基于Elman神经网络的液压泵故障诊断模型研究

针对传统BP网路进行液压泵故障诊断时，网络学习具有收敛速度慢和学习、记忆不稳定的缺陷，提出了将Elman神经网络应用于液压泵故障诊断的新方法，建立了Elman神经网络的应用结构模型，介绍了该网络的训练算法，阐明了液压泵故障诊断的实现过程。通过试验验证了该神经网络收敛速度快，学习记忆稳定，具有很好的学习功能；测试结果表明该诊断方法具有高可靠性，达到了预期的效果，可以用于液压泵故障诊断。

Research on Hydraulic Pump Fault Diagnosis Model Based on Elman Neural Net

Aimed at the disadvantages of low net convergence rate and low stability of learning and memory for the hydraulic pump fault diagnosis by traditional BP neural net. They are the Elman neural net was presented for the hydraulic pump fault diagnosis.Elman neural net structure model was established and the training algorithm and the process of fault diagnosis were described.The better learning function,higher net convergence rate and stability of learning and memory were validated in the new net by test. The diagnosis results indicate the reliability of this method and the expected results are attained, it can be applied to hydraulic pump fault diagnosis.