High-Accuracy Transient Fuel Consumption Model Based on Distance Correlation Analysis

Abstract

Due to lack of a high-accuracy practical fuel consumption model, it is difficult to evaluate and control fuel consumption accurately. It is necessary to establish a more accurate and practical model according to transient motion characteristics of vehicles. To ensure accuracy of model, structural form of the steady-state basic + transient correction module is determined as overall structure of model. Then principal component analysis and cluster analysis are used to classify different driving conditions of vehicles reasonably. Then, based on distance correlation analysis, parameter combination which has strongest correlation with estimation error of steady-state is determined, and transient correction is established. After that, optimal transient correction structures are determined based on Bayesian criterion. Finally, model is tested and results show that mean absolute percentage error of fuel consumption estimation of new model is about 15%, which is lower than classical models. What’s more, new model is simple in structure and calculating speed is similar to VT-Micro model. Therefore, model not only has high accuracy, but also has high calculation speed.