Analysis of compressive fracture of three different concretes by means of 3D-digital image correlation and vacuum impregnation

Fracture of three different concretes under uniaxial compression was investigated. The three different materials were normal concrete, high strength concrete and foamed cement, with uniaxial strength varying between 8 and 92 MPa. The prismatic specimens were loaded between two different types of loading platens, providing maximum boundary restraint (rigid steel platens) or almost zero boundary restraint (Teflon (PTFE) sandwich inserts). Fracture propagation was measured using 3D-digital image correlation where emphasis was placed on measuring crack development just before and at peak stress. The stress at first cracking is dependent on the resolution of the digital image correlation technique, and as such not very well defined. The frictional restraint of the loading platens is recognized in the failure modes, in particular in the direction and structure of the main cracks. Also, the results for normal and high strength concrete underscore previous results: high friction results in a higher uniaxial compressive strength. This finding was not confirmed for foamed cement, where both loading systems gave approximately the same compressive strength. Likely the Teflon inserts do not function properly on the highly porous surfaces of the foamed cement specimens. The direction of the main cracks at peak is defined before the softening regime is entered: cracks have a vertical orientation when loading is applied through Teflon inserts, whereas 'en-echelon' tensile cracks forming inclined shear bands are observed for rigid loading platens. The results from 3D-digital image correlation were confirmed by results from vacuum impregnation with fluorescent epoxy after test termination. Although digital image correlation is capable of showing cracks at relatively early stages of loading, the vacuum impregnation technique shows all the fine detail and crack branches, but this can, unfortunately, only be done once for each specimen. Therefore, combining the two different techniques is more appropriate. (c) 2010 Elsevier Ltd. All rights reserved.