轻量化空时模型融合自注意力机制的航空液压管路故障诊断新方法

摘要
图/表
参考文献(34)
相关文章 (15)

全文: PDF (3590 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要航空发动机外部液压管路结构复杂且管路故障信号中伴随着强大的噪声干扰，导致航空管路故障特征提取十分困难，同时目前诊断模型参数和计算量比较庞大，不适合高效地移植到移动和嵌入式设备。面对上述挑战，本文提出一种轻量化空时模型融合注意力机制的航空发动机液压管路故障诊断新方法，并命名为S-Bi-ATM-Net模型，首先，设计轻量化管路空间特征提取模块，从而融合管路故障数据的空间特征，抽取数据的细粒度特征；然后，设计轻量化管路时间特征提取模块, 从细粒度特征中继续融合管路粗粒度特征，实现粗细粒度特征融合；此外在空时模型中融合自注意力机制进行优化，使得最后的决策更加的聚焦，进一步提高所提模型的诊断精度。基于同一数据集，将所提方法与目前主流方法对比分析发现，所提方法能更准确地识别航空管路不同的故障状态，证明了所提方法的优越性和稳定性。

	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
	杨同光1
	袁晟友1
	周献文2
	韩清凯1
	于晓光3

关键词 ：故障诊断, 空时融合, 轻量化, 注意力机制, 航空管路

Abstract：The structure of the external hydraulic pipeline of aero-engine is complicated and the fault signal of the pipeline is accompanied by strong noise interference, which makes it difficult to extract the fault characteristics of the aviation pipeline. Meanwhile, the parameters and calculation amount of the current diagnostic model are relatively large, which is not suitable for efficient transplantation to mobile and embedded equipment. In the face of these challenges, this paper proposes a new method for fault diagnosis of aero-engine hydraulic pipelines based on lightweight air-time model fusion attention mechanism, and named the S-Bi-ATM-Net model. Then, the lightweight pipeline time feature extraction module is designed, and the coarse-grained features of the pipeline are continuously fused from the fine-grained features to achieve the fusion of coarse-fine-grained features. In addition, the self-attention mechanism is integrated into the space-time model for optimization, which makes the final decision more focused and further improves the diagnostic accuracy of the proposed model. Based on the same data set, comparing and analyzing the proposed method with the current mainstream methods, it is found that the proposed method can more accurately identify different fault states of air pipelines, which proves the superiority and stability of the method.

Key words： diagnostic faults spatial-temporal fusion lightweight self-attention aero-pipeline

收稿日期: 2023-06-13 出版日期: 2024-05-28

引用本文:

杨同光1，袁晟友1，周献文2，韩清凯1，于晓光3. 轻量化空时模型融合自注意力机制的航空液压管路故障诊断新方法[J]. 振动与冲击, 2024, 43(10): 299-310.
YANG Tongguang1, YUAN Shengyou1, ZHOU Xianwen2, HAN Qingkai1, YU Xiaoguang3. New approach to diagnose faults in aero-pipelines based on spatial-temporal model fused with self-attention mechanism. JOURNAL OF VIBRATION AND SHOCK, 2024, 43(10): 299-310.

链接本文:

