Common design structure discovery from CAD models

Abstract

This paper presents a method to solve the problem of common design structure discovery from a large number of CAD models. First, a CAD model is transformed into a Face Adjacency Graph (FAG) and each node of FAG is mapped to a point in two-dimensional plane after representing face shape information with two coordinates. So the shapes of models are directly compared through the point coordinates of FAGs' nodes. Thus, the common design structures are just the frequent appearing subgraphs of FAGs. Second, we develop an algorithm to discovery frequent subgraphs of FAGs. The main steps of the algorithm include: (1)frequent candidate subgraph generation based on merging of last discovered frequent subgraphs; (2)subgraph matching with graph descriptive code. The experiment shows a reasonable result of the discovered common design structures with our approach.