Honeycomb-structured materials are widely used in commercial and military aircraft. Manufacturing defects and damage during operation have become primary safety threats. This has increased the demand for non-destructive testing (NDT) for damage and flaws during aircraft operation and maintenance. Characterizing, or classifying defects, in addition to detecting them, is important. Classifying the liquids trapped in aircraft honeycomb cells is an example. A small amount of ingressed water is often tolerable, whereas a small amount of hydraulic oil may be an early warning of hydraulic system malfunction. This paper proposes an infrared thermography-based NDT technique and a long short term memory recurrent neural network (LSTM-RNN) model which automatically classifies common defects occurring in honeycomb materials. These including debonding, adhesive pooling, and liquid ingress. This LSTM-based algorithm has a greater than 90% sensitivity in classifying water, and hydraulic oil ingress. It has a greater than 70% sensitivity in classifying debonding and adhesive pooling.

LSTM-RNN-based defect classification in honeycomb structures using infrared thermography

Sfarra S.;
2019

Abstract

Honeycomb-structured materials are widely used in commercial and military aircraft. Manufacturing defects and damage during operation have become primary safety threats. This has increased the demand for non-destructive testing (NDT) for damage and flaws during aircraft operation and maintenance. Characterizing, or classifying defects, in addition to detecting them, is important. Classifying the liquids trapped in aircraft honeycomb cells is an example. A small amount of ingressed water is often tolerable, whereas a small amount of hydraulic oil may be an early warning of hydraulic system malfunction. This paper proposes an infrared thermography-based NDT technique and a long short term memory recurrent neural network (LSTM-RNN) model which automatically classifies common defects occurring in honeycomb materials. These including debonding, adhesive pooling, and liquid ingress. This LSTM-based algorithm has a greater than 90% sensitivity in classifying water, and hydraulic oil ingress. It has a greater than 70% sensitivity in classifying debonding and adhesive pooling.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11697/160396
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 27
  • ???jsp.display-item.citation.isi??? 27
social impact