This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).

0933642/0932613 Chow/Balaras

Large-eddy simulation (LES) is one of the most promising numerical techniques for modeling complex turbulent flows in a variety of applications ranging from engineering to geophysical flows. Today, one of the main obstacles to the widespread use of LES for such applications is the lack of high-accuracy numerical techniques that efficiently accommodate complex boundary conditions. The advent of structured adaptive mesh refinement (S-AMR) solvers combined with immersed boundary methods offers new opportunities for LES since they can accommodate complex boundary conditions, efficiently distribute computational nodes and at the same time maintain most of the features of structured solvers. A combined S-AMR/LES approach, however, requires the development of filtering and modeling strategies that can deal with the complexity and flexibility inherent to adaptive grids. This study seeks to overcome these challenges and make LES a robust simulation tool for a large range of practical applications, by developing and testing explicit filtering and reconstruction approaches that allow turbulent fluctuations to accurately cross coarse-fine grid interfaces in S-AMR. Previous attempts showed significant improvement over traditional results. These developments will be easily transferrable to existing codes of respective research communities. In addition, these developments will transform LES use in engineering and geophysics by improving fundamental aspects of turbulence modeling and enabling the simulation of flow in complex geometries that were never before possible due to high numerical errors and high computational cost. The particular focus applications in this study will enable the discovery of new physics of rough-wall boundary layers and set new standards for accuracy/efficiency in computations of atmospheric contaminant dispersion in urban environments. In addition to graduate students, this project will support a postdoctoral researcherat UC Berkeley for one year. Career counseling and grant writing training will be provided through interactions with both PIs. Student class projects will help create a public interactive website to upgrade the current pages in Wikipedia on LES and adaptive mesh refinement. The Wiki websites will aid in spreading nderstanding of the limitations but also of the great potential of LES as a universal simulation tool.

Project Start
Project End
Budget Start
2009-09-01
Budget End
2013-08-31
Support Year
Fiscal Year
2009
Total Cost
$153,289
Indirect Cost
Name
University of Maryland College Park
Department
Type
DUNS #
City
College Park
State
MD
Country
United States
Zip Code
20742