A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Fatigue damage prognosis of orthotropic steel deck based on data-driven LSTM
Abstract Orthotropic steel decks (OSDs) are widely used in steel bridges; however, they suffer from massive fatigue cracks early into their service life, which seriously affects the service security of steel bridges. Thus, it is critical to predicting the cumulative fatigue damage of OSDs accurately. In this study, the characteristics of field-monitored strain data and fatigue damage of typical connection details are investigated, which include rib-to-deck welded joints and rib butt welds. A fatigue damage prognosis method of OSDs based on data-driven long short-term memory (LSTM) was proposed. Further, the sensitive factors of the proposed method were analyzed, and the performance was validated. The characteristics analysis results show that the hourly fatigue damage series is an appropriate data form for the fatigue damage prognosis. The sensitive factors analysis results show that a data size of at least 840 and adding three feature dimensions related to the hour, day, and week, to fatigue damage series, can contribute to the performance of the LSTM model. The validation results show that the proposed method performs well on the fatigue damage prognosis. The mean absolute percentage errors were less than 10% among the two types of fatigue vulnerable details. Data-driven LSTM can significantly maintain the precision of fatigue damage prognosis with time. This study is expected to provide new insights and guidance into the fatigue damage prognosis of OSDs on long-span cable-stayed bridges.
Highlights Eight-week strain monitoring data and fatigue damage were studied. Orthotropic steel deck (OSD) fatigue damage prognosis based on data-driven LSTM. Effects of sensitive factors on precision of fatigue damage prognosis were examined. Proposed prognosis method validated using field-monitored strain data. Data-driven LSTM can maintain the precision of fatigue damage prognosis with time.
Fatigue damage prognosis of orthotropic steel deck based on data-driven LSTM
Abstract Orthotropic steel decks (OSDs) are widely used in steel bridges; however, they suffer from massive fatigue cracks early into their service life, which seriously affects the service security of steel bridges. Thus, it is critical to predicting the cumulative fatigue damage of OSDs accurately. In this study, the characteristics of field-monitored strain data and fatigue damage of typical connection details are investigated, which include rib-to-deck welded joints and rib butt welds. A fatigue damage prognosis method of OSDs based on data-driven long short-term memory (LSTM) was proposed. Further, the sensitive factors of the proposed method were analyzed, and the performance was validated. The characteristics analysis results show that the hourly fatigue damage series is an appropriate data form for the fatigue damage prognosis. The sensitive factors analysis results show that a data size of at least 840 and adding three feature dimensions related to the hour, day, and week, to fatigue damage series, can contribute to the performance of the LSTM model. The validation results show that the proposed method performs well on the fatigue damage prognosis. The mean absolute percentage errors were less than 10% among the two types of fatigue vulnerable details. Data-driven LSTM can significantly maintain the precision of fatigue damage prognosis with time. This study is expected to provide new insights and guidance into the fatigue damage prognosis of OSDs on long-span cable-stayed bridges.
Highlights Eight-week strain monitoring data and fatigue damage were studied. Orthotropic steel deck (OSD) fatigue damage prognosis based on data-driven LSTM. Effects of sensitive factors on precision of fatigue damage prognosis were examined. Proposed prognosis method validated using field-monitored strain data. Data-driven LSTM can maintain the precision of fatigue damage prognosis with time.
Fatigue damage prognosis of orthotropic steel deck based on data-driven LSTM
Deng, Peng-hao (author) / Cui, Chuang (author) / Cheng, Zhen-yu (author) / Zhang, Qing-hua (author) / Bu, Yi-zhi (author)
2023-01-02
Article (Journal)
Electronic Resource
English
Fatigue damage assessment of orthotropic steel deck using dynamic Bayesian networks
| British Library Online Contents | 2019
Evaluation on root-deck fatigue of orthotropic steel bridge deck
| Elsevier | 2013
Fatigue damage assessment of orthotropic steel deck using dynamic Bayesian networks
| British Library Online Contents | 2019
Evaluation on root-deck fatigue of orthotropic steel bridge deck
| Online Contents | 2013