基于改进MeshCNN的机械零件加工特征识别方法  

Machining feature recognition method for mechanical parts based on improved MeshCNN

作　　者：张恩 张胜文[1] 贾家乐 ZHANG En;ZHANG Shengwen;JIA Jiale(School of Mechanical Engineer,Jiangsu University of Science and Technology,Zhenjiang 212003,CHN)

机构地区：[1]江苏科技大学机械工程学院,江苏镇江212003

出　　处：《制造技术与机床》2023年第7期137-142,共6页Manufacturing Technology & Machine Tool

基　　金：江苏省研究生创新项目(SJCX22_1917)。

摘　　要：为了解决传统CAPP系统中特征识别效率低且智能化程度不高的问题,将原始的MeshCNN与Faster RCNN相结合,提出一种基于Mesh-Faster RCNN的机械零件加工特征识别方法,该方法通过将自定义的加工数据集作为神经网络的输入端,获得最优的神经网络模型;然后利用MBD技术对加工模进行标注,通过PMI的信息标注获取待加工特征,将其转化为三角网格数据;在此基础上结合三角网格数据处理的算法,将处理好的加工特征数据导入最优神经网络模型中,完成特征识别过程。最后以某型号船用柴油机部分关键件为例,验证了所提方法的可行性和有效性。In order to solve the problem of low efficiency and low intelligence of feature recognition in traditional computer-aided process planning(CAPP)system,this paper proposed a novel machining feature recognition method for mechanical parts based on Mesh-Faster Region-based CNN(RCNN)by combining original MeshCNN with Faster RCNN.The method obtained the optimal neural network model by taking the customized processing dataset as the input of the neural network.Then MBD technology was used to label the machining model,and the characteristics to be processed were obtained by PMI information annotation,which were transformed into triangular mesh data.On this basis,combined with the algorithm of triangular mesh data processing,the processed machining feature data is imported into the optimal neural network model to complete the feature recognition process.Finally,the feasibility and effectiveness of the proposed method are verified by taking the key parts of the diesel marine engine as an example.

关 键 词：加工特征识别 Mesh-Faster RCNN 三角网格处理 计算机辅助工艺设计 MBD技术 

分 类 号：TP391.72[自动化与计算机技术—计算机应用技术] TP18[自动化与计算机技术—计算机科学与技术]