Gas is adsorbed in the pores of coal matrix and during gas production gas is desorbed from the pore surface and diffuses through the matrix pore structure and flows in the fracture/cleat system to the production well or boreholes. However, coal is highly heterogeneous and anisotropic. How heterogeneity and anisotropy affect the gas storage and especially the diffusion behaviour is not well studied. In this work, a series of measurements were performed on three dry cubic coal samples cut from the same coal block from the Bowen Basin, Australia, using an adsorbing gas, methane. For each sample, gas adsorption experiments with gas flowing from three principal directions were performed. The diffusion data was fitted with a bidisperse diffusion model to obtain diffusion coefficient. The three samples, although from the same coal block, showed difference in adsorption amount and significant difference in effective diffusivity. It was found that the effective macropore diffusivity increased with gas pressure and effective micropore coefficient decreased with gas pressure. The effective diffusivity showed difference among samples and directions, demonstrating coal heterogeneity and anisotropy both have a significant impact on gas diffusion behaviour. However, no generalisation can be obtained with any single pore structure parameter, such as pore size or surface area, as it may be related to all pore and fracture structures at various scales.
Bidisperse model;Langmuir model;Bowen Basin;Coalbed methane;CBM
Xinxin He, Yuanping
To investigate the effects of coal pore structure on the methane‐coal sorption hysteresis, six coal samples were collected. The methane‐coal sorption measurement was performed at 35 °C and pressure up to 5.5 MPa using a high-pressure volumetric analysis system (HPVAS). With the help of N2 physisorption at 77 K and CO2 physisorption at 273 K, basic pore properties including specific surface area (SSA), mode diameters and pore size distribution (PSD) were obtained through classical thermodynamic methods and the advanced density functional theory (DFT). A Fréchet distance index (FDI) based on the resemblance of two curves was proposed to overcome the difficulty in quantitatively evaluating the methane‐coal sorption hysteresis. Quantified heterogeneity of the coal pore structure by five fractal dimensions derived from Frenkel-Halsey-Hill model (DFHH1 and DFHH2), Neimark-Kiselev model (DNK), Wang-Li model (DWL) and Sierpinski model (DSPS) was coupled with the FDI for regression analyses. Results indicate that increasing SSA and stronger first-layer adsorption energy may exacerbate the methane-coal sorption hysteresis, while no satisfactory correlation was observed between the methane-coal sorption hysteresis and the pore volume. Wider Dubinin-Astakhov PSD and bigger mode diameters were found corresponding to smaller FDIs indicating reduced methane-coal sorption hysteresis. Correlation between the FDI and the fractal dimensions revealed a possible positive correlation between the methane-coal sorption hysteresis and the heterogeneity of the coal pore structure, especially for DFHH2 whose applied pore widths were 2.78–385 nm.
Coal; Methane; Physisorption; Fractal dimension; Hysteresis evaluation
Influence of coal type and rank on gas diffusion rate in a wide range of Australian coals
Six orders of magnitude variation in gas diffusion rate in the tested coal samples
The higher the inertinite content, the higher the diffusion rate
The same dispersion parameter for both CO2 and CH4 gases in the same coal
Gas diffusion within the coal matrix plays a key role in determining the rate of natural gas depletion and enhanced coal bed methane production via CO2sequestration (CO2-ECBM) in coal seam gas reservoirs. In this work, we investigated the influence of maceral composition and coal rank on CO2 and CH4diffusion rates of 18 bituminous and sub-bituminous Australian coals. We obtained measures of the gas diffusion rate and the spread of diffusion times. Gas diffusion rate through coal pores was found to vary over 6 orders of magnitude depending on the coal rank and maceral composition. This diffusion rate was independent of pressure in the range 1–5 bar. It increased substantially with inertinite content of the coal in the lower rank and medium rank coals examined. In the high rank coals, the diffusion rate was less sensitive to maceral composition, but alternatively, this may reflect regional variations in the dependence of diffusion rate with maceral composition. The CO2 diffusion rate was faster than the CH4 diffusion rate. The factor describing the spread of diffusion times generally increased with increasing vitrinite content but for a given coal was similar for both CH4 and CO2. This suggests the gases penetrate the same parts of the coal structure.
Based on the experimental data, different synthetic coalbed simulation models were constructed to analyse the impact of CH4 and CO2 diffusion coefficients on ECBM and CO2 sequestration performance. The numerical simulation results showed that CH4 production rate is inversely proportional to the sorption time, if the bulk flow in the cleats does not create any restriction. The results also indicated that CO2breakthrough time is a function of the CO2 sorption time – if the CO2 adsorption is not fast enough, the injected CO2 will be spread into the seam, resulting in an early breakthrough.
Gas diffusion；Coal seam gas；Maceral composition；Coal rank；Sorption time
Bituminous coals from a wide range of sources (including Australia, New Zealand, Europe, China, South America, Canada, the US and South Africa) were characterised by their capacity for gas sorption, rate of gas sorption of CO2 and CH4, and their nanoporosity (pore size distribution less than 50 nm radius) in order to identify the relationships between them. The following new relationships were established:
The rate of gas sorption was unrelated to the capacity for gas sorption. The rate of gas sorption for CH4 and CO2 increased exponentially with the amount of total and accessible porosity in the size range 8–50 nm (with no influence of coal origin on the relationship being discerned), suggesting that the extent of porosity of coals in this size range controls the rates of gas sorption in coals. In contrast, the capacity for gas sorption was only weakly related to pore numbers in this size range, which shows that the number of 8–50 nm pores do not control capacity for gas sorption. Moreover, this difference in relationship shows the number of pores of the size where gas is sorbed predominantly (<5 nm) does not correlate strongly with the number of larger pores.
Both the number of pores and rates of gas sorption tended to increase with inertinite content but the relationship with inertinite content differed for coals from different sources. The inertinite-rich coals from Australia (except those from the Illawarra region) had both the greatest porosity and gas penetration rates, whereas in coals sourced from other regions, although the gas penetration rate increased with inertinite content, the effect was not so strong. The rates of sorption in the inertinite-rich coals also tended to decrease with increasing rank below 0.9% Rv,max. In contrast to the results obtained with kinetic studies, we found no overall trend of capacity for gas sorption with maceral composition, though the Australian bituminous coals generally had greater capacity than the other bituminous coals examined. This suggests that not only the number of 8–50 nm pores in coals sourced from Australia (not those from the Illawarra region) and elsewhere are different, the number density of accessible <5 nm pores (not directly measurable in coals by SANS) may also be systematically different between these coals.
The relationships developed in this study have important implications in predicting coal structure, fundamental understanding of gas transport through coal beds, and explaining the variation of coking properties of coals sourced from Australia and elsewhere.