October 10, 2025

Unsupervised Structural Health Monitoring of Large-Scale Structures Using Frequency Spectrum Including Time Information and Deep Learning

View Full Paper

Ask AI

Bookmark

Anomaly detection was successfully achieved using a deep learning model trained on the frequency response function.
The proposed method eliminates the need for initial parameter settings like mass and stiffness, simplifying structural assessment.
Verification demonstrated accurate detection of structural anomalies from limited observational data in a full-scale steel building.
Incorporating seismic response analysis allows for effective training data generation for deep learning models in health monitoring.

Discussion

No takes yet. Share an insight, caveat, or question.

Proposed method enhances anomaly detection in structures after seismic events, indicating a new approach to assess structural integrity.

AEBA et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark