Next: Experimental Design Up: Microstructural Models and Previous: Concrete Microstructure Model

Simulation Procedure

The simulation procedure used in this study is illustrated for a mortar in Fig. 2. To begin a simulation, the cement PSD is used to establish the interfacial zone thickness, t_ITZ, as being equivalent to the median particle diameter, ignoring any effects of bleeding which would tend to locally augment the value of t_ITZ. Aggregate particles following the aggregate PSD of interest are placed into the concrete volume and systematic point sampling is used to determine the volume fractions of interfacial transition zone (V_ITZ) and bulk (V_bulk) paste for this choice of aggregate PSD and t_ITZ [11]. At this point, for the cement hydration model, the thickness of the single aggregate particle is chosen to match this ratio of V_ITZ / V_bulk for the chosen cement PSD in the cement paste model volume. Cement particles are placed into this computational volume to achieve the desired w/c ratio, and the hydration model is executed to achieve the chosen degree of hydration. The porosity is then measured as a function of distance from the aggregate surface and converted to relative diffusivity values using Eqn. 1. These values are averaged in two subsets, those lying within t_ITZ of the aggregate and those in the "bulk" paste. The ratio of these two diffusivities, D_ITZ / D_bulk, is then used as an input into the original concrete model. Here, random walk (myopic ant) techniques are employed to estimate the diffusivity of the overall concrete system consisting of aggregates with a diffusivity of 0, bulk paste with a diffusivity of 1, and interfacial transition zones with a diffusivity of D_ITZ / D_bulk. The random walker techniques employed for this calculation have been described completely in Refs. [17,18]. After a large number of walkers have each taken a large number of random steps in the concrete, the relative diffusivity of the concrete, D_conc / D_bulk, can be calculated [4,17, 18]. This value can finally be converted into an absolute chloride ion diffusivity for the concrete, D_conc, by multiplying it by D_bulk / D_o determined from the cement-level microstructural model and by D_o, the diffusion coefficient of chloride ions in bulk water, given as 2.0 x 10^-9 m²/s in Ref. [19]. By changing the value of D_o to correspond to that measured for the specific ion of interest, such as sulfate ions, the above techniques can be generalized to other ionic species of relevance in cement-based materials (assuming no chemical effects such as binding and reactions).