Collaborative Research: NISC SI2-S2I2 Conceptualization of CFDSI: Model, Data, and Analysis Integration for End-to-End Support of Fluid Dynamics Discovery and Innovation

流体动力学是一个广泛的领域，跨越了大量的科学和工程问题领域，对包括制造业、气候分析、环境、健康、交通、推进和发电在内的各种重要应用至关重要。为了支持流体动力学研究和应用社区，该项目寻求社区参与，以便构想一个未来的研究所，即计算流体动力学软件基础设施（CFDSI），该研究所将广泛开发，共享和应用计算工具，用于从实验和计算来源生成和分析流体动力学数据。在其概念化之后，CFDSI将创建和扩展用于问题定义，解决和分析计算和实验研究的工具。主要目标是通过一组丰富和可扩展的软件组件促进计算工具和数据资源的共享，这些组件可以集成到广泛的现有流体动力学分析工具中。通过提高工具的效率和易用性，科学家和工程师准确预测和理解复杂流体流动行为的能力将得到增强，这将对流体动力学发挥作用的广泛应用中的设计、创新和发现产生重大影响。CFDSI甚至有可能影响K-12、本科和研究生教育，为学生提供各种各样的流体动力学研究资源。概念化的研究所将通过降低或消除与采用和使用这些资源相关的障碍，为流体动力学研究界提供各种强大的模拟、数据和分析资源。软件基础设施将对流体动力学研究界产生许多积极的影响。为此，CFDSI将在高度可持续的软件开发环境中，将流体动力学的最佳研究与数据科学/分析的最佳研究联系起来。具体而言，CFDSI将：1)加强流体动力学数据资源和CFD建模进展的传播，2)促进流体动力学研究的合作，特别是计算和实验研究人员之间的合作，3)使不同数据源之间的详细比较和计算模型的详细验证成为可能，4)简化了在新的和复杂的应用中使用先进的CFD模型、方法和代码，5)促进了先进的分析，如不确定性量化。数据压缩和优化，6)为学生提供先进的CFD方法和数据资源，包括计算和实验，以加强流体动力学的研究生和本科生教育，7)提高当前和未来CFD软件的可持续性，8)促进对不断增长的流体动力学数据集的管理。这些成果将大大提高流体动力学研究的效率和生产力。特别是，它们将通过以下方式改变流体动力学研究的行为：1)使其更具协作性；2)提高研究结果的可信度；3)使发现成为可能；4)降低追求新研究问题的成本；5)通过降低获取和采用CFD代码和大型数据集的相关障碍，使流体动力学社区多样化并扩大。软件组件将被设计用于分析实验和计算数据库，以及直接集成到CFD代码中。后者将使原位数据分析能够解决数据创建速率（求解器性能）和数据存储速率/容量（I/O资源）之间日益增长的鸿沟。在概念和实施之后，CFDSI将使流体动力学研究更加有效，从而影响流体动力学至关重要的各种应用领域，包括制造业、气候、环境、健康、交通、推进和发电（包括常规、替代和核能源），这反过来将对我们的经济产生重大影响。此外，CFDSI将提供沉浸式模拟和实验的能力，通过在正在进行的模拟或实验中建立与现场数据分析和问题重新定义的链接，将闭合思想、洞察力、发现和设计的循环。最后，CFDSI将影响由偏微分方程（例如固体力学）控制的其他问题领域，作为类似领域特定软件基础设施的模型和起点。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Fluid dynamics is a broad field spanning a large number of science and engineering problem domains that are critical to a wide variety of important applications including manufacturing, climate analysis, environment, health, transportation, propulsion, and power generation. To support the fluid dynamics research and applications community, this project seeks to engage the community in order to conceptualize a future institute, the Computational Fluid Dynamics Software Infrastructure (CFDSI), that will broadly develop, share, and apply computational tools for the generation and analysis of fluid dynamics data from both experimental and computational sources. After its conceptualization, CFDSI will create and extend tools for problem definition, solution, and analysis of both computational and experimental investigations. The primary objective is to facilitate the sharing of computational tools and data resources through a rich and extensible set of software components that can be integrated into a wide range of existing fluid dynamics analysis tools. By improving the efficiency of tools and their ease of use, the ability for scientists and engineers to accurately predict and understand how complex fluid flows behave will be enhanced, having a significant impact on design, innovation, and discovery across the vast range of applications where fluid dynamics plays a role. CFDSI even has the potential to impact on K-12, undergraduate and graduate education by making a wide variety of resources available to students for fluid dynamics investigations.The conceptualized institute will make a wide variety of powerful simulation, data, and analysis resources available to the fluid dynamics research community by lowering or eliminating barriers associated with the adoption and use of these resources. The software infrastructure will have a number of positive impacts on the fluid dynamics research community. To do so, CFDSI will connect the best research in fluid dynamics to the best research in data science/analytics within a highly sustainable software development environment. Specifically, CFDSI will: 1) enhance the dissemination of fluid dynamics data resources and advances in CFD modeling, 2) facilitate collaboration in fluid dynamics research, especially between computational and experimental researchers, 3) enable detailed comparisons between different data sources and detailed validation of computational models, 4) ease the use of advanced CFD models, methods, and codes in new and complex applications, 5) facilitate advanced analytics, such as uncertainty quantification, data compression, and optimization, 6) provide students access to advanced CFD methods and data resources, both computational and experimental, to enhance both graduate and undergraduate education in fluid dynamics, 7) improve the sustainability of current and future CFD software, and 8) facilitate the management of the growing body of fluid dynamics data sets. These outcomes will greatly enhance the effectiveness and productivity of research in fluid dynamics. In particular, they will transform the conduct of fluid dynamics research by: 1) making it more collaborative, 2) enhancing the credibility of research results, 3) enabling discovery, 4) reducing the cost of pursuing new research questions, and 5) diversifying and widening the fluid dynamics community through lowering the barriers associated with accessing and adopting CFD codes and large data sets. Software components will be designed for both analysis of experimental and computational databases as well as direct integration into CFD codes. The latter will enable in situ data analytics to address the growing chasm between data creation rate (solver performance) and data storage rate/volume (I/O resources). After conceptualization and implementation, CFDSI will enable more effective fluid dynamics research and thus impact the wide variety of application domains in which fluid dynamics is critical including manufacturing climate, environment, health, transportation, propulsion, and power generation (including conventional, alternative, and nuclear sources) which will, in turn, strongly impact our economy. Additionally, CFDSI will provide the capability for immersive simulations and experiments that will close the loop on idea, insight, discovery, and design through establishing links to in situ data analytics and problem redefinition during ongoing simulations or experiments. Finally, CFDSI will impact other problem domains governed by partial differential equations (e.g. solid mechanics) by serving as a model and starting point for similar domain-specific software infrastructures.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.