To improve the general understanding of diffusive processes in coal and investigate if representative reservoir properties can be estimated from measurements on crushed coal particles, adsorption and desorption experiments on crushed samples and large, intact core samples are carried out. Diffusion properties are analysed for methane, carbon dioxide and nitrogen up to a pressure of 10 MPa at a temperature of 35 °C. The samples are a low volatile bituminous coal and a medium volatile bituminous coal from Australia's Bowen Basin.A bidisperse model (a simplified version of the Ruckenstein model for micro-pore dominated flow), characterised by a ‘fast’ and a ‘slow’ fraction, is required to describe the observed diffusion behaviour. Results from the modelling show that effective diffusivities of the core samples are typically an order of magnitude lower than for the crushed coals, and the ‘slow’ effective diffusivities are typically at least one order of magnitude lower than the corresponding ‘fast’ effective diffusivities. A pressure dependency of the effective diffusivities is noted but an explicit relationship with gas content could not be established. This may be due to the experimental set-up, which does not eliminate the effect of differential pressure.Calculation of the diffusion path length for the coal core reveals that the diffusion path length in coal may not be constant but decreases with increasing matrix gas content as a result of coal swelling. Furthermore, there is a distinct difference between the diffusion paths derived from the ‘fast’ and the ‘slow’ fractions; the ‘slow’ diffusion path of the coal core is 100s of microns shorter than the ‘fast’ diffusion path. The faster diffusion associated with the longer diffusion path is interpreted to relate to micro-porosity, while slower diffusion via the shorter diffusion path is interpreted to relate to macro-porosity.
Read full abstract