Session: Structural Health Monitoring I & II, Material Characterization by Ultrasonic wave

Paper Number: 170141

170141 - Dataset on Guided Waves From Long-Term Structural Health Monitoring Under Dynamic Conditions 

Abstract: 

Modern society relies on structural health monitoring to assess the condition of mechanical systems and civil infrastructure, ensuring they operate efficiently and safely. While most studies focus on evaluating structures under controlled or stable lab conditions, a significant number of structures function in uncontrolled and dynamic environments over extended periods. Long-term structural health monitoring in such complex conditions presents additional challenges compared to lab-based assessments. Few studies address guided wave structural health monitoring under controlled and dynamic environments, largely due to the lack of a public benchmark dataset.

To bridge this gap, this study introduces a publicly available dataset from a long-term outdoor structural monitoring experiment conducted at the University of Utah, Salt Lake City. Over a span of 4.5 years, approximately 6.4 million guided wave measurements were collected under both regular environmental variations (e.g., daily temperature fluctuations from −12.2℃ to 52.5℃) and irregular changes (e.g., rain and snow). The dataset also captures the effects of sensor drift and installation shifts over time. To facilitate damage detection and severity assessment, thirteen types of damage were intentionally introduced to the monitored structure. Additionally, the dataset includes timestamps, temperature, humidity, air pressure, brightness, and weather conditions to support comprehensive analysis.

To assess the effectiveness of this long-term monitoring dataset, we apply optimal correlation coefficients and deep learning-based unsupervised methods for damage detection. The results show that among the 13 damage types, the four least severe remain undetected, while the other nine can be successfully identified. This publicly available dataset is designed to support researchers in developing more practical structural health monitoring techniques for uncontrolled and dynamic environments.

Presenting Author: KANG YANG University of Florida

Presenting Author Biography: Kang Yang received his B.Sc. degree in 2015 from Southwest Jiaotong University, his M.Sc. degree in 2018 from Xi’an Jiaotong University, and his Ph.D. degree in 2023 from the University of Florida. Now he is a post-doc associate in University of Florida. His main research interests include structural health monitoring, machine learning and artificial intelligence, ultrasonic detection, and anomaly detection.

Authors:

KANG YANG University of Florida
Kaiyi Lei University of Florida
Minzhe Wu University of Florida
Junkai Zhou University of FLoirda
Zhongzheng Ren Zhang University of FLorida
Joel B. Harley University of Florida