Abstract:

：The symmetry information of model is necessary for product innovation design and engineering analysis. However, the existing symmetry detection research for mechanical assembly model is not enough. Therefore, an approach to detecting intra-part symmetry, inter-part symmetry and global symmetry in mechanical assembly model is proposed. Firstly, the feature-manager design tree is traversed to get all parts, which are classified into part sets according to duplicate relations and part surface attributes, for an assembly model. Then the part sets are sorted on the basis of their part numbers to improve symmetry combining efficiency. Next, the feature information within the model is made use of to detect intra-part symmetry and acquire rotational axis and reflectional plane of symmetry. Then the rotation, reflection and translation symmetry among parts can be detected according to centroid and bounding box of the parts. Finally, the global symmetry of an assembly model could be derived after symmetry merging inside and among the part sets. Experimental results show that our method is capable of detecting the intra-part symmetry, the inter-part symmetry and the global symmetry of assembly model accurately.

Key words: global symmetry, inter-part symmetry, intra-part symmetry, part sets classification, symmetry merging, assembly model