| With the rapid development of digital design and manufacturing technology,the Model Based Definition(MBD)has become the mainstream model of current product design and manufacturing.Therefore,the demand for product retrieval for MBD model reuse becomes increasingly prominent.Because the traditional CAD model retrieval method mainly considers the geometric information characteristics of the model,it has been difficult to meet the semantic retrieval requirements of the MBD model.In view of the above,this paper takes the MBD-oriented international standard STEP AP242 application protocol as the research object,and studies the semantic retrieval of MBD product models based on the STEP AP242 knowledge graph.The specific work is as follows:1.Aiming at STEP knowledge graph constructed on the basis of STEP AP242,in which the product's 3D geometric information and non-geometric information are semantically related,a multi-granular semantic cell model oriented to it is proposed.The MAAG-oriented attribute adjacency graph MAAG is constructed to realize the unified expression and correlation of geometric information and non-geometric information such as manufacturing features in the knowledge graph.2.Based on the multi-granular semantic cell model,the semantic classification and matching of MBD product models are performed.Firstly,the serial features of the MBD model are obtained by serializing and encoding the MAAG,and then the MBD model is classified by the improved LSTM.Finally,the semantic matching of the MBD model is realized from a multi-level perspective,which lays the foundation for the implementation of semantic retrieval.3.Based on the above,the semantic retrieval architecture and examples of STEP products are given. |
