Among vibration-based structural health monitoring (SHM), methods based on modal analysis are the most popular and efficient ones for assessing the inherent states of structures. The primary techniques involve tracking changes in modal parameters regularly during the monitoring period. Numerous studies have utilized both modal frequencies and mode shapes for damage detection, as mode shapes help identify damage locations, and the combination is more robust against environmental effects. Obviously, recent advances in machine learning (ML) and artificial intelligence (AI) have revolutionized SHM based on modal analysis. This study aims to develop a clustering technique using variational autoencoder (VAE) for classifying structural damages. To fully expand the entangled spaces that may arise from lower-dimensional representation, the loss function incorporates triplet loss (known as triplet VAE or tVAE) while preserving all dimensions in the encoded space. The proposed approach is thoroughly developed and evaluated using both numerical simulations and experimental investigations of multistory building structures. The effectiveness of detecting changes via VAE-based and tVAE-based modal clustering has been demonstrated under the consideration of environmental variations and uncertainties. Additionally, a procedure is suggested for the field applications and the proposed approach’s capability for structural inspection during long-term monitoring is showcased using a practical scenario.
BoscatoGFragonaraLZCecchiA, et al. Structural health monitoring through vibration-based approaches. Shock Vibration2019: 1–5.
2.
SonySLaventureSSadhuA.A literature review of next-generation smart sensing technology in structural health monitoring. Struct Control Health Monit2019; 26(3): e2321.
3.
SajediSOLiangX.Vibration-based semantic damage segmentation for large-scale structural health monitoring. Comput Aided Civ Infrastruct Eng2020; 35(6): 579–596.
4.
HassaniSDackermannU.A systematic review of advanced sensor technologies for non-destructive testing and structural health monitoring. Sensors2023; 23(4): 2204.
5.
SarmadiHEntezamiAYuenKV, et al. Review on smartphone sensing technology for structural health monitoring. Measurement2023; 223: 113716.
6.
HuangSKChiFC.Development of recursive subspace identification for real-time structural health monitoring under seismic loading. Struct Control Health Monit2023; 2023(1): 1117042.
7.
HuangSKChenJQWengYT, et al. Tracking modal parameters of structures online using recursive stochastic subspace identification under ambient excitations. Buildings2024; 14(4): 964.
8.
WickramasingheWRThambiratnamDPChanTH.Damage detection in a suspension bridge using modal flexibility method. Eng Fail Anal2020; 107: 104194.
9.
AbdulkareemMBakharyNVafaeiM, et al. Application of two-dimensional wavelet transform to detect damage in steel plate structures. Measurement2019; 146: 912–923.
10.
HouRXiaYZhouX.Structural damage detection based on l1 regularization using natural frequencies and mode shapes. Struct Control Health Monit2018; 25(3): e2107.
11.
OzdagliAIKoutsoukosX.Machine learning based novelty detection using modal analysis. Comput Aided Civ Infrastruct Eng2019; 34(12): 1119–1140.
12.
BovsunovskyASuraceC.Structural damage detection based on features insensitive to ambient factors. Theoretical Appl Fract Mech2020; 110: 102780.
13.
FlahMNunezIChaabeneWB, et al. Machine learning algorithms in civil structural health monitoring: a systematic review. Archives Comput Methods Eng2021; 28(4): 2621–2643.
14.
EltounyKGomaaMLiangX.Unsupervised learning methods for data-driven vibration-based structural health monitoring: a review. Sensors2023; 23(6): 3290.
15.
SiowPYOngZCKhooSY, et al. Hybrid machine learning with mode shape assessment for damage identification of plates. Smart Struct Syst2023; 31(5): 485–500.
16.
SabatoADabetwarSKulkarniNN, et al. Noncontact sensing techniques for AI-aided structural health monitoring: a systematic review. IEEE Sens J2023; 23(5): 4672–4684.
17.
AnjumAHrairiMAabidA, et al. Civil structural health monitoring and machine learning: a comprehensive review. Frattura ed Integrità Strutturale2024; 18(69): 43–59.
18.
AlvesVCuryARoitmanN, et al. Structural modification assessment using supervised learning methods applied to vibration data. Eng Struct2015; 99: 439–448.
19.
OhBKKimKJKimY, et al. Evolutionary learning based sustainable strain sensing model for structural health monitoring of high-rise buildings. Appl Soft Comput2017; 58: 576–585.
20.
SajediSOLiangX.A data-driven framework for near real-time and robust damage diagnosis of building structures. Struct Control Health Monit2020; 27(3): e2488.
HuangYBeckJLLiH.Bayesian system identification based on hierarchical sparse Bayesian learning and Gibbs sampling with application to structural damage assessment. Comput Methods Appl Mech Eng2017; 318: 382–411.
23.
HuangYBeckJLLiH.Hierarchical sparse Bayesian learning for structural damage detection: theory, computation and application. Struct Saf2017; 64: 37–53.
24.
WangQADaiYMaZG, et al. Towards probabilistic data-driven damage detection in SHM using sparse Bayesian learning scheme. Struct Control Health Monit2022; 29(11): e3070.
25.
EltounyKSajediSLiangX.High-resolution vision transformers for pixel-level identification of structural components and damage. arXiv preprint arXiv:2308. 03006, 2023.
26.
WordenK.Structural fault detection using a novelty measure. J Sound Vibration1997; 201(1): 85–101.
27.
