CAREER: Open-source GPU-accelerated computational infrastructure for coastal fluid-structure interaction in extreme hydrodynamic conditions

可广泛使用的软件和程序可用于预测沿海结构对气候引起的极端水动力事件的复杂响应，这对于支持气候适应型沿海社区的发展至关重要。然而，目前用于评估海岸结构体系对极端水动力载荷响应的科学工具箱包括缺乏坚实理论基础的经验模型、过时的设计规范和过于简化的数值框架，这些框架在超出其有限范围的情况下表现不佳。该项目旨在通过开发新颖的高保真、基于物理的数值方法和开源、高性能的计算软件基础设施来解决这一不足，用于极端水动力事件与海岸结构之间的流固相互作用模拟。这项研究促进了沿海气候适应能力领域知识的进步。这与美国国家科学基金会致力于推动科学进步，促进气候变化和适应能力方面的突破的承诺是一致的。开发的计算工具的广泛传播有可能通过设计更具气候适应性的沿海基础设施来实现社会和经济效益。此外，它有可能影响涉及流固相互作用的多个科学领域。本研究纳入了几项教育和推广活动，包括对高中教师进行气候适应能力问题培训，让不同的学生群体参与项目，以及改进课程。这些活动提高了人们对气候变化的认识，培养了跨学科和计算思维，促进了多样性和包容性。该项目的技术目标是开发高保真的、基于物理的计算工具，用于极端水动力事件下的海岸流固相互作用。这些工具推进了沿海气候适应能力的数学方法、算法和计算软件。具体而言，该项目引入了以下网络基础设施创新。首先，它采用光滑粒子流体力学来模拟剧烈的自由表面流动和极端的结构变形，包括破碎。这种方法不同于以往的沿海流固相互作用的数值方法，这些方法通常依赖于基于网格的技术或刚体假设来表示固体结构。此外，它采用了一种新颖的压力投影方法，促进了流体和结构域的有效和准确的双向耦合，从而提高了预测精度。研究还深入研究了先进的数值方法来模拟结构损伤和断裂。这些模型包括相场模型、周动力学模型和微平面模型，这些模型具有模拟结构破坏的先进数值能力。除了这些计算发展之外，该项目还使用水槽设施为开发的计算工具提供实验验证。最后，这些计算和数学创新的高潮，以及为土木结构量身定制的复杂预处理模块，结合起来开发一个gpu加速软件平台，通过开源云存储库提供给研究界。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。

The availability of widely accessible software and procedures for predicting the intricate response of coastal structures to climate-induced extreme hydrodynamic events is paramount for supporting the development of climate-resilient coastal communities. However, the current scientific toolbox for evaluating the response of coastal structural systems to extreme hydrodynamic loads consists of empirical models lacking a solid theoretical foundation, outdated design codes, and oversimplified numerical frameworks that misbehave in scenarios beyond their limited scope. This project aims to address this deficiency by pioneering the development of novel high-fidelity, physics-based numerical methodologies, and open-source, high-performance computational software infrastructure for fluid-structure interaction simulation between extreme hydrodynamic events and coastal structures. This research enables the advancement of knowledge in the field of coastal climate resilience. It aligns with NSF's commitment to promoting the progress of science and facilitating breakthroughs in climate change and resilience. The wide dissemination of the developed computational tools holds the potential to deliver societal and economic benefits by enabling the design of more climate-resilient coastal infrastructure. Moreover, it has the potential to impact multiple scientific fields that involve fluid-structure interaction. Integrated into this research are several education and outreach activities that involve training high-school teachers in climate resilience issues, engaging diverse student cohorts in project participation, and enhancing the curriculum. These activities promote climate change awareness, cultivate interdisciplinary and computational thinking, and foster diversity and inclusion.The technical objective of this project is the development of high-fidelity, physics-based computational tools for coastal fluid-structure interaction under extreme hydrodynamic events. These tools advance mathematical methods, algorithms, and computational software on coastal climate resilience. Specifically, the project introduces the following cyberinfrastructure innovations. First, it employs Smoothed Particle Hydrodynamics to simulate violent free-surface flows and extreme structural deformations, including fragmentation. This approach departs from previous numerical methods on coastal fluid-structure interaction that often relied on mesh-based techniques or rigid body assumptions to represent solid structures. Furthermore, it utilizes a novel pressure projection method that facilitates an efficient and accurate two-way coupling of the fluid and structural domains, leading to high predictive accuracy. The research also delves into advanced numerical approaches for modeling structural damage and fracture. These include phase-field, peridynamics, and microplane models, that result in advanced numerical capabilities for simulating structural failure. In addition to these computational developments, the project uses water flume facilities to provide experimental validation for the developed computational tools. Lastly, the culmination of these computational and mathematical innovations, along with a sophisticated pre-processing module tailored to civil structures, are combined to develop a GPU-accelerated software platform made available to the research community through open-source cloud-based repositories.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.