Big Data-driven Statistic Trend Analysis and PCA for Incipient Fault Diagnosis of Rolling Bearings
-
摘要: 针对短时、小样本数据下提取的特征对早期故障敏感度和故障演化过程稳定度低、信息冗余的问题，提出了大数据统计趋势分析和核主元分析方法（Principal component and analysis，PCA）的滚动轴承故障演化特征提取和早期故障诊断方法。采集滚动轴承正常状态到完全失效状态的全寿命振动数据，计算原始数据中不同故障严重程度下的时频统计特征，建立各个统计特征描述的故障演化趋势，分析各个统计特征描述的故障演化特性，初步选择能够敏感且稳定感知故障演化过程的统计特征集，利用PCA分析初选结果中各个统计特征间的相关性和贡献度，进一步剔除冗余特征，最终得到能全面表征故障演化过程的特征。最后，使用滚动轴承全寿命振动数据验证本文所提方法的有效性。实验结果证明，标准差、均值频率、标准差频率等特征能敏感地检测滚动轴承早期内环故障并稳定跟踪其演化过程。Abstract: Aiming at those problems of weak sensitivity for incipient fault, low stability for fault evolution process and information redundancy with short time and small sample, a novel feature extraction and incipient fault diagnosis method for rolling bearings is proposed based on big data-driven statistic trend analysis for fault evolution process and principal component and analysis (PCA) in this paper. A life-time data of rolling bearings in whole lifecycle are collected by various sensors, then statistic characteristics of original vibration data at each echo are calculated, thus those fault evolution trend curves described by all statistic characteristics would be built. Based on fault evolution trend curves, the stability and sensitivity of each feature for fault evolution trends can be calculated, thus the primary features can be selected. The relevancy and contribution of each feature in the primary features by PCA was obtained, thus redundancy feature would be further removed, and those features which can fully reflect fault evolution process will be furthermore choose. Finally, the life-time data of rolling bearings were used to verify the effectiveness of proposed methods. The results have shown that the standard deviation, mean frequency, standard deviation frequency can more sensitively detect incipient inner race of rolling bearings and comprehensive track their evolution process.
-
编号 公式 F18 
F19 
F20 
F21 
F22 
F23 
F24 
F25 
F26 
F27 
F28 
F29 
F30 
注:S(k)为原始振动信号x(n)的频谱; NFT为谱线数; fk为第k个谱线对应的频率值。 
下载: 导出CSV
表 2 5次全寿命实验结果
试验次数 故障位置 故障类型 停机后采集的文件数 1# B 内环故障 530 2# B 外环故障 642 3# A 内环故障 598 4# B 内环故障 511 5# C 外环故障 489
下载: 导出CSV
表 3 30个时频统计特征对故障演化过程的灵敏度和稳定度
统计特征 β τ F1 0.59 0.76 F2 0.58 0.77 F3 0.66 0.76 F4 0.60 0.76 F5 0 0.77 F6 0 0.77 F7 0.44 0.77 F8 0.76 0.76 F9 3.36 0.77 F10 0.78 0.76 F11 0.12 0.77 F12 22.1 0.72 F13 5.00 0.65 F14 2.89 0.67 F15 1.67 0.67 F16 0.29 0.26 F17 0.07 0.33 F18 28.78 0.78 F19 16.29 0.77 F20 0.04 0.30 F21 0 0.29 F22 0 0.72 F23 0 0.78 F24 0 0.70 F25 0 0.56 F26 3.87 0.73 F27 2.82 0.77 F28 0.08 0.21 F29 0.01 0.21 F30 3.91 0.78
下载: 导出CSV
表 4 5次试验的早期故障诊断结果对比
方法 1#试验 2#试验 3#试验 4#试验 5#试验 故障类型 早期故障点 故障类型 早期故障点 故障类型 早期故障点 故障类型 早期故障点 故障类型 早期故障点 基本尺度熵[3] 内环故障 401 外环故障 487 内环故障 424 内环故障 398 外环故障 374 绝对均值[6] 内环故障 423 外环故障 501 外环故障 437 内环故障 410 外环故障 385 归一化均方根[16] 外环故障 380 外环故障 465 内环故障 488 内环故障 379 外环故障 362 Lempel-Ziv[17] 内环故障 398 外环故障 511 内环故障 503 内环故障 384 外环故障 377 能量比[18] 外环故障 465 外环故障 567 内环故障 521 内环故障 423 外环故障 399 时频统计特征 内环故障 372 外环故障 449 内环故障 403 外环故障 357 外环故障 322
