WORKSHOPS: Institute for Mathematics and Its Applications "Initiative in Materials Science"; January 24-26, 1996 and February 1-3, 1996; Minneapolis, Minnesota

9626139 Friedman这项奖金将为分别于1996年1月底和2月初在明尼苏达大学数学及其应用研究所举行的两次为期三天的讲习班提供部分支助。研讨会将集中讨论以下主题：(1)微观结构和薄膜；(2)薄膜的“智能”传感和控制。第一个研讨会将集中讨论薄膜。模拟薄膜的计算方法通常集中在原子尺度或连续尺度上。然而，由于这些完全不同的模型所涉及的时间和长度尺度截然不同，它们目前没有联系起来。这限制了两种模型充分描述工业材料制造过程的能力。增加的计算能力并不能解决这个问题，因为从这些模型中伸缩的策略还不存在。虽然计算能力可能允许扩大规模，但必须开发更复杂的技术来处理时间尺度问题。研讨会将汇集材料科学领域的数学家和材料科学家，他们使用建模和仿真来共同解决微观宏观联系，以了解薄膜中的微观结构演变。一系列的会谈和讨论预计将导致一系列关于如何使用多尺度技术开发模型的建议。实现这些建议的最终目标是：(1)将从头算、计算与连续微观结构/性质模型联系起来的薄膜过程的高保真模拟，以及(2)从这些完整模拟中提取适合实时过程控制的物理非经验模型的方法。第二场研讨会也将从不同的角度关注薄膜。控制器在宏观层面上运行。在这个水平上，模型处于连续尺度，它们的演化由偏微分方程（PDE）决定。PDE的解决方案技术需要当前可用的最强大的计算机。这种细节级别的模型不适合实时过程控制。研讨会将探讨从完整模型中提取主要参数的数学方法的研究思路。理解微观和宏观的联系对于成功地使用反馈控制至关重要。第二个研讨会的一个特别感兴趣的主题是“智能”传感器，即“理解”过程模型并能够直接或间接使用这些模型确定过程参数的传感器。它将强调算法，允许传感器从其他参数中推断出感兴趣的参数，当一个人不能直接感知物理化学模型表明对过程控制至关重要的参数时。

9626139 Friedman This award will provide partial support for two three-day workshops to be held at the Institute for Mathematics and Its Applications of the University of Minnesota in late January and early February, 1996, respectively. The workshops will focus on the following topics: (1) Microstructure and Thin Films and (2) "Smart" Sensing and Control of Thin Films. The first workshop will concentrate on thin films. Computational approaches to modeling thin films have generally focused either on the atomistic scale or on the continuum scale. However, because of the drastically different time and length scales involved in these disparate models, they are not currently linked. This limits the ability of either model to describe the process adequately for industrial materials manufacturing. Increased computational power will not solve this problem because strategies for scaling up or scaling down from these models do not yet exist. Although computational power might allow scaling up in size, more sophisticated techniques must be developed for dealing with time scale issues. The workshop will bring together mathematicians working in materials science and materials scientists who use modeling and simulation in order to jointly address the micro-macro linkage needed to understand microstructure evolution in thin films. A series of talks and discussions are expected to lead to a set of suggestions on how to develop models using multiscaling techniques. The eventual goals in realizing these suggestions are: (1) high-fidelity simulations of thin-film processes which link ab initio, calculations to continuum microstructure/property models, and (2) approaches to extracting physical non-empirical models from these full simulations that are suitable for real-time process control. The second workshop will also focus on thin films, but from a different perspective. Controllers operate at the macro level. At this level the models are at the continuum scale and their evolution is dictated by partial differential equations (PDE). Solution techniques for PDE require the most powerful computers currently available. Models at this level of detail are unsuitable for real-time process control. The workshop will explore research ideas on mathematical approaches for extracting the dominant parameters from the full model. Understanding the micro-macro linkage will be essential for success in using feedback control. A topic of special interest of the second workshop is "smart" sensors, that is, sensors that "understand" the process models and are able to determine process parameters either directly or indirectly using these models. It will emphasize algorithms which allow sensors to infer parameters of interest from other parameters for cases where one cannot directly sense a parameter that the physicochemical models indicates is critical to the control of the process.