基于VMD-CAE的无监督结构损伤识别研究

PDF(3643 KB)

振动与冲击 ›› 2025, Vol. 44 ›› Issue (11) : 309-320.

故障诊断分析

基于VMD-CAE的无监督结构损伤识别研究

王梦倩1，康帅*1，李传飞2,董正方1

作者信息 +

Unsupervised structural damage identification based on VMD-CAE

WANG Mengqian1, KANG Shuai*1, LI Chuanfei2, DONG Zhengfang1

Author information +

文章历史 +

摘要

为了进一步扩展深度学习方法在基于振动信号的结构损伤识别中的应用，提出了一种基于变分模态分解(Variational Modal Decomposition,VMD)和卷积自编码(Convolutional Auto-encoder,CAE)相结合的无监督结构损伤识别方法。首先，利用VMD对振动信号进行分解，去除噪声和一些无关成分的影响，选取与结构自振特性相关的成分作为有效分量；然后通过叠加有效分量作为CAE模型的输入，进而重构信号，通过学习健康样本数据的特征，得到最大重构误差作为判断结构是否损坏的阈值。最后将该方法应用到IASC-ASCE SHM Benchmark结构试验数据和卡塔尔大学看台试验数据，并将结果与其他模型进行了对比，结果表明该方法在两个数据集上的识别结果都更加准确。即使当样本中含有噪声时，也能显著提高噪声样本的识别精度，具有较强的抗噪能力。

Abstract

To further expand the application of deep learning methods in structural damage identification based on vibration signals, a method based on variational modal decomposition (VMD) and convolutional auto-encoder (CAE) is proposed. An unsupervised structural damage identification method. First, use VMD to decompose the vibration signal, remove the influence of noise and some irrelevant components, and select components related to the natural vibration characteristics of the structure as effective components; then superimpose the effective components as the input of the CAE model, and then reconstruct the signal, and through learning Based on the characteristics of healthy sample data, the maximum reconstruction error is obtained as the threshold for judging whether the structure is damaged. Finally, the method was applied to the IASC-ASCE SHM Benchmark structural test data and the Qatar University stand test data, and the results were compared with other models. The results showed that the recognition results of this method were more accurate on both data sets. Even when the sample contains noise, it can significantly improve the recognition accuracy of noise samples and has strong anti-noise ability.

导出引用

王梦倩1, 康帅1, 李传飞2, 董正方1. 基于VMD-CAE的无监督结构损伤识别研究[J]. 振动与冲击, 2025, 44(11): 309-320

WANG Mengqian1, KANG Shuai1, LI Chuanfei2, DONG Zhengfang1. Unsupervised structural damage identification based on VMD-CAE[J]. Journal of Vibration and Shock, 2025, 44(11): 309-320

