Dissertation/Thesis Abstract

Internal delamination detection problems using a combined improved-counterpropagation neural network and genetic algorithm technique
by Phuong, Tran N., M.S., California State University, Long Beach, 2013, 215; 1527487
Abstract (Summary)

Detection of delamination in laminated composites is first formulated as a simulation and second as an optimization problem and solved by the new approach utilizing improved-counterpropagation neural networks and genetic algorithms, respectively. A recently developed improved-layerwise composite laminate theory is extended to model composite laminates with delamination. This new layerwise finite element model is employed to calculate natural frequencies of cross-ply laminates with given delamination patterns placed at different locations. Improved-counterpropagation neural networks are trained to simulate dynamic responses from the finite element analysis. These artificial neural networks are chosen as function approximations which are developed on the available input-output data from a finite element model. Genetic algorithms with mixed type design variables are used to search the optimum delamination patterns associated with the given natural frequencies. Results with remarkable accuracy have been obtained in detecting the internal delamination using a combination of both techniques.

Indexing (document details)
Advisor: Chen, Hsin-Piao
School: California State University, Long Beach
School Location: United States -- California
Source: MAI 52/05M(E), Masters Abstracts International
Subjects: Aerospace engineering, Mechanical engineering, Materials science
Keywords: Laminated composites, Macromechanical behavior, Strain, Stress
Publication Number: 1527487
ISBN: 978-1-303-77380-8
Copyright © 2020 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy