A new vehicle specific power method based on internally observable variables: Application to CO2 emission assessment for a hybrid electric vehicle

Abstract

As an important vehicle activity recognition method, vehicle specific power (VSP) has been widely used for on-road traffic emission modelling since its introduction in 1999. The conventional VSP (VSP_veh) is calculated from externally observable variables (EOVs) on the vehicle level and represents the power that a running vehicle needs to overcome. However, for hybrid electric vehicles (HEVs) with two power sources, vehicle activity is not always directly related to engine emissions. This study introduces the engine level VSP (VSP_eng), which estimates engine power from internally observable variables (IOVs) obtained from the vehicle’s on-board electronic control unit (ECU). An engine bench test is first implemented to validate the estimation algorithm for VSP_eng. A real-world driving emission (RDE) test is then conducted with a HEV in Ningbo city of China to evaluate the performance of VSP_veh and VSP_eng in emission estimation. The results show a strong correlation between emission and VSP_eng (R2 = 0.9783), while a much weaker correlation was found between emission and VSP_veh (R2 = 0.4216). Further analysis indicates that this strong correlation between emission and VSP_eng applies to all driving conditions (urban, rural and highway). The differences between VSP_veh and VSP_eng are then highlighted by a combined correlation analysis where the four work modes of HEV can be graphically identified. Lastly, this study discusses the feasibility and potential benefits of the intelligent and remote vehicle emissions monitoring through the upcoming vehicle to everything (V2X) network.