Research And Application Of Design Intention Reasoning Method Based On STEP Knowledge Graph

With the development of MBD,product information exchange based on MBD model has become more frequent.STEP AP242 is an internationally recognized data exchange standard and contains a wealth of product data information.Design intent is an important part of product information.Design intent information helps to improve product reusability and facilitates the intelligent manufacturing of products.Aiming at the requirement of design intent semantic mining in product information exchange,based on STEP AP242,STEP knowledge map of product 3D geometric information and non-geometric information semantic association is constructed to realize the extraction and exchange of STEP design intent.The key technologies of STEP knowledge map semantic model construction,STEP knowledge extraction for knowledge map construction and STEP knowledge reasoning based on knowledge map are studied.The semantic cell model for design intent and multi-semantic feature knowledge extraction based on reinforcement learning are proposed.Its rule mining and path reasoning technology effectively compensates for the defects of traditional ontology rule definition and reasoning methods.Finally,it combines examples to demonstrate feasibility.The specific research work is as follows:1.Based on STEP AP242 application protocol,a knowledge reason semantic model is constructed to solve the abstractness and inaccuracy of the product semantic information.The formal expressions of the product design intention and other semantic information are given.2.With the knowledge reason semantic model,the knowledge graph of the design intention can be constructed by the application protocol semantic rules and semantic annotations.Based on the mapping of semantic cell to Markov decision model,SWRL rules self-learning based on reinforcement learning is proposed.Reason the design intention depending on serialization decision result of reinforcement learning and relationship of semantic cells.3.Design intent information cannot be fully displayed for the heterogeneous CAD system.Constructing the visualization platform with WebGL technologies.Realize the display of design intent information and reasoning results.