TAFSM

Team for Advanced Flow Simulation and Modeling



Home

Research Overview

Research Highlights

Projects

Takizawa Lab

Undergraduate Research

TAFSM Featured

TAFSM Recognized

Publication, Preprints

Currrent Team Members

Collaborators, Ex-Members

AHPCRC, History

Links

For more information:
  tezduyar@gmail.com

Numerical Simulation of the Dispersion of Air Contamination in a Subway Station

Daniel Clark

Principal Investigator:
Professor Tayfun E. Tezduyar, T*AFSM

Following the poison gas attack in the Tokoyo Subway earlier this year, Professor Tezduyar's group has conducted a numerical simulation of the spread of such a gaseous subsance through a subway station. The simulation was performed on the T3D supercomputer with a computational scheme which simultaneously solved the continuity and Navier-Stokes equations for the velocity field of the incompressible flow and with these results solved an advection-diffusion equation with a single source term to determine the density distribution of the air contaminant in the subway tunnel. The results of this computation were then post-processed to determine the spread of density isosurfaces with time over 80 time steps.

The simulation was initially conducted for a mesh of 187,612 nodes and 1,116,992 finite elements (shown below) with slip allowed along the surface of the tunnel (to avoid a discontinuity with the entering wind at the tunnel mouth) and a no-slip boundary condition applied to the body of the subway station. A wind was assumed to be blowing steadily down the tunnel from one mouth with a natural boundary condition applied at the other mouth, while a suction was also applied to the four roof vents. Currently, the simulation is being repeated for a modified mesh of 138,017 nodes and 795,327 element and staggered entrance ways (shown at bottom). In this case, a wind is assumed to be blowing in one entrance and out of the other with both tunnel mouths now natural boundaries. For both simulations, the following parameters were assumed to be relevent: Re = 350,000, Pr = 0.72, and Le = 1.0.