Crack detection in 2D structures using wavelet transform and the boundary element method

Resumo

Detecting damages in structures is of paramount importance as they can lead to the collapse of structural
elements. In the current scenario, two reliable techniques for non-destructive testing of structures, namely
ultrasound and X-rays, are practically applicable for precise localization of hidden damages. However, such
techniques are costly, time-consuming, and require special procedures to be performed. Currently, research
suggests the possibility of using numerical methods to assist in damage detection in structures. Generally,
numerical methods for damage detection, such as cracks, are based on finite element modeling and comparisons
between the signatures or responses of the structures before and after the occurrence of damage. This article
proposes a new approach to damage identification based on the Wavelet Transform (WT) and the Boundary
Element Method (BEM), making the numerical modeling of structures more accurate, straightforward, and without
requiring the structure's signature before the damage occurs. To validate the adopted methodology, a cantilever
beam containing a crack with different orientations will be modeled using the in-house programs BEMLAB and
BEMCRACKER2D, as well as the algorithms implemented in MATLAB for handling wavelet coefficients and
damage localization. The obtained results indicate that WT is a useful tool for identifying and monitoring damages
in structures.