Next: Summary Up: Main Previous: Computer simulation methods

Results and discussion

Figure 2 shows simulation and experimental results for the relative conductivity Γ, plotted as a function of capillary porosity. These data represent hydration times from about one hour up to approximately 200 hours, with a maximum degree of hydration of about 0.65. Figure 2 shows that over the entire range of capillary porosities studied, the simulation results and the experimental results agree reasonably well, certainly within less than a factor of two. For the highest porosities, there appears to be some systematic disagreement between simulation and experiment, with the experimental data lying above and following a different functional form than the simulation data. The dashed line in Fig. 2, which goes through the high porosity experimental points fairly well, is of the functional form φ^1.5, as explained below.

In the case of simple suspensions of insulating particles in a conducting fluid, Γ is experimentally known to generally follow a power law [13]: Γ = φ^m, where m ≈ 1.5 for spherical particles. A paste of water:cement ratio 0.5 has an initial capillary porosity of φ = 0.62, so that the initial value of Γ might be expected to be about (0.62)^1.5 = 0.48. In this limit, the simulation data appear to obey a power law with a slightly higher power, which is probably an artifact of the finite resolution with which the initially spherical cement particles are represented. The power law form, as well as the simulation, assume only uniform, bulk conduction processes in the conducting fluid. In particular, there are assumed to be no extra conduction processes at particle surfaces. The experimental data seem to agree well with this assumption. It is interesting to note that the experimental conductivities appear to break away from the φ^1.5 power law at a porosity of about 50%. This capillary porosity is realized at a degree of hydration of 0.23, which is very close to the degree of hydration at the final set point. When the cement paste has set, implying a continuous solid phase now exists, the description of the paste as a suspension will certainly not be correct.

The determination of the smallest degrees of hydration in Fig. 2 are also the least precise, implying horizontal error bars on the experimental data, especially in the high porosity region. We are unable to estimate the size of the error bars at this time.

In addition, it is worth pointing out that the rules for the growth of reaction product in the microstructural model are the same during the entire hydration reaction, with no distinction made between early and late product [14] with respect to specific volume ratios between product and cement. The computation of capillary porosity from degree of hydration for the experimental data also depends on this assumption. Any differences in specific volume between early and late hydration will then affect both experimental and theoretical results.