Application of Neural Networks in Ultrasonic Non-Destructive Testing (Review)

Abstract

A review of the current state and experience of the practical application of artificial neural networks in ultrasonic non-destructive testing is presented. Separately, the features of the use of neural networks for the classification of defects according to data obtained by echo-pulse methods and the TOFD are considered. Information is given on the efficiency of defect classification using neural networks for various ultrasonic testing tasks.