下载: 导出CSV
-
[1] 燕晨耀.基于多特征量的滚动轴承退化状态评估和剩余寿命预测方法研究[D].成都: 电子科技大学, 2016 http://cdmd.cnki.com.cn/Article/CDMD-10614-1016169916.htmYan C Y. Degradation assessment and residual life prediction of rolling bearings based on multiple features[D]. Chengdu: University of Electronic Science and Technology of China, 2016(in Chinese) http://cdmd.cnki.com.cn/Article/CDMD-10614-1016169916.htm [2] 黄海凤, 高宏力, 李丹, 等.滚动轴承早期性能退化评估技术研究[J].机械科学与技术, 2017, 36(11):1771-1777 http://journals.nwpu.edu.cn/jxkxyjs/CN/abstract/abstract6874.shtmlHuang H F, Gao H L, Li D, et al. Study on evaluation of incipient performance degradation of rolling bearings[J]. Mechanical Science and Technology for Aerospace Engineering, 2017, 36(11):1771-1777(in Chinese) http://journals.nwpu.edu.cn/jxkxyjs/CN/abstract/abstract6874.shtml [3] Zhang B, Zhang L J, Xu J W. Degradation feature selection for remaining useful life prediction of rolling element bearings[J]. Quality and Reliability Engineering International, 2016, 32(2):547-554 doi: 10.1002/qre.v32.2 [4] 何园园, 张超, 陈帅.自适应随机共振与ELMD在轴承故障诊断中的应用[J].机械科学与技术, 2018, 37(4):607-613 http://journals.nwpu.edu.cn/jxkxyjs/CN/abstract/abstract6997.shtmlHe Y Y, Zhang C, Chen S. The application of self-adaptive stochastic resonance and ELMD in bearing fault diagnosis[J]. Mechanical Science and Technology for Aerospace Engineering, 2018, 37(4):607-613(in Chinese) http://journals.nwpu.edu.cn/jxkxyjs/CN/abstract/abstract6997.shtml [5] Lei Y G, Li N P, Guo L, et al. Machinery health prognostics:A systematic review from data acquisition to RUL prediction[J]. Mechanical Systems and Signal Processing, 2018, 104:799-834 doi: 10.1016/j.ymssp.2017.11.016 [6] Tan X D, Zhang Y, Qiu J, et al. Feature extraction technique for fault prognosis based on fault trend analysis[J]. Journal of Donghua University, 2017, 34(6):784-787 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=dhdxxb-e201706014 [7] Yang F, Habibullah M S, Zhang T Y, et al. Health index-based prognostics for remaining useful life predictions in electrical machines[J]. IEEE Transactions on Industrial Electronics, 2016, 63(4):2633-2644 doi: 10.1109/TIE.2016.2515054 [8] 马伦, 康建设, 孟妍, 等.基于Morlet小波变换的滚动轴承早期故障特征提取研究[J].仪器仪表学报, 2013, 34(4):920-926 doi: 10.3969/j.issn.0254-3087.2013.04.031Ma L, Kang J S, Meng Y, et al. Research on feature extraction of rolling bearing incipient fault based on Morlet wavelet transform[J]. Chinese Journal of Scientific Instrument, 2013, 34(4):920-926(in Chinese) doi: 10.3969/j.issn.0254-3087.2013.04.031 [9] 孙鲜明, 刘欢, 赵新光, 等.基于瞬时包络尺度谱熵的滚动轴承早期故障奇异点识别及特征提取[J].机械工程学报, 2017, 53(3):73-80 http://d.old.wanfangdata.com.cn/Periodical/jxgcxb201703011Sun X M, Liu H, Zhao X G, et al. Singular point recognition and feature extraction for incipient Bearing fault based on instantaneous envelope scalogram entropy[J]. Journal of Mechanical Engineering, 2017, 53(3):73-80(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/jxgcxb201703011 [10] 蒋永华, 李荣强, 焦卫东, 等.