Large Eddy Simulation for City


The governing equations and the manner of turbulence closure in this large eddy simulation (LES) study are from Deardorff (1980). We use a mask method to solve these equations. Mask methods have a great advantage of easily implementing solid obstacles in the computational domain without using boundary fitted grids. They also reduce the computational cost. According to the mask method by Briscolini and Santangelo (1989), numerical integration proceeds in the following steps: (a) free evolution of the velocity field inside the entire computational domain neglecting the forcing term due to obstacles, (b) introduction of the forcing induced by the physical boundary conditions on the obstacle surfaces, and (c) pressure modification by requiring incompressible flow.

The governing equations are approximated by second-order central differences in space. The Adams-Bashforth scheme is applied to time integration. Although the Poisson equation of pressure can be solved by an iterative method such as SOR, it is very time-consuming. Instead, we use the following direct pressure-solver (Raasch and Schroter, 2001): the discretized Poisson equation is Fourier transformed along the x and y directions (2-D), (STEP1). The resulting tridiagonal system of the equations is solved in the z direction (STEP2), and then transformed back from phase space into Cartesian space (STEP3).

Although the parts of the current model are not new, the combination of features results in a very effective model, hereafter LES-CITY, for large eddy simulation of complicated airflow in cities.

Copyright(c) 2005 Kanda lab. All rights reserved.