一种基于状态观测器的大型旋转机械故障诊断研究  

Fault diagnosis of large rotating machinery based on state observer

作　　者：龙昊 侯若玉 LONG Hao;HOU Ruoyu(Hebi College of Vocation and Technology,Hebi 458030,China)

机构地区：[1]鹤壁职业技术学院,鹤壁458030

出　　处：《电子测试》2023年第2期63-66,共4页Electronic Test

摘　　要：针对传统传感器故障在线诊断方法存在的问题,如精度低、效率差等,提出一种基于状态观测器的大型旋转机械故障诊断方法。首先,构建一种开环的状态识别系统,测量真实的输出变量和状态变量并且生成残差;然后,基于状态观测器测出状态变量及一阶导数,并随着状态观测器渐进稳定,缩小误差之和及状态残差值;最后,构建深度残差网络模型,对提取到的不同故障特征进行训练和分类,实现对故障点的准确定位和分类识别。实验结果显示,基于状态观测器的故障诊断方法检测率达到了97.9%,显著优于传统的传感器诊断方法。A state observer based fault diagnosis method for large rotating machinery is proposed to address the problems of low accuracy and poor efficiency in traditional sensor based online fault diagnosis methods.Firstly,construct an open-loop state recognition system,measure the real output variables and state variables,and generate residuals;Then,based on the state observer,measure the state variables and first-order derivatives,and as the state observer gradually stabilizes,reduce the sum of errors and the residual value of the state;Finally,construct a deep residual network model to train and classify the extracted different fault features,achieving accurate localization and classification recognition of fault points.The experimental results show that the detection rate of the fault diagnosis method based on state observer reaches 97.9%,which is significantly better than traditional sensor diagnosis methods.

关 键 词：状态观测器 旋转机械 状态变量 残差网络 故障诊断 传感器 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]