Vol. 19, No. 4, pp. 389-402 (2023)
DEEP LEARNING DAMAGE IDENTIFICATION METHOD FOR STEEL-
FRAME BRACING STRUCTURES USING TIME–FREQUENCY ANALYSIS
AND CONVOLUTIONAL NEURAL NETWORKS
Xiao-Jian Han 1, Qi-Bin Cheng 1 and Ling-Kun Chen 2, 3, 4, *
1 College of Civil Engineering, Nanjing Tech University, Nanjing, 211800 Jiangsu, China
2 College of Civil Science and Engineering, Yangzhou University, Yangzhou, 225127 Jiangsu, China
3 Department of Civil and Environmental Engineering, University of California, Los Angeles, CA 90095 USA
4 School of Civil Engineering, Southwest Jiaotong University, Chengdu, 610031 Sichuan, China
*(Corresponding author: E-mail:This email address is being protected from spambots. You need JavaScript enabled to view it.)
Received: 18 July 2022; Revised: 15 June 2023; Accepted: 27 August 2023
DOI:10.18057/IJASC.2023.19.4.8
 View Article |
Export Citation: Plain Text | RIS | Endnote |
ABSTRACT
Lattice bracing, commonly used in steel construction systems, is vulnerable to damage and failure when subjected to horizontal seismic pressure. To identify damage, manual examination is the conventional method applied. However, this approach is time-consuming and typically unable to detect damage in its early stage. Determining the exact location of damage has been problematic for researchers. Nevertheless, detecting the failure of lateral supports in various parts of a structure using time–frequency analysis and deep learning methods, such as convolutional neural networks, is possible. Then, the damaged structure can be rapidly rebuilt to ensure safety. Experiments are conducted to determine the vibration acceleration modes of a four-storey steel structure considering various support structure damage scenarios. The acceleration signals at each measurement point are then analysed with respect to time and frequency to generate appropriate three-dimensional spectral matrices. In this study, the MobileNetV2 deep learning model was trained on a labelled picture collection of damaged matrix images. Hyperparameter tweaking and training resulted in a prediction accuracy of 97.37% for the complete dataset and 99.30% and 96.23% for the training and testing sets, respectively. The findings indicate that a combination of time–frequency analysis and deep learning methods may pinpoint the position of the damaged steel frame support components more accurately.
KEYWORDS
Damage identification, Bracing system, Deep learning, Convolutional neural networks (CNNs), Time–frequency analysis, MobileNetV2
REFERENCES
[1] H. Shokravi, N. Bakhary, S.R. Koloor, M. Petru, Health Monitoring of Civil Infrastructures by Subspace System Identification Method: An Overview, Applied Sciences. 10 (8) (2020) 1-29.
[2] C. Scuro, P.F. Sciammarella, F. Lamonaca, R.S. Olivito, D.L. Carnì, Iot for Structural Health Monitoring, IEEE Instrumentation and Measurement Magazine.
21 (6) (2018) 4-9 and 14.
[3] K. Geissler, N. Steffens, R. Stein, Basics of the safety-equivalent assessment of
bridges with structural health monitoring, STAHLBAU. 88 (4) (2019) 338-353.
[4] S.O. Sajedi, X. Liang, Uncertainty ‐ assisted deep vision structural health monitoring, Computer-Aided Civil and Infrastructure Engineering. 36 (2) (2021) 126-142.
[5] Y. Kankanamge, Y. Hu, X. Shao, Application of wavelet transform in structural health monitoring, Earthquake Engineering and Engineering Vibration. 19 (02) (2020) 515-532.
[6] O. Yazdanpanah, B. Mohebi, M. Yakhchalian, Seismic damage assessment using improved wavelet-based damage-sensitive features, Journal of Building Engineering. 33 (2020) 101311.
[7] B. Mohebi, O. Yazdanpanah, F. Kazemi, A. Formisano, Seismic damage diagnosis in adjacent steel and RC MRFs considering pounding effects through improved wavelet-based damage-sensitive feature, Journal of Building Engineering. 33(2020) 101847.
[8] T. Tam, D. Dinh, J. Lee, T. Nguyen, An effective deep feedforward neural networks (DFNN) method for damage identification of truss structures using noisy incomplete modal data, Journal of Building Engineering. 30 (2020) 101244.
[9] Y. Xin, J. Li, H. Hao, Damage Detection in Initially Non-linear Structures based on Variational Mode Decomposition, International Journal of Structural Stability & Dynamics. 20 (10) (2020) 2042009.
[10] J. Liu, Z. Lu, M. Yu, Damage identification of non-classically damped shear building by sensitivity analysis of complex modal parameter, Journal of Sound and Vibration. 438 (2019) 457-475.
[11] J. Han, P. Zheng, H. Wang, Structural modal parameter identification and damage diagnosis based on Hilbert-Huang transform, Earthquake Engineering and Engineering Vibration. 13 (1) (2014) 101-111.
[12] A. Paral, D. Roy, A.K. Samanta, A deep learning-based approach for condition assessment of semi-rigid joint of steel frame, Journal of Building Engineering. 34, 2021, 101946.
