The study of crack initiation and propagation is primary to the fracture mechanics domain and essential for developing new materials. Such a study is relevant for refractory ceramics adopted as thermal insulators in high-temperature industrial applications. In particular, mechanical and thermal tests are essential in measuring properties and contribute to the material selection process. Such tests can be assisted using non-intrusive measurement techniques, in particular, by the Digital Image Correlation (DIC) technique, which enables to obtain full-field information concerning the displacement and strain fields in a loaded configuration. Identifying the crack location is a challenge for DIC techniques and requires special treatment of the correlation process to get reasonable resolution around the crack. Global DIC approaches can benefit from mesh refinement strategies to improve the identification of the displacement field and, consequently, the crack's location. In this context, the present article introduces mesh refinement strategies to improve the image correlation, allowing to identify the crack location in cracked specimens better. The adaptivity approach is developed using Correli framework and the Application Programming Interface (API) of the GMsh software, which has several features to control the element mesh size distribution, particularly by controlling mesh size parameters using strain and local gray-level residue. Herein, the mesh strategies are applied in a model composite system submitted to temperature variation. The composite consists of a brass inclusion surrounded by an alumina matrix. The thermal expansion coefficient mismatch between the phases leads to a radial crack pattern when the material is heated. The evaluation of the mesh refinement algorithm is established in terms of reducing global gray-level residual in the correlation procedure, and the enhanced definition of the crack location. The present article contributes to the DIC processing, getting more information on specific regions based on strain and gray level residual information, and also furnishes information to verify computational fracture mechanics tools.
Read full abstract