Research on driving cycles and fuel consumption models for trunk logistics

Trunk logistics serves as a critical link in national economic operations, and its fuel/energy consumption efficiency directly impacts the control of logistics costs and the mitigation of environmental burdens. To address the limitation that existing studies on driving cycles have not fully incorporated slope and load factors, this research proposes a three-dimensional Driving Cycle clustering method based on vehicle speed-slope-load, and develops an effective driving cycle model. On this basis, a Cluster-based Fuel Consumption Model (CFCM) is further established. By analyzing the actual driving data of trunk logistics vehicles in China, this study screens key features through correlation analysis and principal component analysis, and ultimately classifies driving cycles into 18 categories via clustering. The proposed CFCM achieves a mean absolute error (MAE) of 0.5259 g/s and a root mean square error (RMSE) of 0.6305 g/s, both below 10% of the typical fuel flow rate. The prediction errors follow an approximately normal distribution, indicating stable performance and strong generalization capability. Finally, an application example of the proposed CFCM in practical economic route optimization scenarios is presented. The research results can provide data support for energy consumption prediction, route optimization (Green Vehicle Routing Problem), and adaptive energy allocation strategies of trunk logistics vehicles, and hold significant engineering practical value for improving vehicle fuel economy.