基于多域特征与信息融合的叶片裂纹故障诊断  

Fault diagnosis of blade crack based on multi-domain feature and information fusion

作　　者：马天池 沈君贤 宋狄 许飞云[1] Ma Tianchi;Shen Junxian;Song Di;Xu Feiyun(School of Mechanical Engineering,Southeast University,Nanjing 211189,China)

机构地区：[1]东南大学机械工程学院,南京211189

出　　处：《东南大学学报（自然科学版）》2024年第6期1567-1573,共7页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(51975117);江苏省研究生科研与实践创新计划项目资助项目(SJCX23_0054);国家留学基金委员会资助项目。

摘　　要：针对离心风机叶片裂纹故障诊断问题,提出了一种基于多域特征与信息融合的叶片裂纹故障特征提取方法.首先,在时域、频域和时频域特征的基础上,针对叶片裂纹故障信号的幅值调制特点,采用一系列循环域特征,构建多域特征集.其次,使用Laplacian分数、随机森林、ReliefF算法、互信息和信息增益等多种特征选择方法对多域特征集的所有特征进行评分;然后,提出了改进的Dempster-Shafer证据理论方法,并融合多准则下的特征分数向量,得到敏感特征子集;最后,提出了基于蜉蝣算法优化的核主成分分析方法,充分利用多传感器信息,完成叶片裂纹故障敏感特征的提取,实现叶片的裂纹故障诊断.结果表明:所提方法的平均测试准确率达到99.70%,优于其他对比方法,可用于叶片裂纹的故障诊断.Aiming at the fault diagnosis problem of centrifugal fan blade crack,a blade crack fault feature extraction method based on multi-domain feature and information fusion was proposed.Firstly,on the basis of time-domain,frequency-domain and time-frequency-domain features,a series of cyclic-domain features were adopted for the amplitude-modulation characteristics of blade crack faults to construct a multi-domain feature set.Secondly,all the features in the multi-domain feature set were scored using various feature selection methods such as Laplacian score,random forest,ReliefF algorithm,mutual information,and information gain.Thirdly,the improved Dempster-Shafer evidence theory(DST)method was proposed to obtain a subset of sensitive features by fusing the feature score vectors under multiple criteria.Finally,a kernel principal component analysis(KPCA)method based on the optimized mayfly algorithm was proposed to make full use of the multi-sensor information.The extraction of sensitive features of the blade crack faults was completed and the diagnosis of blade crack faults was realized,The results show that the proposed method has an average testing accuracy of 99.70%,which is higher than those of other comparative methods,and is suitable for the fault diagnosis of blade crack.

关 键 词：叶片裂纹 故障诊断 循环域特征 信息融合 DEMPSTER-SHAFER证据理论 核主成分分析 

分 类 号：TH165[机械工程—机械制造及自动化] TH17