Abstract
In order to apply Computational Fluid Dynamics (CFD) techniques to estimate pollutant dispersion in urban areas, it is important to assess the performance of numerical models used to estimate these phenomena. Recently, the MUST (Mock Urban Setting Test) has often been adopted as a test case for evaluating numerical models for micro-scale urban meteorology. This paper presents the results of model evaluation exercises carried out using MUST based on the Architectural Institute of Japan (AIJ) experience. The results of model evaluation exercises conducted by AIJ are broadly consistent with the results obtained by the COST group. However, the variety of results by each computation can be minimized by setting standard computational conditions. All computations including both RANS and LES show good agreement with wind experiment data. In general, LES cases show comparable accuracy to RANS in predicting U and W. However, LES shows better agreement than RANS in TKE and concentration.
