Urban Dispersion CFD Modeling, Fact or Fiction?

The application of Computational Fluid Dynamics (CFD) to the understanding of urban wind flow and dispersion processes has gained increasing attention over recent years. While CFD has already been widely recognized in building and structural design applications (often referred to as Computational Wind Engineering), the rapid growth of this approach has expanded into environmental areas such as wind climate analysis within urban areas, air pollution assessments and fire predictions. In the urban dispersion modeling community, CFD is often viewed as a highly advanced (and perhaps overly complex) computational tool that is of questionable use in simulating meteorological processes that are largely statistical rather than deterministic in character, and where meteorological data for well-defined boundary and initial conditions are usually lacking.

In this paper we will address the strengths and limitations of the CFD approach for urban wind flow and dispersion applications. The focus of this discussion is on the application of this numerical tool for modeling transport and diffusion around buildings and building complexes. LLNL as well as other DOE laboratories have used variants of CFD technology to develop capabilities for emergency preparedness and response from releases of chemical and biological agents within population centers (Lee, et al. 1999). We will discuss the customization of the CFD model physics for the urban application, including turbulence modeling, vegetation canopy, heat and mass transfer, and atmospheric stratification. Among the other topics considered are ingestion of urban databases, grid nesting (coupling with larger scale meteorological models), assimilation of meteorological or gridded data, and initial and boundary conditions.

Examples to support the discussions will be drawn from a recent modeling/field experiment study (Calhoun, et al, 1999) of wind flow around an LLNL on-site office building (figure shown). Field data obtained from this field study have been used to improve and validate LLNL's FEM3MP model. While one single set of field data may not generate definitive conclusions on the superiority of the CFD approach over existing models, it nevertheless suggests that such an approach has a key role in the overall understanding of the urban dispersion problem.

References

Lee, R. L., S. T. Chan, J. M. Leone, D. E. Stevens, 1999: Airflow and dispersion around multiple buildings. UCRL- JC-132241, also in Proceedings of Air Pollution 99, 27-29 July, 1999, San Francisco, Ca

Calhoun, R., S. T. Chan, R. L. Lee, J. M. Leone, Jr., J. Shinn, and D. Stevens, 1999: Flow Patterns Around a Complex Building, A Model Validation Study. UCRL-JC-134876, also in Proccedings of 11th Conf. On Applications of Air Pollution, 9-14 January, 2000, Long Beach, Ca

This work was performed under the auspices of the U. S. Department of Energy by Lawrence Livermore National Laboratory under contract no. W-7405-ENG-48.