Neural networks in damage detection of composite laminated plates

  • Authors:

  • L. Roseiro;U. Ramos;R. Leal

  • Affiliations:

  • Departamento de Engenharia Mecânica, CEMUC, Instituto Superior de Engenharia de Coimbra, Coimbra, Portugal;Departamento de Engenharia Mecânica, CEMUC, Instituto Superior de Engenharia de Coimbra, Coimbra, Portugal;Departamento de Engenharia Mecânica, CEMUC, Universidade de Coimbra, Coimbra, Portugal

  • Venue:
  • NN'05 Proceedings of the 6th WSEAS international conference on Neural networks
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this work a methodology for damage detection on laminated composite plates involving the use of piezoelectric sensors and artificial neural networks is present. The presence of damage in the laminated composite plate leads to changes in its structural characteristics, causing variations in electrical potential of sensors. A feed-forward type neural network, trained by Levenberg-Marquardt algorithm is used in order to locate and quantify damage on the laminated plate using data obtained from piezoelectric sensors. A higher order finite element formulation allowing the response of the laminated composite plates is used to obtain the changes on electrical potential. A numerical example shows the feasibility of the proposed procedure.