ICF10B

FRACTURE OF CONCRETE CYLINDERS SUBJECTED TO TORSION G. Lilliu and J.G.M. van Mier Microlab, Faculty of Civil Engineering and Geosciences, Delft University of Technology P.O. Box 5048, 2600 GA Delft, The Netherlands ABSTRACT In order to validate the 3D version of the 2D ’Delft’ lattice model some torsion tests were conducted on concrete cylinders, both notched and un-notched. A biaxial loading machine with rotational and vertical loading axes was used. Tests were conducted under different axial restraints. It was observed that a perturbation of the stress state induced by the torque may affect significantly the crack pattern. Under axial compressive stress, after a spiralling crack, an axial crack was detected in un-notched specimens. In notched cylinders, the failure surface is conical if an axial load is present, while it appears flat and localized within the notch height in case no axial load is applied. KEYWORDS Concrete, 3D fracture, torsion, multiaxial stress. INTRODUCTION At the beginning of the 1990’s a 2D lattice model was developed in Delft [1]. In the past ten years this model has been used for simulating experiments on concrete specimens subjected to plane stress. Despite the realistic crack patterns, the response of the model in terms of load-displacement diagrams has always been far too brittle when compared with the experimental results. A possible cause is that fracture processes are 3D phenomena. In fact, the presence of randomly distributed inclusions in the material (aggregates) produces curvatures of the crack surface in all directions. As a consequence, the dissipated energy increases. For this reason, a 3D version of the lattice model was developed [2] and torsion tests were designed to validate the numerical simulations. Several attempts have been made in the past to design a set-up for conducting torsion experiments on cylinders [3,4]. The main reason for conducting torsion tests was that they could likely give insight in the mode III fracture energy, GIII f . The same tests, conducted on notched cylinders, allow measurements of mode II fracture energy, GII f . In notched cylinders the shear stresses induced by the applied torque should produce a planar crack and, as a result, GIII f andGII f should coincide. However, it seems that the condition of pure shear stress cannot be achieved in practice. There is experimental evidence that shear fracture tests are very sensitive to axial restraints, such as the axial compressive force arising from presence of friction at the supports. In the torsion tests it has been observed that the failure surface can change drastically from planar to conical [5].

RkJQdWJsaXNoZXIy MjM0NDE=