流体机械

2021, v.49;No.586(04) 60-65+78

[打印本页] [关闭]
本期目录(Current Issue) | 过刊浏览(Past Issue) | 高级检索(Advanced Search)

基于往复压缩机轴心轨迹特征的故障诊断方法研究
Research on fault diagnosis method based on trajectory features of axial center of reciprocating compressor

张旭东;张进杰;茆志伟;
ZHANG Xudong;ZHANG Jinjie;MAO Zhiwei;Beijing Key Laboratory of Health Monitoring Control and Fault Self-recovery for High-end Machinery Beijing University of Chemical Technology;Compressor Health and Intelligent Monitoring Center of National Key Laboratory of Compressor Technology,Beijing University of Chemical Technology;

摘要(Abstract):

往复压缩机的故障诊断一直以来都是研究工作的重点问题,对其关键运动部件的监测诊断更是研究的热点之一。当往复压缩机发生运动部件磨损、撞缸等故障时,活塞杆的运行状态会发生改变,进而引起活塞杆轴心轨迹的变化。因此,提出了一种基于活塞杆轴心轨迹的往复压缩机智能诊断方法。首先,文章利用改进的离散点轮廓包络方法提取活塞杆轴心轨迹特征,同时提取活塞杆信号的时、频域特征;然后利用ReliefF方法计算轴心轨迹特征与时、频域特征的特征权重,选取大于各自权重均值的特征分别组成轴心轨迹敏感特征集和时频域敏感特征集;最后,提取2个特征集的Related-Similar(RS)特征,将其融合后作为BP神经网络的训练特征集训练故障诊断模型。利用往复压缩机磨损故障和撞缸故障的数据对模型进行验证,结果表明,模型能较好地识别这两种故障,并且故障恶化时的识别准确率均达到了99%以上。
The fault diagnosis of reciprocating compressors has always been the focus of research,and the monitoring and diagnosis of its core moving parts is one of the research hotspots.When the moving parts of the reciprocating compressor are worn out and collide with the cylinder,the operating state of the piston rod will change,which in turn will cause the change of the axial trajectory of the piston rod.Therefore,An intelligent diagnosis method for reciprocating compressors is proposed based on the axial trajectory of the piston rod.Firstly,in the present study,the improved discrete point contour envelope method was used to extract the piston rod axis trajectory features,and extract the piston rod signal time and frequency domain features as well;Secondly,the ReliefF method was used to calculate the feature weights of the axis trajectory feature and the time and frequency domain feature.Furthermore,the feature greater than the average of their respective weight was selected to form the axis trajectory sensitive feature set and the time-frequency domain sensitive feature set;Finally,the Related-Similar(RS) features of the two feature sets were extracted,and the fusion was used as the training feature set of the BP neural network to train the fault diagnosis model.The wear fault and the cylinder collision data of the reciprocating compressor were used to verify the model.The results show that the model could better identify these two faults,and the recognition accuracy rate when the fault deteriorates achieved more than 99%.

关键词(KeyWords): 轴心轨迹;智能诊断;活塞杆;往复压缩机;特征融合
axis trajectory;intelligent diagnosis;piston rod;reciprocating compressor;feature fusion

Abstract:

Keywords:

基金项目(Foundation): 双一流建设专项经费资助项目(ZD1601);; 压缩机技术国家重点实验室(压缩机技术安徽省实验室)开放基金资助项目(SKL-YSJ201808,SKL-YSJ201911)

作者(Authors): 张旭东;张进杰;茆志伟;
ZHANG Xudong;ZHANG Jinjie;MAO Zhiwei;Beijing Key Laboratory of Health Monitoring Control and Fault Self-recovery for High-end Machinery Beijing University of Chemical Technology;Compressor Health and Intelligent Monitoring Center of National Key Laboratory of Compressor Technology,Beijing University of Chemical Technology;

参考文献(References):

扩展功能
本文信息
服务与反馈
本文关键词相关文章
本文作者相关文章
中国知网
分享