期刊
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
卷 64-65, 期 -, 页码 403-412出版社
ACADEMIC PRESS LTD- ELSEVIER SCIENCE LTD
DOI: 10.1016/j.ymssp.2015.04.007
关键词
Elastic, homogeneous plate; Lamb waves; Guided wave monitoring/measurement; Dispersion curve; Artificial Neural Networks; Hybrid computational systems
资金
- Polish National Science Center, AGH [UMO-2011/01/B/ST8/ 07210, 18.18.130.384]
A new hybrid computational system for material identification (HCSMI) is presented, developed for the identification of homogeneous, elastic, isotropic plate parameters. Attention is focused on the construction of dispersion curves, related to Lamb waves. The main idea of the system HCSMI lies in separation of two essential basic computational stages, corresponding to direct or inverse analyses. In the frame of the first stage an experimental dispersion curve DCexp is constructed, applying Guided Wave Measurement (GWM) technique. Then, in the other stage, corresponding to the inverse analysis, an Artificial Neural Network (ANN) is trained 'off line'. The substitution of results of the first stage, treated as inputs of the ANN, gives the values of identified plate parameters. In such a way no iteration is needed, unlike to the classical approach. In such an approach, the distance between the approximate experimental curves DCexp and dispersion curves DCnum obtained in the direct analysis, is iteratively minimized. Two case studies are presented, corresponding either to measurements in laboratory tests or those related to pseudo-experimental noisy data of computer simulations. The obtained results prove high numerical efficiency of HCSMI, applied to the identification of aluminum plate parameters. (C) 2015 Elsevier Ltd. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据