基于 CAE-LSTM 的航发轴承故障诊断方法

Fault Diagnosis Method for Aero-engine Bearings Based on CAE-LSTM

摘要

随着计算机深度学习理论的发展及航空领域对关键装备智能化故障诊断与运维需求的提升，基于深度学习的航空发动机运行状态的监测与评估方法成为了飞机安全运行的重要保障。由于航空发动机机械结构复杂，轴承在高温、高压等恶劣环境下高速运行，其故障特征信息存在多尺度、非线性等问题，使得故障信号难以有效识别及分析诊断。因此，本文提出了一种基于CAE-LSTM的航发轴承故障诊断方法，首先利用改进的卷积自编码器（Convolutional Autoencoder，CAE）对高维振动信号进行降维和特征提取，然后将提取到的特征输入到长短期记忆网络（Long Short-Term Memory，LSTM）分类器中进行故障类型识别，从而提升轴承故障分类的准确性和鲁棒性。实验结果表明本文提出的方法能够有效地学习航发轴承传感信号序列中的动态特征，提高航发轴承故障诊断的精确性和智能性。

Abstract

With the development of computer deep learning theory and the enhancement of the demand for intelligent fault diagnosis and operation and maintenance of key equipment in the field of aviation，the monitoring and evaluation method of the operating state of aero-engine based on deep learning has become an important guarantee for the safe operation of aircraft.Due to the complexity of the mechanical structure of the aero-engine and the high-speed operation of the bearings in a harsh environment such as high temperature and high pressure，there are problems such as multi-scale and non-linearity in the fault characteristic information，which makes it diffi- cult to effectively identify and analyze and diagnose the fault signals.Therefore，this paper proposes a CAE-LSTM-based fault diagno- sis method for airframe bearings，which firstly utilizes the improved Convolutional Autoencoder（CAE）to perform downscaling and feature extraction for high-dimensional vibration signals，and then inputs the extracted features into the Long Short-Term Memory （LSTM）classification network.Memory（LSTM）classifier for fault type identification，thus improving the accuracy and robustness of bearing fault classification.The experimental results show that the method proposed in this paper can effectively learn the dynamic features in the aircraft bearing sensing signal sequence and improve the accuracy and intelligence of aircraft bearing fault diagnosis.

关键词

航空发动机 / 卷积自编码器 / 长短期记忆网络 / 特征提取 / 战障诊断

Key words

aero-engine / convolutional autoencoder / long short-term memory / feature extraction / fault diagnosis

参考文献

[1] Jin Y, Zhou X, Quan X, et al. Topological structures of vibration re- sponses for dual-rotor aeroengine[J]. Mechanical Systems Signal Processing, 2024, 208:1-22,doi:10.1016/j.ymssp.2023.111053.

[2] Chongyang Wang, Zihang Li, Haoze Wang, et al. Nonlinear model- ing and vibration response analysis of rod fastening rotor-bearing system with curvic coupling[J]. Tribology International, 2024, 199:1-17,doi:10.1016/j.triboint.2024.109984.

[3] Xu G, Guo R, Yin Z, et al. MCRCNet:a bearing fault diagnosis method for unknown faults based on transfer learning[J]. Applied Sciences. 2025, 15:1-20,doi:10.3390/app15020921.

[4] Yin Z, Zhang F, Xu G, et al. Multi-scale rolling bearing fault diag- nosis method based on transfer learning[J]. Applied Sciences, 2024, 14(3):1-20,doi:10.3390/app14031198.

[5] LIU J F, YU X, WAN II B, et al. Fault diagnosis method of rolling bearing using MFMD and Transformer-CNN[J]. Journal of Aero- space Power, 2023, 38(6):1446-1456.

[6] Long Wen, Guang Yang, Longxin Hu, et al. A new unsupervised health index estimation method for bearings early fault detection based on Gaussian mixture model[J]. Engineering Applications of Artificial Intelligence, 2024, 128:1-14,doi:10.1016/j.engappai.2023.107562.

[7] Lingli Cui, Zhilin Dong, Hai Xu, et al. Triplet attention-cnhanced residual tree-inspired decision network:a hierarchical fault diagnosis model for unbalanced bearing datasets[J]. Advanced Engineering Informatics, 2024,59:102322,doi:10.1016/j.aei.2023.102322.

[8] LIU J, TAN Y T, YANG N. A study on bearing fault diagnosis based on ACYCBD-MTF-MobileViT under strong noise[J]. Jour- nal of Vibration and Shock, 2024, 43(24):34-47.

