CDS&E: Better by Design: Establishing Modeling and Optimization Techniques for Producing New Classes of Biomimetic Nanomaterials

CDS

• 批准号: 1761068
• 负责人: Wenxiao Pan
• 金额: $ 55.47万
• 依托单位: University of Wisconsin-Madison
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-09-15 至 2023-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1761068&HistoricalAwards=false
• 关键词: CDS& Better; Design; Establishing; Modeling

Novel nanomaterials with precisely tailored characteristics can enable innovation in areas ranging from manufacturing to energy storage and drug delivery but designing such materials can be a challenge. This project develops modeling and optimization techniques that will enable researchers to use desired properties to drive materials and process selection. The researcher harnesses the power of mesoscale modeling techniques and computational methods based on Bayesian machine learning and stochastic optimization to search the vast universe of options to identify promising candidates. This approach gives the community an enabling predictive tool for analysis and design of polymer-based nanomaterials. The knowledge gained from this research will be broadly disseminated through publications, conference presentations and by organizing symposia. Educational and outreach programs will be developed to train a diverse STEM workforce and to broaden participation of underrepresented students in the fields of engineering and computational science. Integrating a mesoscale coarse-graining method with stochastic optimization provides an enabling tool in soft materials and advances knowledge about design exploration in high-dimensional search spaces and design optimization under uncertainty. The goal of this project is to significantly reduce the cost of simulating the molecular self-assembly process and the characteristics of assembled materials. The researcher will develop a mesoscale model that describes the dynamics of self-assembly and simulates and predicts the structures and mechanical properties of assembled materials. A coarse-grained approach balances the need for accuracy in material properties, which is the basis for optimization, and the computational efficiency needed to make the optimization feasible. The computational framework, which includes the mesoscale modeling, classification, and optimization steps, will be validated by comparison with experiments on peptoids. This research will enable inverse design of peptoid-based biomimetic nanomaterials with precisely tailored structures and properties for applications such as chemical/biological sensors, biomimetic nanodevices and water/ion transport membranes. The computational methodology will be shared through GitHub and as LAMMPS subroutines and the computer codes will be released to the scientific community as open-source software.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.

具有精确定制特性的新型纳米材料可以在从制造到储能和药物输送等领域实现创新，但设计此类材料可能是一项挑战。该项目开发建模和优化技术，使研究人员能够使用所需的属性来驱动材料和工艺选择。研究人员利用基于贝叶斯机器学习和随机优化的中尺度建模技术和计算方法的力量来搜索大量的选项，以确定有希望的候选人。这种方法为社区提供了一个分析和设计聚合物基纳米材料的有利预测工具。从这项研究中获得的知识将通过出版物、会议介绍和组织专题讨论会广泛传播。将制定教育和推广计划，以培训多元化的STEM劳动力，并扩大工程和计算科学领域代表性不足的学生的参与。将中尺度粗粒化方法与随机优化相结合，为软材料提供了一种使能工具，并提高了关于高维搜索空间中的设计探索和不确定性下的设计优化的知识。该项目的目标是显著降低模拟分子自组装过程和组装材料特性的成本。研究人员将开发一个介观模型，描述自组装的动力学，模拟和预测组装材料的结构和力学性能。粗粒度的方法平衡了对材料属性精度的需求，这是优化的基础，以及使优化可行所需的计算效率。计算框架，其中包括中尺度建模，分类和优化步骤，将通过与peptoids的实验比较进行验证。这项研究将使基于肽的仿生纳米材料的逆向设计具有精确定制的结构和性能，用于化学/生物传感器，仿生纳米器件和水/离子传输膜等应用。计算方法将通过GitHub和LAMMPS子程序共享，计算机代码将作为开源软件发布给科学界。该奖项反映了NSF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。

项目成果

Multifidelity Cross-Entropy Estimation of Conditional Value-at-Risk for Risk-Averse Design Optimization

用于风险规避设计优化的条件风险价值的多保真交叉熵估计

• DOI: 10.2514/6.2020-2129
• 发表时间: 2020
• 期刊: AIAA Scitech 2020 Forum
• 作者: Chaudhuri, A.;Peherstorfer, B.;Willcox, K.
• 通讯作者: Willcox, K.

A multigrid preconditioner for spatially adaptive high-order meshless method on fluid–solid interaction problems

用于解决流固相互作用问题的空间自适应高阶无网格方法的多重网格预处理器

• DOI: 10.1016/j.cma.2022.115506
• 发表时间: 2022
• 期刊: Computer Methods in Applied Mechanics and Engineering
• 影响因子: 7.2
• 作者: Ye, Zisheng;Hu, Xiaozhe;Pan, Wenxiao
• 通讯作者: Pan, Wenxiao

Context-Aware Surrogate Modeling for Balancing Approximation and Sampling Costs in Multifidelity Importance Sampling and Bayesian Inverse Problems

用于平衡多保真度重要性采样和贝叶斯逆问题中的近似和采样成本的上下文感知代理建模

• DOI: 10.1137/21m1445594
• 发表时间: 2023
• 期刊: SIAM/ASA Journal on Uncertainty Quantification
• 作者: Alsup, Terrence;Peherstorfer, Benjamin
• 通讯作者: Peherstorfer, Benjamin

Physics-informed regularization and structure preservation for learning stable reduced models from data with operator inference

基于物理的正则化和结构保存，用于通过算子推理从数据中学习稳定的简化模型

• DOI: 10.1016/j.cma.2022.115836
• 发表时间: 2023
• 期刊: Computer Methods in Applied Mechanics and Engineering
• 影响因子: 7.2
• 作者: Sawant, Nihar;Kramer, Boris;Peherstorfer, Benjamin
• 通讯作者: Peherstorfer, Benjamin

Data-driven low-fidelity models for multi-fidelity Monte Carlo sampling in plasma micro-turbulence analysis

• DOI: 10.1016/j.jcp.2021.110898
• 发表时间: 2021-03
• 期刊: J. Comput. Phys.
• 作者: J. Konrad;Ionut-Gabriel Farcas;B. Peherstorfer;A. Siena;F. Jenko;T. Neckel;H. Bungartz