INFLUENCE OF VARIOUS COMPUTATIONAL CONDITIONS IN RANS MODEL ON THE PREDICTION ACCURACY OF CONCENTRATION DISTRIBUTIONS

Recently, applications of CFD (Computational Fluid Dynamics) are expanding to various environmental issues such as pollutant/thermal dispersion in urban areas. The outdoor environment sub-committee of the Architectural Institute of Japan have conducted several benchmark tests for obtaining basic information and knowledge in order to provide the extended practical guidelines of CFD, which can be applied to more broad environmental issues than the present ones. This paper reports the results which considered the sensitivity of various computational conditions in the RANS (Reynolds-Averaged Navier-Stokes equations) model to the prediction accuracy of pollutant concentration distributions for two different configurations.