On-Road Fuel Consumption Algorithm Based On Floating Car Data For Light-Duty Vehicles

With the rapid social and economic development followed by the continued growth of automobile ownership, the road transportation has become a significant consumer of energy during the development of modern city. In order to evaluate the fuel consumption by motor vehicles in urban dynamic transportation networks from the perspective of raising the level of traffic management for saving energy, it is necessary to accurately quantify the fuel consumption. In this background, the development of fuel consumption algorithms and models has become a fast-evolving research topic in the field of transportation. Particularly, it becomes one of key problems to determine the average fuel consumption level of road networks by using the available dynamic traffic data. Dynamic traffic data mainly contain the average road speeds, which are used to evaluate traveling conditions of motor vehicles in a widely ranged road network. Such data include two categories, the floating car data with real-time mean speeds of roads and the mean speeds of roads derived by the detector data, which is still an on-going research topic at present. It is noted that existing fuel consumption algorithms and models can only calculate mean fuel consumption level of the road network by using the second-by-second speed data. However, it is extremely difficult, if not impossible, to collect the second-by-second speed data for all motor vehicles in a large-scale road network, thus such an approach is not practical and cannot meet the requirement in evaluating the mean fuel consumption level of road network in an dynamic manner. In this context, after summarizing existing fuel consumption algorithms and models, the research in this thesis develops an algorithm for calculating the average fuel consumption level of light-duty vehicles on expressways dynamically using the mean speeds of roads.First, it introduces a new concept in the fuel consumption model, named Vehicle Specific Power (VSP), a variable that has a better logical relationship and statistical pattern than the speed and acceleration. After a description of the calculation of VSP and its relationship with the fuel consumption, the thesis proposes an approach in developing the algorithm for calculating the fuel consumption through VSP. Further, it collects the basic data for developing the fuel consumption algorithm and designs the process of the algorithm development. Then, it establishes the on-road vehicle fuel consumption algorithm based on floating car data, which can calculate the average fuel consumption level on expressways dynamically by providing the mean speeds on the expressways. Subsequently, it validates the proposed fuel consumption algorithm by using the speed data and fuel consumption data of light-duty vehicles running on the expressways. The results show that the average relative error of calculating the total fuel consumption of light-duty vehicles running on expressways using the proposed fuel consumption algorithm is 4.13%, with a maximum relative error of 6.07%. When the validation is pursued for the fuel consumption under binned speeds at a 10 km/h interval for two Jetta vehicles, the average relative errors are 3.46% and 9.50%. Finally, the thesis applies the proposed fuel consumption algorithm using the one-minute floating car data of West 3^rd Ring Road in Beijing by selecting the fuel consumption per 100 kilometers as the economic indicator to the calculation of the fuel consumption levels on the West 3^rd Ring Road and West 4^th Ring Road in Beijing dynamically. And it compares the calculation results by the proposed algorithm with the statistics from Beijing, and found that they match very well.