Transfer learning-based data anomaly detection for structural health monitoring

Structural Health Monitoring - Tập 22 Số 5 - Trang 3077-3091 - 2023
Qiuyue Pan1,2,3, Yuequan Bao1,2,3, Hui Li1,2,3
1Key Lab of Smart Prevention and Mitigation of Civil Engineering Disasters of the Ministry of Industry and Information Technology, Harbin Institute of Technology, Harbin, China
2Key Lab of Structures Dynamic Behavior and Control of the Ministry of Education, Harbin Institute of Technology, Harbin, China
3School of Civil Engineering, Harbin Institute of Technology, Harbin, China

Tóm tắt

The structural health monitoring (SHM) data of civil infrastructure are inevitably contaminated due to sensor faults, environmental noise interference, and data transmission failures. Anomalous data severely disturb the subsequent structural modal identification, damage identification, and condition assessment. Therefore, it is critical to detect and clean SHM data before data analysis. This paper proposes a novel approach for data anomaly detection based on transfer learning, that makes full use of the similarity of the anomalous patterns across different bridges and shares the knowledge incorporated in a deep neural network to achieve high-accuracy data anomaly identification for bridge groups. In the proposed approach, first, a multivariate database for a source bridge is built by plotting and labeling the raw sequential data. Then, a convolutional neural network (CNN) for data anomaly classification is designed and trained with the database in different conditions. The original CNN with the highest accuracy is transferred to a new bridge with enhancement training using a small part of the target bridge data. To validate the performance of the proposed method, the multivariate SHM data for two real long-span bridges are employed, including the acceleration, strain, displacement, humidity, and temperature data. The results demonstrate that transfer learning leads to a better classification capacity for the case of scarce labeled training data compared with the original network.

Từ khóa


Tài liệu tham khảo

10.1016/j.ymssp.2012.08.026

10.1177/1475921719877579

10.1016/j.comnet.2007.02.001

10.3390/jimaging4020036

10.1111/mice.12141

10.1016/j.eng.2018.11.030

10.1177/1475921719883202

Zhai G, 2022, Smart Struct Syst, 29, 237

10.1177/14759217211006485

Pang G, 2021, ACM Comput Surv, 54

10.1145/1541880.1541882

10.1016/j.sigpro.2013.12.026

10.1016/j.ymssp.2011.05.017

10.1049/iet-gtd.2016.1465

Thiyagarajan K, 2017 12th IEEE conference on industrial electronics and applications (ICIEA), 1926

10.1016/j.snb.2019.127428

10.1109/ICC.2018.8422402

10.1155/2017/4184196

Lasaga D, 2018, KDD 2017 workshop on anomaly detection in finance, 114

Srivastava N, 2015, International conference on machine learning, 843

10.1177/1475921718757405

10.1002/stc.2296

10.1111/mice.12528

10.1177/1475921720924601

10.1016/j.comcom.2020.05.034

10.1037/h0074898

10.1016/j.infsof.2011.09.007

10.1109/FUZZ-IEEE.2013.6622530

10.1007/978-3-319-11179-7_34

10.1016/j.conbuildmat.2017.09.110

10.1016/j.autcon.2020.103199

10.1111/mice.12755

10.1109/5.726791

He K, 2016 IEEE conference on computer vision and pattern recognition (CVPR), 770

Ioffe S, 2015, Proceedings of the 32nd international conference on machine learning, 448

Glorot X, 2010, Proceedings of the thirteenth international conference on artificial intelligence and statistics, 9, 249

Kingma DP, Ba J. Adam: a method for stochastic optimization. arXiv:14126980 [cs], http://arxiv.org/abs/1412.6980 (2017, accessed 25 April 2022).

10.1109/TKDE.2009.191

10.1109/TNN.2010.2091281

Wang J, 2021, Introduction to transfer learning

Long M, 2015, International conference on machine learning, 97

10.1016/j.aiopen.2021.08.002

Thông tin
Thông tin xuất bản

Structural Health Monitoring

Tập 22 Số 5

3077-3091

Thông tin tác giả