http://jvs.sjtu.edu.cn/CN/ 或 http://jvs.sjtu.edu.cn/CN/Y2024/V43/I10/299

[1]汪博,高培鑫,马辉,等.航空发动机管路系统动力学特性综述[J].航空学报,2022, 43(05): 139- 162. WANG Bo, GAO Pei-xin, MA Hui, et al. A review on dynamic characteristics of aero-engine pipeline system[J]. Acta Aero- nautica et Astronautica Sinica, 2022,43(05):139-162. [2] 郭长虹,高静,王阔强等.循环冲击作用下缺陷飞机液压管路微泄漏故障诊断[J].液压与气动, 2020(07):28-35. GUO Chang-hong, GAO Jing, WANG Kuo-qiang, et al. Fault Diagnosis of Micro-leakage in Hydraulic Pipeline of Defected Aircraft Under Cyclic Impact[J]. Chinese Hydraulics＆Pneumatics, 2020, (7):28-35． [3] 权凌霄,李东,刘嵩等.膨胀环频域特性影响因素分析[J].浙江大学学报,2016,50(06):1065-1072. QUAN Ling-xiao, LI Dong, LIU Song, et al. Analysis of Influence Factors of Frequency Domain Characteristics of Expansion Ring[J]. Journal of Zhejiang University, 2016, 50(6): 1065-1072． [4] H. Liu, D. Jia, X. Wang, X. Liu, D. Gao, Clamp assembly stress measurement of aero-engine and assembly parameter control method, Journal of Aerospace Power, 35 (2020) 368-377. [5] D. Jana, J. Patil, S. Herkal, S. Nagarajaiah, L. Duenas-Osorio, CNN and Convolutional Autoencoder (CAE) based real-time sensor fault detection, localization, and correction, Mechanical Systems and Signal Processing, 169 (2022) 108723. [6] 高晔,孙伟,朴玉华,等.基于实测频响函数反推管路卡箍支承刚度与阻尼[J].航空动力学报, 2019, 34(03):664-670. GAO Ye, SUN Wei, PIAO Yu-hua, et al. Inverse identification of support stiffness and damping of hoop based on measured FRF[J]. Journal of Aerospace Power, 2019, 34(3): 664-670. [7] GAO Ye, SUN Wei. Inverse identification of the mechanical parameters of a pipeline hoop and analysis of the effect of preload[J]. Frontiers of Mechanical Engineering, 2019, 14(3): 358-368. [8] 刘中华,贾铎,王鑫,等.航空发动机卡箍装配应力试验与装配参数控制方法[J]. 航空动力学报, 2020, 35(02):368-377. LIU Zhong-hua, JIA Duo, WANG Xin, et al. Clamp assembly stress test of aero-engine and assembly parameter control method[J]. Journal of Aerospace Power, 2020,35(2): 368-377. [9] Li Zhe-zhu,Gao Pei-xin, Zhao Da-zhe, et al. Fault diagnosis and location of the aero-engine hydraulic pipeline based on Kalman filter[J]. Advances in Mechanical Engineering, 2017, Vol. 9(12)1-9. [10] M. Sujataa, M. Madana, K. Raghavendra, et al. Unraveling the cause of an aircraft accident[J]. Engineering Failure Analysis. 2019 (97) 740–758. [11] 窦金鑫,于晓光,杨同光等.斜裂纹航空液压直管振动特性分析[J].推进技术,2022,43(02):266-274. DOU Jin-xin, YU Xiao-guang, YANG Tong-guang, et al. Vibration Characteristics Analysis of An Aeronautical Hydraulic Straight Pipe with Slant Crack[J]. Journal of Propulsion Technology,2022,43(02):266-274. [12] Zahid Mehmooda, Asad Hameeda, Ali Javeda, et al. Analysis of premature failure of aircraft hydraulic pipes[J]. Engineering Failure Analysis. 2020 (109) 104356. [13] Hinton G E and Salakhutdinov R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313504. [14] 钱思宇,秦东晨,陈江义,等.基于卷积神经网络的领域适配模型的多工况迁移的轴承故障诊断[J].振动与冲击,2022,41(24):192-200. QIAN si-yu, QIN dong-chen, CHEN jian-yi, et al. Bearing fault diagnosis based on a domain adaptation model of convolutional neural network under multiple working conditions[J]. Journal of Vibration and Shock, 2022, 41(24): 192-200. [15] Z. Ye, J. Yu, Deep morphological convolutional network for feature learning of vibration signals and its applications to gearbox fault diagnosis, Mechanical Systems and Signal Processing, 161 (2021) 107984. [16] J. Shi, D. Peng, Z. Peng, Z. Zhang, K. Goebel, D. Wu, Planetary gearbox fault diagnosis using bidirectional-Convolu- Tional LSTM networks, Mechanical Systems and Signal Processing, 162 (2022) 107996. [17] 卞文彬, 邓艾东, 刘东川，等.基于改进深度残差收缩网络的风电机组滚动轴承故障诊断方法[J].机械工程学报, 2023, 59(12):202-214. BIAN Wenbin，DENG Aidong，LIU Dongchuan，et al．Fault Diagnosis Method of Wind Turbine Rolling Bearing Based on Improved Deep Residual Shrinkage Network[J]. Journal of Mechanical Engineering, 2023,59(12):202-214. [18] 王志颖,李天福,许文纲，等.降噪混合注意力变分自编码器及其在轴向柱塞泵故障诊断中的应用[J/OL].