SBIR Phase I: Domain-Unified Modeling for Electro-Mechanical Component Libraries

这个小型企业创新研究第一阶段项目提出了一个框架，通过领域统一的产品模型来支持多领域工程环境中的产品开发、分析和决策。特殊的、甚至是正式的(基于标准的)产品数据模型受到数据冗余和一致性问题的影响，需要多个视图和属性来支持特定于领域的应用程序需求。几何表示方案中信息的隐含性质及其访问的困难和延迟加剧了这些问题。其结果是昂贵的翻译、手动验证和缓慢的设计迭代，导致漫长而昂贵的产品开发周期。这项工作的智力价值在于，所发展的形式主义能够支持产品模型数据的多个视图的同时生成和维护，以及它们之间的一致性的加强。在该项目中，代理几何脚本的一种新颖使用使得能够跟踪几何表示中数值几何属性对隐式信息的依赖关系。这项工作的更广泛影响来自于通过产品数据建模的国际标准(ISO 10303-210)将领域统一的建模方法应用于航空航天和国防工业中印刷电子组件的封装组件。应用程序的苛刻性质和较长的产品生命周期导致了非常大的组件库和模型，它们必须支持跨多个领域(电气、热、机械、制造、计量等)的设计和分析流程。因此，组件模型中数据的冗余性和一致性问题尤为严重。如果成功，该项目将产生的工具将为美国工业带来更高的技术能力和生产力，并将为科学和工程界提供宝贵的见解，使其了解在多领域工程设计环境中面临的数据一致性问题的可处理性。

This Small Business Innovation Research Phase I project proposes a framework to support product development, analysis, and decision making in multi-domain engineering environments through domain-unified product models. Ad hoc and even formal (standards-based) product data models, subject to the need for multiple views and attributes to support domain-specific application requirements, suffer from data redundancy and consistency problems. These problems are exacerbated by the implicit nature of information in geometric representation schemes and the difficulty and latency of its access. The result is expensive translations, manual verification, and slow design iterations leading to lengthy and expensive product development cycles. The intellectual merit of the work lies in the ability of the developed formalism to support the simultaneous generation and maintenance of multiple views of product model data, and the enforcement of consistency between them. In this project, a novel use of proxy geometric scripts enables tracking dependencies of numeric-valued geometric attributes on implicit information in a geometric representation. The broader impact of this work accrues from the application of the domain-unified modeling methodology to packaged components for printed electronic assemblies in the aerospace and defense industry though an international standard for product data modeling (ISO 10303-210). The demanding nature of the application and long product life-cycles result in very large component libraries and models that must support design and analysis processes across multiple domains (electrical, thermal, mechanical, manufacturing, metrology, etc.). Hence, the problem of redundancy and consistency of data in components models is particularly severe. If successful, this project will result in tools that will bring increased technical capability and productivity to American industry, and will provide the scientific and engineering community valuable insight into tractability of the data consistency problems being faced in multi-domain engineering design environments.