DOI QR코드

DOI QR Code

Structural health monitoring data anomaly detection by transformer enhanced densely connected neural networks

  • Jun, Li (Centre for Infrastructure Monitoring and Protection, School of Civil and Mechanical Engineering, Curtin University) ;
  • Wupeng, Chen (School of Civil Engineering, Guangzhou University) ;
  • Gao, Fan (School of Civil Engineering, Guangzhou University)
  • 투고 : 2022.06.11
  • 심사 : 2022.10.30
  • 발행 : 2022.12.25

초록

Guaranteeing the quality and integrity of structural health monitoring (SHM) data is very important for an effective assessment of structural condition. However, sensory system may malfunction due to sensor fault or harsh operational environment, resulting in multiple types of data anomaly existing in the measured data. Efficiently and automatically identifying anomalies from the vast amounts of measured data is significant for assessing the structural conditions and early warning for structural failure in SHM. The major challenges of current automated data anomaly detection methods are the imbalance of dataset categories. In terms of the feature of actual anomalous data, this paper proposes a data anomaly detection method based on data-level and deep learning technique for SHM of civil engineering structures. The proposed method consists of a data balancing phase to prepare a comprehensive training dataset based on data-level technique, and an anomaly detection phase based on a sophisticatedly designed network. The advanced densely connected convolutional network (DenseNet) and Transformer encoder are embedded in the specific network to facilitate extraction of both detail and global features of response data, and to establish the mapping between the highest level of abstractive features and data anomaly class. Numerical studies on a steel frame model are conducted to evaluate the performance and noise immunity of using the proposed network for data anomaly detection. The applicability of the proposed method for data anomaly classification is validated with the measured data of a practical supertall structure. The proposed method presents a remarkable performance on data anomaly detection, which reaches a 95.7% overall accuracy with practical engineering structural monitoring data, which demonstrates the effectiveness of data balancing and the robust classification capability of the proposed network.

키워드

과제정보

The support from the National Natural Science Foundation of China Project No. 52178279 and Guangzhou Basic and Applied Basic Research Foundation project, is acknowledged.

