Fault Diagnosis of a Bearing using Feature Extraction Method based on CEEMD Algorithm and CNN
-
摘要: 轴承动力学行为具有非线性的特点,导致其振动信号特征与运行状态之间存在较强的非线性关系;且振动信号的特征提取与选择往往需要大量的先验知识,导致特征的设计难以准确反映不同的运行状态。针对以上问题,提出一种基于互补集合经验模态分解(Complementary ensemble empirical mode decomposition,CEEMD)与卷积神经网络(Convolutional neural networks,CNN)的特征提取方法,从振动信号时频图中自适应提取其敏感特征,反映设备运行状态。首先采用CEEMD算法分解得到振动信号的固有模态函数(Intrinsic mode function,IMF)分量,构造各个IMF时频图,并采用CNN提取时频图的特征;然后,将提取到的特征与小波包分频带能量值相结合,组建特征指标向量,用于构建轴承故障诊断模型。将该方法应用于不同负载、不同故障深度的轴承试验中,结果表明该方法能够在多种工况下有效地提高故障识别率。
-
关键词:
- 特征提取 /
- 互补集合经验模态分解 /
- 卷积神经网络 /
- 自适应
Abstract: The dynamic behavior of a bearing is nonlinear, resulting in the strong nonlinear relationship between its vibration signal features and operation state. The feature extraction and selection of vibration signals often require a large amount of prior knowledge, which makes it difficult to accurately reflect different operation states for feature design. In order to solve the above problems, a feature extraction method based on Complementary Ensemble Empirical Mode Decomposition (CEEMD) algorithm and Convolutional Neural Networks (CNN) is proposed to extract sensitive features of self-adaptive vibration signals from their time-frequency diagram. These features can effectively reflect the operating status of the bearing. Firstly, the feature extraction method uses the CEEMD algorithm to obtain the Intrinsic Mode Function (IMF) components of vibration signals and then constructs each IMF time-frequency diagram; it also uses CNN to extract the features of time-frequency diagram. Secondly, it combines the extracted features with the wavelet packet frequency band energy values to form feature index vectors, which are used to build the bearing's fault diagnosis model. The experimental results show that this method can effectively improve the fault diagnosis rate. -
表 1 轴承样本数分布表
故障尺寸/
mils负载 正常 滚动体故障 内圈故障 外圈故障 0 0 300 1 300 2 300 3 300 7 0 100 100 100 1 100 100 100 2 100 100 100 3 100 100 100 14 0 100 100 100 1 100 100 100 2 100 100 100 3 100 100 100 21 0 100 100 100 1 100 100 100 2 100 100 100 3 100 100 100 合计 1 200 1 200 1 200 1 200 -
[1] 李笑梅, 贺德强, 谭文举, 等.基于EEMD和RBFNN的列车滚动轴承故障诊断研究[J].铁道科学与工程学报, 2017, 14(5):1056-1062 doi: 10.3969/j.issn.1672-7029.2017.05.024Li X M, He D Q, Tan W J, et al. Research on fault diagnosis of train rolling bearing based on EEMD and RBFNN[J]. Journal of Railway Science and Engineering, 2017, 14(5):1056-1062(in Chinese) doi: 10.3969/j.issn.1672-7029.2017.05.024 [2] Zhang P J, Neti P. Detection of gearbox bearing defects using electrical signature analysis for doubly fed wind generators[J]. IEEE Transactions on Industry Applications, 2014, 51(3):2195-2200 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=c6075c669895d33601ef4120ada37b9b [3] Qiao W, Lu D G. A survey on wind turbine condition monitoring and fault diagnosis——Part Ⅰ:components and subsystems[J]. IEEE Transactions on Industrial Electronics, 2015, 62(10):6536-6545 doi: 10.1109/TIE.2015.2422112 [4] 杨明玉, 谭书平, 张举.基于数学形态学的暂态行波方向保护方案[J].华北电力大学学报, 2006, 33(6):1-5 doi: 10.3969/j.issn.1007-2691.2006.06.001Yang M Y, Tan S P, Zhang J. Scheme for transient traveling wave-based directional protection using mathematical morphology[J]. Journal of North China Electric Power University, 2006, 33(6):1-5(in Chinese) doi: 10.3969/j.issn.1007-2691.2006.06.001 [5] 陈茂英, 杨明玉, 张新国.基于小波变换的EHV输电线路单端暂态电压保护算法的探讨[J].继电器, 2005, 33(2):17-20 http://d.old.wanfangdata.com.cn/Periodical/jdq200502005Chen M Y, Yang M Y, Zhang X G. Study of a wavelet transform based one-terminal transient voltage signals protection algorithm for EHV transmission line[J]. Relay, 2005, 33(2):17-20(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/jdq200502005 [6] Mallat S. A wavelet tour of signal processing[M]. San Diego:Academic Press, 1998. [7] 钱伟, 王海斌, 杨江, 等.飞机发电机故障诊断的多特征参数组合分析[J].