Smart edge computing framework for in-line signal detection and classification

Gullapalli, Anirudh, Aburakhis, Taha, Featherston, Carol ORCID: https://orcid.org/0000-0001-7548-2882, Pullin, Rhys ORCID: https://orcid.org/0000-0002-2853-6099 and Kundu, Abhishek ORCID: https://orcid.org/0000-0002-8714-4087 2024. Smart edge computing framework for in-line signal detection and classification. Presented at: 2024 51st Annual Review of Progress in Quantitative Nondestructive Evaluation, Denver, CO, USA, 21-24 July 2024. Proceedings of the 2024 51st Annual Review of Progress in Quantitative Nondestructive Evaluation. ASME, 10.1115/QNDE2024-138485

Full text not available from this repository.

Abstract

Composite structures’ complex, anisotropic nature poses challenges for damage monitoring. Non-intrusive inspection methods are crucial for continuous monitoring, aiding the shift to predictive maintenance. Onboard monitoring systems must efficiently acquire and analyze signals, correlating them with damage metrics for automated assessment of structural integrity and optimal maintenance planning. Real-time monitoring faces hurdles in handling large data volumes due to computing constraints. This paper proposes a cyber-physical architecture for acquiring and classifying ultrasonic guided wave signals. It employs a sparse transducer array and an edge device for signal processing. Experiments involved exciting a 12-layer carbon fiber composite panel with tone-burst sinusoidal signals. The responses to these excitations were captured and subjected to soft-threshold-based wavelet denoising to extract the structural acoustic response in the ultrasonic frequency band. Subsequently, the conditioned signals were transformed into time-frequency scalograms, which were then employed to train a multi-class classification algorithm operating on the edge device for effective real-time signal classification in an in-service setting.

Item Type:	Conference or Workshop Item (Paper)
Date Type:	Published Online
Status:	Published
Schools:	Engineering
Publisher:	ASME
ISBN:	9780791888162
Funders:	EPSRC
Last Modified:	18 Nov 2024 11:00
URI:	https://orca.cardiff.ac.uk/id/eprint/173660

Actions (repository staff only)

Edit Item