Rolling bearing fault diagnosis based on data-level and feature-level information fusion  

基于数据级和特征级信息融合的滚动轴承故障诊断

作　　者：Shu Yongdong Ma Tianchi Lin Yonggang 舒永东;马天池;林勇刚(浙江大学流体动力与机电系统国家重点实验室,杭州310027;东南大学机械工程学院,南京211189)

机构地区：[1]State Key Laboratory of Fluid Power&Mechatronic Systems,Zhejiang University,Hangzhou 310027,China [2]School of Mechanical Engineering,Southeast University,Nanjing 211189,China

出　　处：《Journal of Southeast University(English Edition)》2024年第4期396-402,共7页东南大学学报（英文版）

基　　金：The National Natural Science Foundation of China(No.U22A20178);National Key Research and Development Program of China(No.2022YFB3404800);Jiangsu Province Science and Technology Achievement Transformation Special Fund Program(No.BA2023019).

摘　　要：To address the limitation of single acceleration sensor signals in effectively reflecting the health status of rolling bearings,a rolling bearing fault diagnosis method based on the fusion of data-level and feature-level information was proposed.First,according to the impact characteristics of rolling bearing faults,correlation kurtosis rules were designed to guide the weight distribution of multi-sensor signals.These rules were then combined with a weighted fusion method to obtain high-quality data-level fusion signals.Subsequently,a feature-fusion convolutional neural network(FFCNN)that merges the one-dimensional(1D)features extracted from the fused signal with the two-dimensional(2D)features extracted from the wavelet time-frequency spectrum was designed to obtain a comprehensive representation of the health status of rolling bearings.Finally,the fused features were fed into a Softmax classifier to complete the fault diagnosis.The results show that the proposed method exhibits an average test accuracy of over 99.00%on the two rolling bearing fault datasets,outperforming other comparison methods.Thus,the method can be effectively utilized for diagnosing rolling bearing faults.针对单一加速度传感器信号难以充分反映滚动轴承健康状态的问题,提出了一种基于数据级和特征级信息融合的滚动轴承故障诊断方法.首先,根据滚动轴承故障的冲击特性,设计了相关峭度规则来指导多传感器信号的权重分配,结合加权融合方法获得高质量的数据级融合信号;随后,设计了一个特征融合卷积神经网络(FFCNN),对从融合信号中提取的一维(1D)特征和从小波时频谱中提取的二维(2D)特征进行融合,获得滚动轴承健康状态的充分表征;最后,将融合后的特征输入Softmax分类器,完成故障诊断.结果表明,所提方法在2个滚动轴承故障数据集上平均测试准确率均高于99.00%,优于其他对比方法,可用于滚动轴承的故障诊断.

关 键 词：fault diagnosis information fusion correlation kurtosis feature-fusion convolutional neural network 

分 类 号：TH165[机械工程—机械制造及自动化] TH17