多头注意力驱动的航空高速轴承故障诊断方法

摘要
图/表
参考文献(26)
相关文章 (15)

全文: PDF (1967 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要航空发动机运行速度高、工况变化大、结构复杂且干扰噪声大，导致微弱故障特征往往存在于多子空间中，目前基于数据驱动的诊断模型尚不足以可靠捕捉不同子空间中丰富的特征信息。针对上述问题，提出一种基于信号特征的多头注意力诊断方法MADM（A Multi-head Attention Diagnosis Method Based on Signal Features），可实现高速非平稳工况下航空轴承故障状态的识别和诊断。该方法首先通过卷积模块和双向GRU模块对原始振动信号进行特征提取，然后引入多头注意力模块，使得网络同时注意并融合不同表示子空间的信息以提高故障特征的显著性水平。最后，利用全连接模块和Softmax分类器对提取的特征进行高速轴承故障诊断。实验结果表明提出的MADM该诊断方法可实现转速为12000转/分以上、剥落面积最小为0.5 mm2的航空轴承高精度可靠诊断，且优于目前主流的深度诊断方法。

	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
	王兴1
	张晗1
	朱家正1
	林建波1
	杜朝辉2

关键词 ：航空轴承, 多头注意力, 故障诊断, 深度学习

Abstract：Due to the high running speed, large variation of operating conditions, complex structure and large interference noise, the weak fault features often exist in multiple subspaces. At present, the data-driven diagnosis model is not enough to reliably capture rich feature information in different subspaces. In order to solve the above problems, MADM(A Multi-head Attention Diagnosis Method Based on Signal Features) is proposed to identify and diagnose the fault state of aeronautical bearings under high-speed and non-stationary working conditions. In this method, the features of the original vibration signals are extracted by convolution module and bi-directional GRU module, and then the multi-head attention module is introduced to make the network pay attention to and fuse the information of different representation subspaces at the same time to improve the significance level of fault features. Finally, the full connection module and Softmax classifier are used to diagnose the high-speed bearing fault. The experimental results show that the proposed MADM diagnosis method can realize high precision and reliable diagnosis of aeronautical bearings with rotational speed of more than 12000 rpm and the minimum spalling area of 0.5 mm2, and is better than the current mainstream depth diagnosis methods.

Key words： Aviation bearing multi-head attention fault diagnosis deep learning

收稿日期: 2022-02-16 出版日期: 2023-02-28

引用本文:

王兴1，张晗1，朱家正1，林建波1，杜朝辉2. 多头注意力驱动的航空高速轴承故障诊断方法[J]. 振动与冲击, 2023, 42(4): 295-305.
WANG Xing1, ZHANG Han1, ZHU Jiazheng1, LIN Jianbo1, DU Zhaohui2. A fault diagnosis method for aviation high-speed bearings driven by multi-head attention. JOURNAL OF VIBRATION AND SHOCK, 2023, 42(4): 295-305.

链接本文:

