Autoencoders and Convolutional Neural Networks Applied to Damage Detection

FLÁVIO BARBOSA; LUCAS RESENDE; RAFAELLE FINOTTI; ALEXANDRE CURY; CÁSIO MOTTA; HERNÁN GARRIDO; MARTÍN DOMIZIO; CURADELLI, OSCAR

Milán

Conferencia; 10th International Conference on Experimental Vibration Analysis for Civil Engineering Structures; 2023

EVACES 2023

Structural Health Monitoring (SHM) is a growing eld incivil engineering and has relevance for detecting changes in the state ofstructures, including identifying possible damage conditions. SHM strategiescommonly employ Articial Intelligence (AI) techniques on raw dynamicdata measured from structures to perform classications or extractfeatures from the original data. Among the AI algorithms for SHM, autoencoder,and convolutional neural networks have been identied aspromising solutions, being the focus of this article. Both algorithms areapplied to identify eight damage scenarios in a beam starting from thetime histories of the tested structure, pointing out the advantages anddisadvantages of each algorithm. The authors tested the beam throughmonitoring based on image processing using a high-speed camera.