Pore structure characterization and permeability prediction of coal samples based on SEM images

Abstract

The pore structure of coal reservoirs determines the reserves of coalbed methane, and the gas permeability determines the level of the production capacity. In this study, the SEM images of coal samples were analyzed by various means. First, the grayscale threshold of binarization of the coal sample image is determined by a suitable algorithm. Comparing several different algorithms, the porosity based on Yen algorithm is closer to the results of vacuum saturation method and nuclear magnetic resonance (NMR) (8.35% vs. 7.51% vs. 8.92%). Further, the pore size distribution (PSD) of the coal sample is obtained according to discrete and continuous algorithms. By comparing the SEM results with the NMR results, it is found that the calculation results based on the continuous algorithm (CPSD) are better than the discrete algorithm (DPSD) and closer to the NMR results. For the effect of scale, we found that the image resolution has a certain influence on the minimum pore size characterized, such as sample C1: 0.29 μm ( × 1000) vs 0.58um ( × 500). At high resolution, more micro-pores are observed. Further, we predict the permeability of coal samples based on SEM images. It is found that the calculation results based on the continuous algorithm and the Hagen–Poiseuille equation are closer to the measured values (e.g., 16.97 (DPSD) vs. 0.45(CPSD) vs. 0.59 mD (Lab), sample C2, magnification of × 1000). In general, this method can effectively evaluate the pore structure characteristics and permeability of coal samples.