| Summary: | © 2018 EAGE. Estimations of the elastic properties of a core sample are extremely sensitive to the quality of the CT-scan image. A coarse image resolution does not allow separating the grains and estimating their roughness. The size of images with fine resolution is either below representative volume or so representatively large that require huge computational resources for the numerical simulation of static loading tests. We present a new multi-scale numerical methodology that combines the geostatistics, computational topology and numerical upscaling. The essence of the approach is to estimate the distribution of the grain surface roughness, the distance between them and cement material by microscopic images (SEM), then to calculate the effective parameters of cement distribution for an equivalent model with flat contacts between grains. Finally, to use this effective cement to fill the interfaces between grains in a digital model built on CT images of medium resolution by computational topology.
|
|---|