Crack Identification by Digital Image Correlation Method Using Crack Shape as Prior Information

Abstract

Abstract A new crack identification method that estimates the cracks in invisible locations based on the surface deformation measured by digital image correlation (DIC) is developed. An inverse problem is setup to estimate such invisible cracks from surface deformations. The inverse problem has an ill-condition because of noise contained in surface deformations. Our proposed regularization method uses prior information and Expectation a Posteriori (EAP) estimation. Prior information includes candidate crack shapes and surface deformations due to cracks. The candidate crack shapes are created by determining a crack's starting point and propagating it based on the force at its perimeter (ligament). A prior distribution is the surface deformations due to the candidate crack shapes. The likelihood distribution is a surface deformation measured by the DIC method. A posterior distribution is defined from the prior and likelihood distributions. In this study, the estimated result is the expected value of the posterior distribution. The validation test was performed, and the result shows that the proposed method superior to the conventional L1-norm regularization method.