Petascale computations for complex turbulent flows at high Reynolds number

该奖项促进了使用IBM开发的名为Blue Waters的大型新计算资源的科学研究，并计划在伊利诺伊大学部署。它为佐治亚理工学院、华盛顿大学、圣地亚哥超级计算机中心、德克萨斯大学奥斯汀分校的主要研究人员、Blue Waters项目团队和供应商技术团队之间的技术协调提供旅费。提出的研究采用直接数值模拟（DNS）来解决与三种高雷诺数湍流相关的研究问题：湍流非预混甲烷火焰；各向同性湍流中的标量混合和粒子色散以及沟道中紊流的壁面流动。在每种情况下，中心重点是达到足够高的雷诺数，以探索由此产生的复杂物理。由于计算时间的要求，这些模拟在高达8192立方分辨率或同等分辨率的网格上进行，在现有设施上是不可行的，并且超出了实验可能的范围。研究人员开发的先进算法将允许模拟解决能源生产（湍流燃烧）和消费（运输）技术核心的基础研究问题。湍流燃烧是美国大部分能源生产的核心，而壁面湍流是运输过程中固有的能量损失的核心。在这些应用中，对湍流的详细动力学的理解可能会提高对湍流影响的预测能力，从而通过改进设计来提高性能。因此，提议的DNS及其后续研究将对燃烧和运输系统的工程产生主要影响。这反过来又影响能源消耗和污染。

This award facilitates scientific research using the large new computational resource named Blue Waters being developed by IBM and scheduled to be deployed at the University of Illinois. It provides travel funds to support technical coordination between the principal investigators at Georgia Institute of Technology, University of Washington, San Diego Supercomputer Center, University of Texas at Austin, the Blue Waters project team and vendor technical team. The proposed research employs direct numerical simulations (DNS) to address research questions associated with three high Reynolds number turbulent flows: a turbulent non-premixed methane flame; scalar mixing and particle dispersion in isotropic turbulence; and turbulent wall-bounded flow in a channel. In each case, the central emphasis is on reaching sufficiently high Reynolds number to explore the resulting complex physics. Due to the compute time requirements, these simulations, which are on grids up to 8192-cubed resolution or equivalent, are not feasible on existing facilities and exceed what is experimentally possible. The advanced algorithms developed by the researchers will allow the simulations to address fundamental research issues at the heart of energy production (turbulent combustion) and consumption (transportation) technologies.Turbulent combustion is at the center of a large fraction of US energy production, and wall-bounded turbulence is central to the energy losses inherent in transportation. An understanding of the detailed dynamics of the turbulence in these applications will likely increase predictive capabilities of the effects of turbulence leading to enhanced performance through improved design. Thus, the primary impact of the proposed DNS and the subsequent research that it will enable will be on the engineering of combustion and transportation systems. This in turn affects energy consumption and pollution.