【摘要】 针对航空发动机中盘类零件的特征自动识别问题，提出一种基于图和规则的盘类零件车削特征自动识别方法。根据盘类零件的结构特点，对盘类零件的车削特征进行分类；通过边、点以及点的凹凸性来描述各部分之间的逻辑关系，构建盘类零件的属性邻接图，根据轮廓与端面的邻接关系对属性邻接图进行初步分解；使用启发式规则和轮廓重构规则，得到槽特征和孔特征的特征子图，与预定义特征库进行子图匹配，完成车削特征的识别；最后，在QJCAM平台中完成车削特征识别模块的开发，并通过零件实例验证方法的有效性。结果表明：所提方法识别出盘类零件11个车削特征，其中包含5个槽特征和1个孔特征，总用时1 s;应用传统子图同构的方法识别该盘类零件用时1.5 s,可见所提方法通过启发式规则筛选特征子图，减少了特征匹配的计算量，有效提高了特征识别的效率。更多还原

【Abstract】 A graph and rule-based method for automatic recognition of turning features of disk-type parts in aero-engine was proposed to solve the problem of automatic feature recognition of disk-type parts in aero engines.According to the structural characteristics of the disk-type parts, the turning features of the disk-type parts were classified.The logical relationships between different parts were described using edges, points and the concavity/convexity of points, an attribute adjacency graph was constructed for disk-type parts, and the attribute adjacency graph was preliminarily decomposed according to the adjacency relationship between the profile and the end face.Then, heuristic rules and contour reconstruction rules were used to obtain feature subgraphs of groove features and hole features, and the subgraphs were matched with the predefined feature database to complete the identification of turning features.Finally, the development of the turning feature recognition module was completed in the QJCAM platform, and the effectiveness of the method was verified by part examples.The results show that the proposed method identifies 11 turning features of the disk-type part, including 5 groove features and 1 hole feature, with a total time of 1 s.In contrast, using the traditional subgraph isomorphism method to identify the same disk-type part takes 1.5 s.This indicates that the proposed method reduces the computational load of feature matching by screening feature subgraphs through heuristic rules, thereby effectively improving the efficiency of feature recognition.更多还原