摘要
针对传统的前馈神经网络误差反传学习算法具有容易陷入局部极小值,收敛速度慢以及容易导致震荡等缺点,提出了一种基于Elman神经网络的故障诊断方法。介绍了该网络的训练算法,阐明了故障诊断的系统结构及诊断步骤。在实验研究部分,根据波焊炉的现场条件及运行经验,确定了波焊炉的征兆集和故障集,详细介绍了数据处理方法及故障诊断的实现过程。MATLAB仿真实验验证了该神经网络收敛速度快、学习记忆稳定,可用于波焊炉故障诊断。
Classic neural networks mostly train the network with error back propagation algorithm. But the back propagation algorithm often gets into the minimum value and its convergence speed is slow, Aiming at the limitation of traditional fault diagnosis technology, this paper presents a new method based on Elman neural network for fault diagnosis. In the experiment research part, a symptom set and a fault set have been established according to wave soldering machine site condition and operational experience. The simulation results show that the fault diagnosis based on the Elman neural network is accurate, which has better learning function, higher net convergence speed and stability of memory. The Matlab simulation shows that it is feasible and can be applied to wave soldering machine fault diagnosis.
出处
《太原理工大学学报》
CAS
北大核心
2008年第3期277-280,共4页
Journal of Taiyuan University of Technology
基金
山西省自然科学基金资助项目(20051037)
太原理工大学校科技发展基金资助项目(190-12901244)
关键词
ELMAN神经网络
波焊炉
故障诊断
Elman neural network
wave soldering machine
fault diagnosis
