基于人工神经网络的冷挤压工艺设计系统的知识自动获取研究  

ANN-based knowledge auto-acquisition of cold extrusion process planning system

作　　者：高军[1] 赵国群[1] 季廷炜[1] 王广春[1] 

机构地区：[1]山东大学材料科学与工程学院,济南250061

出　　处：《塑性工程学报》2005年第5期28-33,共6页Journal of Plasticity Engineering

基　　金：山东省自然科学基金资助项目(Q2002F01)

摘　　要：提出了用人工神经网络补充和修改冷挤压工艺设计系统知识库内容的模型框架,根据人工神经网络映射的特征编码提出并实现了一种专家系统自动获取冷挤压零件特征的方法,并详细讨论了算法过程,最后结合典型实例给出了知识自动获取的过程。所提出的知识自动获取方法的优点在于不仅可以通过人工神经网络把冷挤压零件的成形过程整体地存入知识库中,而且由于所选用的人工神经网络映射模式是集中反馈式、分工序循环进行的。因此,通过该方法还可以获取冷挤压成形过程中每步工序半成品件的特征,然后存入冷挤压工艺设计系统的知识库。In this paper, a frame model and implementary method, which is used to reinforce and amend the knowledge base of expert system by means of artificial neural networks （ANN）, was brought forward. A new approach to acquire features of cold extrusion part was put forward and completed according to features codes of ANN mapping. The algorithm was discussed, and a typical example was used to show the process of knowledge auto-acquisition. Because the mapping mode of ANN feed back in every working procedure, not only can the whole forming process of the cold extrusion part be acquired and stored in the database, but also the features of each forming procedure.

关 键 词：冷挤压工艺设计 知识自动获取 人工神经网络 专家系统 

分 类 号：TG376[金属学及工艺—金属压力加工]