on the vehicle weight,which includes three components,the simplified FR model,the simplified VSP distribution model,and the simplified FF model.The three simplified models can not only independently describe the relationship between the physical meaning of each parameter and the vehicle weight,but also achieve the interconnection and transformation under different scenarios.The main research accomplishments are described as follows:First,the simplified FR model based on the vehicle weight is developed.It is found that the FRs of different vehicle types have some similarities,ie.The FRs are related to the VSP and speed,as well as some differences,i.e.the FRs increase with the increase of the vehicle weight(except for some special VSP bins).The relationship between the FRs and the vehicle weights can be described as the polynomial functions,of which the parameters are related to the VSP and speed.The goodness-of-fit coefficients(R2)are greater than 0.8 for 70%VSP bins,and the VSP bins with R2 of less than 0.8 represent the operating modes of braking,idling,negative VSP,and low VSP.The FR can be derived from the vehicle weight without the fuel consumption test employing the simplified FR model,which applies to the scenario when sufficient vehicle activity data are available.Second,the simplified VSP distribution model based on the vehicle weight is developed.It is found that the VSP distributions of different vehicle types have some similarities,i.e.the VSP distribution of different average travel speeds can be characterized by the Gauss function(R2 falls into 0.9 to 1),the overlap rates between the modeled and the measured VSP distributions are mostly between 70%and 100%,the symmetrical axis moves to right and the width becomes wider,as well as some differences,i.e.with the increase of the average travel speed the VSP distribution' peak value decreases:the per-unit vehicle weight's contribution to the VSP distribution' peak value,symmetrical axis and the width of different vehicle types decrease with the increase of the vehicle weight.The relationship between the VSP distributions and the vehicle weights can be described as the polynomial functions,of which the parameters are related to the average travel speed.The goodness of fit coefficients(R2)ranges from 0.6 to 1,and the R2 decreases with the increase of the average travel speed.The average overlap rate between the modeled and the measured for the speed of 2-30 km/h is 80%,and 32-70 km/h is 70%.The VSP distribution can be derived from the vehicle weight without the vehicle activity test employing the simplified VSP distribution model,which applies to the scenario when sufficient FR data are available.Finally,the simplified FF model based on the vehicle weight is developed,which is the core of this study.It is found that the FFs of different vehicle types have some similarities,i.e.the measured data and the FFs derived by the traditional model can be characterized by the negative power function,and the goodness-of-fit coefficients(R2)are greater than 0.98(except for measured data of a few trucks),as well as some differences,i.e.the concept of Fuel Consumption Factor Unit(FFU)is proposed,which is defined as the ratio of FFs of different vehicle types to FFs of the light-duty passenger car,and the FFUs of per-unit vehicle weight decrease with the increase of the vehicle weight.The relationship between the FFs and the vehicle weights can be described as the negative power functions,of which the parameters are related to the average travel speed.The goodness-of-fit coefficients(R2)of the model are mostly not less than 0.95.The relative errors between the modeled and measured FFs are mostly from-20%to 20%.The FFs of the passenger cars are overestimated compared with those of trucks.The FF can be derived from the vehicle weight without the fuel consumption test or the vehicle activity test employing the simplified FF model,which can also simplify the application process in practical applications.The simplified model is employed to estimate the vehicle fuel consumption on the expressways in Beijing during the morning and evening peaks.It is found that the simplified model overestimates traffic fuel consumption by only 4.22%and 4.31%respectively in the morning and evening peak of the workday compared with the traditional model.In the case study,the test of the fuel consumption and the vehicle activity and the evaluation process can be simplified by 87.5%with a sacrifice of only 5%accuracy.The contribution of this study includes:1)use the vehicle weight as the representative variable for different vehicle types,which overcomes the current limitation that model building,model application and theoretical study only focus on a specific vehicle type,thus achieving the pupose of simplifying the traditional fuel consumption model;2)propose the concept of the FFU,base on which the FFs of other vehicle types be counted and converted by the light-duty passenger car.