Stochastic Multiscale Computational Design Methodology

这项研究的目的是发展一种随机多尺度计算设计方法，以实现健壮和可靠的多尺度工程设计。系统遵循基于模拟的设计范例，使用多尺度分析。这项研究将为设计跨不同应用领域的分层材料和产品系统提供一种数学上严格和方法上可行的方法，例如与材料、能源和医学相关的领域。与大多数只考虑参数不确定性的现有工作不同，这项研究将提供一种统一的概率方法来量化和传播材料和产品设计领域的其他关键不确定性来源。这将通过有效地整合来自多尺度模拟和多尺度物理实验的信息来实现。通过利用多尺度分析中的层次尺度分解结构，将开发出一套独特适用于多尺度设计的高性能设计算法，从而扩展了不确定性量化、统计灵敏度分析、元建模和多学科设计优化的最新研究进展。如果成功，这项拟议的研究将为设计分层材料和产品系统提供一种通用方法；更广泛的应用可能会使军队和医疗、能源、消费电子和汽车行业的广泛国内行业受益。在一个多学科研究团队的共同努力下，该项目将揭示多尺度模拟、实验和设计之间的复杂集成问题。这项研究的成果将在机械科学、材料科学和工程设计之间建立新的界限。将通过全面传播计算机代码和研究成果以及通过课堂教学接触到广泛的受众。西北大学新设立的“预测科学与工程设计”跨学科研究生集群项目，美国国家科学基金会“纳米机械、纳米材料和微/纳米制造”夏季研究所，以及未被充分代表的群体的积极参与，都是传播我们研究成果的手段。

The objective of this research is to develop a stochastic multiscale computational design methodology to enable design of robust and reliable multiscale ?engineered? systems following the paradigm of simulation-based design using multiscale analysis. The research will provide a mathematically rigorous and methodologically viable approach for designing hierarchical materials and product systems across diverse application domains such as those associated with material, energy, and medicine. Unlike most existing work that only considers parametric uncertainty, this research will provide a unified probabilistic approach to quantify and propagate other critical sources of uncertainties across both the material and product design domains. This will be accomplished by effectively integrating information from both multiscale simulations and physical experiments at multiple scales. By exploiting the hierarchical scale decomposition structure in multiscale analysis, a set of high performance design algorithms that are uniquely suited for multiscale design will be developed, thereby extending the state-of-art research developments in uncertainty quantification, statistical sensitivity analysis, metamodeling, and multidisciplinary design optimization. If successful, the proposed research will provide a generic methodology for designing hierarchical materials and product systems; the broader application is likely to benefit the military and a wide range of domestic industries in the medical, energy, consumer electronics, and automotive businesses. With the joint effort from a multidisciplinary research team, the project will reveal the complex integration issues between multiscale simulations, experiments, and design. The fruition of this research will forge new boundaries between mechanical science, materials science, and engineering design. A broad audience will be reached through comprehensive dissemination of computer codes and research findings, as well as through classroom teaching. The newly established graduate interdisciplinary cluster program in "Predictive Science and Engineering Design" at Northwestern University, the NSF Summer Institute on "Nanomechanics, Nanomaterials, and Micro/Nano-Manufacturing," and the active participation of underrepresented groups are all means to disseminate our research findings.