Estudo da cinemática do ensaio de fratura pelo método da cunha com auxílio da técnica de correlação de imagens digitais

Abstract

Castable refractories are indispensable in the basic industry due to their ability to maintain physicochemical properties at high temperatures. Owing to the quasi-brittle nature of these materials, crack initiation and propagation over prolonged service times are inevitable, and improving their durability can be achieved through the development of compositions that maximize resistance to crack propagation. Fracture energy, directly related to this resistance, can be obtained by the Wedge Splitting Test (WST). The Digital Image Correlation (DIC) technique is frequently used with the WST, as it enables non-contact full-field displacement measurements on the specimen. In this work, the repeatability of fracture energy was investigated for seven nominally identical specimens, and the spurious contribution introduced when energy is determined exclusively from testing machine data was quantified. For three WSTs exhibiting the highest repeatability, the efficiency of the experimental setup in transferring work to the specimens, the kinematics of the Load Transfer System (LTS), and the fracture process on both analyzed faces were also investigated. Results showed that measurements based on the vertical actuator displacement systematically overestimate fracture energy by approximately 8.4% relative to CMOD (Crack Mouth Opening Displacement)-based measurements obtained via DIC, with the overall system efficiency stabilizing at 92-93% during the final loading stages. The kinematic analysis demonstrated that the loose-component LTS exhibits globally stable behavior, with negligible sliding at the contacts and roller motion asymmetries below 5% of the final CMOD. Crack propagation showed a two-stage behavior: spatial advance of the crack front followed by progressive crack opening in the post-peak regime.