Improved source definition methods based on numerical simulation for predicting vehicle exhaust transport in street canyons

Correctly predicting vehicle exhaust transport in street canyons is crucial for public health. To provide the vehicle exhaust distribution with high spatial resolution, computational fluid dynamics (CFD) was employed to calculate vehicle exhaust transport. To overcome the limitation of the existing line source definition method, this investigation proposed two improved source definition methods, namely, a point source definition method and a virtual vehicle model. First, a field test was conducted in a real street canyon to validate a benchmark model for source definition, which constructed the vehicle geometry in the geometric model for CFD simulation. A case study was then performed in a street canyon to assess the proposed improved methods. Carbon monoxide (CO) was chosen as the vehicle exhaust, and the results from the proposed improved methods were compared with those from the existing line source definition method and the benchmark model. In the studied case, the proposed improved methods and the existing line source definition method were all able to predict the primary trend of CO transport. Compared to the existing line source definition method, the two proposed improved methods provided better predictions of the peak value of CO concentration. The point source definition method and the virtual vehicle model exhibited an improvement in overall accuracy by 8 % and 10 %, respectively, for prediction of the pedestrian-level average CO concentration along the pedestrian road on the high-concentration side. The virtual vehicle model slightly overperformed the point source-definition method by defining virtual vehicle cells in volumes occupied by vehicles.