Abstract:Aiming at the problem that assembly data of assembly products usually scattered in the CAD model and the assembly process document, thus different types of assembly semantics information cannot be interconnected due to the heterogeneity, a modeling method for assembly semantic information based on knowledge graph is proposed. Combing with traditional data mapping and natural language processing in artificial intelligence, the intelligent integration of geometric information from the CAD model and assembly process information from the document is realized. Firstly, the concepts, entity classes and relation classes representing different levels of semantic features in the knowledge graph-based assembly semantics model (KG–ASM) are described. Secondly, according to the CAD model file and assembly process document, the instance nodes of the KG–ASM are automatically constructed based on data extraction, semantics mapping and natural language processing techniques. Finally, the assembly process of the speed adjusting mechanism in aviation product as an example is given to verify the feasibility of the KG-ASM and its construction method.