针对航空发动机滚动轴承在低转速状态下故障难检测的问题，提出了一种基于Transformer框架的深度支持向量描述方法用于检测低转速滚动轴承的故障。首先，构建了基于Transformer模型的振动特征提取主干网络。然后，将所提取的特征输入一个三层自编码器结构，用于计算网络模型的损失函数。为减少网络计算量，提高训练速度，在预处理中将滚动轴承的振动加速度时域信号通过快速傅里叶变换(FFT)得到的 频谱结果作为网络的输入，且仅依靠正常数据完成模型的训练。最后，在带机匣的航空发动机转子试验器和某型真实的航空发动机上分别进行了试验验证。结果表明，所提方法能够准确的实现对低转速滚动轴承故障的检测，且检测精度分别为93%和100%，充分表明本文方法具有很好的异常检测能力及应用价值。
Abstract
Aiming at the problem that the faults of aero-engine rolling bearings are difficult to detect at low speed, a deep support vector description method based on Transformer framework was proposed to detect the faults of low speed rolling bearings. Firstly, a vibration feature extraction backbone network based on Transformer model is constructed. Then, the extracted features are input into a three-layer autoencoder structure, which is used to calculate the loss function of the network model. In order to reduce the amount of network calculation and improve the training speed, the frequency spectrum results obtained by Fast Fourier Transform (FFT) of the vibration acceleration time domain signal of rolling bearing were used as the input of the network in the preprocessing, and the model training was completed only by normal data. Finally, the experiments were carried out on an aero-engine rotor tester with casing and a real aero-engine. The results show that the proposed method can accurately detect the faults of low speed rolling bearings, and the detection accuracy is 93% and 100%, respectively, which fully indicates that the proposed method has good anomaly detection ability and application value.
关键词
低转速 /
滚动轴承 /
深度异常检测 /
Transformer /
航空发动机
{{custom_keyword}} /
Key words
Low speed /
Rolling bearing /
Deep anomaly detection /
Transformer /
Aero-engine
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1]. 尉询楷, 杨立, 战立光, 等. 航空发动机预测与健康管理[M]. 北京: 国防工业出版社, 2014.
Wei Xunkai,YANG Li,ZHAN Liguang, et al. Aero-engine forecasting and health management[M]. Beijing: National Defense Industry Press, 2014.
[2]. 郝如江,卢文秀,褚福磊.声发射检测技术用于滚动轴承故障诊断的研究综述[J].振动与冲击,2008(03):75-79+181.
HAO Rujiang,LU Wenxiu,CHU Fulei .Review of acoustic emission detection technology for rolling bearing fault diagnosis[J]. JOURNAL OF VIBRATION AND SHOCK, 2008(03):75-79+181.
[3]. 柯燕亮,王华庆,唐刚,等.基于Teager峰值能量的低转速轴承故障特征提取方法[J].振动与冲击,2017,36(11):124-128+133.DOI:10.13465/j.cnki.jvs.2017.11.019.
KE Yan-liang, WANG Hua-qing, TANG Gang, et al. Fault feature extraction of low speed roller bearing based on the Teager peak energy. JOURNAL OF VIBRATION AND SHOCK, 2017, 36(11): 124-128.
[4]. Condition Monitoring of Low Speed Slewing Bearings Based on Ensemble Empirical Mode Decomposition Method[J]. Transactions of the Korean Society for Noise and Vibration Engineering,2013,23(2).
[5]. Mishra, C.; Samantaray, A.; Chakraborty, G. Rolling Element Bearing Fault Diagnosis under Slow Speed Operation Using Wavelet De-Noising. Meas. J. Int. Meas. Confed. 2017, 103, 77–86. [CrossRef]
[6]. Wang, S.; Niu, P.; Guo, Y.; Wang, F.; Li, W.; Shi, H.; Han, S. Early Diagnosis of Bearing Faults Using Decomposition and Reconstruction Stochastic Resonance System. Measurement 2020, 158, 107709
[7]. Diego Sandoval,Urko Leturiondo,Francesc Pozo,Yolanda Vidal. Low-Speed Bearing Fault Diagnosis Based on Permutation and Spectral Entropy Measures[J]. Applied Sciences,2020,10(13).
[8]. Han, T.; Liu, Q.; Zhang, L.; Tan, A. Fault Feature Extraction of Low Speed Roller Bearing Based on Teager Energy Operator and CEEMD. Meas. J. Int. Meas. Confed. 2019, 138, 400–408.
[9]. 李壮,沙云东,栾孝驰,等.基于模拟机匣的中介轴承微弱故障特征提取技术[J].科学技术与工程,2021,21(35):15262-15269.
Li Zhuang, Sha Yundong, Luan Xiaochi, et al. Weak Fault Feature Extraction Technology of Intermediate Bearing Based on analog casing[J]. Science Technology and Engineering,2021,21(35):15262-15269.
[10]. DosoViTskiy A, Beyer L, Kolesnikov A, Weissenborn D, Zhai XH, Unterthiner T, Dehghani M, Minderer M, Heigold G, Gelly S,Uszkoreit J, Houlsby N. An image is worth 16x16 words: Transformers for image recognition at scale. In: Prof. of the 9th Int’l Conf. on Learning Representations. OpenReview.net, 2021. 1–22.
[11]. Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez NA, Kaiser L, Polosukhin I. Attention is all you need. In: Proc. of the31st Int’l Conf. on Neural Information Processing Systems. Long Beach: Curran Associates Inc., 2017. 6000–6010. [doi:10.5555/3295222.3295349]
[12]. Ruff L , Vandermeulen R A , N Görnitz, et al. Deep One-Class Classification[C]// International Conference on Machine Learning. 2018.
[13]. Chalapathy R , Menon A K , Chawla S . Anomaly Detection using One-Class Neural Networks[J]. 2018.
[14]. Radford, A., Metz, L., and Chintala, S. Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks. arXiv:1511.06434, 2015.
[15]. 陈果.转子-滚动轴承-机匣耦合系统中滚动轴承故障的动力学分析[J].振动工程学报,2008,21(06):577-587.DOI:10.16385/j.cnki.issn.1004-4523.2008.06.008.
CHEN Guo.Dynamic analysis of ball bearing faults in rotor-ball bearing-stator coupling system[J]. Journal of Vibration Engineering,2008,21(06):577-587.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
PDF(3349 KB)