ZhouHFNiYQKoJM.Structural damage alarming using auto-associative neural network technique: exploration of environment-tolerant capacity and setup of alarming threshold. Mech Syst Signal Process2011; 25(5): 1508–1526.
28.
SilikANooriMGhiasiR, et al. Dynamic wavelet neural network model for damage features extraction and patterns recognition. J Civil Struct Health Monit2023; 13(4): 925–945.
29.
HsuTYLohCH.Damage detection accommodating nonlinear environmental effects by nonlinear principal component analysis. Struct Control Health Monit2010; 17(3): 338–354.
30.
LohCHChenCHHsuTY.Application of advanced statistical methods for extracting long-term trends in static monitoring data from an arch dam. Struct Health Monit2011; 10(6): 587–601.
31.
YuanZZhuSChangC, et al. An unsupervised method based on convolutional variational auto-encoder and anomaly detection algorithms for light rail squat localization. Constr Build Mater2021; 313: 125563.
32.
MaXLinYNieZ, et al. Structural damage identification based on unsupervised feature-extraction via variational auto-encoder. Measurement2020; 160: 107811.
33.
KimMSongJ.Near-real-time identification of seismic damage using unsupervised deep neural network. J Eng Mech2022; 148(3): 04022006.
34.
ZhangYXieXLiH, et al. An unsupervised tunnel damage identification method based on convolutional variational auto-encoder and wavelet packet analysis. Sensors2022; 22(6): 2412.
35.
PollastroATestaGBilottaA, et al. Semi-supervised detection of structural damage using variational autoencoder and a one-class support vector machine. In: IEEE access, 2023.
36.
CoraçaEMFerreiraJVNóbregaEG.An unsupervised structural health monitoring framework based on Variational Autoencoders and Hidden Markov Models. Reliab Eng Syst Saf2023; 231: 109025.
37.
KramerMA.Nonlinear principal component analysis using autoassociative neural networks. AIChE J1991; 37(2): 233–243.
EL DahlinA. Attribute embedding for variational auto-encoders: regularization derived from triplet loss. In: KTH, School of electrical engineering and computer science (EECS), 2022.
42.
HofferEAilonN.Deep metric learning using triplet network. In: Similarity-based pattern recognition: third international workshop, SIMBAD 2015, Copenhagen, Denmark, October 12–14, 2015. Proceedings 3 (pp. 84–92). Springer International Publishing.
NguyenBMorellCDe BaetsB.Distance metric learning for ordinal classification based on triplet constraints. Knowledge-Based Syst2018; 142: 17–28.
45.
WangYWuXMLiQ, et al. Large margin meta-learning for few-shot classification. In: Workshop on Meta-Learning (MetaLearn 2018)@ NeurIPS 2018. Neural Information Processing Systems Foundation, 2018.
46.
WangYWuXMLiQ, et al. Large margin few-shot learning. arXiv preprint arXiv:1807.02872, 2018.
ShiWShiDOrouskhaniM, et al. Adaptive few-shot deep metric learning. Int J Electr Inform Eng2021; 15(7): 289–295.
49.
ÖztürkŞÇelikEÇukurT. Content-based medical image retrieval with opponent class adaptive margin loss. Inform Sci2023; 637: 118938.
50.
Van OverscheePDe MoorB. N4SID: subspace algorithms for the identification of combined deterministic-stochastic systems. Automatica1994; 30(1): 75–93.
51.
Van OverscheePDe MoorB. Subspace identification for linear systems: theory—implementation—applications. Berlin, Germany: Springer Science & Business Media, 2012.
52.
ZhangPHeZCuiC, et al. Operational modal analysis of offshore wind turbine tower under ambient excitation. J Marine Sci Eng2022; 10(12): 1963.
53.
LiHBuSWenJR, et al. Synthetical modal parameters identification method of damped oscillation signals in power system. Appl Sci2022; 12(9): 4668.
VerhaegenM.Identification of the deterministic part of MIMO state space models given in innovations form from input-output data. Automatica1994; 30(1): 61–74.
56.
FarrarCRBakerWEBellTM, et al. Dynamic characterization and damage detection in the I-40 bridge over the Rio Grande (No. LA-12767-MS). Los Alamos National Lab.(LANL), Los Alamos, NM, USA, 1994.
57.
KhandanAMSumnerESeracinoR, et al. Effect of temperature on modal characteristics of steel-concrete composite bridges: field testing. In: 4th International conference on structural health monitoring on intelligent infrastructure (SHMII-4), Zurich, Switzerland, 2009.
58.
NandanHSinghMP.Effects of thermal environment on structural frequencies: part I–A simulation study. Eng Struct2014; 81: 480–490.
59.
TitsNWangFHaddadKE, et al. Visualization and interpretation of latent spaces for controlling expressive speech synthesis through audio analysis. arXiv preprint arXiv:1903.11570, 2019.
60.
CaicedoJMDykeSJJohnsonEA.Natural excitation technique and eigensystem realization algorithm for phase I of the IASC-ASCE benchmark problem: simulated data. J Eng Mech2004; 130(1): 49–60.
61.
ChenJDLohCH.Tracking modal parameters of building structures from experimental studies and earthquake response measurements. Struct Health Monit2017; 16(5): 551–567.
62.
ChenJDLohCH.Two-stage damage detection algorithms of structure using modal parameters identified from recursive subspace identification. Earthquake Eng Struct Dyn2018; 47(3): 573–593.
