Study of Three-Dimensional Digital Core Reconstruction Based on Multiple-Point Geostatistics in a Cylindrical Coordinate System

Image inpainting and the extraction of different components from the electrical image log images are performed. After generating the training image, determining the appropriate template, multigrid, and event, calculating the conditional probability distribution function and defining the search tree, a three-dimensional digital core is reconstructed. A method for reconstructing three-dimensional digital cores based on multiple-point geostatistics in a cylindrical coordinate system is proposed in this paper. The electrical image log image is generated in a cylindrical coordinate system. The training image and conditional probability distribution function based on the cylindrical coordinate system are different from those based on a rectangular coordinate system. A three-dimensional digital core is evaluated based on an analysis of the local porosity and pore size. The experimental results show that three-dimensional digital core reconstruction based on multiple-point geostatistics in a cylindrical coordinate system is similar to the real core by statistical characteristics and connectivities. A three-dimensional digital core with high reliability can be used as a basis for numerical simulations in rock physics research.