No abstract

A fracture mechanics model for alkali-silica reaction (ASR) is presented that deals with the case of a concrete made up of dense spherical aggregates. Chemistry and diffusion (of ions and gel) are not modelled. The focus is put on the mechanical consequences of the progressive replacement of the aggregates by a less dense gel. A ring-shaped crack then appears in the cement paste depending on the pressure build-up, according to an incremental energy criterion. The stored elastic energy and deformation of each configuration are determined assuming that each aggregate is embedded in an infinite cement paste matrix, through Finite Element Analysis. We note a very different behaviour of aggregates of different sizes. Adding the contributions of different aggregates leads to an estimate of the free expansion of a concrete of given aggregate size distribution. Parameters of the model are identified, providing a good fit to experiments taken from Multon's work.

International audienceWe present a comparative study of the ability of some micromechanics estimates to predict the overall properties of heterogeneous materials. We focus mainly on cracked materials, for which this task is difficult and many estimates fail. We study particularly the interaction direct derivative estimate, proposed by Zheng and Du, which is an approximation of the generalized self-consistent scheme, but has the very convenient property to be always explicit. A modified version of this estimate, called full-range IDD by Zheng and Du, yields good results when comparing all poromechanical coefficients predicted by the estimate to finite element simulations of a 2D cracked material in plane strain, up to crack density factors of 1 for aligned cracks and 0.60 for randomly oriented cracks. The accuracy of finite element computations of the overall moduli is also commented by plotting the convergence of the average of the properties as well as the confidence intervals on these averages

International audienceIn this article, we introduce a new micromechanical model for alkali-silica reaction. Our idea was to build a model with the following characteristics. First, the model has to be simple enough to be used to compute damage under loading and chemical attack at the level of each element in a structure code. Second, its parameters must be easy to identify on available alkali-silica reaction lab experiments. We have chosen to model the behavior of concrete containing aggregates such that most of the damage occurs in the cement paste. Using micromechanics and an energy criterion, the model remains analytical except for the minimization of the energy. The parameters were identified on Multon's triaxial experiments and good results were obtained for compressive loadings up to 10 MPa. ?? 2014 Elsevier Ltd