Artificial Intelligence-Based Structural Health Monitoring of Bridges and Tunnels: A Comprehensive Review

Abstract

The rapid deterioration of civil infrastructure, particularly tunnels and bridges, poses a global socio-economic challenge. As an alternative to traditional inspection methods, Structural Health Monitoring (SHM) has become a key part of identifying deterioration of components and ensuring public safety and serviceability. The last decade has seen the introduction of Artificial Intelligence (AI) to SHM, which has included Machine Learning (ML) and Deep Learning (DL). These methods have revolutionised the discipline and shifted the focus of SHM from a manual and reactive practice to the automation and prediction of maintenance works. This paper provides a detailed overview of the applications of AI in the SHM of bridges and tunnels. The research surveyed 99 of the most cited papers in the discipline and focused on the contributions those papers made to the field. The use of DL, particularly Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM), has provided significant advances. The reported research has a 95% success rate for the detection of structural damage in high complexity situations. This review has organized improvements in SHM along the lines of vibration-based anomaly detection, vision-based defect recognition, and multi-modal data fusion. This review has also focused on the combination of AI with rapidly emerging and advanced technologies, such as Digital Twins (DT), Physics-Informed Neural Networks (PINNs), and Explainable AI (XAI). This review provides an innovative research guide for SHM systems that focuses on SHM systems that use AI and modern supports to address challenges such as sparse data and class imbalance.