摘要
本文介绍了开发钢铁生产主线板坯缺陷的诊断系统的必要性 ;并提出了一种基于神经网络的板坯缺陷诊断及预报算法。利用具有不同激励函数的分组方法提高网络性能 。
This paper introduces the importance of developing the diagnostic system in steel slab production. It proposes an algorithm for diagnosis and prediction based on artificial neural network. The network performance is much enhanced by using the method with different stimulating functions. The algorithm of random grading can efficiently avoid falling into local minimums.
出处
《微型电脑应用》
2001年第1期34-35,22,共3页
Microcomputer Applications
基金
国家教育部博士点基金项目
关键词
神经网络
钢铁板坯
板坯连铸
BP算法
缺陷诊断
re-judgement artificial neural network BP algorithm random grade algorithm
