Fault Diagnosis Method of Rolling Bearing Combining VMD with t-SNE
-
摘要: 滚动轴承故障诊断普遍采用有监督学习的方式，针对有标签数据难以获取的问题，提出一种VMD分解与t-SNE流形学习相结合的滚动轴承故障诊断方法。利用VMD分解将滚动轴承原始振动信号分解为若干本征模态分量（IMF）；计算每个模态分量的时频特性指标组成高维故障特征，通过t-SNE对故障进行二次特征提取，获取低维敏感特征并将其作为K-means分类器的输入，实现故障类型的识别。将该方法应用到滚动轴承故障诊断中并与VMD + PCA、原始时频特征+ t-SNE两种方法进行对比，结果表明VMD + t-SNE方法以无监督学习的方式实现了故障诊断的去标签化和自适应性，同时提高了故障诊断的准确性。Abstract: Aiming at the problem that the supervised learning is commonly adopted in fault diagnosis for rolling bearing, while the labeled data are often difficult to obtain, a fault diagnosis method of rolling bearing combining VMD (Variational Mode Decomposition) with t-SNE (t-distributed Stochastic Neighbor Embedding) is proposed. The original vibration signal of the rolling bearing is decomposed into the several intrinsic mode functions (IMF) by VMD decomposition; the time-frequency characteristics of each IMF is calculated to form high-dimensional fault features, and the secondary feature extraction is performed by t-SNE to form low-dimensional sensitive features; low-dimensional sensitive features are input to the K-means classifier for fault type identification. The method is applied to the fault diagnosis of rolling bearing and compared with VMD + PCA, original time-frequency features + t-SNE. The results show that the VMD + t-SNE method realizes the de-labeling and adaptability of fault diagnosis in the form of unsupervised learning, while improving the accuracy of fault diagnosis.
-
Key words:
- VMD /
- t-SNE /
- rolling bearing /
- feature extraction /
- fault diagnosis
-
表 1 滚动轴承信号原始特征向量
特征维数 时频特性 特征维数 时频特性 1 最大值 11 中心频率 2 均值 12 频率方差 3 峰峰值 13 E30/E 4 均方根 14 E31/E 5 波形因子 15 E32/E 6 峰值因子 16 E33/E 7 裕度因子 17 E34/E 8 脉冲因子 18 E35/E 9 峭度因子 19 E36/E 10 歪度因子 20 E37/E 
下载: 导出CSV
表 2 不同K值下IMF分量中心频率
K 中心频率/Hz 2 1 624 3 820 3 1 600 2 760 3 831 4 1 506 1 774 3 315 3 848 5 1 554 1 674 2 756 3 758 3 895 6 1 546 1 664 2 740 3 376 3 804 3 929
下载: 导出CSV
表 3 不同方法分类结果对比
诊断方法 DBI 准确率 原始特征+t-SNE 0.374 93.75% VMD+PCA 0.309 96.25% VMD+t-SNE 0.157 99.37%
下载: 导出CSV
-
[1] 阳建宏, 黎敏, 丁福焰, 等.滚动轴承诊断现场实用技术[M].北京:机械工业出版社, 2015Yang J H, Li M, Ding F Y, et al. Rolling bearing diagnostic field practical technology[M]. Beijing:Mechanical Industry Press, 2015(in Chinese) [2] 徐乐, 邢邦圣, 郎超男, 等.LMD能量熵和SVM相结合的滚动轴承故障诊断[J].机械科学与技术, 2017, 36(6):915-918 doi: 10.13433/j.cnki.1003-8728.2017.0615Xu L, Xing B S, Lang C N, et al. Fault diagnosis of rolling bearing combined LMD energy entropy and SVM[J]. Mechanical Science and Technology for Aerospace Engineering, 2017, 36(6):915-918(in Chinese) doi: 10.13433/j.cnki.1003-8728.2017.0615 [3] 钟先友, 曾良才, 赵春华, 等.基于CITD和同态滤波解调的滚动轴承故障诊断方法[J].机械科学与技术, 2014, 33(1):37-42 https://journals.nwpu.edu.cn/jxkxyjs/article/id/5070Zhong X Y, Zeng L C, Zhao C H, et al. Rolling element bearings fault diagnosis method based on CITD and homomorphic filtering demodulation[J]. Mechanical Science and Technology for Aerospace Engineering, 2014, 33(1):37-42(in Chinese) https://journals.nwpu.edu.cn/jxkxyjs/article/id/5070 [4] 林旭泽, 蔡艳平, 王新军.自适应的EEMD及其在滚动轴承故障诊断中的应用[J].机械科学与技术, 2016, 35(11):1727-1732 doi: 10.13433/j.cnki.1003-8728.2016.1115Lin X Z, Cai Y P, Wang X J. Adaptive EEMD and its applications to rolling bearing fault diagnosis[J]. Mechanical Science and Technology for Aerospace Engineering, 2016, 35(11):1727-1732(in Chinese) doi: 10.13433/j.cnki.1003-8728.2016.1115 [5] 刘嘉敏, 彭玲, 刘军委, 等.遗传算法VMD参数优化与小波阈值轴承振动信号去噪分析[J].机械科学与技术, 2017, 36(11):1695-1700 doi: 10.13433/j.cnki.1003-8728.2017.1110Liu J M, Peng L, Liu J W, et al. Denoising analysis of bearing vibration signal based on genetic algorithm and wavelet threshold VMD[J]. Mechanical Science and Technology for Aerospace Engineering, 2017, 36(11):1695-1700(in Chinese) doi: 10.13433/j.cnki.1003-8728.2017.1110 [6] 唐贵基, 王晓龙.