Experimental investigations on the size effect of fracture energy for concretes of hydraulic structures

Abstract

Numerical simulation using non-linear constitutive laws turns out to be a necessity for the case of hydraulic structures affected by the alkali-aggregate reaction (AAR). The fracture energy constitutes an important input parameter for these nonlinear laws and requires an experimental characterization step. Because of the size of the hydraulic structures, it is a question here of characterizing a representative fracture energy and independently of all size effects phenomena. As such, this work aims to characterize the asymptotic fracture energy of two different mixtures of an existing hydraulic facility, having maximum aggregate sizes of 38 mm and 76 mm. It constitutes a first phase of a project whose final objective is to characterize the impact of the AAR on the asymptotic fracture energy. The experimental campaign consists of 71 specimens divided between two mixtures with different maximum aggregate sizes: 38 and 76 mm and two different compositions: without and with alkali overdose. 29 of the 71 specimens were already tested on two different experimental setups: a threepoint bending test (RILEM TPB) and a wedge splitting test (WST). In order to filter the size effects, the method known as SBEM for Simplified Boundary Effect Method was used [1], based on a bilinear distribution of local fracture energy due to a boundary effect. The analysis of the results of the specimens tested confirms the presence of the size effect for the tested mixtures and the positive impact of the maximum aggregate size on the fracture energy. The results show that the fracture energy from a given test is proportional to the area of the cracked concrete surface. It was possible to extract, using the SBEM method, an asymptotic fracture energy independent of the size of the specimens and which would represent the fracture energy of a RILEM TPB test or a WST test on a very large specimen. This asymptotic fracture energy has a magnitude of an order of 2 to 3 times greater than that obtained from a TPB RILEM test, or that predicted by the semiempirical equations based on this standardized test. J. Lemery, M. Ben Ftima and M. Leclerc