Análise experimental e numérica do ensaio de compressão diametral sob diferentes configurações

Abstract

Tensile strength is a fundamental mechanical property in the analysis of fragile materials, mainly due to the susceptibility to failure when subjected to critical loads. There are three experimental methods that determine the effects of traction, being the tests of: direct traction, traction by diametrical compression and traction by flexion. Among them, the most used is the second, which consists in the application of uniaxial compression loading on cylindrical samples, producing transversal tensile stresses, uniform along the vertical diameter of the specimen. Despite this, the test has parameters that influence the final result of resistance of the materials, such as the width of the load distribution strip and the test speed, depending on the adopted solution, elastic properties of the materials can be changed. Thus, the present work aims to analyze the failure mechanism generated in the tensile test by diametrical compression when the sections and materials of load distribution flights are modified. For this, an experimental program was developed with 45 specimens of concrete under draft with a strength of 45 MPa. Of these, 10 specimens were used per flight configuration, namely: wooden flight (CMD), metallic flight 10 mm (CMT10), metallic flight 15 mm (CMT15) and metallic flight 20 mm (CMT20) and the others tested to axial compression. Among the experimental results, it can be highlighted that the CMD and CMT20 configurations showed no difference between their results. The values obtained by the CMD configuration directly coincide with the values proposed by the 6118 standard (ABNT, 2014). The rupture shape of the specimens showed great variation in the shape and degree of damage, which did not allow to determine a representative failure pattern. The starting point of the cracks on the face does not coincide with the exact center of the specimen in numerous tests, which diverges from the theory that supports the experiment. Numerical analyzes were also carried out under twodimensional (2D) and three-dimensional (3D) systems through the association of the method of discrete elements composed of bars (LDEM), together with the finite element method (MEF). Among the numerical responses, the stress-strain curves and the rupture shapes are similar to the experimental behavior, the failure with metallic strips always occurs close to the contact edge and with wooden slivers close to the effective width, in the 3D model the crack propagates from the faces to the center of the body as in 2D simulations, and also in length they spread from the edges to the center of the specimen. Keywords: Brazilian tensile test, loading strips, DEM associated with MEF