基于人工神经网络的冲裁件缺陷诊断专家系统  

Artificial Neural Network Based Expert System for Diagnosing of Blanking Parts' Defects

作　　者：汪锐[1] 郑晓丹[1] 何丹农[1] 张立人[2] 黄皆捷[2] 

机构地区：[1]上海交通大学塑性成形工程系,上海200030 [2]上海通用汽车有限公司,上海201206

出　　处：《上海交通大学学报》2001年第7期977-980,共4页Journal of Shanghai Jiaotong University

基　　金：上海市科学技术发展基金资助项目 (9911110 11

摘　　要：说明了冲裁件成形缺陷分析人工神经网络专家系统的总体结构框架 .系统分为综合信息处理模块和缺陷分析推理模块两大部分 ,有针对性地综合提取了冲裁件缺陷信息加以分类 ,采用了双3层 BP神经网络来分析和处理输入参数 ,并对系统知识表示、知识获取、知识库的构造及推理机策略的选用进行了深入的阐述 .同时还对根据系统原理编写的 BDAES(v1 .0 )The paper showed the function frame of artificial neural network based expert system for the analysis of blanking parts' defects. Including the integrated information processing module and defect analysis reasoning module, the system pertinently selects important information on the defects of blanking parts, makes use of two three-layer BP neural networks to analyze and treat with the input parameters. It demonstrates the knowledge representation, knowledge acquisition, knowledge database structure and the selection of reasoning mechanism. The article also introduced the BDAES (v1.0) which was established according to the mechanism of the system.

关 键 词：人工神经网络 专家系统 冲裁件 缺陷诊断 综合信息处理 缺陷分析推理 

分 类 号：TG386.2[金属学及工艺—金属压力加工] TP182[自动化与计算机技术—控制理论与控制工程]