Although digital core technology offers unparalleled advantages and is used to investigate microstructure and fluid flow mechanism in sandstone and carbonate reservoirs, it is not fully used for shale reservoirs. Direct imaging techniques, such as focused ion beam-scanning electron microscopy and nano-scale X-ray computed tomography are limited to very small sample sizes or certain resolution levels, which is insufficient for observing nanoscale pores; experiment time cost also impedes the practical application of these techniques. Conventional indirect reconstruction techniques, such as simulated annealing method and process-based reconstruction algorithm, cannot effectively describe multi-scale pore distribution and connectivity.

To identify the characteristics of pore spaces in gas shale, this study proposes a reconstruction approach based on basic pore analysis from scanning election microscope images, Dual-Region strategy, and Very Fast Simulated

Annealing algorithm. Quantitative pore analysis provides fundamental data and validation basis for the subsequent

3D reconstruction procedure. The Dual-Region strategy is proposed to achieve multi-scale pore reconstruction that considers matrix porosity and organic or mineral porosity. By creating different templates for different regions, the

2D digital cores are first generated. To improve simulation efficiency, the Very Fast Simulated Annealing algorithm replaces the conventional Simulated Annealing algorithm. In addition, the normalization technique of the constraint functions is used to take into account the characteristics of different core samples.

To illustrate the proposed approach and simulation procedure, three reconstruction cases for shale are presented in this paper. By means of observation and indicator validation, the results show that real pore distribution features are realized better by using this new approach than by using the conventional reconstruction method.

This content is only available via PDF.
You can access this article if you purchase or spend a download.