ニューラルネットワークと計算力学に基づく欠陥同定 : 超音波非破壊検査への応用

概要

論文の詳細を見る
Quantitative nondestructive evaluation (QNDE) has become increasingly important in various engineering fields. This paper describes an application of the hierarchical neural network to defect identification with the ultrasonic method. The present method consists of three subprocesses. First, sample data of identification parameters vs dynamic responses of displacements at several monitoring points on surface are calculated by the dynamic finite-element method. Second, the errorbackpropagation neural network is trained using the sample data. Finally, the trained network is utilized for defect identification. The present method is successfully applied to the identification of size and location of a surface defect in a solid. The dependence of network performance on the distribution of learning data is also discussed.
1994-01-25