参考文献

[1] Lecun Y, Bottou L, Bengio Y, et al. Gracicnt-base learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.
[2] 骆勇鹏, 王林堃, 廖飞宇, 等. 基于一维卷积神经网络的结构损伤识别[J]. 地震工程与工程振动, 2021, 41(4): 145-156.
LUO Yong-peng, WANG Lin-kun, LIAO Fei-yu, et al. Vibration-based structural damage identification by 1-dimensional convolutional neural network[J]. Earthquake Engineering andEngineering Dynamics, 2021, 41(4): 145-156.
[3] 罗旭欣, 陈龙, 梁韬, 等. 基于迁移卷积神经网络的桥梁结构损伤识别方法[J]. 铁道科学与工程学报, 2024.
LUO Xu-xin, CHEN Long, LIANG Tao, et al. Structural damage identification forbridges based on transfer learning and 1-D convolutional neural network[J]. Journal of Railway Science and Engineering, 2024.
[4] 李雪松, 马宏伟, 林逸洲. 基于卷积神经网络的结构损伤识别[J]. 振动与冲击, 2019, 38(1):159-167.
LI Xue-song, MA Hong-wei, LIN Yi-zhou. Structural damage identification based on convolutionneural network[J]. Journal of Vibration and Shock, 38(1): 159-167.
[5] Teng S, Liu A R, Chen B C, et al. Bridge progressive damage detection using unsupervised learning and self-attention mechanism[J]. Engineering Structures, 2024, 301: 117278.
[6] Ni F T, Zhang J, Noori M N. Deep learning for data anomaly detection and data compression of a long-span suspension bridge[J]. Computer-Aided Civil and Infrastructure Engineerin, 2020,35(7), pp.685-700.
[7] Pathirage C S N, Li J, Li L, et al. Structural damage identification based on autoencoder neural networks and deep learning[J]. Engineering Structures, 2018, 172: 13-28.
[8] Wang Z, Cha Y J. Unsupervised deep learning approach using a deep auto-encoder with a one-class support vector machine to detect damage[J]. Structural Health Monitoring, 2021, 20: 406-425.
[9] Finotti R P, Barbosa F S, Cury A A, et al. Novelty Detection Using SparseAuto-Encoders to Characterize Structural Vibration Responses[J]. Arabian Journal for Science and Engineering, 2022, 47: 13049-13062.
[10] Lee K, Jeong S, Sim S, et al. Field experiment on a PSC-I bridge for convolutional autoe ncoder-based damage detection[J]. Structural Health Monitoring, 2021, 20: 1627-1643.
[11] Rautela M, Senthilnath J, Monaco E, et al. Delamination prediction in composite panels using unsupervised-featurelearning methods with wavelet-enhanced guided wave represen tations[J]. Composite Structures, 2022, 291: 115579.
[12] Li T, Hou H, Zheng K K, et al. Automated method for structural modal identification based on multivariate variational mode decomposition and its applications in damage characteristics of subway tunnels[J]. Engineering Failure Analysis, 2024, 163: 108499.
[13] 王盟, 翁顺, 余兴胜, 等. 基于时变模态振型小波变换的结构损伤识别方法[J]. 振动与冲击, 2021, 40(16): 10-19.
WANG Meng, WENG Shun, YU Xing-sheng, et al. Structural damage identification based on time-varying modal mode shape of wavelet transformation[J]. Journal of Vibration and Shock, 2021, 40(16): 10-19.
[14] 刘习军, 王正飞, 张素侠. 基于振动响应相关性的简支梁桥损伤识别方法[J]. 实验力学, 2019, 34(01): 29-37.
LIU Xi-jun, WENG Zheng-fei, ZHANG Su-xia. On the damage identification method of simply-supported girder bridge based on vibration response correlation[J]. Journal of Experimental Mechanics, 2019, 34(01): 29-37.
[15] Huang N E, Shen Z, Long S R, et al. The empirical mode decomposition and the Hubert spectrum for nonlinear and non-stationary time series analysis[J]. PROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES, 1998, 454: 903-995.
[16] Bagheri A, Ozbulut O E, Harris D K.Structural system identification based on variationalmode decomposition[J]. Journal of Sound and Vibration, 2018,417: 182-197.
[17] Zheng X B, Yang Y, Hu N Q, et al. A novel empirical reconstruction Gauss decomposition method and its application in gear fault diagnosis[J]. Mechanical Systems and Signal Processing, 2024, 210: 111174.
[18] Park S, Kim S, Choi J H. Gear faultdiagnosis using transmission error andensembleempirical mode decomposition[J]. Mechanical Systems and Signal Processing, 2018, 108: 262-275.
