CAREER: A scalable multiscale modeling framework to explore soot formation in reacting flows

该奖项全部或部分由《2021年美国救援计划法案》（公法117-2）资助。煤烟是在不完全燃烧过程中形成的一种含碳颗粒物质，对公众健康和福利有重大不利影响，是气候变化的重要强迫因子。为了准确地理解和减轻煤烟的影响，我们需要了解与煤烟形成和生长相关的所有过程——从煤烟在原子水平（又名原子尺度）的开始到其在现实世界燃烧系统中在设备水平（又名设备尺度）的成熟。不幸的是，如此详细的多尺度建模仍然是一项艰巨的任务。这导致在预测和控制烟尘排放及其对气候和公共健康的影响方面存在重大的知识缺口和重大的不确定性。该项目将创建一个模型框架，将燃烧系统的小规模原子建模与大规模工程建模结合起来。该项目将为燃烧产生的烟尘排放建模提供更好的预测能力，从而实现更清洁的燃烧系统。该项目还将在原子水平上详细了解煤烟的特性，从而更好地了解煤烟对地球的影响。这样做，该项目将服务于NSF的使命，以促进科学的进步和推进国家的健康，繁荣和福利。在这个项目中所做的技术工作的直接影响是双重的。首先，它将导致对烟灰起源的物理更完整的理解，并详细了解烟灰在现实世界中的演变。第二，建立的多物理场和多尺度建模框架将为燃烧烟尘的理论探索开辟新的视野。本项目中开发的多尺度桥接策略可以适用于需要多尺度和多物理场探索的其他问题。随着技术的发展，该项目还将与一家艺术博物馆合作开展外展活动，鼓励基于事实和数据的社区讨论诸如煤烟过程的复杂性、煤烟对社会的影响、环境政策和环境正义等问题。此外，还将有高中生参与的活动，这将有助于促进科学计算并鼓励学生从事STEM研究。这个项目将利用新颖的计算方法在不同的尺度上连接不同的物理领域。在原子尺度上，该项目将使用分子动力学等技术来解开烟灰起源的物理和化学。这些模型的结果以及实际烟尘颗粒的高分辨率电子显微镜图像将使用机器学习技术进行分析，以创建一个新的随机烟尘建模框架。这种烟尘建模框架将保留从原子尺度模型中获得的详细知识，同时有效地在连续尺度模拟中运行，例如燃烧装置的反应计算流体动力学（CFD）模拟。随机烟尘模型将使用一种新的欧拉-拉格朗日混合方法与详细而精确的湍流化学和辐射模型相结合。这种混合欧拉-拉格朗日方法将提供独特的混合数据-任务并行性和自动负载平衡，从而为多尺度、多物理场反应流求解器提供高效和可扩展的框架，用于详细探索煤烟过程。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。

This award is funded in whole or in part under the American Rescue Plan Act of 2021 (Public Law 117-2).Soot, a carbonaceous particulate matter formed during incomplete combustion, has significant adverse effects on public health and welfare and is an important forcing agent in climate change. To accurately understand and mitigate the effects of soot, we need to understand all the processes related to soot formation and growth - from the inception of soot at an atomic level (aka atomic scale) to its maturation in real-world combustion systems at the device level (aka device scale). Unfortunately, such detailed multiscale modeling remains a daunting task. This leads to a significant gap of knowledge and significant uncertainty in the prediction and control of the emission of soot and its effects on the climate and public health. This project will create a framework of models to combine small-scale atomistic modeling with larger-scale engineering modeling of combustion systems. The project will enable a better predictive capability for modeling soot emission from combustion which will lead to cleaner combustion systems. The project will also provide a detailed insight into the properties of soot at an atomic level enabling a better understanding of the effects of soot on the planet. In so doing, the project will serve NSF's mission to promote the progress of science and to advance the national health, prosperity, and welfare. The direct impacts of the technical work done in this project are two-fold. First, it will lead to a more complete understanding of the physics of soot inception and a detailed insight into the evolution of soot in the real world. Second, the developed multiphysics and multiscale modeling framework will open up a new horizon in the theoretical exploration of soot in combustion. The multiscale bridging strategies developed in this project can be adapted to other problems that require multiscale and multiphysics explorations. Along with the technical development, the project will also conduct outreach activities in collaboration with an art museum to encourage the community in fact- and data-based discourse on issues such as complexities of soot processes, the effect of soot on the society, environmental policies, and environmental justice, etc. Additionally, there will be activities involving high school students that will help promote scientific computing and encourage students to pursue STEM research. This project will bridge different domains of physics across different scales by utilizing novel computational approaches. At the atomic scale, this project will use techniques such as molecular dynamics to unravel the physics and chemistry of the soot inception. The results from these models along with high-resolution electron microscopic images of actual soot particles will be analyzed using machine learning techniques to create a novel stochastic soot modeling framework. This soot modeling framework will retain the detailed knowledge gained from atomic-scale models while efficiently operating at continuum-scale simulations such as in reacting computational fluid dynamics (CFD) simulations of combustion devices. The stochastic soot model will be combined with detailed and accurate turbulent chemistry and radiation models using a novel hybrid Eulerian-Lagrangian approach. This hybrid Eulerian-Lagrangian approach will provide a unique hybrid data-task parallelism and automatic load balancing leading to an efficient and scalable framework for multiscale, multiphysics reacting flow solver for detailed exploration of soot processes.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.