Experimental Tests and Numerical Modelling of Hexagonal Concrete Specimens: Fid-Bau Portal

Experimental Tests and Numerical Modelling of Hexagonal Concrete Specimens

Bennett, T. / Jefferson, A. D.

Abstract This paper describes a novel experimental investigation into concrete cracking under complex and rotating loading paths. Hexagonally shaped specimens are loaded in compression until some damage has occurred and then unloaded and rotated before being loaded again until failure. The results from tests on specimens loaded in two directions are compared with control tests undertaken under monotonic loading. The development of the test setup and sample preparation procedure for these novel experiments is described. The results show that the diffuse cracking accumulated during the initial loading direction has the effect of lowering the peak load carrying capacity of the rotated specimen relative to the peak load obtained under monotonic loading. The results of the experiments performed are subsequently used as benchmark tests for numerical simulations, to test the ability of a constitutive model, as implemented in a finite element program, to simulate cracking under complex and rotating load conditions. Numerical simulations of the experiments are performed here using a recently developed plastic-damage-contact constitutive model (Jefferson (2003) Int J Solids Str 40:5973–5999; Jefferson (2003) Int J Solids Str 40:6001–6022; Jefferson (2004) Comput Concrete 1(3):261–284). This model couples the familiar damage and plasticity model frameworks, taking advantage of their relative strengths in modelling tensile and compressive stress states, respectively. In addition, this model employs a contact function to simulate crack closure and aggregate interlock behaviour. The numerical results manage to capture the essential features of the experimental behaviour, in particular, a reduction of the peak load attained as a result of damage initiated in the original loading direction. However, the constitutive model is unable to fully capture the unloading response. A relatively high value for the specific fracture energy Gf was employed in the simulations to emulate the relatively ductile response of the experiments. This relatively ductile response of the specimens is conjectured to be due to friction on formed crack planes, and currently local plasticity on these planes is not simulated within the model framework.