 |
For one of the French fly ashes, an extensive experimental data set of adiabatic heat signatures is available. For a variety of water-to-cement (w/c) ratios by mass, and fly ash contents, concretes have been prepared and characterized by monitoring their temperature rise over time when hydrating under adiabatic conditions. This data set can be used for a direct comparison with model predictions. This fly ash is a relatively simple one, in that it can be approximated as containing only three components: silica (S), an aluminosilicate (AS), and inert material. In addition, from SEM and image analysis, it appears that for this fly ash, most of the particles (unlike typical cement particles) are monophase. Thus, we can use the program distfapart.c in Appendix B to distribute the fly ash phases randomly amongst monophase particles to achieve the appropriate volume fractions (determined to be 0.33 for AS, 0.39 for S, and 0.28 for inert material based on the oxide composition determined for the fly ash). Computationally, this approach is much easier than using the detailed autocorrelation analysis to distribute the phases, as is typically performed for a cement [1].
The dissolution probability of the aluminosilicate and the reaction probability of the silica (parameter PPOZZ) were adjusted to provide a ``best'' fit to the experimental adiabatic heat signature data at w/c=0.65 and 50% fly ash content. These values were then held constant in simulations for other w/c ratios and fly ash content combinations. A heat of hydration value of 800 J/g AS reacted was used in the model for the aluminosilicate to stratlingite conversion. This value was determined based on experiments in which the fly ash alone was hydrated in a solution of lime (CH) and the system temperature rise monitored. An activation energy of 83.14 kJ/mole (the same as that determined for the pozzolanic reaction of silica fume [9,10]) was used in the model for all of the reactions involving one or more components of the fly ash. Example comparisons of the experimental and model temperature rise vs. time curves are provided in Figures 2, 3, and 4. In general, the agreement between model and experiment is always within a few degrees Celsius, as had been observed previously for systems with and without silica fume additions [10]. The higher model temperatures at longer times are perhaps due to more water being incorporated into the cement-fly ash hydration products than that indicated by the stoichiometries in Figure 1. Since the hydration is being executed under sealed curing conditions, it will effectively terminate when all remaining porosity is empty (as opposed to water-filled) due to chemical shrinkage. Another possibility is that certain hydration products (e.g., ettringite) are unstable at higher temperatures. If their deterioration reactions are endothermic in nature, the predicted temperature rise based on the computer model would exceed that observed experimentally. Because this fly ash is rather simple in structure and contains only a few distinct phases, further validation of the 3-D microstructural model with other fly ashes will definitely still be needed, as outlined in the next section of this report.
|
 |
 |