Embedding local driving behaviour in regional emission models to increase the robustness of on-road emission inventories

Abstract

This study presents the development of operating mode (opmode) distributions derived from local drive cycle construction methods developed based on real-world GPS data collection, and their impacts on average-speed emission factors (EFs). A data collection campaign was conducted between March and July 2018 whereby 82 research participants were recruited to record daily driving behaviors in the Greater Toronto and Hamilton Area (GTHA) for a period of one week. A drive cycle construction methodology was employed to build representative drive cycles based on micro-trips. The constructed drive cycles were compared with the interpolated drive cycles derived from the default database of the USEPA MOVES model. The results indicate that the MOVES default opmode distributions lead to higher average-speed EFs than the ones derived from local data. The difference between two drive cycle construction methods was also evaluated by comparing the variability in opmode distributions and the resulting average speed EFs. We observed that EFs were similar within each speed category, and the variation in cumulative opmode distributions was highest for an average speed of 40 mph. Moreover, a Monte Carlo Simulation was conducted to generate EF distributions based on local opmodes, further illustrating that local drive cycles generated significantly lower emission estimates than those based on the default database of MOVES. Finally, the minimum number of GPS data points required to develop a local opmode database with adequate variability was determined, illustrating that 4400–19,300 s were needed to generate robust distributions for different speed categories and road types.