Application of deep autoencoder model for structural condition monitoring

doi:10.21629/JSEE.2018.04.22

Journal of Systems Engineering and Electronics ›› 2018, Vol. 29 ›› Issue (4): 873-880.doi: 10.21629/JSEE.2018.04.22

• Reliability • Previous Articles

Application of deep autoencoder model for structural condition monitoring

Chathurdara Sri Nadith PATHIRAGE¹(), Jun LI²(), Ling LI^1,*(), Hong HAO²(), Wanquan LIU¹()

¹ Department of Computing, Curtin University, Perth WA 6102, Australia
² Centre for Infrastructural Monitoring and Protection, Curtin University, Perth WA 6102, Australia

Received:2017-01-03 Online:2018-08-01 Published:2018-08-30
Contact: Ling LI E-mail:c.nadithpa@postgrad.curtin.edu.au;junli@curtin.edu.au;L.Li@curtin.edu.au;hong.hao@curtin.edu.au;W.Liu@curtin.edu.au
About author:PATHIRAGE Chathurdara Sri Nadith was born in 1987. He received his B.S. degree in computer science in 2009 and is currently stduying for his Ph.D. degree in the Department of Computing, Curtin University, Australia. He worked as a specialist software engineer at London Stock Exchange group. His research interests lie in the field of artificial intelligence in computer vision, concepts of human visual perception via combined intellect of image processing, sparse sensing and deep learning. E-mail: c.nadithpa@postgrad.curtin.edu.au|LI Jun was born in 1984. He received his Ph.D. degree from Hong Kong Polytechnic University in 2012. He is currently a senior lecturer in Department of Civil Engineering/Centre for Infrastructural Monitoring and Protection at Curtin University. His research interests include structural health monitoring and signal processing techniques. E-mail: junli@curtin.edu.au|LI Ling was born in 1965. She obtained her B.S. degree in computer science from Sichuan University, China, and Ph.D. degree in computer engineering from Nanyang Technological University, Singapore. She is now an associate professor in the Department of Computing at Curtin University in Australia. Her research interests are mainly in computer vision and graphics, and artificially intelligent beings. E-mail: L.Li@curtin.edu.au|HAO Hong was born in 1962. He received his B.E. degree from Tianjin University, China, and Ph.D. degree from the University of California at Berkeley. He is a john curtin distinguished professor and the director of Centre for Infrastructural Monitoring and Protection at Curtin University, Australia. His research interests include earthquake engineering, blast engineering and structural condition monitoring. E-mail: hong.hao@curtin.edu.au|LIU Wanquan was born in 1965. He received his B.S. degree in applied mathematics from Qufu Normal University, China, in 1985, M.S. degree in control theory and operation research from Chinese Academy of Sciences in 1988, and the Ph.D. degree in electrical engineering from Shanghai Jiaotong University, in 1993. He once held the ARC fellowship, U2000 fellowship and JSPS fellowship and attracted research funds from different resources over 2.4 million Australian dollars. He is currently an associate professor in the Department of Computing at Curtin University and is in editorial board for nine international journals. His current research interests include large-scale pattern recognition, signal processing, machine learning, and control systems. E-mail: W.Liu@curtin.edu.au
Supported by:
the Australian Research Council;This work was supported by the Australian Research Council

Abstract

Abstract:

Damage detection in structures is performed via vibration based structural identification. Modal information, such as frequencies and mode shapes, are widely used for structural damage detection to indicate the health conditions of civil structures. The deep learning algorithm that works on a multiple layer neural network model termed as deep autoencoder is proposed to learn the relationship between the modal information and structural stiffness parameters. This is achieved via dimension reduction of the modal information feature and a non-linear regression against the structural stiffness parameters. Numerical tests on a symmetrical steel frame model are conducted to generate the data for the training and validation, and to demonstrate the efficiency of the proposed approach for vibration based structural damage detection.

Key words: auto encoder, non-linear regression, deep auto encoder model, damage identification, vibration, structural health monitoring

Chathurdara Sri Nadith PATHIRAGE, Jun LI, Ling LI, Hong HAO, Wanquan LIU. Application of deep autoencoder model for structural condition monitoring[J]. Journal of Systems Engineering and Electronics, 2018, 29(4): 873-880.

