Representative elementary volume evaluation of coal microstructure based on CT 3D reconstruction

CT scanning is an essential means to evaluate coal microstructure for its non-destructive quantization and visualization. However, the original images obtained by CT scanning are computationally expensive and challenging to analyze. Due to the self-similarity of the coal microstructure, analyzing the representative elementary volume (REV) of digital rock can help to establish representative results. In this work, we analyze the REV of eight coal samples using micro-CT with resolutions of 0.7, 1.5, and 9.5 μm. Furthermore, the relationships between model sizes and pore structure properties are systematically established in terms of heterogeneity, anisotropy, porosity, fractal dimension, pore size, tortuosity and permeability. Results show that most coal samples have a definite REV for pore structure parameters under high image resolution, while it is difficult to determine the REV for describing fluid flow. Specifically, the anisotropy, porosity, fractal dimension, and mean pore sizes tend to be stable, and the maximum pore sizes increase with increasing model sizes. There is no distinct rule between the tortuosity and model sizes, while the tortuosity gradient error gradually stabilizes with the increase of model sizes. It is challenging to determine permeability REV for the permeability of most samples showing a gradually increasing trend as the model size increases. In addition, the image resolution further affects the REV. High-resolution images tend to have a definite REV due to the detailed and complex pore-fracture system. The comprehensive evaluation of REV based on multiple microstructure parameters helps improve the rationality and reliability of coal microstructure reconstruction.