测控技术, 2017, 36(7):47-51, 58 doi: 10.3969/j.issn.1000-8829.2017.07.012Qian W, Wang H B, Yang J, et al. Combination analysis on multi-characteristic parameters of aircraft generator fault diagnosis[J]. Measurement & Control Technology, 2017, 36(7):47-51, 58(in Chinese) doi: 10.3969/j.issn.1000-8829.2017.07.012 [8] 谢光军, 李俭川, 胡茑庆, 等.涡轮泵实时故障检测的多特征参量自适应阈值综合决策算法[J].中国机械工程, 2005, 16(13):1184-1187 doi: 10.3321/j.issn:1004-132X.2005.13.014Xie G J, Li J C, Hu N Q, et al. Multi-feature adaptive threshold compositive decision-making algorithm for real-time fault detection of turbopump[J]. China Mechanical Engineering, 2005, 16(13):1184-1187(in Chinese) doi: 10.3321/j.issn:1004-132X.2005.13.014 [9] 向丹, 岑健.基于EMD熵特征融合的滚动轴承故障诊断方法[J].航空动力学报, 2015, 30(5):1149-1155 http://d.old.wanfangdata.com.cn/Periodical/hkdlxb201505016Xiang D, Cen J. Method of roller bearing fault diagnosis based on feature fusion of EMD entropy[J]. Journal of Aerospace Power, 2015, 30(5):1149-1155(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/hkdlxb201505016 [10] 孙曙光, 于晗, 杜太行, 等.基于多特征融合与改进QPSO-RVM的万能式断路器故障振声诊断方法[J].电工技术学报, 2017, 32(19):107-117 http://d.old.wanfangdata.com.cn/Periodical/dgjsxb201719012Sun S G, Yu H, Du T H, et al. Vibration and acoustic joint fault diagnosis of conventional circuit breaker based on multi-feature fusion and improved QPSO-RVM[J]. Transactions of China Electrotechnical Society, 2017, 32(19):107-117(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/dgjsxb201719012 [11] Fukushima K, Miyake S. Neocognitron:a new algorithm for pattern recognition tolerant of deformations and shifts in position[J]. Pattern Recognition, 1982, 15(6):455-469 doi: 10.1016/0031-3203(82)90024-3 [12] Sun J, Cao W F, Xu Z B, et al. Learning a convolutional neural network for non-uniform motion blur removal[C]//Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA: IEEE, 2015: 769-777 [13] Yeh J R, Shieh J S, Huang N E. Complementary ensemble empirical mode decomposition:a novel noise enhanced data analysis method[J]. Advances in Adaptive Data Analysis, 2010, 2(2):135-156 doi: 10.1142/S1793536910000422 [14] 孙灵芳, 徐曼菲, 朴亨, 等.基于改进CEEMD的超声检测信号自适应降噪[J].振动与冲击, 2017, 36(20):225-232 http://d.old.wanfangdata.com.cn/Periodical/zdycj201720034Sun L F, Xu M F, Piao H, et al. Self-adaptive noise denoising for ultrasonic detection signal based on improved CEEMD[J]. Journal of Vibration and Shock, 2017, 36(20):225-232(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zdycj201720034 [15] Wu Z H, Huang N E. Ensemble empirical mode Decomposition:a noise-assisted data analysis method[J]. Advances in Adaptive Data Analysis, 2009, 1(1):1-41 http://d.old.wanfangdata.com.cn/Periodical/dianzixb201305033 [16] 吴小涛, 杨锰, 袁晓辉, 等.基于峭度准则EEMD及改进形态滤波方法的轴承故障诊断[J].振动与冲击, 2015, 34(2):38-44 http://d.old.wanfangdata.com.cn/Periodical/zdycj201502008Wu X T, Yang M, Yuan X H, et al. Bearing fault diagnosis using EEMD and improved morphological filtering method based on kurtosis criterion[J]. Journal of Vibration and Shock, 2015, 34(2):38-44(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zdycj201502008 [17] 黄刚劲, 范玉刚, 黄国勇.CEEMD与广义形态差值滤波结合的故障诊断方法研究[J].华中师范大学学报:自然科学版, 2017, 51(3):304-308, 316 http://d.old.wanfangdata.com.cn/Periodical/hzsfdxxb201703007Huang G J, Fan Y G, Huang G Y. Research on fault diagnosis method combining CEEMD with generalized morphology difference filter[J]. Journal of Central China Normal University:Natural Sciences, 2017, 51(3):304-308, 316(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/hzsfdxxb201703007 [18] Bouvrie J. Notes on convolutional neural networks[R/OL]. Cambridge: Massachusetts Institute of Technology, 2006[2016-09-15]. http://cogprints.org/5869/1/cnn_tutorial.pdf [19] Case Western Reserve University Bearing Data Center. Bearing data center fault test data[EB/OL]. http://www.eecs.case.edu/laboratory/bearing