Investigation on the combining application of artificial neural network and wavelet analysis in ultrasonic testing

Guohua Chen, Xiangbo He, Changhuan Xie

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

1 Citation (Scopus)

Abstract

Ultrasonic Testing (UT) has been applied widely in many professional fields for its advantages, but the accuracy of analysis is interfered seriously by the existed subjective or objective factors. The research on the application of Artificial Neural Network (ANN) in Ultrasonic Testing, which has the function of pattern recognition, has aroused the engineering attention in many fields. Combined with the developing state of the art having been applied in UT, we carried out studies in the following aspects in the paper. How to organized the structural parameters of ANN, how to train and test the neural network, how to utilize ANN to process signal in defect recognition, all of these are discussed. And two typical models in the application of the combination of ANN and Wavelet analysis method are analyzed, One is based on taking the Wavelet analysis method in signal processing as pretreatment to get the input vectors of ANN while the other is based on taking the Wavelet analysis method as the excitation function of ANN. In addition, some shortcomings and problems in defect quantitative recognition are also discussed. Finally, some suggestions on how to improve ANN technique combined with Wavelet analysis in practical application are presented, and the prospect of ANN technique applied in UT is also predicted.
Original languageEnglish
Title of host publicationProceedings of the International Symposium on Test and Measurement
Pages1448-1452
Number of pages5
Volume2
Publication statusPublished - 16 Oct 2003
Externally publishedYes

ASJC Scopus subject areas

  • General Engineering

Fingerprint

Dive into the research topics of 'Investigation on the combining application of artificial neural network and wavelet analysis in ultrasonic testing'. Together they form a unique fingerprint.

Cite this