参数优化变分模态分解方法在滚动轴承早期故障诊断中的应用[J].西安交通大学学报, 2015, 49(5):73-81 http://d.old.wanfangdata.com.cn/Periodical/xajtdxxb201505012Tang G J, Wang X L. Parameter optimized variational mode decomposition method with application to incipient fault diagnosis of rolling bearing[J]. Journal of Xi'an Jiaotong University, 2015, 49(5):73-81(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/xajtdxxb201505012 [7] 李奕江, 张金萍, 李允公.基于VMD-HMM的滚动轴承磨损状态识别[J].振动与冲击, 2018, 37(21):61-67 http://d.old.wanfangdata.com.cn/Periodical/zdycj201821011Li Y J, Zhang J P, Li Y G. Wear state recognition of rolling bearings based on VMD-HMM[J]. Journal of Vibration and Shock, 2018, 37(21):61-67(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zdycj201821011 [8] 王新, 闫文源.基于变分模态分解和SVM的滚动轴承故障诊断[J].振动与冲击, 2017, 36(18):252-256 http://d.old.wanfangdata.com.cn/Periodical/zdycj201718037Wang X, Yan W Y. Fault diagnosis of roller bearings based on the variational mode decomposition and SVM[J]. Journal of Vibration and Shock, 2017, 36(18):252-256(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zdycj201718037 [9] 程静, 王维庆, 樊小朝, 等.基于二值双谱和模糊聚类的风电轴承故障诊断[J].振动、测试与诊断, 2018, 38(4):765-771 http://d.old.wanfangdata.com.cn/Periodical/zdcsyzd201804017Cheng J, Wang W Q, Fan X C, et al. Bearing fault pattern recognition of wind turbine based on two-value bispectrum feature-fuzzy clustering method[J]. Journal of Vibration, Measurement & Diagnosis, 2018, 38(4):765-771(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zdcsyzd201804017 [10] 梁海英, 许昕, 潘宏侠, 等.基于EEMD与模糊核聚类的供输弹系统早期故障识别[J].机械设计与研究, 2019, 35(1):192-195 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jxsjyyj201901042Liang H Y, Xu X, Pan H X, et al. Early fault diagnosis based on EEMD-KFCM for supply and delivery system[J]. Machine Design & Research, 2019, 35(1):192-195(in Chinese) http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jxsjyyj201901042 [11] 张勇, 钟建伟, 周文辉, 等.基于K-means聚类分析的变压器故障诊断[J].湖北民族学院学报(自然科学版), 2018, 36(4):452-455 http://d.old.wanfangdata.com.cn/Periodical/hbmzxyxb-zrkx201804020Zhang Y, Zhong J W, Zhou W H, et al. Transformer fault diagnosis based on K-means clustering analysis[J]. Journal of Hubei University for Nationalities (Natural Science Edition), 2018, 36(4):452-455(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/hbmzxyxb-zrkx201804020 [12] 戚晓利, 叶绪丹, 蔡江林, 等.基于变分模态分解与流形学习的滚动轴承故障特征提取方法[J].振动与冲击, 2018, 37(23):133-140 http://d.old.wanfangdata.com.cn/Periodical/zdycj201823020Qi X L, Ye X D, Cai J L, et al. Fault feature extraction method of rolling bearings based on VMD and manifold learning[J]. Journal of Vibration and Shock, 2018, 37(23):133-140(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zdycj201823020 [13] Gisbrecht A, Mokbel B, Hammer B. Linear basis- function t-SNE for fast nonlinear dimensionality reduction[C]//Proceeding of the 2012 International Joint Conference on Neural Networks (IJCNN). Brisbane, QLD: IEEE, 2012: 1-8 https://www.researchgate.net/publication/261087350_Linear_basis-function_t-SNE_for_fast_nonlinear_dimensionality_reduction [14] Dragomiretskiy K, Zosso D. Variational mode decomposi-tion[J]. IEEE Transactions on Signal Processing, 2014, 62(3):531-544 doi: 10.1109/TSP.2013.2288675 [15] Hinton G E, Roweis S T. Stochastic neighbor embedd- ing[C]//Proceedings of the 15th International Conference on Neural Information Processing Systems. Vancouver, British Columbia, Canada: MIT Press, 2002: 833-840 [16] The Case Western Reserve University Bearing Data Center Bearing data center seeded fault test data[EB/OL].[2016-06-10]. http://csegroups.case.edu/bearingdatacenter/pages/download-data-file -
点击查看大图
图(8) / 表(3)
计量
- 文章访问数: 303
- HTML全文浏览量: 163
- PDF下载量: 35
- 被引次数: 0
