Simulation of cement degradation processes by means of coupled water flow and reactive solute transport modeling

Resumo

Given the risk inherent to radioactive waste repositories and that cementitious materials have achieved widespread use in recent years in the construction of the multi-barrier systems that protect the environment from potential leakage, the estimation of the durability of such materials is of vital importance for it is tied to the safety of those structures in the long term. The present thesis is aimed towards that goal, intending to provide insight into the phenomena involved in the evolution of the degradation of cementitious materials by means of simulations carried out with reactive transport models. Two not entirely independent lines of research converge to this end: on the one hand, the conceptualization of concrete as a mineral aggregate (as such liable to deterioration) as well as a porous material, in whose voids transport processes develop, and on the other hand, the use of reactive solute transport models, which has been increasingly and more widely adopted in this field, since they provide a mechanistic approach to address the complex chemical and transport phenomena involved in the processes of cement degradation. The first of such guidelines acquires a concrete meaning in the digital-image-based generation of three-dimensional representations of hardened cement microstructure, as well as in the appropriate methods available to describe ionic diffusion and the accuracy of the constitutive laws (e.g., porosity/permeability, porosity/diffusivity, etc.) that were developed for these materials. With respect to the other guideline, the application of reactive transport models to the field of cementitious materials is relatively recent. It is thus important to trace the evolution of such procedures up to this date and identify how they were (and are) affected by the idiosyncratic properties of these materials. These objectives not only crystallized in the description of the milestones of the transition that has led to the current state of the art but also in the critical appraisal of current models, as has been attempted here by examining the most appropriate methods for modeling the diffusion of electrically charged particles. Addressing this particular matter is one of the original contributions of this work and, in that sense, results showed that Fick's law-based approaches to diffusive transport might not be accurate enough to model the degradation of cementitious materials (see Paper III). At the core of this thesis, results obtained by means of the reactive transport program CORE2D, specifically adapted to evaluate the evolution of the physical properties of concrete, were found to satisfactorily match empirical data from experiments published in the scientific literature (see, e.g., Paper I). Moreover, in concert with digital-image-based cement hydration programs, reactive transport models can become an extremely useful tool in the design process of concrete and cement-based grout mixes. The latter case (see Paper II), which can be considered critical since it implies a compromise between two conflicting requirements (the workability and injection capacity at early stages, on the one hand, durability, on the other), is illustrative of the scope of this methodology.