Numerical simulations represent a unique predictive tool for understanding the three-dimensional flow fields and associated concentration distributions from contaminant releases in complex urban settings (Britter and Hanna 2003). Utilization of the most accurate urban models, based on fully three-dimensional computational fluid dynamics (CFD) that solve the Navier-Stokes equations with incorporated turbulence models, presents many challenges. We address two in this work; first, a fast but accurate way to incorporate the complex urban terrain, buildings, and other structures to enforce proper boundary conditions in the flow solution; second, ways to achieve a level of computational efficiency that allows the models to be run in an automated fashion such that they may be used for emergency response and event reconstruction applications. We have developed a new integrated urban dispersion modeling capability based on FEM3MP (Gresho and Chan 1998, Chan and Stevens 2000), a CFD model from Lawrence Livermore National Lab. The integrated capability incorporates fast embedded boundary mesh generation for geometrically complex problems and full three-dimensional Cartesian adaptive mesh refinement (AMR). Parallel AMR and embedded boundary gridding support are provided through the SAMRAI library (Wissink et al. 2001, Hornung and Kohn 2002). Embedded boundary mesh generation has been demonstrated to be anmore » automatic, fast, and efficient approach for problem setup. It has been used for a variety of geometrically complex applications, including urban applications (Pullen et al. 2005). The key technology we introduce in this work is the application of AMR, which allows the application of high-resolution modeling to certain important features, such as individual buildings and high-resolution terrain (including important vegetative and land-use features). It also allows the urban scale model to be readily interfaced with coarser resolution meso or regional scale models. This talk will discuss details of the approach and present results for some example calculations performed in Manhattan in support of the DHS Urban Dispersion Program (UDP) using some of the tools developed as part of this new capability.« less
[1]
Stevens T. Chan,et al.
High-Resolution CFD Simulation of Airflow and Tracer Dispersion in New York City
,
2005
.
[2]
An approximate projection method for incompressible flow
,
2002
.
[3]
J. Ferziger,et al.
A ghost-cell immersed boundary method for flow in complex geometry
,
2002
.
[4]
P. M. Gresho,et al.
Projection 2 Goes Turbulent - and Fully Implicit
,
1998
.
[5]
D E Stevens,et al.
Validation of Two CFD Urban Dispersion Models using High Resolution Wind Tunnel Data
,
2001
.
[6]
John Paul Iselin,et al.
A comparison of contaminant plume statistics from a Gaussian puff and urban CFD model for two large cities
,
2005
.
[7]
R. Britter,et al.
FLOW AND DISPERSION IN URBAN AREAS
,
2003
.
[8]
Scott R. Kohn,et al.
Managing application complexity in the SAMRAI object‐oriented framework
,
2002,
Concurr. Comput. Pract. Exp..
[9]
S. T. Chan,et al.
An Evaluation of Two Advanced Turbulence Models for Simulating the Flow and Dispersion Around Buildings
,
2004
.