CDS&E: Studying Multiscale Fluid Turbulence via Open Numerical Laboratories

1507469（Meneveau）该项目的目标是通过生成新的高保真数据，并将这些新数据提供给整个国家科学界，从而发现湍流中的新的基本现象。湍流的基本特征取决于在湍流中观察到的多个时间和长度尺度，将被探索，并且能够进行这种分析的软件工具将被开发并提供给研究界。拉格朗日混沌和向湍流的过渡将使用大量计算机模拟产生的数据集进行研究，这些数据集将被吸收到网络上可访问的开放数值实验室中。众所周知，向湍流的过渡是通过一系列具有高度时空复杂性的现象发生的。然而，关于拉格朗日混沌如何在湍流的多个尺度上表现出来，以及如何更好地表征湍流的过渡行为，人们知之甚少。湍流在小尺度和大尺度上都含有涡状结构。该项目旨在回答关于多尺度涡旋结构的层次组织及其与能量级联的关系的基本问题。合作项目开发了一个开放的、计算数据支持的湍流实验室，整个时空数据可供更广泛的研究团体使用。该项目将进一步开发4D流数据集的多尺度版本，有效地为数据增加了第5维（尺度）。虽然约翰霍普金斯大学目前可用的资源集中在各向同性湍流上，但拟议的努力将把数据扩展到其他比各向同性湍流更复杂的情况，比如通道流动数据。数据的可用性和分析工具的可用性将使湍流研究发生重大变化，通过民主化获取高质量的模拟数据，可能会改变研究人员在这一领域的工作方式。教育模块，主要基于Matlab，也将允许学生查询模拟数据集前所未有的细节。

1507469(Meneveau)The goal of this project is to discover new fundamental phenomena in turbulent flow by generating new high fidelity data and by making these new data available to the whole scientific community of the Nation. Basic features of turbulence that depend on the multiple time and length scales observed in turbulent flows will be explored, and software tools that enable such analysis will be developed and offered to the research community. Lagrangian chaos and transition to turbulence will be studied using datasets arising from massive computer simulations that are ingested into web-accessible, open numerical laboratories. It is known that the transition to turbulence occurs through a sequence of phenomena characterized by high spatio-temporal complexity. However, very little is known about how Lagrangian chaos manifests itself at the multiple scales of turbulence, and how transitional behavior of turbulence may be better characterized. Turbulence also is known to contain vortical structures at the smallest scales, and at large-scales. This project aims to answer fundamental questions about the hierarchical organization of vortical structures at multiple scales and their relationship to the energy cascade. The co-PIs have developed an open, computational data-enabled turbulence laboratory where the entire space-time data is available to the wider research community. The project will further develop multiscale versions of the 4D flow datasets, effectively adding a 5th dimension (scale) to the data. While the currently available resource at Johns Hopkins is focused on isotropic turbulence, the proposed effort will extend the data to other more complex cases than isotropic turbulence, like channel flow data. Availability of data and availability of analysis tools will enable a step change in turbulence research, possibly transforming the way researchers work in this area by democratizing access to high-quality simulation data. Educational modules, mostly based on Matlab, will also allow students to query simulation datasets in unprecedented detail.