Abstract

The performance of unsteady Reynolds-Averaged Navier–Stokes equations (URANS) for simulations of flow and dispersion fields around isolated cubical buildings has been examined in this study. URANS results were compared with those obtained from steady-RANS (SRANS) computations and experiments. The comparison determines not only the applicability of URANS simulations, but also the contribution of unsteady large-scale fluctuations to pollutant dispersion around buildings. Three different source locations, i.e. upwind, rooftop and downwind releases, were considered for pollutant dispersion around the building. It was found that the improvement of the predicted concentration field achieved by URANS largely depends on the source location. Although this improvement was not as significant in the upwind and rooftop release cases, the prediction accuracy achieved by URANS was substantially improved for the downwind release case, for which, the unsteady-RANS simulations yielded larger estimates of the momentum and concentration diffusions behind the building than SRANS did, improving the accuracy of the estimation of the mean concentration.