The rapid growth of sensing technologies has enabled large-scale deployment of structural health monitoring (SHM) systems and produced massive data streams. However, sensor malfunctions and environmental disturbances introduce anomalies that impede reliable structural assessment. This paper proposes a transfer learning (TL)-based data anomaly detection method using vision transformer (ViT) model. This methodology can effectively identify various types of anomalies while requiring only minimal labeled data. Furthermore, the integration of the ViT model with the class anchor clustering open set recognition method addresses the issue of incomplete training sample categories. The method begins with the generation of acceleration data through numerical simulation, followed by data augmentation techniques to construct 10 representative types of anomalies that reflect common time-domain anomaly patterns. These time-series signals are then transformed into scatter-time images and used as input to pre-train a benchmark and transferable ViT for anomaly detection (BAT-ViT). Subsequently, a small amount of labeled data from the target bridge is used to fine-tune the BAT-ViT for real-world application. To demonstrate the robustness and transferability of the proposed method, SHM data from two different large-span bridges are utilized for validation across bridge structures. The results demonstrate that with only 100 labeled data samples from the actual bridge, the fine-tuned model achieves an accuracy of 95.6% through TL, significantly outperforming conventional convolutional neural network and long short-term memory models. Moreover, the model also achieves over 92% accuracy in open set recognition, even when using a limited transferred dataset. The method delivers robust and adaptable performance, especially when training data are scarce or resolution settings vary, and greatly reduces the need for manual data labeling.
ZhangCWMousaviAAMasriSF, et al. Vibration feature extraction using signal processing techniques for structural health monitoring: a review. Mech Syst Signal Process2022; 177: 109175.
2.
GongFLeiXXiaY.Real-time damage identification of hinge joints in multi-girder bridges using recursive least squares solution of the characteristic equation. Eng Struct2024; 300: 117147
3.
DanDZhengWXuZ.Research on monitoring index of transverse cooperative working performance of assembled multi-girder bridges based on displacement spectrum similarity measure. Structures2023; 48: 1322–1332.
4.
FanCCZhengYXWangBL, et al. Damage identification method for tied arch bridge suspender based on quasi-static displacement influence line. Mech Syst Signal Process2023; 200: 110518
5.
GongFXiaYLozanoF, et al. Experimental analysis of structural nonlinear damping ratio induced by bolt joint friction. Eng Struct2024; 314: 118220
6.
XiaYGongFZShangZQ, et al. Damping ratio analysis of cable-stayed bridges based on attenuation response pattern extraction from monitoring data. China J Highw Transp2022; 36(7): 90–101.
7.
MaTWangNHanCJ, et al. Asphalt pavement residual life assessment based on smartrock sensors. IEEE Trans Instrum Meas2023; 72: 3267372
8.
JianXXiaYChatziE, et al. Bridge influence surface identification using a deep multilayer perceptron and computer vision techniques. Struct Health Monit2024; 23(3): 1606–1626.
9.
KaramiKAminiF.Decreasing the damage in smart structures using integrated online DDA/ISMP and semi-active control. Smart Mater Struct2012; 21(10): 105017.
10.
DengYJuHWZhongGQ, et al. A general data quality evaluation framework for dynamic response monitoring of long-span bridges. Mech Syst Signal Process2023; 200: 110514.
11.
ZhangLYSunLMShangZQ.Real-time reliability assessment based on acceleration monitoring for bridge. Sci China Technol Sci2016; 59(8): 1294–1304.
12.
MattsonSGPanditSM.Statistical moments of autoregressive model residuals for damage localisation. Mech Syst Signal Proc2006; 20(3): 627–645.
13.
ZhouYSunLM.Effects of high winds on a long-span sea-crossing bridge based on structural health monitoring. J Wind Eng Ind Aerodyn2018; 174: 260–268.
14.
CuryACrémonaCDidayE.Application of symbolic data analysis for structural modification assessment. Eng Struct2010; 32(3): 762–775.
15.
XiaYWangPSunLM.Neutral axis-based health monitoring and condition assessment techniques for concrete box girder bridges. Int J Struct Stab Dyn2019; 19(1): 17.
16.
FengYLiuJLiuW.Coincidence of the rao test, wald test and GLRT for anomaly detection in hyperspectral imagery. Signal Process2019; 169(7): 107416.
17.
JungISBergesMGarrettJH, et al. Exploration and evaluation of AR, MPCA and KL anomaly detection techniques to embankment dam piezometer data. Adv Eng Inf2015; 29(4): 902–917.
18.
AkouemoHNPovinelliRJ.Probabilistic anomaly detection in natural gas time series data. Int J Forecast2016; 32(3): 948–956.
19.
ZhangMGuoJLiX, et al. Data-driven anomaly detection approach for time-series streaming data. Sensors2020; 20(19): 5646.
20.
FraktABKarlWC.A multiscale hypothesis testing approach to anomaly detection and localization from noisy tomographic data. IEEE Trans Image Process1998; 7(6): 825–837.
21.
ZhangYWangXDingZ, et al. Anomaly detection of sensor faults and extreme events based on support vector data description. Struct Control Health Monit2022; 29(10): 3047.
22.
JianXZhongHXiaY, et al. Faulty data detection and classification for bridge structural health monitoring via statistical and deep-learning approach. Struct Control Health Monit2021; 28: e2824.
23.
SunLShangZXiaY, et al. Review of bridge structural health monitoring aided by big data and artificial intelligence: from condition assessment to damage detection. J Struct Eng2020; 146(5): 22.
