基于EMD-样本熵的矿业风机轴承故障特征提取  

Extraction of Mining Wind Turbine Bearing Fault Characteristics Based on EMD-Sample Entropy

作　　者：孙猛 张军[1] 王磊[2] Sun Meng;Zhang Jun;Wang lei(School of Artificial Intelligence;School of Mechatronics Engineering,Anhui University of Science and Technology,Huainan,Anhui 232001,China)

机构地区：[1]安徽理工大学人工智能学院,安徽淮南232001 [2]安徽理工大学机械工程学院,安徽淮南232001

出　　处：《黑龙江工业学院学报(综合版)》2025年第1期117-122,共6页Journal of Heilongjiang University of Technology(Comprehensive Edition)

基　　金：潞安化工集团科技项目(项目编号:23A8103107C)。

摘　　要：针对煤矿通风机工作特殊造成采集到的信号成分复杂的现象,采用小波阈值降噪、经验模态分解(EMD)和样本熵的方法对风机轴承振动信号进行故障特征提取。先对采集到的不同损伤类型的滚动轴承振动信号进行小波阈值降噪处理,通过分析降噪前后的信噪比和均方根误差选出降噪效率最高的小波函数;再对降噪信号进行EMD分解,并对分解后得到的各IMF分量计算其样本熵。通过分析表明,不同工况滚动轴承的样本熵特征向量具有明显差别,验证了该方法用于轴承信号故障特征提取的可行性。Given the complexity of the signal components collected due to the unique working conditions of coal mine ventilation fans,methods such as wavelet threshold denoising,Empirical Mode Decomposition(EMD)and sample entropy are employed for the extraction of fault characteristics from bearing vibration signals.Firstly,conduct wavelet threshold denoising on the vibration signals of rolling bearings with various types of damages collected.By analyzing the signal-to-noise ratio and root mean square error before and after denoising,select the wavelet function that achieves the highest denoising efficiency.Then perform Empirical Mode Decomposition(EMD)on the noise-reduced signal,and calculate the sample entropy of each Intrinsic Mode Function(IMF)component obtained after decomposition.Analysis indicates that the sample entropy characteristic vectors of rolling bearings under different operating conditions exhibit significant distinctions,thus validating the feasibility of this method for extracting fault features from bearing signals.

关 键 词：风机轴承 小波阈值降噪 EMD 样本熵 

分 类 号：TH17[机械工程—机械制造及自动化]