Stochastic reconstruction of chalk from 2D images

We study the stochastic reconstruction of the microstructure of chalk, from limited morphological information that may be readily extracted from 2D images of the pore space. Backscatter Scanning Electron Microscope images of a North Sea chalk sample are analyzed to determine de-scriptors of pore space morphology, such as the void-phase autocorrelation function, and void- and solid-phase chord distribution functions. This information is used to constrain the stochastic reconstruction of the chalk sample in 2D and 3D by a simulated annealing method. Quantitative analysis of 2D reconstructions using different morphological constraints reveals that imposing chord distribution functions results only in minor improvement over what is achieved by using the void-phase autocorrelation function as the only constraint. This result is further verified by geometric and topological characterization of a 3D replica of the sample generated using only autocorrelation function constraints. Pore and throat size distributions determined by 3D pore space partitioning methods, are consistent with mercury porosimetry results. The predicted permeability and formation factor are shown to be in very good agreement with experimentally determined values.