Feasibility of a data-constrained prediction of shale microstructures

The microstructures of hydrocarbon reservoir shales and sandstones are critical in determining their permeability, mechanical, electrical and other physical properties. However, the available experimental techniques in characterising shale 3D microstructures are time consuming and sample destructive. This paper outlines our effort investigating the feasibility of determining shale microstructures using multiple computedtomography (CT) data sets with different X-ray beam energies, together with a data-constrained microstructure (DCM) modelling approach. A sample composition data set was generated using an experimental tomography data set as a template for a shale rock sample with quartz, kaolinite and mica (illite). X-ray transmission through the sample was numerically simulated to produce X-ray projection images from the sample at multiple incident angles. This was repeated with several beam energies. The simulated projection images were used to reconstruct the 3D distribution of X-ray attenuation coefficient inside the sample for each energy. The compositional microstructures were predicted with our DCM software using the above 3D X-ray attenuation maps as constraints. Our preliminary comparisons indicate that the DCM-predicted compositional microstructure is qualitatively consistent with the known original microstructure. The accuracy of the prediction would depend on the CT reconstruction accuracy and the level of noise. Further work is in progress in quantitative evaluation of the predictions under various experimental conditions and rock microstructures.