Next: Maturity Method Up: Computational Tools Previous: Percolation Theory

IMAGE RECONSTUCTION

Since for the prediction of performance, it is necessary to model the microstructural development of cement paste in three dimensions, realistic three-dimensional starting images of cement particles in water are needed. While SEM and X-ray imaging can provide segmented images of cement particles in two dimensions [15], such as that shown in Figure 1, a direct three-dimensional measurement is impractical. Fortunately, three-dimensional image reconstruction techniques originally developed for two-phase porous materials [29, 30,31,32,33] can be readily adapted to distribute the five cement phases (tricalcium silicate (C3S)1, dicalcium silicate (C2S), tricalcium aluminate (C3A), tetracalcium aluminoferrite (C4AF), and gypsum) amongst a collection of digitized spherical cement particles. It should be noted that the phases in an actual cement material are not limited to these five and that even these five do not typically exist in their pure form, but this simplification is commonly made for characterization and modelling purposes.

The image reconstructions rely on the fact that for an isotropic material, the autocorrelation function of a phase is the same in two and three dimensions. Thus, one can measure the autocorrelation functions in two dimensions for an actual image of the cement particles and employ them in a three-dimensional reconstruction. In the reconstructed three-dimensional image, the particle size distribution (PSD), phase volume fractions, and phase surface area fractions of the real cement powder can all be preserved [7].


Next: Maturity Method Up: Computational Tools Previous: Percolation Theory