Machine Learning-Based Damage Detection in Steel Frame Structures Using Vibration and Acoustic Emission Data

Early and accurate detection of structural damage in steel frame buildings is essential for maintaining safety and reducing maintenance costs. This study proposes a machine learning-based framework that leverages both vibration signals and acoustic emission (AE) data for automated damage identification. The methodology involves preprocessing raw sensor data, extracting relevant time-frequency features, and training classification models using convolutional neural networks (CNN) and support vector machines (SVM). The integrated use of AE and vibration data improves the sensitivity of damage detection by capturing both microstructural changes and dynamic behavior. Experimental results on steel frame specimens subjected to simulated loading conditions demonstrate that the proposed approach achieves high classification accuracy and robustness across different damage scenarios. The findings highlight the potential of AI-driven sensor fusion for real-time structural health monitoring (SHM) in steel infrastructure, offering a scalable solution for intelligent maintenance and resilience assessment.

کلیدواژه ها:

Structural Health Monitoring (SHM) ، Steel Frame Structures ، Damage Detection ، Machine Learning ، Acoustic Emission (AE) ، Vibration Analysis ، Sensor Fusion ، Convolutional Neural Networks (CNN) ، Support Vector Machines (SVM) ، Time-Frequency Features