机械工程学报:1-11[2023-08-17]. WANG Zhiying，LI Tianfu，XU Wengang, et al. Denoising Mixed Attention Variational Auto-encoder for Axial Piston Pump Fault Diagnosis[J/OL]. Journal of Mechanical Engineering, 1-11 [2023- 08-17]. [19] Zhuang Ye, Yu Jian-bo. Deep morphological convolutional network for feature learning of vibration signals and its applications to gearbox fault diagnosis[J], Mechanical Systems and Signal Processing, 2021, (161) 107984. [20]侯召国,王华伟,熊明兰,等.基于迁移学习与加权多通道融合的齿轮箱故障诊断[J].振动与冲击,2023,42(09):236-246. HOU Zhao-guo，WANG Hua-wei，XIONG Ming-lan, et al. Gearbox fault diagnosis based on transfer learning and weighted multi-channel fusion[J]. Journal of vibration and shock, 2023,42(09):236-246. [21]贾思祥,孙丁一,毛刚，等.基于对抗熵的转子系统跨工况故障诊断方法[J/OL].机械工程学报,2023(15):110-120[2023-08-17]. JIA Sixiang，SUN Dingyi，MAO Gang, et al. Adversarial Entropy Based Fault Diagnosis Method for Rotor System Across Different Working Conditions[J/OL]. Journal of Mechanical Engineering, 2023(15):110-120[2023-08-17]. [22]邵海东,李伟,刘翊，等.时变转速下基于双阈值注意力生成对抗网络和小样本的转子-轴承系统故障诊断[J/OL].机械工程学报,2023(12):215-224. SHAO Haidong，LI Wei，LIU Yi，YANG Bin，Fault Diagnosis of Rotor-bearing System under Time-varying Speeds by Using Dual-threshold Attention-embedded GAN and Small Samples [J/OL]. Journal of Mechanical Engineering,2023(12): 215-224. [23] 常春,梅检民,赵慧敏,等.基于CUM3-CNN的柴油机高压油路故障诊断[J].振动与冲击,2023,42(03):174-180. CHANG Chun，MEI Jian-min，ZHAO Hui-min, et al. Fault diagnosis of diesel engine high pressure oil circuit based on CUM3-CNN[J]. Journal of vibration and shock, 2023, 42(03): 174-180. [24] 赵志宏,李春秀,窦广鉴,等.基于MTF-CNN的轴承故障诊断研究[J].振动与冲击,2023,42(02):126-131. ZHAO Zhi-hong，LI Chun-xiu，DOU Guang-jian, et al. Bearing fault diagnosis method based on MTF-CNN[J]. Journal of vibration and shock, 2023,42(02):126-131. [25] Sun Yan-ning, Zhuang Zi-long, Xu Hong-wei, et al. Data- driven modeling and analysis based on complex network for multimode recognition of industrial processes[J]. Journal of Manufacturing Systems, 2021: 1-10. [26] SHE Dao-ming, JIA Ming-ping. A Bi-GRU method for remaining useful life prediction of machinery. Measurement, 2021, 167: 108277. [27] 赵志宏,吴冬冬,窦广鉴,等.一种基于CNN-Bi GRU孪生网络的轴承故障诊断方法[J].振动与冲击,2023,42(06): 166-171+ 211. ZHAO Zhi-hong，WU Dong-dong，DOU Guang-jian, et al. Bearing fault diagnosis method based on a CNN-Bi-GRU Siamese network[J]. Journal of vibration and shock, 2023, 42 (06): 166-171+211. [28] Sun Yan-ning, QIN Wei, ZUANG Li-long, et al. An adaptive fault detection and root-cause analysis scheme for complex industrial processes using moving window KPCA and information geometric causal inference[J]. Journal of Intelligent Manufacturing, 2021: 1-15. [29] 林俊亭,王帅,刘恩东,等.基于模糊聚类和CNN-BIGRU的轨道电路故障预测方法[J/OL].振动.测试与诊断:1-8 [2023-06-01]. LIN Jun-ting，WANG Shuai，LIU En-dong, et al. Track circuit fault prediction method based on fuzzy clustering and CNN-BIGRU [J/OL]. Journal of vibration, measurement & diagnosis, :1-8 [2023-06-01]. [30] Yao Qi-hang, WANG Ru-xin, FAN Xiao-mao, et al. Multi-class Arrhythmia detection from 12-lead varied-length ECG using Attention-based Time-Incremental Convolutional Neural Network[J]. Information Fusion. 2020,53 (02) 174–182. [31]雷春丽,薛林林,夏奔锋,等.采用马尔科夫转移场和图注意力网络的滚动轴承故障诊断方法[J/OL].振动工程学报:1-10[2023-06-01]. LEI Chun li, XUE Lin lin, XIA Ben feng, et al. Rolling Bearing Fault Diagnosis Method Based on Markov Transition Field and Graph Attention Network. Journal of vibration engineering[J], 1-10[2023-06-01]. [32] Iandola F N, Han S, Moskewicz M W, et al. Squeeze-Net: Alex-Net-level accuracy with 50x fewer parameters and＜0. 5MB model size. ICLR 2017 Conference Official Review, 2016, 11 (1):112-120. [33] ZHANG Chuan-fang, PENG Kai-xiang, Dong Jie, et al. A comprehensive operating performance assessment framework based on distributed Siamese gated recurrent unit for hot strip mill process[J]. Applied Soft Computing 132 (2023) 109889. [34] Wang Jin-rui, Li Shun-ming, An Zeng-hui, et al. Batch- normalized deep neural networks for achieving fast intelligent fault diagnosis of machines[J], Neurocomputing 329 (2019) 53–65.