Figures/Tables 9

Fig 1

Table 1

Fig 2

Fig 3

Fig 4

Table 2

Table 3

Fig 5

Fig 6

References 21

1	YUNC B, YIJ H, BAHNGE Y. Joint damage assessment of framed structures using a neural networks technique. Engineering Structures, 2001, 23 (5): 425- 435. doi: 10.1016/S0141-0296(00)00067-5
2	BAKHARY N, HAO H, DEEKSA J. Damage detection using artificial neural network with consideration of uncertainties. Engineering Structures, 2007, 29 (11): 2806- 2815. doi: 10.1016/j.engstruct.2007.01.013
3	CHEN Y K, FENG Z G, ZHANG S, et al. Research on reconfigurable triple-module redundancy space on-board computer. Computer Measurement & Control, 2017, 25(2): 201-203. (in Chinese)
4	HẢSTAD J, GOLDMANN M. On the power of small-depth threshold circuits. Computational Complexity, 1991, 1 (2): 113- 129. doi: 10.1007/BF01272517
5	HE R X, CHEN N J. Fault tolerance of reconfigurable computing system. Journal of Mianyang Teachers'College, 2017, 36(5): 90-97. (in Chinese)
6	BENGIO Y, LECUN Y. Scaling learning algorithms towards AI. Large-Scale Kernel Machines, 2007, 34 (5): 1- 41.
7	HAN J, LEUNG E, LIU L B, et al. A fault-tolerant technique using quadded logic and quadded transistors. IEEE Trans. on Very Large Scale Integration Systems, 2015, 23(8): 1562-1566.
8	ROHANIPOOR M R, GHAVAMI B, RAJI M. Design of fault tolerant digital integrated circuits based on quadded transistor logic. Proc. of the 8th International Conference on Information and Knowledge Technology, 2016: 188-192.
9	ZHANG J B, CAI J Y, MENG Y F. A design technology of fault tolerance circuit system facing complex electromagnetic environments. Journal of Xi'an Jiaotong University, 2017, 51(2): 53-59. (in Chinese)
10	SOLOVIEV A, STEMPKOVSKY A, KALEEV D, et al. Methods of increasing the fault-tolerance of control unit by introducing hardware redundancy. Proc. of the IEEE Internet Technologies and Applications, 2015: 37-40.
11	SHOKER A. Exploiting universal redundancy. Proc. of the 15th IEEE International Symposium on Network Computing and Applications, 2016: 199-203.
12	SALAKHUTDINOV R, HINTON G. Semantic hashing. International Journal of Approximate Reasoning, 2009, 50 (7): 969- 978. doi: 10.1016/j.ijar.2008.11.006
13	SHEIKH A T, EL-MALEH A H, ELRABAA M E S, et al. A fault tolerance technique for combinational circuits based on selective-transistor redundancy. IEEE Trans. on Very Large Scale Integration Systems, 2017, 25(1): 224-237.
14	BENGIO Y. Learning deep architectures for AI. Foundations & Trends in Machine Learning, 2009, 2 (1): 1- 127.
15	HINTONG E, SALAKHUTDINOVR R. Reducing the dimensionality of data with neural networks. Science, 2006, 313 (5786): 504- 507. doi: 10.1126/science.1127647
16	VON NEUMANN J. Probabilistic logics and the synthesis of reliable organisms from unreliable components. Princeton: Princeton University Press, 1956.
17	VINCENT P, LAROCHELLE H, LAJOIE I, et al. Stacked denoising autoencoders: learning useful representations in a deep network with a local denoising criterion. Journal of Machine Learning Research, 2010, 11 (12): 3371- 3408.
18	FRISWELL M, MOTTERSHEADJ E. Finite element model updating in structural dynamics. Dordrecht: Kluwer Academic Publishers, 2013.
19	LI J, LAWS S, DING Y. Substructure damage identification based on response reconstruction in frequency domain and model updating. Engineering Structures, 2012, 41 (3): 270- 284. doi: 10.1016/j.engstruct.2012.03.035
20	LI J, HAO H, LO J V. Structural damage identification with power spectral density transmissibility: numerical and experimental studies. Smart Structures & Systems, 2015, 15 (1): 15- 40. doi: 10.12989/sss.2015.15.1.015
21	HINTONG E, OSINDERO S, TEHY W. A fast learning algorithm for deep belief nets. Neural Computing, 2006, 18 (7): 1527- 1554. doi: 10.1162/neco.2006.18.7.1527

Element	Dimension
Height of each storey/m	0.3
Length of the beam/m	0.5
Cross-section of the column/mm$^2$	49.98×4.85
Cross-section of the beam/mm$^2$	49.89×8.92
Mass density of the column/(kg/m$^3$)	7 850
Mass density of the beam/(kg/m$^3$)	7 734.2

Phase	Dataset size	Validation error(MSE)/%	Error(MSE)/%
Pre-training layer 1	1 200	23.31	5.46
Pre-training layer 2	1 200	39.50	9.26
Fine-tuning	1 200	0.39	0.09
Testing	600	—	0.77

Phase	Dataset size	Validation error(MSE)/%	Error(MSE)/%
Pre-training layer 1	1 200	3.84	0.90
Pre-training layer 2	1 200	2.87	0.67
Pre-training layer 3	1 200	2.86	0.69
Fine-tuning	1 200	2.20	0.54
Testing	600	—	2.9

[1]	Fitri Yakub, Shamsul Sarip, Andika Aji Wijaya, and Yasuchika Mori. Enhancing of nominal characteristic trajectory following control for motion systems#br# [J]. Journal of Systems Engineering and Electronics, 2017, 28(6): 1221-1235.
[2]	Tong Zhang, Jun Zhang, and Junjie Hu. Dynamic model and simulation of aseismatic system about laser gyro strapdown inertial measure assembly based on MSTMM [J]. Journal of Systems Engineering and Electronics, 2012, 23(5): 741-751.
[3]	Hongbing Sun, Shenfang Yuan, Xia Zhao, Hengbao Zhou, and Dong Liang. Technology of structure damage monitoring based on multi-agent [J]. Journal of Systems Engineering and Electronics, 2010, 21(4): 616-622.
[4]	Shi Quan, Tian Juan & Man Qiang. Damage effect analysis and experiment for electronic equipment in impact vibration environment [J]. Journal of Systems Engineering and Electronics, 2009, 20(6): 1384-1388.

Application of deep autoencoder model for structural condition monitoring

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

Share this article

Figures/Tables 9

References 21

Related Articles 4

Recommended Articles

Metrics

Comments