Abstract
This study focuses on the performance of a Reynolds-averaged Navier-Stokes simulation (RANS) for air flow and gas dispersion in urban districts. The present numerical simulation is based on a finite volume method with an unstructured grid system to resolve the flow structure in a complex geometry. To evaluate the performance of RANS, the results are compared to those obtained by wind-tunnel experiments. The wind-tunnel experiments are conducted for gas dispersion from point sources in a complex residential area. The results show that RANS tends to underestimate the mean velocity and the turbulent intensities behind the relatively-high buildings. In addition, the mean concentration is strongly sensitive to the turbulent Schmidt number, which is the selective parameter appearing in the model of turbulent scalar fluxes, and RANS can be used to predict the concentration only when the optimum empirical value of the turbulent Schmidt number is selected.
