SUMMARY

Adaptive reuse of archived parametric finite element analysis (FEA) models is a common form of FEA model reuse that involves integrating new information into an existing model to reapply the model to a new problem. Adaptive reuse scenarios arise in component-based design when new components are used in variations of existing designs, or changes in loading conditions or arrangements of components in a design cause a new failure mode or phenomenon to be the subject of the analysis. The use of keyword-based searches to retrieve models and supporting documents from electronic repositories for adaptive reuse results in imprecise searches that return irrelevant documents and miss important ones. The objective of this research is to develop a retrieval method for FEA models and supporting documents that results in more precise retrieval from open, dynamic engineering repositories, based on a modeler’s intended reuse. This research results in an automated retrieval method applicable to FEA models of component-based designs. The development of the method is based on studies using various types of electronic chip packages. The retrieval method is developed to support adaptive reuse of FEA models and uses an automated estimation of reusability to rank retrieved FEA models. Key contributions include: - Development and validation of an automated reusability-based selection method for FEA models of component-based designs - Development and validation of a classification-based retrieval algorithm based on description logic (ALE) subsumption hierarchies - Development and validation of conceptual data models for representing FEA models and associated information to enable automated retrieval and ranking based on reusability - Experimental studies on query performance and the computational feasibility of the retrieval algorithms