ニューラルネットワークと実験データによるCFRP積層材の損傷同定

概要

論文の詳細を見る
To enable the wider use of CFRP, damage especially in the laminated direction such as transverse cracking, delamination or fiber-matrix debonding should be easily and economically detected and a reasonable test method for detecting damage nondestructively should also be devised. We present the application of a hierachical neural network to damage identification in a CFRP laminated beam and discuss the accuracy and the efficiency of this method. It is found that the neural network is a useful and practical nondestructive method for the first stage approximation of damage identification in the CFRP laminated beam. Although the network is developed through an iterative calculation, this network is suited for field measurement because the damage can be identified by simple operations and multiplications.
一般社団法人日本機械学会の論文
1996-10-25