摘要
提出一种基于自组织特征映射神经网络(SOFM)的零件加工尺寸类型识别方法.首先从三维CAD软件中获取包含零件特征数据的XML文件,并从文件中提取相应的加工特征及尺寸.然后以零件加工特征作为SOFM的输入层的神经元,经处理后作为SOFM的输入向量,利用SOFM自学习和自组织能力对输入向量进行训练.训练好的网络可以实现对零件加工尺寸类型进行较好的识别.最后通过对某零件的尺寸类型识别,验证了所提方法对平面、内孔、外圆和定位四类典型加工尺寸类型识别的有效性.
An automatic machining dimension type recognition method based on self-organizing feature map （SOFM） neural network was presented. Firstly, the extensible markup language （XML） file including data of component features was acquired from 3D CAD software, and the machining features and dimensions were extracted from the XML files. The machining features of component were treated as neurons in the input layer of SOFM network, which would be considered as input vectors of SODM after processing. By employing the self-studying and self-organizing capability of SOFM, the input vectors were trained and the trained network could implement well recognition of machining dimension type. Finally, though recognition of dimension type of some component, the validity of the proposed method for four typical machining dimension types, namely plane, inner hole, outer circle and orientation, were demonstrated.
出处
《工程设计学报》
CSCD
北大核心
2008年第5期341-346,共6页
Chinese Journal of Engineering Design
基金
2005年度教育部科学技术研究重点项目(205122)
广西自然科学基金资助项目(桂科自0640166)
关键词
自组织特征映射
加工尺寸类型
可扩展标记语言
self-organizing feature map
machining dimension type
extensihle markup language
