时变工况下行星轮轴承特征分布拟合与智能故障诊断

摘要
图/表
参考文献(24)
相关文章 (6)

全文: PDF (2039 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要

行星齿轮箱中行星轮轴承运动复杂，振动信号成分多样且呈非线性，时变工况下特征频率随时间变化，人工识别故障特征更为困难。针对该问题，提出基于对抗变分自编码的智能故障诊断模型。获取样本时频图来揭示样本中包含的时变特征；利用模型中的变分自编码器自适应提取时频图特征，为赋予特征显式意义，引入多维独立高斯分布并进行采样，根据类别信息对样本点进行变换，使其服从新的多维独立高斯分布，确保样本点中的每个元素都有自己的专属分布；通过对抗机制，使特征逼近变换后的对应类别的分布样本，且服从新的多维独立高斯分布，从而实现用已知的分布拟合未知的特征分布，同时通过控制分布强化不同类别特征间的差异性，改善特征的模式识别性能；利用优化后的特征对分类器进行训练并识别测试样本特征。模型经行星齿轮箱实验台数据进行了验证。研究表明，该模型能够使提取的特征服从给定的先验分布，通过控制分布强化了不同类别特征间的差异性，提高了特征的聚合性能，有效诊断了行星轮轴承故障，与自编码和变分自编码器相比，表现出一定的优越性。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	赵川
	冯志鹏

关键词 ：行星轮轴承, 智能故障诊断, 多维独立高斯分布, 对抗变分自编码, 时变工况

Abstract：The intricate kinematics of planet bearings in planetary gearboxes leads to complex and even non-linear vibration signals.Besides, the characteristic frequencies change with time under time-varying working condition, which makes the fault diagnosis of planet bearings manually more difficult.For addressing these issues, an intelligent fault diagnosis method was proposed based on the adversarial variational auto-encoder.Firstly, the time-frequency representation of a sample was obtained to reveal the time-varying properties.Then, the variational auto-encoder of the model was utilized to extract features of a time-frequency image.In order to make the latent features exposed in an explicit meaning, a multivariate independent Gaussian distribution was introduced, and samples were collected from the distribution.After that, the samples were transformed into new ones according to the label information and made to follow a new multivariate independent Gaussian distribution to ensure each element in the sample has its own distribution.Through an adversarial game, the latent features were made to have the same distribution as the corresponding transformed samples and follow the new distribution so that the unknown features distribution could be fitted by a given prior distribution.And the discriminability of the features among different classes was enhanced by controlling the distribution to improve their performance for pattern identification.Finally, a classifier was trained and tested by the optimized features.The method was validated via some planetary gearbox data sets.The results show that the model enables extracted features among different classes to follow an explicit distribution, improves their clustering performance for pattern identification and effectively diagnoses planet bearing faults.It outperforms to a certain extent the traditional auto-encoder and variational auto-encoder.

Key words： planet bearing intelligent fault diagnosis multivariate independent Gaussian distribution adversarial variational auto-encoder time-varying condition

收稿日期: 2020-04-06 出版日期: 2021-07-28

引用本文:

赵川，冯志鹏. 时变工况下行星轮轴承特征分布拟合与智能故障诊断[J]. 振动与冲击, 2021, 40(14): 252-260.
ZHAO Chuan， FENG Zhipeng. Features distribution fitting and intelligent fault diagnosis of planet bearings under time-varying condition. JOURNAL OF VIBRATION AND SHOCK, 2021, 40(14): 252-260.

链接本文:

