Computational Fluid Dynamics CFD Modeling of Buildingscale Dispersion

Computational Fluid Dynamics (CFD) Modeling of Building-scale Dispersion Shuming Du September 12, 2002 Air Resources Board California Environmental Protection Agency Working Draft - Do Not Cite or Quote

What are CFD Models? l l l CFD models numerically solve basic equations of fluid dynamics, similar to what MM 5 and RAMS do for mesoand regional scales CFD models focus on detail features in a small scale setting, e. g. , flow around individual buildings A dispersion model can be built into CFD model or standalone 4 stand-alone is used in ozone/PM modeling, i. e. , generate flow field first and then use it to drive dispersion model Working Draft - Do Not Cite or Quote

Motivation: Why CFD Models? l Sometimes we need to estimate concentrations from sources just across the street Receptor l Regulatory models are not capable of simulating the extremely complex wind fields and dispersion of pollutants in these conditions Working Draft - Do Not Cite or Quote

GASFLOW l Many commercial and public-domain models are available, GASFLOW is chosen because: 4 Available to the public 4 Can be used in both outdoor and indoor environments 4 Option to use particle model to calculate dispersion therefore avoiding problems associated with K-theory Working Draft - Do Not Cite or Quote

Preliminary Results l Model has not been tested l Some features are not consistent with what we know l Dispersion model is not correct, still needs improvement l The purpose of the following plots is merely to show what a CFD model can do Working Draft - Do Not Cite or Quote

Preliminary Results l Two one-story 15 m x 6 m x 5 m buildings, 16 m apart l Two source locations Wind Source 2 Source 1 Working Draft - Do Not Cite or Quote

Preliminary Results l Modeling domain: side-view Working Draft - Do Not Cite or Quote

Preliminary Results l Modeling domain: top-view Working Draft - Do Not Cite or Quote

Preliminary Results l Wind field: side-view Working Draft - Do Not Cite or Quote

Preliminary Results l Wind field: side view Working Draft - Do Not Cite or Quote

Preliminary Results l Modeling domain: top-view (at z = 25 cm) Working Draft - Do Not Cite or Quote

Preliminary Results l Modeling domain: top-view (at z = 425 cm) Working Draft - Do Not Cite or Quote

Preliminary Results l Modeling domain: top-view (at z = 745 cm) Working Draft - Do Not Cite or Quote

Preliminary Results l Diffusion from source 1 (top view) Working Draft - Do Not Cite or Quote

Preliminary Results l Diffusion from source 1 (side view) Working Draft - Do Not Cite or Quote

Preliminary Results l Diffusion from source 1 (angled view) Working Draft - Do Not Cite or Quote

Preliminary Results l Diffusion from source 2 (top view) Working Draft - Do Not Cite or Quote

Preliminary Results l Diffusion from source 2 (side view) Working Draft - Do Not Cite or Quote

Preliminary Results l Diffusion from source 2 (angled view) Working Draft - Do Not Cite or Quote

Future Work l l l Modify the model to address the problems revealed in the preliminary work Correct the dispersion algorithm to reflect the latest development of Lagrangian particle modeling Test the model against EPA wind tunnel experiment and real-world data collected at Logan Memorial Junior High School and at CE-CERT, UC Riverside Working Draft - Do Not Cite or Quote

Future Work (continued) l This work may eventually enhance our modeling capability in dealing with dispersion at building scale, for example, estimating concentrations caused by chrome plating facilities in Barrio Logan Working Draft - Do Not Cite or Quote