24.
DuniaRQinSJEdgarTF, et al. Identification of faulty sensors using principal component analysis. Aiche J1996; 42(10): 2797–2812.
25.
FriswellMIInmanDJ.Sensor validation for smart structures. J Intell Mater Syst Struct1999; 10(12): 973–982.
26.
ChengZWangSZhangP, et al. Improved autoencoder for unsupervised anomaly detection. Int J Intell Syst2021; 36(12): 7103–7125.
27.
MohamedM.Comparative evaluation of VAEs, VAE-GANs and AAEs for anomaly detection in network intrusion data. EMITTER Int J Eng Technol2023; 11(2): 160–173.
28.
KerschenGDe BoePGolinvalJC, et al. Sensor validation using principal component analysis. Smart Mater Struct2005; 14(1): 36–42.
29.
MaoJXWangHSpencerBF.Toward data anomaly detection for automated structural health monitoring: exploiting generative adversarial nets and autoencoders. Struct Heal Monit2021; 20(4): 1609–1626.
30.
RumelhartDEHintonGEWilliamsRJ.Learning representations by back propagating errors. Nature1986; 323(6088): 533–536.
31.
KrizhevskyASutskeverIHintonG.ImageNet classification with deep convolutional neural networks. Commun ACM. 2017; 60(6): 84–90.
ZhangHLinJHuaJ, et al. Data anomaly detection for bridge SHM based on CNN combined with statistic features. J Nondestruct Eval2022; 41(1): 1–13.
34.
DengYJuHZhongG, et al. Data quality evaluation for bridge structural health monitoring based on deep learning and frequency-domain information. Struct Health Monit2023; 22(5): 23.
35.
ZhangJZhangJWuZ.Long-short term memory network-based monitoring data anomaly detection of a long-span suspension bridge. Sensors2022; 22(16): 17.
36.
BaoYTangZLiH, et al. Computer vision and deep learning–based data anomaly detection method for structural health monitoring. Struct Health Monit2018; 8: 401–421.
37.
TangZChenZBaoY, et al. Convolutional neural network-based data anomaly detection method using multiple information for structural health monitoring. Struct Control Health Monit2019; 26(1): e2296.1–e2296.22.
38.
VaswaniAShazeerNParmarN, et al. Attention is all you need. arXiv preprint arXiv:1706.03762, 2017.
39.
ZhouHZhangSPengJ, et al. Informer: beyond efficient transformer for long sequence time-series forecasting. arXiv preprint arXiv:2012.07436, 2020.
40.
DosovitskiyABeyerLKolesnikovA, et al. An image is worth 16x16 words: transformers for image recognition at scale. arXiv preprint arXiv: 2010.11929, 2020.
41.
TangZGuoJWangY, et al. Automated seismic event detection considering faulty data interference using deep learning and Bayesian fusion. Comput Aided Civ Infrastruct Eng2025; 40(14): 1910–1931.
42.
SatyanarayananM.The emergence of edge computing. Computer2017; 50(1): 30–39.
43.
MondalTGChouJYFuYG, et al. A hybrid deep neural network compression approach enabling edge intelligence for data anomaly detection in smart structural health monitoring systems. Smart Struct Syst2023; 32(3): 179–193.
44.
YuXZhaoYQCuiSW, et al. Efficient edge intelligence for onboard data anomaly classification in wireless structural health monitoring using knowledge distillation on low-cost IoT nodes. Struct Health Monit2025: 20.
45.
PanSJYangQA.A survey on transfer learning. IEEE Trans Knowl Data Eng2010; 22(10): 1345–1359.
46.
PanQBaoYLiH.Transfer learning-based data anomaly detection for structural health monitoring. Struct Health Monit2023; 22(5): 3077–3091.
47.
DeloGBunceACrossEJ, et al. When is a bridge not an aeroplane? Part II: a population of real structures. In: RizzoPMilazzoA (eds) European workshop on structural health monitoring. China: Springer International Publishing, 2023, pp. 965–974.
48.
YanoMOFigueiredoEda SilvaS, et al. Foundations and applicability of transfer learning for structural health monitoring of bridges. Mech Syst Signal Process2023; 204: 110766.
49.
GengCHuangSChenS.Recent advances in open set recognition: a survey. IEEE Trans Pattern Anal Mach Intell2020; 43(10): 3614–3631.
50.
BendaleABoultTE.Towards open set deep networks. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 27–30 Jun 2016, Las Vegas, NV, USA pp. 1563–1572. IEEE.
51.
KimYNChoJSKimJK, et al. Development of open set recognition-based multiple damage recognition model for bridge structure damage detection. KSCE J Civ Environ Eng Res2022; 42(1): 117–126.
52.
LiSXuanJZhangQ, et al. Open set transfer learning for bearing defect recognition based on selective momentum contrast and dual adversarial structure. Adv Eng Inf2024; 62: 102641.
53.
DengYZhaoYJJuHW, et al. Abnormal data detection for structural health monitoring: state-of-the-art review. Develop Built Environ2024; 17: 100337.
54.
PanSJTsangIWKwokJT, et al. Domain adaptation via transfer component analysis. IEEE Trans Neural Netw Learn Syst2011; 22(2): 199–210.
55.
Class anchor clustering: a loss for distance-based open set recognition. In: 2021 IEEE winter conference on applications of computer vision (WACV), Waikoloa, HI, USA, 05–09 Jan 2021, pp. 3569–3577.
