Modeling Flow around Bluff Bodies and Predicting Urban Dispersion Using Large Eddy Simulation

详细信息    查看全文

• 作者：Yu-Heng Tseng ; Charles Meneveau ; and Marc B. Parlange
• 刊名：Environmental Science & Technology
• 出版年：2006
• 出版时间：April 15, 2006
• 年：2006
• 卷：40
• 期：8
• 页码：2653 - 2662
• 全文大小：652K
• 年卷期：v.40,no.8(April 15, 2006)
• ISSN：1520-5851

文摘

Modeling air pollutant transport and dispersion in urbanenvironments is especially challenging due to complexground topography. In this study, we describe a large eddysimulation (LES) tool including a new dynamic subgridclosure and boundary treatment to model urban dispersionproblems. The numerical model is developed, validated,and extended to a realistic urban layout. In such applicationsfairly coarse grids must be used in which each buildingcan be represented using relatively few grid-points only. Bycarrying out LES of flow around a square cylinder andof flow over surface-mounted cubes, the coarsest resolutionrequired to resolve the bluff body's cross section whilestill producing meaningful results is established. Specifically,we perform grid refinement studies showing that at least6-8 grid points across the bluff body are required forreasonable results. The performance of several subgridmodels is also compared. Although effects of the subgridmodels on the mean flow are found to be small, dynamicLagrangian models give a physically more realistic subgrid-scale (SGS) viscosity field. When scale-dependence istaken into consideration, these models lead to more realisticresolved fluctuating velocities and spectra. These resultsset the minimum grid resolution and subgrid modelrequirements needed to apply LES in simulations of neutralatmospheric boundary layer flow and scalar transportover a realistic urban geometry. The results also illustratethe advantages of LES over traditional modeling approaches,particularly its ability to take into account the complexboundary details and the unsteady nature of atmosphericboundary layer flow. Thus LES can be used to evaluateprobabilities of extreme events (such as probabilities ofexceeding threshold pollutant concentrations). Somecomments about computer resources required for LES arealso included.