Mode density function simulation algorithm for three-dimensional reconstruction of porous media

Abstract

The invention discloses a mode density function simulation algorithm for three-dimensional reconstruction of a porous media. A three-dimensional structure of the porous media can be reconstructed with a computer technology and a mathematic modeling method by utilizing a two-dimensional image. The invention aims to achieve three goals of accurate three-dimensional porous media reconstruction. Firstly, a three-dimensional reconstruction result is consistent with a low-order statistic property of a training image; secondly, the three-dimensional reconstruction result (three orthogonal tangent planes) is similar to a morphological characteristic of the training image; and thirdly, the algorithm has stability. The invention has main innovations that a mode density function concept and an iterative algorithm, namely, the mode density function simulation algorithm taking the mode density function as a target function and fusing multiple stages of grids are provided; a neighborhood statistic method and an adjacent grid inverse point selection method are provided for speed problems of key points in the reconstruction algorithm, and a temperature control mechanism is simplified; and that the three-dimensional structure of the porous media can be accurately reconstructed by utilizing the two-dimensional training image is verified by utilizing a testing experiment of a rock core sample.

Description

technical field [0001] The invention relates to a technique for randomly simulating a three-dimensional (3D) structure of a porous medium from a two-dimensional (2D) training image by using mathematical statistics and computer modeling methods, and belongs to the technical field of image processing. Background technique [0002] The 3D structure is helpful for in-depth study of the microstructural properties of porous media, but due to various reasons, it is impossible or difficult to obtain the 3D structure of porous media in actual production. A 3D structure with similar features can be reconstructed from a 2D image with the help of computer technology and mathematical modeling methods. [0003] Around this technology, many scholars have made a lot of efforts and explorations over the years, and a variety of more classic reconstruction algorithms have also been born. According to the reconstruction method, the current 3D reconstruction algorithms can be divided into two t...

AI Technical Summary

Problems solved by technology

[0002] The 3D structure is helpful for in-depth study of the microstructural properties of porous media, but due to various reasons, the 3D structure of porous media cannot be obtained directly or is difficult to obtain in actual production

However, this type of reconstruction algorithm only emphasizes the visual similarity between the reconstruction result and the original training image, and chooses to ignore the consistency in the statistical characteristics of the two.

[0008] The process-based method is also a 3D space direct reconstruction algorithm. This algorithm simulates the geological formation process, but the complex geological formation process and conditions make the reconstruction speed of this type of algorithm very long and difficult to control.

Images