Adaptive multilayer perceptron networks for detection of cracks in anisotropic laminated plates

Citation
Yg. Xu et al., Adaptive multilayer perceptron networks for detection of cracks in anisotropic laminated plates, INT J SOL S, 38(32-33), 2001, pp. 5625-5645
Citations number
36
Categorie Soggetti
Mechanical Engineering
Journal title
INTERNATIONAL JOURNAL OF SOLIDS AND STRUCTURES
ISSN journal
00207683 → ACNP
Volume
38
Issue
32-33
Year of publication
2001
Pages
5625 - 5645
Database
ISI
SICI code
0020-7683(200108)38:32-33<5625:AMPNFD>2.0.ZU;2-U
Abstract
In this study, an adaptive multiplayer perceptron (MLP) technique is propos ed for the detection of cracks in anisotropic laminated plates. The displac ement response on the surface of plate, excited by a time-harmonic line loa d, is used as the input of the MLP. The crack parameters that specify the l ocation and size of the cracks in the anisotropic laminated plates are take n as the output of the MLP. The MLP model is first trained to establish the nonlinear relationship between the scattered surface displacement response and the corresponding location and size of the cracks. The scattered displ acement responses required in training samples are calculated from the stri p element method (SEM). To facilitate this training process, the correlatio n analysis for the outputs of neurons in the hidden layers of the MLP model is carried out to optimize the MLP architecture. A modified back-propagati on learning algorithm with a dynamically adjusted learning rate and an addi tional jump factor is developed to speed up the convergence of the MLP mode l in the training process. The concept of orthogonal array is adopted to ge nerate the representative combinations of the crack parameters, which signi ficantly reduces the number of samples while maintaining the completeness o f sample data. The well-trained model is then used to reconstruct the crack parameters by feeding in the measured displacement response on the plate s urface. These reconstructed crack parameters are further examined by compar ing their resulting displacement response from the SEM forward calculation with the measured displacement response. If the comparison is satisfactory, the reconstructed crack parameters would be considered to be true and the computation ends. Otherwise, the MLP model would go another round of re-tra ining process until the satisfactory reconstruction is obtained. The propos ed technique was verified numerically using an anisotropic laminated plate [C0/G + 45/G-45](s) with four types of horizontal cracks. The verification includes the detection for both the location and the size of cracks using t he simulated response data with and without noise. (C) 2001 Elsevier Scienc e Ltd. Ah rights reserved.