Collaborative Research: Robust and Scalable Methods for Simulation and Data-Driven Modeling of Particulate Flows

协作研究：用于颗粒流模拟和数据驱动建模的稳健且可扩展的方法

• 批准号: 1821334
• 负责人: Denis Zorin
• 金额: $ 15万
• 依托单位: New York University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-08-01 至 2021-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1821334&HistoricalAwards=false
• 关键词: Collaborative; Research; Robust; Scalable; Methods

Complex fluids are mixtures of homogeneous fluids and deformable or rigid particles or fibers. These fluids are ubiquitous in biological systems and engineering applications and accurate prediction of their behavior is important for a broad set of problems such as hemodynamics modeling, simulation of intracellular processes, and microfluidics device modeling. Currently, accurate models of complex fluids require highly expensive simulations at microscopic or mesoscopic level, resolving the behavior of each particle (e.g., a deforming cell or fiber) and their interactions. In contrast, macroscopic continuum models for complex fluids (when available) allow for a far more efficient computation. However, these models often lack accuracy, and cannot capture all important aspects of the flow behavior in realistic settings; hindering construction of fast predictive models. The goal of the project is to enable construction of such models in a data-driven way. The investigators will develop a framework based on microscopic numerical simulations. The challenge in developing efficient and accurate continuum computational models of such fluids is the hard-to-analyze transition from microscopic to macroscopic parameters. In this work, the investigators focus on (1) robust and scalable solvers for particulate flow in 2D and 3D; and (2) the utilization of our framework to explore data-driven macroscopic models for the complex flows in 2D, with an initial exploration of extensions to 3D. The team will develop the algorithms for accurate solution of large-scale 3D flows with a high volume fraction of immersed deformable particles handling viscosity contrast, complex geometric boundaries, and long simulation times. Key mathematical components for this include boundary integral formulation and computation, kernel-independent quadrature schemes for singular and near-singular integration on surfaces in 3D, contact handling and close interaction, and improving efficiency of time-stepping schemes. In the second component of the project, the investigators will apply this framework to explore macroscopic characteristics of complex flows and their dependence on microscopic parameters and local geometry; this will create a foundation for data-driven continuum models for complex fluids.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.

复杂流体是均匀流体和可变形或刚性颗粒或纤维的混合物。这些流体在生物系统和工程应用中无处不在，并且准确预测它们的行为对于诸如血液动力学建模、细胞内过程的模拟和微流体装置建模等广泛的问题是重要的。目前，复杂流体的精确模型需要在微观或介观水平上进行非常昂贵的模拟，从而解决每个粒子的行为（例如，变形细胞或纤维）和它们的相互作用。 相比之下，复杂流体的宏观连续模型（如果可用）允许更有效的计算。然而，这些模型往往缺乏准确性，并且不能捕捉在现实环境中的流动行为的所有重要方面，阻碍了快速预测模型的构建。该项目的目标是以数据驱动的方式构建此类模型。研究人员将开发一个基于微观数值模拟的框架。 在开发高效和准确的连续计算模型，这样的流体的挑战是难以分析的过渡，从微观到宏观参数。在这项工作中，研究人员专注于（1）2D和3D颗粒流的鲁棒性和可扩展性求解器;以及（2）利用我们的框架来探索2D复杂流的数据驱动宏观模型，并初步探索3D扩展。 该团队将开发用于大规模3D流的精确解决方案的算法，这些流具有高体积分数的浸没可变形颗粒，可处理粘度对比度，复杂的几何边界和长时间的模拟。 这方面的关键数学组成部分包括边界积分公式和计算，核独立的求积方案的奇异和近奇异积分的表面在3D中，接触处理和密切的相互作用，并提高效率的时间步进计划。在该项目的第二部分，研究人员将应用该框架来探索复杂流动的宏观特征及其对微观参数和局部几何形状的依赖性;这将为复杂流体的数据驱动连续模型奠定基础。该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估来支持。

项目成果

A robust solver for elliptic PDEs in 3D complex geometries

用于 3D 复杂几何形状中椭圆偏微分方程的强大求解器

• DOI: 10.1016/j.jcp.2021.110511
• 发表时间: 2021
• 期刊: Journal of Computational Physics
• 影响因子: 4.1
• 作者: Morse, Matthew J.;Rahimian, Abtin;Zorin, Denis
• 通讯作者: Zorin, Denis

Optimizing contact-based assemblies

• DOI: 10.1145/3478513.3480552
• 发表时间: 2021-12
• 期刊: ACM Transactions on Graphics (TOG)
• 作者: Davi C. Tozoni;Yu-Nong Zhou;D. Zorin

Scalable simulation of realistic volume fraction red blood cell flows through vascular networks

• DOI: 10.1145/3295500.3356203
• 发表时间: 2019-09
• 期刊: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis
• 作者: Libin Lu;Matthew J. Morse;Abtin Rahimian;G. Stadler;D. Zorin