Collaborative Research: A Langevin Subgrid Scale Closure and Discontinuous Galerkin Exascale Large Eddy Simulation of Complex Turbulent Flows

点击翻译按钮获取中文摘要

1603131/1603589/1604142 Givi/Mavriplis/Girimaji Turbulent combustion is encountered in nearly all energy conversion devices, such as internal combustion engines, gas turbines, boilers, and gas burners. Turbulent combustion involves many complex physical and chemical phenomena. The ability to predict accurately the turbulent combustion process is critical to the design and optimization of combustion devices for achieving low emissions and high efficiency. This research will develop accurate computer models and numerical algorithms that can reveal the fundamental turbulent combustion processes and allow industry to design clean combustion devices, thus benefiting the environment. Extensive code development and computer simulations will be conducted. Additionally, this project will also include numerous education and outreach activities, including the development of an Interdisciplinary Computational Science Minor to provide additional courses on high-performance computing, student training in the multidisciplinary field of Computational Modeling and Simulation, K-12 outreach through the INVESTING NOW and CAMP-SOAR programs, and recruitment of students from minority and under-represented groups through the EXCEL program.In an effort to increase the accuracy of turbulent combustion simulation, this award provides funding to develop a new Langevin subgrid scale (SGS) closure and to implement it with a new discontinuous Galerkin numerical scheme for large eddy simulation (LES) of turbulent flows. The former provides accurate modeling of the SGS transport for a wide range of turbulent flows, including compressible and chemically reactive. The combined methodology will be put together in a package in which the available computational cores are utilized in a dynamic, adaptive manner. This is an important concept in high performance computing necessary for massively parallel simulations up to petascale, and towards (future) exascale. The new LES tool will be employed for predictions of several turbulent flows. The computational requirements for the proposed LES, in its most sophisticated form and utilizing the highest intended resolution, will be several orders of magnitude less than that required for direct numerical simulations. When successfully completed, this research will have a significant impact on turbulent combustion research. It will be extremely valuable for both industry and government agencies. The outcome of this work can also positively impact other disciplines, such as climate and atmospheric modeling or bioengineering.