Study On Semantic Knowledge Retrieval And Design Reuse Of MBD Model

Aiming at the two problems existing in traditional 3D retrieval methods,that is,it is difficult to accurately acquire the retrieval intention of designers and that the 3D model retrieved lacks reusable product design knowledge.Based on the design specification of MBD,this paper makes a deep research on the relationship between product 3D model and semantic knowledge.Further analyzed how to construct high-level semantic information from low-level shape retrieval information.Explored the relationship between MBD technology and semantic-based three-dimensional model retrieval technology.Based on the design specification of MBD,an automatic marking method of 3D product model is proposed and validated.The above methods lay a foundation for the research of high-level semantic retrieval and design reuse of MBD model.The main work of this paper is summarized as follows:(1)A Base on feature in layered MBD ontology is constructed.The construction principle of domain ontology is summarized and analyzed.According to this principle,a layered MBD model ontology based on features is constructed.It is based on the ontology layer of geometric structure,takes the common feature ontology as the bridge,and takes the PMI information ontology layer as the top layer of the hierarchical semantic network.Based on STEP intermational standard,the ontology layer of common features builds process-based entity shape modeling elements.On this basis,it extended and modified the requirement of MBD model semantic retrieval to meet the requirement of effectively supporting MBD model retrieval.(2)A mapping mechanism between hierarchical MBD model ontology and high-level semantics is constructed.Firstly,normalizing the common feature ontology layer and PMI information ontology layer in feature-based hierarchical MBD model ontology,defining ontology element and sequence set,introducing weighted Hamming distance,using ant colony algorithm to calculate the shortest Hamiltonian circle,using this as standard sequence,constructing one-to-one correspondence between sequence set and MBD model,and realizing MBD model.The mapping mechanism between type and high-level semantics.(3)An automatic semantic annotation method for MBD model based on feature recognition is proposed.This method combines the mapping mechanism between the layered MBD model ontology and the high-level semantics,and builds the AAG of the part model based on the fusion of key information.The ontology element information is integrated into vertex attributes and the vertex sequence is reconstructed by the adjacency matrix of the attribute adjacency graph.The maximum common subgraph is obtained by dynamic coding and distance matching,and the similarity between MBD models is obtained.Finally,the clustering method is used to realize automatic semantic annotation of MBD model.Based on this method,some small-scale experiments are carried out and good results are achieved.The experimental results show that the proposed method can realize automatic semantic annotation of MBD model and improve the degree of automation of semantic tagging.(4)A multi-mode MBD model reuse prototype system is developed.Based on the above research results,a multi-mode MBD model reuse prototype system is developed for common mechanical parts,which is divided into model-based reuse and semantic-based reuse.The similarity algorithm of MBD model proposed in this paper provides technical support for model-based reuse,while the automatic semantic annotation method of MBD model based on mapping mechanism between hierarchical MBD model ontology and high-level semantics provides possibility for semantic-based reuse function.In this paper,an example is given to verify the effectiveness of the multi-mode MBD model reuse prototype system.