Data-driven method for estimating emission factors of multiple pollutants from excavators based on portable emission...

This study aims to explore the factors that influence the emission characteristics of multiple pollutants from non-road mobile machinery (NRMM) under real-world conditions and to establish a data-driven method for calculating accurate emission factors. This research focused on NRMM excavators meeting the third-stage emission standards and identified the actual work characteristics of 108 excavators in different scenarios based on a self-developed testing system for 368,000 h. Additionally, a portable emission testing system (PEMS) was used to study the instantaneous emission characteristics under different driving styles and modes for 10 EC210 excavators with the largest engineering construction inventory. The results showed that the average time proportions of idling, working, and moving modes for excavators were 21 %, 66 %, and 13 %, respectively. The results also revealed that the instantaneous emission rates of multiple pollutants varied significantly under different driving styles and modes. Driving style affected the hydraulic pump power change rate through hydraulic pilot pressure, and engine load surge caused turbocharger response delay and in-cylinder combustion deterioration, which affected pollutant emissions. Driving mode affected the emission characteristics of idling, high-speed idling, moving, and working modes of excavators through the external characteristics corresponding to the engine speed gear set. The data-driven method for calculating emission factors differed from the traditional method for most indicators to varying degrees. In terms of fuel-based emission factors (EFfs), except for the EFf_NOx indicator, which was 7.859 % higher than the traditional method, the other three indicators were significantly lower than the traditional method. In terms of power-based emission factors (EFps), except for EFp_PM and EFp_PN, the other two indicators were much higher than the traditional method. EFp_CO and EFp_NOx were 7.93 % and 20.332 % higher than the traditional method, respectively. It is recommended to use the data-driven method based on the actual driving data distribution to provide scientific support for accurately establishing the emission inventory.