Multiple taxonomy-based design knowledge representation for mechanical fasteners.

摘要

The overall motivation of this research is to incorporate formalisms and structures of science into the domain of conceptual design. Structuring the body of knowledge includes a vocabulary of well-defined terms to precisely describe concepts, classifications of concepts as taxonomies, and formalisms to represent the relevant corresponding body of knowledge.; As designed artifacts increase in complexity and require multidisciplinary knowledge and as product development process is becoming increasingly collaborative, design repositories have emerged as media where product development community can retrieve and reuse design knowledge. Design repositories represent the context for this research whereas mechanical fastening represents the application domain. The focus of this research is on defining a methodology for developing vocabularies, formal representation of design concepts, and taxonomies for the fastening domain.; Design concepts are mapping of the functionalities to structural descriptions based on physical principles. Thus, to formally capture the concept of fasteners, the research included an investigation of all these principles that achieve this function resulting in a high level classification of the solutions. This step was followed by in-depth analysis of the fastening domain to delineate the types of knowledge generated and retrieved throughout the life cycle of a product. The result was a controlled vocabulary of the domain that constitutes the domain's knowledge and its semantics. Structural models of fasteners were defined based on an analysis of common fasteners.; The next step was to structure the body of knowledge by defining taxonomy for the domain. However, initial attempts showed that current methods are not adequate to capture the complexities and relationships of design concepts. A framework for handling this task was then defined based on combining and extending several techniques. The resulting framework utilized multiple multi-faceted Subsumtion hierarchies based on conceptual indexing. This allowed structuring the knowledge into sub-taxonomies which were bridged by the use of mediators comprising articulations. The final framework was semantically enhanced by introducing the notion of spectrum to accommodate for the attribute-value ranges locally.