Researchers developed AI-based algorithms for real-time structural assessment of bridges

Bridges play a crucial role in India's infrastructure, with approximately 13,500 of them spread across the nation. These structures naturally deteriorate over time due to environmental factors like temperature fluctuations and exposure to water and air. Traditionally, the assessment of bridge conditions has relied on visual inspections, a method considered insufficient by experts. It is limited in its ability to detect all structural issues and time-consuming as it involves manual analysis of numerous photographs.

Recent advances in instrumentation, data analysis, and AI tools, including Deep Learning (DL), offer significant potential for Structural Health Monitoring (SHM) of bridges and other structures. These technologies make it more convenient to identify, measure, comprehend, and even predict the development of defects over time.

In this direction, the Indian Institute of Technology Mandi has collaborated with INRIA in France to develop advanced techniques in Artificial Intelligence (AI) and signal processing. Their main goal is to enhance the accuracy of predicting the structural health of bridges and/or other structures. It is worth noting that temperature fluctuations can significantly impact a bridge's dynamic characteristics, especially in the case of prestressed concrete and cable-stayed bridges. Therefore, it is essential to account for these temperature effects in both real-time and AI-based Structural Health Monitoring.

The research team at IIT Mandi has developed an innovative Structural Health Monitoring (SHM) approach based on Deep Learning (DL). Their AI algorithms have the capability to identify and isolate structural damage by analyzing the recorded ambient dynamic responses, all without requiring human intervention. The method is based on data-driven techniques, e.g. Machine Learning, AI, and Bayesian statistical inference and estimates a bridge's health predicting its remaining life.

IIT Mandi's algorithm underwent rigorous validation on an actual bridge located in a cold region with extreme annual and daily temperature variations. The results demonstrated the effectiveness in identifying structural damage caused by various factors. Furthermore, in a related study, the researchers employed advanced filtering techniques to assess the condition of structural components. This technique allows for the separate assessment of each component's health, aiding in the evaluation of overall structural integrity. By utilizing computer simulations and conducting comprehensive tests, the researchers verified the method's performance, even when subjected to background noise and varying levels of damage severity.