참고문헌

  1. Abdelghani, M. and Friswell, M.I. (2004), "Sensor validation for structural systems with additive sensor faults", Struct. Health Monitor., 3(3), 265-275. https://doi.org/10.1177/1475921704045627
  2. Arul, M. and Kareem, A. (2020), "Data anomaly detection for structural health monitoring of bridges using shapelet transform". https://doi.org/10.48550/arXiv.2009.00470
  3. Bao, Y., Tang, Z., Li, H. and Zhang, Y. (2018), "Computer vision and deep learning-based data anomaly detection method for structural health monitoring", Struct. Health Monitor., 18(2), 401-421. https://doi.org/10.1177/1475921718757405
  4. Bao, Y., Chen, Z., Wei, S., Xu, Y., Tang, Z. and Li, H. (2019), "The state of the art of data science and engineering in structural health monitoring", Engineering, 5(2), 234-242. https://doi.org/10.1016/j.eng.2018.11.027
  5. Chen, W.H., Lu, Z.R., Lin, W., Chen, S.H., Ni, Y.Q., Xia, Y. and Liao, W.Y. (2011), "Theoretical and experimental modal analysis of the Guangzhou New TV Tower", Eng. Struct., 33(12), 3628-3646. https://doi.org/10.1016/j.engstruct.2011.07.028
  6. Chenglin, Z., Xuebin, S., Songlin, S. and Ting, J. (2011), "Fault diagnosis of sensor by chaos particle swarm optimization algorithm and support vector machine", Expert Syst. Applicat., 38(8), 9908-9912. https://doi.org/10.1016/j.eswa.2011.02.043
  7. Cross, E.J., Koo, K.Y., Brownjohn, J.M.W. and Worden, K. (2013), "Long-term monitoring and data analysis of the Tamar Bridge", Mech. Syst. Signal Process., 35(1), 16-34. https://doi.org/10.1016/j.ymssp.2012.08.026
  8. Fan, G., Li, J. and Hao, H. (2019), "Lost data recovery for structural health monitoring based on convolutional neural networks", Struct. Control Health Monitor., 26(10), p. e2433. https://doi.org/10.1002/stc.2433
  9. Fan, G., Li, J. and Hao, H. (2021a), "Dynamic response reconstruction for structural health monitoring using densely connected convolutional networks", Struct. Health Monitor., 20(4), 1373-1391. https://doi.org/10.1177/1475921720916881
  10. Fan, G., Li, J., Hao, H. and Xin, Y. (2021b), "Data driven structural dynamic response reconstruction using segment based generative adversarial networks", Eng. Struct., 234, 111970. https://doi.org/10.1016/j.engstruct.2021.111970
  11. Fu, Y., Peng, C., Gomez, F., Narazaki, Y. and Spencer Jr, B.F. (2019), "Sensor fault management techniques for wireless smart sensor networks in structural health monitoring", Struct. Control Health Monitor., 26(7), p. e2362. https://doi.org/10.1002/stc.2362
  12. Hou, J., Jiang, H., Wan, C., Yi, L., Gao, S., Ding, Y. and Xue, S. (2022), "Deep learning and data augmentation based data imputation for structural health monitoring system in multi-sensor damaged state", Measurement, 196, 111206. https://doi.org/10.1016/j.measurement.2022.111206
  13. Huang, Y., Beck, J.L., Wu, S. and Li, H. (2016), "Bayesian compressive sensing for approximately sparse signals and application to structural health monitoring signals for data loss recovery", Probabil. Eng. Mech., 46, 62-79. https://doi.org/10.1016/j.probengmech.2016.08.001
  14. Huang, G., Liu, Z., Van Der Maaten, L. and Weinberger, K.Q. (2017), "Densely connected convolutional networks", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, July.
  15. Ibarguengoytia, P.H., Sucar, L.E. and Vadera, S. (2007), "Real time intelligent sensor validation", IEEE Power Eng. Rev., 21(9), 63-64. https://doi.org/10.1109/MPER.2001.4311630
  16. Kerschen, G., De Boe, P., Golinval, J.C. and Worden, K. (2004), "Sensor validation using principal component analysis", Smart Mater. Struct., 14(1), p. 36. https://doi.org/10.1088/0964-1726/14/1/004
  17. Krawczyk, B. (2016), "Learning from imbalanced data: open challenges and future directions", Progress Artif. Intell., 5(4), 221-232. https://doi.org/10.1007/s13748-016-0094-0
  18. Kullaa, J. (2013), "Detection, identification, and quantification of sensor fault in a sensor network", Mech. Syst. Signal Process., 40(1), 208-221. https://doi.org/10.1016/j.ymssp.2013.05.007
  19. Lei, X., Xia, Y., Wang, A., Jian, X., Zhong, H. and Sun, L. (2023), "Mutual information based anomaly detection of monitoring data with attention mechanism and residual learning", Mech. Syst. Signal Process., 182, 109607. https://doi.org/10.1016/j.ymssp.2022.109607
  20. Lin, Y.Z., Nie, Z.H. and Ma, H.W. (2017), "Structural damage detection with automatic feature-extraction through deep learning", Comput.-Aided Civil Infrastr. Eng., 32(12), 1025-1046. https://doi.org/10.1111/mice.12313
  21. Lo, C., Bai, Y., Liu, M. and Lynch, J.P. (2015), "Efficient Sensor Fault Detection Using Group Testing", ArXiv, abs/1501.04152. https://doi.org/10.1109/DCOSS.2013.57
  22. Mao, J., Wang, H. and Spencer Jr, B.F. (2020), "Toward data anomaly detection for automated structural health monitoring: Exploiting generative adversarial nets and autoencoders", Struct. Health Monitor., 20(4), 1609-1626. https://doi.org/10.1177/1475921720924601
  23. Ni, Y.Q., Xia, Y., Liao, W.Y. and Ko, J.M. (2009), "Technology innovation in developing the structural health monitoring system for Guangzhou New TV Tower", Struct. Control Health Monitor., 16(1), 73-98. https://doi.org/10.1002/stc.303
  24. Rabatel, J., Bringay, S. and Poncelet, P. (2011), "Anomaly detection in monitoring sensor data for preventive maintenance", Expert Syst. Applicat., 38(6), 7003-7015. https://doi.org/10.1016/j.eswa.2010.12.014
  25. Smarsly, K. and Law, K.H. (2014), "Decentralized fault detection and isolation in wireless structural health monitoring systems using analytical redundancy", Adv. Eng. Software, 73, 1-10. https://doi.org/10.1016/j.advengsoft.2014.02.005
  26. Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I. and Salakhutdinov, R. (2014), "Dropout: a simple way to prevent neural networks from overfitting", J. Mach. Learn. Res., 15(1), 1929-1958.
  27. Tang, Z., Chen, Z., Bao, Y. and Li, H. (2019), "Convolutional neural network-based data anomaly detection method using multiple information for structural health monitoring", Struct. Control Health Monitor., 26(1), p. e2296. https://doi.org/10.1002/stc.2296
  28. Thiyagarajan, K., Kodagoda, S. and Van Nguyen, L. (2017), "Predictive analytics for detecting sensor failure using autoregressive integrated moving average model", Proceedings of 2017 12th IEEE Conference on Industrial Electronics and Applications (ICIEA), Siem Reap, Cambodia, June.
  29. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, L. and Polosukhin, I. (2017), "Attention is all you need", arXiv. https://doi.org/10.48550/arXiv.1706.03762
  30. Wan, H.P. and Ni, Y.Q. (2018), "Bayesian modeling approach for forecast of structural stress response using structural health monitoring data", J. Struct. Eng., 144(9). p. 04018130. https://doi.org/10.1061/(asce)st.1943-541x.0002085
  31. Wu, B., Huang, Y. and Li, H. (2015), Sparse reconstruction of flaw signal from noisy ultrasonic data: A bayesian framework.
  32. Xia, Y., Chen, B., Zhou, X.Q. and Xu, Y.L. (2013), "Field monitoring and numerical analysis of Tsing Ma Suspension Bridge temperature behavior", Struct. Control Health Monitor., 20(4), 560-575. https://doi.org/10.1002/stc.515
  33. Yi, T.H., Li, H.N., Song, G. and Guo, Q. (2016), "Detection of shifts in GPS measurements for a long-span bridge using CUSUM chart", Int. J. Struct. Stabil. Dyn., 16(04), p. 1640024. https://doi.org/10.1142/s0219455416400241
  34. Yuen, K.V. and Mu, H.Q. (2012), "A novel probabilistic method for robust parametric identification and outlier detection", Probabil. Eng. Mech., 30, 48-59. https://doi.org/10.1016/j.probengmech.2012.06.002