Structural health monitoring using unsupervised quantum machine learning

Abstract

This study presents a novel approach using unsupervised Quantum Machine Learning (QML) for Structural Health Monitoring (SHM). The proposed methodology involves extracting features from raw acceleration signals and encoding them into quantum states for a subsequent analysis in a quantum classifier. By training the model with known intact scenarios, an anomaly score function is evaluated to identify deviations from normal behavior aiming to detect potential structural anomalies. The framework is validated through experimental applications on a two-story laboratory frame and on a real-scale railway bridge, demonstrating encouraging results in anomaly detection, localization, and quantification. Through experimentation and numerical analyses, this study advances on the edge of SHM research, laying the foundation for future exploration at the intersection of Quantum Computing and Civil Engineering.