Based On The Instance Of The Injection Mold Design Technology Research

Injection mould design still remains more an art than a science. In this thesis, technologies of case-based reasoning(CBR) computer aided injection mould design are studied.It's a key condition to build a comprehensive case library for application of CBR system. Mould cases are investigated, and rules of features information describing and abstracting are proposed, and also, a layer model to represent the case of injection mould is built. Further more, rules of case library designing, organizing and distributed storing are investigated, in which, an E-R(Entity-relation ) model of plastic parts and injection mould is constructed.Case retrieve is an important sector for CBR system to realize its utility. The case-retrieve algorithms are investigated and a newer one is proposed. Firstly, a distance of feature item is introduced and unitized for measurement of mould features similarity. Then, the unitized similarity algorithm of mould case is presented with the unitized distance of feature items. Weighted values of features are introduced based on the ratio of feature devotions, so that the undetermined artificial factors are avoided in apportioning them with experiments.The revision of retrieved results is of great importance for CBR system to realize its self-improvement and ampliation. Several revision algorithms are investigated and a new one based on standard component is proposed. For the storage of revised cases, rules are presented for the self-improvement of case library, that is to reserve high standardized cases and disuse low standardized ones.As an application, the CBR mould design system developed in the study is introduced to show the effectiveness of algorithms presented in the thesis.