A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Machine Learning Approach for Damage Detection of Railway Bridges: Preliminary Application
In the framework of structural health monitoring, a ‘model-free’ approach has been gaining increasing attention to describe the structural behavior without building a numerical model but adopting the so-called artificial neural networks (ANNs) that must be trained on data (e.g., accelerations) recorded on the existing structure. However, the lack of complete ‘real-life’ applications is the biggest current shortcoming to the use of ANNs. Indeed, the application of ANNs is often limited to medium-scale structures. The few applications on full scale structures are rarely run for a time interval sufficient to recognize structural deterioration due to material degradation and/or damage due to external loads. Moreover, the formulation of a methodology to design the sensor network serving a neural network is necessary.
In the context of a regional research project for Risks and Safety Management of Infrastructures at Regional Scale (GRISIS), ANNs are used by authors to develop a methodology for damage detection of railway bridges where the repetitive load condition, due to the passage of trains, is particularly favorable to the purpose. The potential damage detection is investigated referring to the structural accelerations recorded at a limited number of points of the structure. At this stage of the project, recorded data are not yet available, so an auxiliary numerical model of the structure is considered.
Machine Learning Approach for Damage Detection of Railway Bridges: Preliminary Application
In the framework of structural health monitoring, a ‘model-free’ approach has been gaining increasing attention to describe the structural behavior without building a numerical model but adopting the so-called artificial neural networks (ANNs) that must be trained on data (e.g., accelerations) recorded on the existing structure. However, the lack of complete ‘real-life’ applications is the biggest current shortcoming to the use of ANNs. Indeed, the application of ANNs is often limited to medium-scale structures. The few applications on full scale structures are rarely run for a time interval sufficient to recognize structural deterioration due to material degradation and/or damage due to external loads. Moreover, the formulation of a methodology to design the sensor network serving a neural network is necessary.
In the context of a regional research project for Risks and Safety Management of Infrastructures at Regional Scale (GRISIS), ANNs are used by authors to develop a methodology for damage detection of railway bridges where the repetitive load condition, due to the passage of trains, is particularly favorable to the purpose. The potential damage detection is investigated referring to the structural accelerations recorded at a limited number of points of the structure. At this stage of the project, recorded data are not yet available, so an auxiliary numerical model of the structure is considered.
Machine Learning Approach for Damage Detection of Railway Bridges: Preliminary Application
Lecture Notes in Civil Engineering
Rainieri, Carlo (editor) / Fabbrocino, Giovanni (editor) / Caterino, Nicola (editor) / Ceroni, Francesca (editor) / Notarangelo, Matilde A. (editor) / Bilotta, A. (author) / Testa, G. (author) / Capuano, C. (author) / Chioccarelli, E. (author)
International Workshop on Civil Structural Health Monitoring ; 2021 ; Naples, Italy
2021-08-25
13 pages
Article/Chapter (Book)
Electronic Resource
English
Neural Networks for Damage Detection in Steel Railway Bridges
| British Library Conference Proceedings | 1995
An unsupervised machine learning approach for real-time damage detection in bridges
| Elsevier | 2024