[13] C. Pathirage, J. Li, L. Li, H. Hao, W. Liu, R. Wang, Development and application of a deep learning-based sparse autoencoder framework for structural damage identification, Structural Health Monitoring. 18 (1) (2019) 103-122.
[14] C. Pathirage, J. Li, L. Li, H. Hao, W. Liu, R. Wang, P.H. Ni, Structural damage identification based on autoencoder neural networks and deep learning, Engineering Structures. 172 (2018) 13-28.
[15] S. Hakim, H.A. Razak, Frequency Response Function-based Structural Damage Identification using Artificial Neural Networks-A review, Research Journal of Applied Sciences Engineering & Technology. 7 (9) (2014) 1750-1764.
[16] S. Hakim, H.A. Razak, S.A. Ravanfar, Ensemble neural networks for structural damage identification using modal data, International Journal of Damage Mechanics. 25 (3) (2016) 400-430.
[17] P. Seventekidis, D. Giagopoulos, A. Arailopoulos, O. Markogiannaki, Structural Health Monitoring using deep learning with optimal FE model generated data, Mechanical Systems and Signal Processing. 145 (2020) 106972.
[18] T. Liu, H. Xu, M. Ragulskis, M. Cao, W. Ostachowicz, A Data-Driven Damage Identification Framework Based on Transmissibility Function Datasets and One- Dimensional Convolutional Neural Networks: Verification on a Structural Health Monitoring Benchmark Structure, Sensors. 20 (4) (2020) 1059.
[19] E. Figueiredo, I. Moldovan, A. Santos, P. Campos, C.W. Costa, Finite Element-Based Machine-Learning Approach to Detect Damage in Bridges under Operational and Environmental Variations, Journal of Bridge Engineering. 24 (7) (2019) 1432.
[20] Nadith P, Li J, Ling L, et al. Structural damage identification based on autoencoder neural networks and deep learning. Engineering Structures, 2018, 172: 13-28.
[21] Gordan M, Ismail Z, Razak H A, et al. Data mining-based damage identification of a slab-on-girder bridge using inverse analysis. Measurement, 2019, 151.
[22] Ali R, Cha Y J. Subsurface damage detection of a steel bridge using deep learning and uncooled micro-bolometer. Construction and Building Materials, 2019, 226(30): 376-387.
[23] Kourehli S S, Ghadimi R. Vibration analysis and identification of breathing cracks in beams subjected to single or multiple moving mass using online sequential extreme learning machine. Inverse Problems in Science and Engineering. 2019, 27(08): 1057-1080.
[24] Teng Z Q, Teng S, Zhang J Q, et al. Structural Damage Detection Based on Real-Time Vibration Signal and Convolutional Neural Network. Applied Sciences, 2020, 10(14): 4720.
[25] Luke, H. D. The origins of the sampling theorem. IEEE Communications Magazine, 1999, 37(4), 106-108.
[26] Jerri, A. J.. The Shannon sampling theorem—Its various extensions and applications: A tutorial review. Proceedings of the IEEE, 1977, 65(11), 1565-1596.
[27] Song, Z., Liu, B., Pang, Y., Hou, C., & Li, X.. An improved Nyquist–Shannon irregular sampling theorem from local averages. IEEE transactions on information theory, 2012, 58(9), 6093-6100.
[28] Vaidyanathan, P. P.. Generalizations of the sampling theorem: Seven decades after Nyquist. IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications, 2001, 48(9), 1094-1109.
[29] Paige, C. C.. Computational variants of the Lanczos method for the eigenproblem. IMA Journal of Applied Mathematics, 1971, 10(3), 373-381.
[30] Golub, G. H., & Underwood, R.. The block Lanczos method for computing eigenvalues. In Mathematical software. 1977, 361-377. Academic Press.
[31] Chollet F. Xception: Deep learning with depthwise separable convolutions. Proceedings of the IEEE conference on computer vision and pattern recognition. 2017: 1251-1258..
[32] Sandler M, Howard A, Zhu M, et al. Mobilenetv2: Inverted residuals and linear bottlenecks。Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 4510-4520.
[33] Nguyen H. A lightweight and efficient deep convolutional neural network based on depthwise dilated separable convolution, Journal of Theoretical and Applied Information Technology. 98 (15) (2020) 2937-2947.
[34] Ananthanarayana T, Ptucha R, Kelly S C. Deep learning based fruit freshness classification and detection with CMOS image sensors and edge processors. Electronic Imaging, 2020, 2020(12): 172-1-172-7.
[35] Nguyen H. Fast object detection framework based on mobilenetv2 architecture and enhanced feature pyramid. J. Theor. Appl. Inf. Technol, 2020, 98(05). 812-824.
[36] Buiu C, Dănăilă V R, Răduţă C N. MobileNetV2 ensemble for cervical precancerous lesions classification. Processes, 2020, 8(5): 595-624.
[37] Li, W., Chen, C., Zhang, M., Li, H., & Du, Q.. Data augmentation for hyperspectral image classification with deep CNNs. IEEE Geoscience and Remote Sensing Letters, 2018, 16(4), 593-597.
[38] Torrey, L., & Shavlik, J.. Transfer learning. In Handbook of research on machine learning applications and trends: algorithms, methods, and techniques. 2010, 242-264. IGI global.