Voting Neural Network Classifier for Detection of Fatigue Damage in Aircrafts - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

Voting Neural Network Classifier for Detection of Fatigue Damage in Aircrafts

Résumé

An ANN based method for detection and localization of fatigue damage in aircraft structures is presented in the paper. Damage indices are calculated from Lamb-wave measurements conducted by the network of piezoelectric transducers. Data gathered by the sensors is used as an input to the proposed voting neural network classifier. A set of neural network electors of different architecture cooperates to achieve consensus concerning the state of each monitored path. Sensed signal variations in the ROI, detected by the networks at each path, are used to assess the state of the structure as well as to localize detected damage and to filter out ambient changes. The classifier has been extensively tested on large data sets acquired in the tests of specimens with artificially introduced notches as well as the results of numerous fatigue experiments. Effect of the classifier structure and test data used for training on the results is evaluated. It is shown that the developed classifier performs better than individual ANNs in terms of damage detection. The classifier structure, composed of different networks working together, yields an increased reliability, mainly due to the lower impact of the initial weights distribution on the final result.
Fichier principal
Vignette du fichier
0250.pdf (3.41 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01022042 , version 1 (10-07-2014)

Identifiants

  • HAL Id : hal-01022042 , version 1

Citer

Ziemowit Dworakowski, Lukasz Ambrozinski, Krzysztof Dragan, Tadeusz Stepinski, Tadeusz Uhl. Voting Neural Network Classifier for Detection of Fatigue Damage in Aircrafts. EWSHM - 7th European Workshop on Structural Health Monitoring, IFFSTTAR, Inria, Université de Nantes, Jul 2014, Nantes, France. ⟨hal-01022042⟩
117 Consultations
195 Téléchargements

Partager

Gmail Facebook X LinkedIn More