基于多网络模型的工程机械液压系统故障诊断研究

Fault diagnosis of construction machinery hydraulic system based on multi-network model

提出一种针对工程机械液压系统的多网络模型的故障诊断方法。该网络模型以广义回归神经网络(General regression neural network,GRNN)为基础,引入全局递归的反馈机制,构建动态GRNN模型。该方法首先为多个目标故障建立同等数量的动态GRNN目标故障模型,计算每个目标故障模型的检测阈值;然后,将测试故障样本代入每个目标故障模型中,当其残差平方和在对应阈值范围内即可确定故障类型。实验结果表明:多网络模型的故障诊断方法准确地诊断出95%以上的系统故障。

A fault diagnosis approach of construction machinery hydraulic system based on multi-network model was proposed.A dynamic general regression neural network（GRNN） model was established by introducing the global feedback to the GRNN.As a dynamic model with global recursion,dynamic GRNN model is feasible to identify nonlinear system.Firstly,multiple dynamic GRNN model was established for multiple target faults and a test threshold for each dynamic GRNN model was computed.Secondly,the sum of residuals’ square was developed to test model’s residual so as to determine the fault type.The results show that the test faults of 95% are correctly identified.