低炭素社会構築のためのCCS

[3404] Effects of segmentation methods on physical and hydrologic properties for 3-D digital rock model

^○曹聡¹, 蒋飛¹, 辻健¹ (1.九州大学)

司会:末包哲也(東京工業大学)

Keywords:Segmentation, Digital rock, Lattice Boltzmann, Porosity

To fulfill CO2 storage process, appropriate reservoirs (e.g., rock texture) should be selected out previously. Instead of establishing stereo rock model and measure relevant rock parameters in laboratory, foregoing researchers have established the concept of digital rock. Using the digital rock approach, we can calculate physical and hydrologic properties under various conditions. Furthermore, we can compute chemical reactions in pore space of the digital rock model. In first step of these analyses, we construct 3D digital rock from CT value extracted from micro-CT scan. Although several methods for rock segmentation were proposed, they are not well evaluated. Here we adopt 3 segmentation methods (single threshold, Otsu threshold, watershed transform) to process original 2D CT images and then reconstruct 3D digital rock models. We study the segmentation induced differences of physical and hydrologic parameters of rock, including porosity, pore size distribution, permeability, and coordination number distribution of sandstone sample. Since each segmentation method may specialize in processing particular kind of rock image, selecting out optimal segmentation method will help reveal most accurate characteristics of various types of rock, from which researchers can decided if certain kind of rock will be suitable for storing CO2.