论文部分内容阅读
介绍了一种基于Elman神经网络的通风机故障诊断的诊断原理,学习算法以及技术路线。通过对现场信号特征数据的采集以及归一化处理,对Elman神经网络选取最优的结构与参数,实现了煤矿主通风机故障类型的智能分类与诊断。与传统BP神经网络诊断结果相比较,Elman神经网络综合诊断性能更优。最后通过风机的故障诊断实例表明:Elman神经网络在提高学习速度上有了很大的改进,并且有效地抑制了传统神经网络容易陷于局部极小的缺陷,缩短了自主学习的时间,是风机故障诊断的有效方法。
A diagnostic principle, learning algorithm and technical route of ventilator fault diagnosis based on Elman neural network are introduced. Through the collection and normalization of the signal characteristic data in the field, the optimal structure and parameters of the Elman neural network are selected to realize the intelligent classification and diagnosis of the main ventilator fault types. Compared with the traditional diagnosis results of BP neural network, Elman neural network has better performance in comprehensive diagnosis. Finally, the fan fault diagnosis shows that Elman neural network can improve the learning speed greatly, and effectively restrain the traditional neural network from being trapped in local minima and shortening the time of autonomous learning. Effective method of diagnosis.