An Interpretable Deep Learning Method for Identifying Extreme Events under Faulty Data Interference
Structural health monitoring systems continuously monitor the operational state of structures, generating a large amount of monitoring data during the process. The structural responses of extreme events, such as earthquakes, ship collisions, or typhoons, could be captured and further analyzed. Howev...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-05-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/9/5659 |
| _version_ | 1797602952526954496 |
|---|---|
| author | Jiaxing Guo Zhiyi Tang Changxing Zhang Wei Xu Yonghong Wu |
| author_facet | Jiaxing Guo Zhiyi Tang Changxing Zhang Wei Xu Yonghong Wu |
| author_sort | Jiaxing Guo |
| collection | DOAJ |
| description | Structural health monitoring systems continuously monitor the operational state of structures, generating a large amount of monitoring data during the process. The structural responses of extreme events, such as earthquakes, ship collisions, or typhoons, could be captured and further analyzed. However, it is challenging to identify these extreme events due to the interference of faulty data. Real-world monitoring systems suffer from frequent misidentification and false alarms. Unfortunately, it is difficult to improve the system’s built-in algorithms, especially the deep neural networks, partly because the current neural networks only output results and do not provide an interpretable decision-making basis. In this study, a deep learning-based method with visual interpretability is proposed to identify seismic data under sensor faults interference. The transfer learning technique is employed to learn the features of seismic data and faulty data with efficiency. A post hoc interpretation algorithm, termed Gradient-weighted Class Activation Mapping (Grad-CAM), is embedded into the neural networks to uncover the interest regions that support the output decision. The in situ seismic responses of a cable-stayed long-span bridge are used for method verification. The results show that the proposed method can effectively identify seismic data mixed with various types of faulty data while providing good interpretability. |
| first_indexed | 2024-03-11T04:22:51Z |
| format | Article |
| id | doaj.art-08bc83ea138d43c6922b261dd486d183 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-11T04:22:51Z |
| publishDate | 2023-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-08bc83ea138d43c6922b261dd486d1832023-11-17T22:37:04ZengMDPI AGApplied Sciences2076-34172023-05-01139565910.3390/app13095659An Interpretable Deep Learning Method for Identifying Extreme Events under Faulty Data InterferenceJiaxing Guo0Zhiyi Tang1Changxing Zhang2Wei Xu3Yonghong Wu4Faculty of Civil Engineering and Mechanics, Kunming University of Science and Technology, Kunming 650500, ChinaFaculty of Civil Engineering and Mechanics, Kunming University of Science and Technology, Kunming 650500, ChinaFaculty of Civil Engineering and Mechanics, Kunming University of Science and Technology, Kunming 650500, ChinaFaculty of Civil Engineering and Mechanics, Kunming University of Science and Technology, Kunming 650500, ChinaFaculty of Civil Engineering and Mechanics, Kunming University of Science and Technology, Kunming 650500, ChinaStructural health monitoring systems continuously monitor the operational state of structures, generating a large amount of monitoring data during the process. The structural responses of extreme events, such as earthquakes, ship collisions, or typhoons, could be captured and further analyzed. However, it is challenging to identify these extreme events due to the interference of faulty data. Real-world monitoring systems suffer from frequent misidentification and false alarms. Unfortunately, it is difficult to improve the system’s built-in algorithms, especially the deep neural networks, partly because the current neural networks only output results and do not provide an interpretable decision-making basis. In this study, a deep learning-based method with visual interpretability is proposed to identify seismic data under sensor faults interference. The transfer learning technique is employed to learn the features of seismic data and faulty data with efficiency. A post hoc interpretation algorithm, termed Gradient-weighted Class Activation Mapping (Grad-CAM), is embedded into the neural networks to uncover the interest regions that support the output decision. The in situ seismic responses of a cable-stayed long-span bridge are used for method verification. The results show that the proposed method can effectively identify seismic data mixed with various types of faulty data while providing good interpretability.https://www.mdpi.com/2076-3417/13/9/5659structural health monitoringextreme event detectionlong-span bridgetransfer learninginterpretable deep learning |
| spellingShingle | Jiaxing Guo Zhiyi Tang Changxing Zhang Wei Xu Yonghong Wu An Interpretable Deep Learning Method for Identifying Extreme Events under Faulty Data Interference Applied Sciences structural health monitoring extreme event detection long-span bridge transfer learning interpretable deep learning |
| title | An Interpretable Deep Learning Method for Identifying Extreme Events under Faulty Data Interference |
| title_full | An Interpretable Deep Learning Method for Identifying Extreme Events under Faulty Data Interference |
| title_fullStr | An Interpretable Deep Learning Method for Identifying Extreme Events under Faulty Data Interference |
| title_full_unstemmed | An Interpretable Deep Learning Method for Identifying Extreme Events under Faulty Data Interference |
| title_short | An Interpretable Deep Learning Method for Identifying Extreme Events under Faulty Data Interference |
| title_sort | interpretable deep learning method for identifying extreme events under faulty data interference |
| topic | structural health monitoring extreme event detection long-span bridge transfer learning interpretable deep learning |
| url | https://www.mdpi.com/2076-3417/13/9/5659 |
| work_keys_str_mv | AT jiaxingguo aninterpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference AT zhiyitang aninterpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference AT changxingzhang aninterpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference AT weixu aninterpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference AT yonghongwu aninterpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference AT jiaxingguo interpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference AT zhiyitang interpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference AT changxingzhang interpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference AT weixu interpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference AT yonghongwu interpretabledeeplearningmethodforidentifyingextremeeventsunderfaultydatainterference |