Multiscale pore structure characterization based on SEM images

The micropore structure’s permeability contribution to total of the heterogeneous reservoir with multiscale pore structures is critical for evaluation but still not well understood. This paper proposes a structure characterization method based on high-resolution SEM images quantitatively analyse structures’ content and their contributions via six steps. First, image-based rock typing implemented classify into different types using random forest algorithm. Second, 3D model macropore every reconstructed applying MPS method. Third, each calculated LBM, corresponding REV this estimated. Four, an upscaling process carried out divide many cells whose length determined by maximum structures. Five, cell coarse grid LBM except some that are randomly selected as assigned directly according types. Finally, evaluated comparing before after assuming target impermeable. result shows varies significantly its permeability, content, spatial distribution, structure.