CFD Modeling of Near-Roadway Air Pollution

Currently, there is an increasing interest in modeling the dispersion of atmospheric pollutants using computational fluid dynamics (CFD), to characterize the influence of the traffic-generated emissions on the temporal and spatial variability in air pollutant concentrations in the near-roadway environment. To advance in this task, we modeled the dispersion of total suspended particles (TSP), over a flat terrain, within a neutrally stratified and fully developed atmospheric boundary layer. We included the effect of turbulence and deposition on particle size distribution downstream. We found that TSP concentration downwind exhibits a single profile when expressed in terms of three dimensionless numbers: normalized concentration, normalized distance, and emission speed ratio. Using this generic character of the results, we determined the average short- and long-term TSP concentration, modeling successive short-term intervals in which it could be assumed a pseudo-steady-state behavior. Results exhibited correlation levels of R-2 > 0.85 for daily and R-2 > 0.94 for monthly averages when compared with measured TSP concentrations downwind two unpaved roadways. Results also showed that the implemented CFD model resolved the two main issues with Gaussian models (currently the most used air quality model): over-prediction of pollutant concentrations near the roadway and problems dealing with wind speeds < 1 ms(-1).