ABSTRACT
The simulation of transport properties is implemented in two ways: 1) through capillary pores only; or 2) through capillary pores and hydration products (more precisely CSH gel). Because the size of the sparse matrix A is N times N the matrix is not stored explicitly but only implicitly in the vectors cx, cy, and cz which store the conductance coefficients of the bonds among voxel faces in the x, y, and z directions, respectively. The system of equations is solved by conjugate gradient algorithm. The bottleneck in this solver routine is the multiplication of matrix A with an arbitrary vector. In order to speed up this
1 INTRODUCTION
Transport in cementitious materials plays a crucial role in both the degradation process of building materials and the containing of hazardous wastes. A good quality of the concrete cover enhances the durability and reliability of concrete structures. One parameter that is of paramount importance is the effective (macroscopic) transport coefficient. A relevant and reliable method is needed in order to obtain effective diffusion coefficient (Def) with an acceptable accuracy. Various experimental methods exist but they are either time consuming or have too many drawbacks. Simulating a microstructure evolution during hydration is an advantageous (fundamental) starting point to model the morphological nature of the effective diffusion (transport) coefficient. A virtual 3D microstructure created with an available hydration models provides a basis for the analysis of the morphological influence, including the porosity, tortuosity, constrictivity and the pore water content effect onto the effective diffusion coefficient. Such an approach contributes to a better understanding of the phenomenology and thus improves the predicting reliability of the coupled transport models.
