Case Representation and Retrieval under a Case-based Reasoning System for Fixture Design

Abstract

Fixtures play a key role in accurately locating and clamping workpiece. Fixture design is a complex process involving many factors and traditionally the designer's experience plays a big role. Recently case-based reasoning(CBR) based fixture design method becomes popular and it can ease personal experience dependence. However existing methods mainly use simple manufacturing parts' geometric feature in case representation and retrieval and this leads to inaccuracy and inefficiency in their real application. In view of that, a new comprehensive retrieval model for fixture design case representation is proposed and it incorporates geometry, material, processing and functional features. The new model uses object-oriented unified modeling language(UML) case representation to enable information to be processed in a structural and modular way. A parallel retrieval algorithm, combining geometrical shape and its topological relationship, has been subsequently presented. The similarity of geometry feature of two parts is calculated by searching a maximal similar connected subgraph with node corresponding constraints through geometric topological graph, and a concrete similarity formula about other factors has also been given. Finally, a case study illustrates the efficiency of this method.