基于VMD-KPCA的轴承设备在线监控  

Online Monitoring of Bearing Equipment Based on VMD-KPCA

作　　者：惠兴胜 

机构地区：[1]中国电信股份有限公司烟台分公司

出　　处：《工程机械》2024年第4期107-116,I0017,共11页Construction Machinery and Equipment

摘　　要：针对轴承运行工况复杂且采集的信号受噪声影响较大,难以提取有效的故障特征并准确进行故障监测的情况,提出基于变分模态分解(VMD)和核主元分析(KPCA)融合的方法对轴承设备进行在线监控。该方法利用相关系数表征VMD得到各本征模态函数与原振动信号之间的关联程度,选择实相关的分量进行重构,提取重构信号的时、频域特征输入到KPCA模型中,结合热尔连T平方统计量(T2)和平方预测误差统计量(SPE)对轴承设备进行实时故障监测。根据已有的公开试验轴承振动数据集,通过试验分析发现,提出的方法对轴承设备的监控效果较好,能有效地监测出轴承设备发生的故障。In view of the complexity of the operating conditions of bearings and the significant influence of noise on the collected signals,which makes it difficult to extract effective fault features and accurately monitor faults,a method based on the fusion of variational mode decomposition(VMD)and kernel principal component analysis(KPCA)is proposed for online monitoring of bearing equipment.The method uses correlation coefficients to characterize VMD to obtain the degree of correlation between each intrinsic mode function and the original vibration signal,selects the real correlated components for reconstruction,extracts the time-domain and frequency-domain features of the reconstructed signals and inputs them into the KPCA model,carries out real-time fault monitoring of bearing equipment through the combination of Hotelling T-squared statistic(T 2)and squared prediction error statistic(SPE).Based on the available bearing vibration datasets of the public tests,it is found through the experimental analysis that the proposed method has a good monitoring effect on the bearing equipment and can effectively monitor faults of the bearing equipment.

关 键 词：变分模态分解 核主元分析 统计量 轴承 故障监测 

分 类 号：TH133.3[机械工程—机械制造及自动化]