Feature Recognition Based On Graphs And Rules For Rotational Parts

From the development of feature recognition technology to today,a variety of feature recognition methods have gradually formed.Among them,the most used one is the feature recognition method based on boundary matching.Despite this,feature recognition technology still does not meet the urgent needs of today's smart manufacturing field.The complexity of the part model and the particularity of the machining process make the engineering staff not dare to relax in the process of finding the most common feature recognition algorithm.Based on the graph-based research method,this paper combines the feature decomposition and the predefined feature library.Feature matching,identifying typical feature types in a rotary part model.The main research work of the thesis is as follows:(1)Aiming at the problem that the topological information of the attributed adjacency graph of the rotating part is not enough,by defining the characteristics of the feature-face node,the concept of the feature base-face and the feature segmentation and the method of segmentation of the feature boundary are proposed to extend the original attributed adjacency graph.Constructing an extended attributed adjacency graph lays the foundation and providing an important theoretical support for the decomposition of features.(2)Aiming at the fact that the feature decomposition in the previous graph-based feature recognition process is not thorough enough,the traditional feature decomposition method is extended,and the feature decomposition problem of groove features and convex features in rotary parts is studied.Providing a feature decomposition algorithm based on feature base-face.The method can simultaneously support the decomposition of positive and negative features,thereby ensuring the integrity of the processing features during feature decomposition,and avoiding the lack of partial processing features in the decomposition process.(3)In order to realize the extraction of processing features after feature decomposition,a simple graph isomorphism algorithm based on adjacency matrix is proposed.The algorithm marks the node numbers in descending order of the degree sequence of the nodes,and then performs the adjacency matrix matching of the features.In addition,the knowledge library of processing features is created by combining feature heuristic rules,which improves the efficiency of feature matching.(4)Based on the above theoretical research basis,a prototype system for identifying the machining feature of rotary parts was developed on Inte3 D,a three-dimensional process planning platform with independent intellectual property rights based on ACIS and HOOPS libraries.The system realizes the recognition of typical features and partial convex features in the rotary class.In addition,the key content and algorithm of this paper are verified by modeling and analysis engineering examples.