基于CEEMDAN和排列熵的滚动轴承故障诊断方法  

Fault Diagnosis Method for Rolling Bearings Based on CEEMDAN and Permutation Entropy

作　　者：白丽丽 韩振南[1] 任家骏[1] 秦晓峰[1] BAI Lili;HAN Zhennan;REN Jiajun;QIN Xiaofeng(College of Mechanical Engineering,Taiyuan University of Technology,Taiyuan 030024,China)

机构地区：[1]太原理工大学机械工程学院

出　　处：《轴承》2019年第11期54-59,共6页Bearing

基　　金：国家自然科学基金项目(50775157,51805355);山西省基础研究项目(2012011012-1)

摘　　要：针对滚动轴承振动信号非线性和非平稳性等特征影响故障类型及严重程度识别准确性的问题,提出了一种基于完整的自适应噪声集成经验模态分解和排列熵的故障特征提取方法,并结合支持向量机自动分类识别的功能形成了一种高效准确的智能故障诊断方法。首先,将振动信号分解为一系列的本征模态函数;然后,计算前n个IMF的排列熵值并形成一个多尺度的特征矩阵;最后,通过粒子群寻优的支持向量机对该特征矩阵进行模式识别,诊断出轴承振动信号所对应的故障类型及故障严重程度。试验数据分析表明,该方法能有效识别滚动轴承振动信号中隐藏的故障状态,准确率高且稳定性好。The vibration signal of rolling bearings often presents non-linear and non-stationary characteristics,which affects accuracy of identifying fault types and severity.A fault feature extraction method is proposed based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)and Permutation Entropy(PE),and combined with automatic classification and recognition function of SVM(Support Vector Machine)to form an efficient and accurate intelligent fault diagnosis method.Firstly,the vibration signal is decomposed into a series of intrinsic mode functions(IMFs).Then,the PE values of the first n IMFs are calculated to generate multi-scale characteristic matrix.Finally,the mode recognition for feature matrix is carried out by SVM optimized by particle swarm(PSOSVM)to diagnose fault type and fault severity corresponding to vibration signal of the bearings.The experimental data analysis show that the fault state implied in vibration signal of rolling bearings is recognized effectively by using the method with high accuracy and good stability.

关 键 词：滚动轴承 故障诊断 CEEMDAN 排列熵 

分 类 号：TH133.33[机械工程—机械制造及自动化] TH17