Locating the crack tip of a surface-breaking crack: Part I. Line crack
Journal
Ultrasonics
Journal Volume
36
Journal Issue
7
Pages
803-811
Date Issued
1998
Author(s)
Abstract
Four distinct algorithms to locate the crack tip of a surface-breaking crack using only the arrival time information of the first diffracted waves are described and compared. To illustrate these algorithms, a line crack in a half-plane is considered. The first two algorithms are based mainly on elementary geometric arguments, where the crack tip is formulated as the intersecting point of two ellipses (algorithm 1) and/or three circles (algorithm 2). The other two algorithms are formulated as optimization problems, where cost functions based upon the arrival time data of diffracted waves are constructed. The unknown crack tip coordinates are then determined by minimizing the cost functions through the Lagrange multiplier method (algorithm 3) or the simplex method (algorithm 4). In the numerical experiments, the exact arrival times are superimposed by Gaussian error with different levels to simulate the real extracted arrival times from experimental signals. The numerical optimization method (algorithm 4) is found to have the best performance with respect to noise, as well as for accuracy. Moreover, the recovery of the crack length is much more robust than the orientation and depth. © 1998 Elsevier Science B.V. All rights reserved.
Four distinct algorithms to locate the crack tip of a surface-breaking crack using only the arrival time information of the first diffracted waves are described and compared. To illustrate these algorithms, a line crack in a half-plane is considered. The first two algorithms are based mainly on elementary geometric arguments, where the crack tip is formulated as the intersecting point of two ellipses (algorithm 1) and/or three circles (algorithm 2). The other two algorithms are formulated as optimization problems, where cost functions based upon the arrival time data of diffracted waves are constructed. The unknown crack tip coordinates are then determined by minimizing the cost functions through the Lagrange multiplier method (algorithm 3) or the simplex method (algorithm 4). In the numerical experiments, the exact arrival times are superimposed by Gaussian error with different levels to simulate the real extracted arrival times from experimental signals. The numerical optimization method (algorithm 4) is found to have the best performance with respect to noise, as well as for accuracy. Moreover, the recovery of the crack length is much more robust than the orientation and depth. © 1998 Elsevier Science B.V. All rights reserved.
Type
journal article