http://jvs.sjtu.edu.cn/CN/ 或 http://jvs.sjtu.edu.cn/CN/Y2021/V40/I14/252

［1］GUI Y, HAN Q K, CHU F L.A vibration model for fault diagnosis of planetary gearboxes with localized planet bearing defects［J］.Journal of Mechanical Science and Technology, 2016, 30(9):4109-4119.
［2］ZHAO C, FENG Z P, WEI X K, et al.Sparse classification based on dictionary learning for planet bearing fault identification［J］.Expert Systems with Applications, 2018, 108:233-245.
［3］FENG Z P, CHEN X W, LIANG M.Joint envelope and frequency order spectrum analysis based on iterative generalized demodulation for planetary gearbox fault diagnosis under nonstationary conditions［J］.Mechanical Systems and Signal Processing, 2016, 76/77:242-264.
［4］CHEN X W, FENG Z P.Iterative generalized time-frequency reassignment for planetary gearbox fault diagnosis under nonstationary conditions［J］.Mechanical Systems and Signal Processing, 2016, 80:429-444.
［5］CHEN X W, FENG Z P.Time-frequency analysis of torsional vibration signals in resonance region for planetary Gearbox fault diagnosis under variable speed conditions［J］.IEEE Access, 2017, 5:21918-21926.
［6］FENG K, WANG K S, NI Q.A phase angle based diagnostic scheme to planetary gear faults diagnostics under non-stationary operational conditions［J］.Journal of Sound and Vibration, 2017, 408:190-209.
［7］GUAN Y P, LIANG M, NECSULESCU D S.A velocity synchro-squeezing transform for fault diagnosis of planetary gearboxes under nonstationary conditions［J］.Journal of Mechanical Engineering Science, 2017, 231(15):2868-2884.
［8］HU Y, TU X T, LI F C.Joint high-order synchro squeezing transform and multi-taper empirical wavelet transform for fault diagnosis of wind turbine planetary gearbox under non-stationary conditions［J］.Sensors, 2018, 18:1-18.
［9］CHEN S Q, DU M G, PENG Z K.Fault diagnosis of planetary gearbox under variable-speed conditions using an improved adaptive chirp mode decomposition［J］.Journal of Sound and Vibration, 2020, 468:1-20.
［10］ZHANG W, LI C H, PENG G L, et al.A deep convolutional neural network with new training methods for bearing fault diagnosis under noisy environment and different working load［J］.Mechanical Systems and Signal Processing, 2018, 100:439-453.
［11］SHAO H D, JIANG H K, ZHANG H Z, et al.Rolling bearing fault feature learning using improved convolutional deep belief network with compressed sensing［J］.Mechanical Systems and Signal Processing, 2018, 100:743-765.
［12］JIA F, LEI Y G, GUO L, et al.A neural network constructed by deep learning technique and its application to intelligent fault diagnosis of machines［J］.Neurocomputing, 2018, 272:619-628.
［13］DING C C, ZHAO M, LIN J.Transient feature extraction of encoder signal for condition assessment of planetary gearboxes with variable rotational speed［J］.Measurement, 2020, 151:1-17.
［14］MAKHZANI A, SHLENS J, JAITLY N.Adversarial autoencoders［Z］.arXiv: 1511.05644v2［cs.LG］, 2016.
［15］HONG Y, HWANG U, YOO J.How generative adversarial networks and their variants work: an overview［J］.ACM Computing Surveys, 2019, 52(1):1-10.
［16］MAO W, LIU Y M, DING L.Imbalanced fault diagnosis of rolling bearing based on generative adversarial network: a comparative study［J］.IEEE Access, 2019,7:9515-9530.
［17］XU C Y, DAI Y F, LIN R J.Social image refinement and annotation via weakly-supervised variational auto-encoder［J］.Knowledge-Based Systems, 2020, 192:1-9.
［18］WANG Z R, WANG J, WANG Y R.An intelligent diagnosis scheme based on generative adversarial learning deep neural networks and its application to planetary gearbox fault pattern recognition［J］.Neurocomputing, 2018, 310:213-222.
［19］ZHAO D F, LIU S L, GU D.Enhanced data-driven fault diagnosis for machines with small and unbalanced data based on variational auto-encoder［J］.Measurement Science and Technology, 2020, 31(3):1-9.
［20］WEN L, GAO L, LI X.A new deep transfer learning based on sparse auto-encoder for fault diagnosis［J］.IEEE Transactions on Systems Man Cybernetics-systems, 2019, 7(1):136-144.
［21］LI Y J, ZHANG T.Introduction to deep learning and case studies［M］.Beijing: Machinery Industry Press, 2016.
［22］KINGMA D P, WELLING M.Auto-encoding variational bayes［Z］.arXiv preprint arXiv:1312.6114, 2014.
［23］OTSU N.A threshold selection method from gray-level histograms［J］.IEEE Transactions on Systems Man & Cybernetics, 2007, 9(1):62-66.
［24］ZHAO C, FENG Z P.Application of multi-domain sparse features for fault identification of planetary gearbox［J］.Measurement, 2017, 104:169-179.