[19] Chen X H, Cheng G, Shan X L, et al. Research of weak fault feature information extraction of planetary gear based on ensemble empirical mode decomposition and adaptive stochastic resonance[J]. Measurement, 2015, 73:55-67.
[20] Mohanty S, Gupta K K, Raju K S. Hurst based vibro-acoustic feature extraction of bearingusing EMD and VMD[J].Measyrement, 2018, 117: 200-220.
[21] Diao Y S, Lv J D, Wang QX, et al. Structural damage identification based on variationalmode decomposition-Hilbert transform and CNN[J]. Journal ofCivil Structural Health Monitoring, 2023, 13: 1415-1429.
[22] 赵亚军, 窦远明, 张明杰. 基于变分模态分解的模态参数识别研究[J]. 振动与冲击, 2020, 39(2): 115-122.
ZHAO Ya-jun, DOU Yuan-ming, ZHANG Ming-jie. Modal parameter identification based on variational mode decomposition[J]. Journal of Vibration and Shock, 2020, 39(2): 115-122.
[23] 王秋潇. 基于模态响应与深度学习的结构损伤识别研究[D]. 青岛: 青岛理工大学, 2022.
WANG Qiu-xiao. Study on structural damage identification based on modal response and deep learning[D]. Qingdao: Qingdao Technological University, 2022.
[24] Xu Z F, Li C, Yang Y. Fault diagnosis of rolling bearing of wind turbines based on the variational mode decomposition and deep convolutional neural networks[J]. Applied Soft Computing 2020, 95: 106515.
[25] 吕宏政, 陈仁文, 张祥, 等. 基于VMD交叉样本熵的旋翼桨叶故障诊断方法[J]. 电子测量技术, 2019, 42(9): 5.
LV Hong-zheng, CHEN Ren-wen, ZHANG Xiang, et al. Faultdiagnosis method of rotor blade based on VMD and cross-sample entropy[J]. Electronic Measurement Technology, 2019, 42(9):5.
[26] Liao J, Zheng J B, Chen Z B. Research on the Fault Diagnosis Method of an Internal Gear Pump Based on a Convolutional Auto-Encoder and PSO-LSSVM[J]. Sensors, 2022, 22(24): 9841.
[27] 韩敏, 姜涛, 冯守渤. 基于VMD循环随机跳跃状态网络的时间序列长期预测[J]. 控制与决策, 2020, 35(09): 2175-2181.
HAN Min, JIANG Tao, FENG Shou-bo. Long-term prediction of time series based on VMDcycicreservoir with random jumps network[J]. Control and Decision, 2020, 35(09): 2175-2181.
[28] Dragomiretskiy K, Zosso D. Variational mode decomposition[J]. IEEE Transactions OnSignal Process, 2014, 62(03):531-544.
[29] 唐贵基, 王晓龙. 参数优化变分模态分解方法在滚动轴承早期故障诊断中的应用[J]. 西安交通大学学报, 2015, 49(05): 73-81.
TANG Gui-ji, WANG Xiao-long. Parameter Optimized Variational Mode Decomposition Method with Application to lncipient Fault Diagnosis of RollingBearinc[J]. Journal of Xi'an Jiaotong University, 2015, 49(05): 73-81.
[30] 张培霄, 尹晓红, 李少远, 等. 基于VMD-CNN-LSTM的农业大棚园区用电负荷短期预测[J]. 信息与控制, 2024, 53(02): 238-249.
ZHANG Pei-xiao, YIN Xiao-hong, LIShao-yuan, et al. The Short-term Forecasting of Power Load in AgriculturalGreenhouses Based on VMD-CNN-LSTM[J]. Information and Control, 2024,53(02): 238-249.
[31] 杨银枪, 康帅, 王自法, 等. 基于卷积自编码和相关函数的钢框架损伤识别研究[J]. 工业建筑, 2024.
YANG Yin-qiang, KANG Shuai, WANG Zi-fa, et al. Structural Damage Identification based on Convolutional Autoencoder and Correlation Function[J]. Industrial Construction, 2024.
[32] Dyke S J, Bernal D, Beck J, et al. Experimental phase II of the structural health monitoring benchmark problemC]//Proceedings of the l6th ASCE engineering mechanicsconference, 2003.
[33] Ching J, Beck J L. Bayesian analysis of the phase II IASC-ASCE structural health monitoring experimental benchmark data[J]. Journal of Engineering Mechanics, 2004, 130(10): 1233-1244.
[34] 李行健. 基于响应时频图和深度学习的结构损伤识别研究[D]. 青岛: 青岛理工大学,2023.
LI Hang-jian. Study on structural damageidentime based on responsetime frequening diagram and deeplearning[D]. Qingdao: Qingdao Technological University, 2023.
[35] Avci O, Abdeljaber O, Kiranyaz S, etal. Convolutional Neural Networks forReal-Timeand Wireless Damage Detection[J]. Dynamics of Civil Structures,2020, 2: 129-136.
[36] Liu Q, Nie P, Dai H L, et al. Research on the Identification of Bridge Structural Damage Using Variational Mode Decomposition and Convolutional Self-Attention Neural Networks[J]. Applied Sciences, 2023, 13(21): 12082.
[37] Wang C, Pan X, Qi T Y, et al.Damage Identification of Simple Supported Bridges Under Moving Loads Based on Variational Mode Decomposition and Deep Learning[J]. International Journal of Structural Stability and Dynamics, 2024.