Crack Parameter Characterization by a Neural Network

Date
1996
Authors
Takadoya, M.
Achenbach, J.
Guo, G. C.
Kitahara, M.
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Research Projects
Organizational Units
Journal Issue
Department
Abstract

A neural network with binary outputs is presented to determine the angle and the depth of a surface-breaking crack from ultrasonic backscattering data. The estimation procedure is divided into two steps: (1) The angle of the crack is estimated in the range from 10 to 70 degrees with a precision of 5 degrees. To improve the accuracy of estimation, information on the integral of the backscattered signal is utilized. (2) When the angle of the crack has been estimated, the depth of the crack is determined with a precision of 0.5mm in the range from 2.0mm to 4.0mm. This determination is achieved by employing sets of neural networks corresponding to various angles of the crack.

Comments
Description
Keywords
Citation
DOI