应用EMD和双谱分析的故障特征提取方法[J].振动、测试与诊断, 2017, 37(2):338-342 http://d.old.wanfangdata.com.cn/Periodical/zdcsyzd201702021Jiang Y H, Li R Q, Jiao W D, et al. Feature extraction method based on empirical mode decomposition and bispectrum analysis[J]. Journal of Vibration, Measurement & Diagnosis, 2017, 37(2):338-342(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zdcsyzd201702021 [11] 郭泰, 邓忠民, 徐萌.基于粒子群优化的改进EMD算法在轴承故障特征提取中的应用[J].振动与冲击, 2017, 36(16):182-187 http://d.old.wanfangdata.com.cn/Periodical/zdycj201716028Guo T, Deng Z M, Xu M. An improved EMD algorithm based on particle swarm optimization and its application to fault feature extraction of bearings[J]. Journal of Vibration and Shock, 2017, 36(16):182-187(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/zdycj201716028 [12] Qiu M, Chen L, Li Y C, et al. Bearing tribology:principles and applications[M]. Berlin, Heidelberg:Springer, 2017 [13] 李文峰, 戴豪民, 许爱强.时域新指标和PNN在滚动轴承故障诊断中的应用[J].机械科学与技术, 2016, 35(9):1382-1386 http://journals.nwpu.edu.cn/jxkxyjs/CN/abstract/abstract6495.shtmlLi W F, Dai H M, Xu A Q. New time domain index and probabilistic neural network and their application in fault diagnosis of rolling bearing[J]. Mechanical Science and Technology for Aerospace Engineering, 2016, 35(9):1382-1386(in Chinese) http://journals.nwpu.edu.cn/jxkxyjs/CN/abstract/abstract6495.shtml [14] 傅鹏程, 王昌明, 段俊斌, 等.基于故障演化分析的齿轮故障预测特征选择[J].深圳大学学报理工版, 2015, 32(4):434-440 http://d.old.wanfangdata.com.cn/Periodical/szdxxb201504016Fu P C, Wang C M, Duan J B, et al. A feature selection for fault prognosis based on fault evolution analysis[J]. Journal of Shenzhen University Science and Engineering, 2015, 32(4):434-440(in Chinese) http://d.old.wanfangdata.com.cn/Periodical/szdxxb201504016 [15] 曾庆虎.机械动力传动系统关键部件故障预测技术研究[D].长沙: 国防科学技术大学, 2010 http://cdmd.cnki.com.cn/Article/CDMD-90002-1011279952.htmZeng Q H. Fault prognostics technologies research for key components of mechanical power and transmission systems[J]. Changsha: National University of Defense Technology, 2010(in Chinese) http://cdmd.cnki.com.cn/Article/CDMD-90002-1011279952.htm [16] El-Thalji I, Jantunen E. Dynamic modelling of wear evolution in rolling bearings[J]. Tribology International, 2015, 84:90-99 doi: 10.1016/j.triboint.2014.11.021 [17] 窦东阳, 赵英凯.基于EMD和Lempel-Ziv指标的滚动轴承损伤程度识别研究[J].振动与冲击, 2010, 29(3):5-8 doi: 10.3969/j.issn.1000-3835.2010.03.002Dou D Y, Zhao Y K. Fault severity assessment for rolling element bearings based on EMD and Lempel-Ziv index[J]. Journal of Vibration and Shock, 2010, 29(3):5-8(in Chinese) doi: 10.3969/j.issn.1000-3835.2010.03.002 [18] Li R Y, Sopon P, He D. Fault features extraction for bearing prognostics[J]. Journal of Intelligent Manufacturing, 2012, 23(2):313-321 doi: 10.1007/s10845-009-0353-z -
点击查看大图
图(9) / 表(4)
计量
- 文章访问数: 375
- HTML全文浏览量: 169
- PDF下载量: 66
- 被引次数: 0
