重型减速器在线监测与故障诊断研究  

Research on Online Monitoring and Fault Diagnosis of Heavy-duty Reducer

作　　者：余洋 陈云 孔进 Yu Yang;Chen Yun;Kong Jin(Ningxia Tiandi Benniu Industrial Group Co.,Ltd.,Yinchuan 750001,China)

机构地区：[1]宁夏天地奔牛实业集团有限公司,银川750001

出　　处：《煤矿机械》2024年第10期187-191,共5页Coal Mine Machinery

摘　　要：针对重型减速器故障诊断方法中故障特征不显著、易受外界干扰等问题,提出一种基于多维特征选择和BP神经网络的故障诊断方法。首先设计并布置重型减速器故障模拟实验台,使用多源传感信息采集平台进行信号采集;然后基于采集的振动信号,提出双树复小波变换和多维特征选择的轴承故障诊断方法作为底层诊断逻辑,并搭建BP神经网络模型进行故障识别。结果表明,该方法的诊断准确率为94%,能够满足在实际工况中对重型减速器的故障诊断需求。A fault diagnosis method based on multidimensional feature selection and BP neural network was proposed to address the issues of insignificant fault characteristics and susceptibility to external interference in the diagnosis of heavy-duty reducer faults.Firstly,a simulation test bench for heavyduty reducer faults was designed and deployed,and a multi-source sensor information acquisition platform was used for signal collection.Then,based on the collected vibration signals,a fault diagnosis method for bearing faults using dual-tree complex wavelet transform and multidimensional feature selection was proposed as the underlying diagnostic logic,and a BP neural network model was constructed for fault identification.The results show that this method achieves a diagnostic accuracy of 94%,which meets the requirements for fault diagnosis of heavy-duty reducer under practical operating conditions.

关 键 词：减速器 故障诊断 双树复小波变换 BP神经网络 

分 类 号：TH132.46[机械工程—机械制造及自动化] TP277[自动化与计算机技术—检测技术与自动化装置]