CAD assembly descriptors for knowledge capitalization and model retrieval

Today, there exists a huge amount of digital data easily downloadable from Internet and/or simply accessible from large databases. Despite this rise, the methods to retrieve and search for specific data have not been sufficiently studied and developed, notably when considering 3D contents. Thus, it is sometime more efficient to define new 3D shapes starting from scratch rather than to try to make use of existing ones hardly identifiable within those databases. This is particularly true when considering CAD assembly models often resulting from a long and time-consuming modeling phase within the Product Development Process. Thus, having new methods, models and tools to capitalize, retrieve and reuse CAD assembly models would help saving a lot of time. This paper addresses such a difficult problem of finding a method to characterize and structure CAD assemblies so as to be able to search for similar ones. A framework has been designed for the retrieval of globally and/or partially similar assembly models according to different user-specified search criteria. It is based on an assembly descriptor, called the Enriched Assembly Model (EAM), which encodes all the required data automatically extracted from the geometry and structure of the CAD models. The data are organized in several layers thus enabling multi-level structuring and queries. It also allows fuzzy queries, which can be further refined.