Three-dimensional reconstruction of porous media using super-dimension-based adjacent block-matching algorithm

Area of Science:

• Geosciences
• Materials Science
• Computational Modeling

Background:

• Porous media are crucial in industrial applications, necessitating accurate physical property understanding.
• Super-dimensional (SD) reconstruction algorithms create 3D porous media structures from 2D images but have limitations.
• Existing SD algorithms suffer from inexact porosity, long run times, artifacts, and suboptimal accuracy.

Purpose of the Study:

• To address limitations of current super-dimensional (SD) reconstruction algorithms for porous media.
• To present a novel SD method with improved accuracy and efficiency.
• To enhance the characterization of porous media structures.

Main Methods:

• Developed a special template with adjacent and central blocks for reconstruction.
• Implemented block matching and porosity control mechanisms.
• Validated the method against existing SD, multipoint statistics, and simulated annealing methods on silica and sandstone samples.

Main Results:

• The proposed method overcomes limitations of traditional SD algorithms.
• Achieved improved accuracy in reconstructing porous media structures.
• Demonstrated a 4-6 speedup factor compared to the traditional SD method.

Conclusions:

• The novel SD reconstruction method significantly enhances accuracy and efficiency.
• This approach provides a more reliable tool for analyzing porous media.
• The findings contribute to better understanding and application of porous materials.