Abstract

Acoustic emission analysis is a non-destructive testing method used for identifying defect locations in a material. The location problem can be solved analytically, but the corresponding program differs from case to case. For this reason an empirical model was used which has a universal structure of a neural network. It is able to learn from known examples to find unknown location of a detected acoustic emission burst. The model is based on a non-parametric regression algorithm which receives data from piezoelectric sensors and converts them into the acoustic emission source coordinates. Fiber glass composite plates were used in experimental acoustic emission testing. The results of experiment have shown that this method can be used for identifying defect locations without solving location problem analytically.

Keywords: Ultrasonic Testing

Source: Conference Proceedings of the 4th International Conference of the Slovenian Society for Nondestructive Testing, 24 - 25 April 1997, Ljubljana, Slovenia.
Pages:31 - 40

Publisher:
Slovenian Society for Nondestructive Testing, Faculty of Mechanical Engineering, Askerceva 6, SLO 1000 Ljubljana, Slovenia. Tel +386 61 177 12 03, Fax +386 61 218 567, Email: janez.grum@fs.uni-lj.si.