Collaborative Research: CDS&E: data-enabled dynamic microstructural modeling of flowing complex fluids

包括印刷电子、光子学、可穿戴传感器和太阳能电池在内的许多重要技术都是通过涉及聚合物材料复杂流动的工艺制造的。在许多情况下，聚合物分子的排列和相对位置（微观结构）的细节在其性能中起着关键作用，这些特征敏感地取决于用于制造它们的流动。该奖项将利用实验方法的最新发展，产生丰富的数据集，包括复杂聚合物材料的流动和微观结构。这些数据的可用性为开发和应用机器学习、数据科学和聚合物物理学提供了机会，以开发预测数学模型，最终实现涉及复杂材料的先进制造工艺的设计。利用现代数据驱动方法对流动复杂流体进行建模的转型进展需要：(1)涉及不同流动范围内流动和微观结构的时间演变的大型实验数据集，以及(2)利用这些数据集的新建模框架。该奖项推进了这两个主题，旨在开发一种工具，以快速开发流体结构和应力的预测模型，目前尚无基于第一原理的模型。将使用丰富而真实的实验数据集，这些数据集来自复杂流动中复杂流体的同时流动和空间分辨散射测量。新的建模框架将集成机器学习、数据同化、降维和数据驱动的动态建模，并根据流动复杂流体的物理特性进行建模。这些方法将应用于材料加工中的一个重大挑战问题：发现非稀释取向粒子分散体的新物理描述，为描述流动与粒子相互作用耦合的新第一性原理模型的构建提供信息。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。

Many important technologies including printed electronics, photonics, wearable sensors, and solar cells are manufactured with processes that involve complex flows of polymeric materials. In many cases, the details of the alignment and relative positions (microstructure) of the polymer molecules play a key role in their performance, and these features depend sensitively on the flows used to manufacture them. This award will exploit the recent development of experimental methods that engender rich data sets incorporating both flow and microstructure for complex polymeric materials. The availability of these data provides the opportunity to develop and apply machine learning, data science, and polymer physics toward development of predictive mathematical models that will ultimately enable design of advanced manufacturing processes involving complex materials. Transformational progress in exploiting modern data-driven methods toward modeling flowing complex fluids requires (1) large experimental data sets involving time-evolution of flow and microstructure in a diverse range of flows, and (2) new modeling frameworks to exploit these data sets. This award advances these two themes toward development of a tool to rapidly develop predictive models of fluid structure and stress in material classes for which no first-principles-based models currently exist. Rich and realistic experimental data sets will be used, which come from simultaneous flow and spatially-resolved scattering measurements for complex fluids in complex flows. The new modeling frameworks will integrate machine learning, data assimilation, dimension reduction, and data-driven dynamic modeling, informed by the physics of flowing complex fluids. These methods will be applied to a grand challenge problem in materials processing: discovery of new physical descriptions of non-dilute orientable particle dispersions, informing the construction of new first principles models to describe the coupling of flow with particle interactions.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.