[9] Hou L, Yi H, Jin Y, et al. Inter-shaft bearing fault diagnosis based on an aero engine system:a benchmarking dataset study[J]. Journal of Dynamics,Monitoring and Diagnostics, 2023, 4(2):228-242.

[10] Lijie Zhang, Bin Wang, Pengfei Liang, et al. Semi-supervised fault diagnosis of gearbox based on feature pre-extraction mechanism and improved generative adversarial networks under limited labeled samples and noise environment[J]. Advanced Engineering Infor- matics, 2023, 58:1-14,doi:10.1016/j.aei.2023.102211.

[11] Cao Z, Dai J, Xu W, et al. Sparse Bayesian learning approach for compound bearing fault diagnosis[J]. IEEE Transactions on Indus- trial Informatics, 2024, 20(2):1562-1574.

[12] ZHANG B W, PANG X Y, CHENG B A, et al. Fault diagnosis method for aeroengine bearings based on PIRD-CNN[ J]. Journal of Vibration and Shock, 2024, 43(18):201-207＋231.

[13] Chen Z, Xu B, Zhang Z. Rolling bearing fault diagnosis based on recurrence plot[J]. IEEE Access, 2024, 12:149710-149721,doi: 10.1109/access.2024.3472454.

[14] Xu W, He J,G.Li, et al. Bearing fault feature extraction via adap- tive frequency matching and coupling analysis[J]. IEEE Sensors Journal, 2024, 24(23):39565-39577.

[15] Zhang X, Liu J, Zhang X, et al. Multiscale channel attention-driven graph dynamic fusion learning method for robust fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2024, 20 (9): 11002-11013.

[16] Chen H, Wang X B, Li J, et al. Dynamic focusing network for semisupervised mechanical fault diagnosis of rotating machinery[J]. IEEE Transactions on Industrial Informatics, 2024, 20(10): 11575-11586.

[17] An Y, Zhang K, Liu Q, et al. Rolling bearing fault diagnosis meth- od hase on periodic sparse attention and ISTM[J]. IFEE Sensors Journal, 2022, 22(12):12044-12053.

[18] LIU F, CHEN R W, XING K L, et al. Fast fault diagnosis algorithm for rolling bearing based on transfer learning and deep residual net- work[J]. Journal of Vibration and Shock, 2022, 41(3):154-164.

[19] Yang L, Wen C, Zhou Z. Autoencoder-multidimensional taylor net- work for intelligent fault diagnosis under privacy preservation[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73: 1-17,doi:10.1109/TIM.2024.3472785.

[20] Fan H, Ren Z, Cao X, et al. A GTI ＆ Ada-Act LMCNN method for intelligent fault diagnosis of motor rotor-bearing unit under variable conditions[J]. IEEE Transactions on Instrumentation and Measure- ment, 2024, 73:1-14,doi:10.1109/TIM.2024.3351265.

[21] JIANG B D, ZHANG B W, LIANG Y G, et al. Research on intelli- gent fault diagnosis method for aero-engine bearing based on multi-scale feature extraction of displacement information[J]. Aviation Maintenance ＆ Engineering, 2024,(8):23-27,doi:10.19302/j.cnki.1672-0982.2024.08.025.

[22] 刘俊锋, 俞翔, 万海波, 等. 基于 MFMD 和 Transformer-CNN 的滚动轴承故障诊断方法[J]. 航空动力学报, 2023, 38(6): 1446-1456.

[23] 刘杰, 谭玉涛, 杨娜. 强噪声下基于 ACYCBD-MTF-Mobi- leViT 的轴承故障诊断研究[J]. 振动与冲击, 2024, 43(24): 34-47.

[24] 张搏文, 庞新宇, 程宝安, 等. 基于 PIRD-CNN 的航空发动机轴承故障诊断方法研究[J]. 振动与沛击, 2024, 43(18):201-207＋231.

[25] 刘飞, 除仁文, 邢凯玲, 等. 基于䢌移学习与深度残差网络的滚动轴承快速故障诊断算法[J]. 振动与冲击, 2022, 41(3): 154-164.

[26] 姜波舟, 张搏文, 梁亚刚, 等. 基于位移信息多尺度特訨提取的航空发动机轴承智能敋障诊断方法研究[J]. 航空维修与工程, 2024,(8):23-27,doi:10.19302/j.cnki.1672-0982.2024.08.025.

基金资助

辽宁省科技重大专项项目(2024JH1/1170043)