Abstract
This work is a continuation of the process of developing a model for predicting the potential of Portland cements to exhibit expansion due to delayed ettringite formation, as presented in part 1 [1]. The model was derived basing on the kinetics theory applied to cement chemistry and evaluates two primary conditions of, the presence of sufficient active alumina in cement and the requirements for sulphate to alumina ratio applied against the countering microstructural effects, to determine the underlying kinetic characteristics for DEF in the system. The primary purpose of the model is determination of the intrinsic presence of the potential for DEF in cements.
In the second part of this work herein presented, the model is validated and tested for robustness by its application to a wide range of existing experimental data for DEF studies, taken from various global sources available in the literature. Over 164 data sets taken from various independent works, were used in the validation. It was found that the model was in strong agreement with experiment data for both expansive and non-expansive cements. However, some few outputs showed disagreements between model predictions and experimental observations. Most of the important disparities appear to occur with cements of high Fe2O3/Al2O3 ratios and may be related to inadequacy of the Bogue formula used in the model. These revealed aspects where future refinement of the model will be required. The model at this stage is applicable to plain Portland cements with a future possibility of improvements to examine more complex systems containing pozzolanic materials.