Temperature compensation for guided wave signals using denoising autoencoders
e-Journal of Nondestructive Testing · 2026-07 · 31 (7)
DOI: 10.58286/33241 ↗At a glance
Journal article in e-Journal of Nondestructive Testing, 2026-07. Written by Matthew Newson and 1 co-authors.
Abstract (original language)
Ultrasonic guided-waves are widely used in Structural Health Monitoring (SHM) due to their long propagation range and high sensitivity to defects. However, their analysis is challenging because of their multimodal and dispersive behaviours, and most baseline-based approaches are highly sensitive to Environmental and Operational Conditions (EOC), which can lead to false alarms. Existing methods such as Baseline Signal Stretch (BSS) and Optimal Baseline Selection (OBS) [1] partially address this issue, but BSS is effective only over a narrow temperature range, while OBS requires a large number of reference signals, increasing implementation complexity and cost. This work introduces a compensation strategy based on a denoising autoencoder that incorporates a description of the target state into the latent space. Two types of target representations are considered: one using directly the target temperature, and the other using the target signal itself. The method is evaluated on multiple laboratory structures, including aluminium and steel alloy plates. The results show that the proposed approach maintains a low prediction error under variable conditions and independently of the temperature difference between the input and the target.
Cite
APA
Newson, M., & Fisher, C. (2026). Temperature compensation for guided wave signals using denoising autoencoders. e-Journal of Nondestructive Testing, 31(7). https://doi.org/10.58286/33241
Vancouver
Newson M, Fisher C. Temperature compensation for guided wave signals using denoising autoencoders. e-Journal of Nondestructive Testing. 2026;31(7). doi:10.58286/33241
BibTeX
@article{Newson2026,
author = {Newson, Matthew and Fisher, Clément},
title = {Temperature compensation for guided wave signals using denoising autoencoders},
journal = {e-Journal of Nondestructive Testing},
year = {2026},
volume = {31},
number = {7},
doi = {10.58286/33241}
}
EUVIMED