http://jvs.sjtu.edu.cn/CN/ 或 http://jvs.sjtu.edu.cn/CN/Y2023/V42/I4/295

[1] 吕超, 程弓, 刘云清. 基于BDPCA聚类算法的航空发动机故障数据标记 [J]. 振动与冲击, 2020, 39(09): 35-41.
LV Chao, Cheng Gong, Liu Yunqing Aeroengine fault data tag using on improved BDPCA algorithm [J] Vibration and shock, 2020, 39 (09): 35-41
[2] JIAO J, ZHAO M, LIN J, et al. Hierarchical discriminating sparse coding for weak fault feature extraction of rolling bearings [J]. Reliability Engineering & System Safety, 2018, 184(APR.): 41-54.
[3] 张晗, 王兴, 田毅, 等. 稀疏低秩协同正则优化及航空轴承故障诊断 [J]. 西安交通大学学报, 2021, 55(11): 46-58.
Zhang Han, Wang Xing, Tian Yi, et al. Collaborative sparse low rank regularization for aero-engine bearing fault diagnosis [J] Journal of Xi'an Jiaotong University, 2021, 55 (11): 46-58
[4] 张晗, 杜朝辉, 方作为, 等. 基于稀疏分解理论的航空发动机轴承故障诊断 [J]. 机械工程学报, 2015, 51(01): 97-105.
Zhang Han, Du Chaohui, Fang as, et al. Sparse decomposition based aero-engine’s bearing fault diagnosis [J] Journal of mechanical engineering, 2015, 51 (01): 97-105.
[5] RZADKOWSKI R, ROKICKI E, PIECHOWSKI L, et al. Analysis of middle bearing failure in rotor jet engine using tip-timing and tip-clearance techniques [J]. Mechanical Systems and Signal Processing, 2016, 76-77: 213-227.
[6] ZHANG H, CHEN X F, DU Z H, et al. Nonlocal sparse model with adaptive structural clustering for feature extraction of aero-engine bearings [J]. Journal of Sound and Vibration, 2016, 368: 223-248.
[7] ZHANG H, CHEN X, ZHANG X, et al. Aero-engine bearing fault detection: A clustering low-rank approach [J]. Mechanical Systems and Signal Processing, 2020, 138: 106529.
[8] KHAN S, YAIRI T. A review on the application of deep learning in system health management [J]. Mechanical Systems and Signal Processing, 2018, 107: 241-265.
[9] LEI Y, YANG B, JIANG X, et al. Applications of machine learning to machine fault diagnosis: A review and roadmap [J]. Mechanical Systems and Signal Processing, 2020, 138: 106587.
[10] ZHAO D Z, ZHENG H, DUAN S Q, et al. Aero-engine Bearing Fault Diagnosis Based on Deep Neural Networks; proceedings of the 11th International Conference on Mechanical and Aerospace Engineering (ICMAE), Athens, GREECE, F Jul 14-17, 2020 [C]. 2020.
[11] WANG B J, ZHANG X L, SUN C, et al. A Quantitative Intelligent Diagnosis Method for Early Weak Faults of Aviation High-speed Bearings [J]. Isa Transactions, 2019, 93: 370-383.
[12] 王奉涛, 薛宇航, 王洪涛, 等. GLT-CNN方法及其在航空发动机中介轴承故障诊断中的应用 [J]. 振动工程学报, 2019, 32(06): 1077-1083.
Wang Fengtao, Xue Yuhang, Wang Hongtao, et al. GLT-CNN and its application of aero-engine intermediary bearing fault diagnosis [J] Journal of vibration engineering, 2019, 32 (06): 1077-1083.
[13] LIN X S, LI B W, YANG X Y, et al. Fault Diagnosis of Aero-engine Bearing Using a Stacked Auto-Encoder Network; proceedings of the IEEE 4th Information Technology and Mechatronics Engineering Conference (ITOEC), Chongqing, PEOPLES R CHINA, F Dec 14-16, 2018 [C]. 2018.
[14] 张向阳, 陈果, 郝腾飞, 等. 基于机匣信号的滚动轴承故障卷积神经网络诊断方法 [J]. 航空动力学报, 2019, 34(12): 2729-2737.
Zhang Xiangyang, Chen Guo, Hao Tengfei, et al. Convolutional neural network diagnosis method of rolling bearing fault based on casing signal [J] Journal of Aerospace Power, 2019, 34 (12): 2729-2737.
[15] HSIAO P W, CHEN C P, IEEE. EFFECTIVE ATTENTION MECHANISM IN DYNAMIC MODELS FOR SPEECH EMOTION RECOGNITION [C]// 2018 IEEE International Conference on Acoustics, Speech and Signal Processing. Calgary : IEEE ,2018.
[16] LI X, ZHANG W, DING Q. Understanding and improving deep learning-based rolling bearing fault diagnosis with attention mechanism [J]. Signal Processing, 2019, 161: 136-154.
[17] YANG Z B, ZHANG J P, ZHAO Z B, et al. Interpreting network knowledge with attention mechanism for bearing fault diagnosis [J]. Applied Soft Computing, 2020, 97: 106829.
[18] XU Z F, LI C, YANG Y. Fault diagnosis of rolling bearings using an Improved Multi-Scale Convolutional Neural Network with Feature Attention mechanism [J]. Isa Transactions, 2021, 110: 379-393.
[19] ZHANG X, CONG Y, YUAN Z, et al. Early Fault Detection Method of Rolling Bearing Based on MCNN and GRU Network with an Attention Mechanism [J]. Shock and Vibration, 2021, 2021(3): 1-13.
[20] WANG H, LIU Z L, PENG D D, et al. Understanding and Learning Discriminant Features based on Multiattention 1DCNN for Wheelset Bearing Fault Diagnosis [J]. Ieee Transactions on Industrial Informatics, 2020, 16(9): 5735-5745.
[21] YIJIA H, HUAN W, ZHILIANG L, et al. Multi-Scale CNN based on Attention Mechanism for Rolling Bearing Fault Diagnosis [J]. 2020 Asia-Pacific International Symposium on Advanced Reliability and Maintenance Modeling (APARM), Vancouver :IEEE,2020: 5 pp.-5 pp.
[22] VASWANI A, SHAZEER N, PARMAR N, et al. Attention Is All You Need [J]. arXiv, 2017.
[23] CHUNG J, GULCEHRE C, CHO K H, et al. Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling [J]. Eprint Arxiv, 2014.
[24] RUDER S. An overview of gradient descent optimization algorithms arXiv [J]. arXiv, 2016: 12 pp.-12 pp.
[25] WEI Z, GAOLIANG P, CHUANHAO L, et al. A New Deep Learning Model for Fault Diagnosis with Good Anti-Noise and Domain Adaptation Ability on Raw Vibration Signals [J]. Sensors, 2017, 17(3): 425.
[26] ZHAO Z B, LI T F, WU J Y, et al. Deep learning algorithms for rotating machinery intelligent diagnosis: An open source benchmark study [J]. Isa Transactions, 2020, 107: 224-255.