Large-eddy Simulation of Turbulent Flow in an Idealized Street Canyon

Abstract

The use of CFD as a research tool has been adopted successfully in many ways to predict the turbulent flow nature and pollutant dispersion properties in urban environment. Apart from the advance advantages offer by this method, some problems can be arising on respect to accurate prediction of data and uncertainties of assumption that are made in numerical modeling. Therefore, validation and generic sensitivity analyses are important to provide guidance for the execution and evaluation of the CFD studies. This study performs a series of Large-eddy simulations (LESs) to investigate the flow field within and above a two-dimensional idealized street canyon. Simulations were conducted for various domain sizes with a unity aspect ratio i.e. street width to building height ratio. Three computational domains were 2H x H x 6H (streamwise (x), spanwise (y) and vertical (z) directions), 6H x H x 6H and 10H x H x 6H define as Case 1, Case 2 and Case 3 respectively. The results were validated against experimental data obtained in wind tunnel studies. The mean velocities for each run cases estimated by LES are in good agreement with those obtained by wind-tunnel experiments. However, profiles of the Reynold shear stress as well as standard deviation for the streamwise and vertical velocity components show large discrepancies at all measured locations for each cases at building height. Apparently, it was indicating that increasing the domain size does not fully rectify the under predicted turbulent statics in the street canyons. This is most likely due to coherent structures that have developed above a canyon cannot properly simulated in a small domain size as designated without any consideration of other compromise parameters such as grid resolution.