Magnetostrictive Sensor-Based Smart Monitoring System for Bridge Health Using Machine Learning

In recent years, the need for continuous and real-time monitoring of civil infrastructures has intensified due to increasing incidences of structural degradation and the limitations of conventional inspection methods. Structural Health Monitoring (SHM) systems, integrated with smart sensors and machine learning, have emerged as promising solutions to address these challenges. This paper presents a comprehensive framework for SHM using a Galfenol-based magnetostrictive sensor capable of converting mechanical vibrations into electrical signals, thereby enabling self-powered monitoring. The experimental results demonstrate that among supervised classifiers, XGBoost achieved the highest accuracy of 82%, while DBSCAN led the unsupervised category with a Silhouette Coefficient of 0.49, indicating effective cluster separation. The findings validate the effectiveness of the proposed system in identifying healthy and unhealthy states with minimal data labeling. The novelty of this work lies in the integration of a nonlinear magnetostrictive energy harvester with a machine learning-based SHM framework, offering a battery-free, scalable, and adaptive solution for real-time infrastructure monitoring.