Investigations on shear transfer by aggregate interlock with a unique test setup (TorAx)

Abstract

The structural behaviour of concrete is controlled through shear and normal stress transfer in cracked and uncracked concrete. To assess aggregate interlock (mode II), different test setups have been developed in the past decades and used for this purpose with more or less disadvantages. In test methods with finite shear length, the maximum transferred shear stresses due to aggregate interlock were only reached at the edges of the ligament length. In consequence, the experimental average shear strength is significantly smaller than the actual shear strength of a material point, since failure is induced at the edges of the specimen or ligament length. Therefore, in tests with finite shear length, the experimentally obtained value of shear strength is biased by the ligament length. In contrast, due to the continuous shear length, the experimentally determined average shear stress from a torsion test of circular concrete specimens is equal to the sought-after unbiased shear strength of a representative material point.Based on the experience of small-scale tests, the authors developed a new large-scale test setup for testing of circular concrete specimens with torsional- and axial forces (TorAx), that overcomes the shortcomings of current test methods. The setup allows imposing mode II kinematics by applying torsion and/or compression/tension to either pre-cracked or non-cracked concrete specimens. For the investigation of aggregate interlock, the test setup allows for a continuous shear length at specimens to avoid edge and size effects.This paper shows test results investigating the shear transfer by aggregate interlock with a torsional test setup and gives an overview on the newly developed TorAx test setup.