Measurement and simulation of relative permeability of coal to gas and water
MetadataShow full item record
AbstractCoalbed methane is an important energy resource. The commercial use of a coalbed methane reservoir requires gas and water production predictions for the life of the reservoir and infill drilling response using a suitable reservoir simulator. A key element in simulation is the two-phase gas-water relative permeability, which is the subject of this research. Gas-water relative permeability of two coal samples: Grande Prairie and Goldsource were determined by the Johnson-Bossler-Naumann (JBN) method, which is widely used for petroleum reservoir rocks. These coal samples have low porosity (4% and 12%) and very low permeability (< 0.1 mD). The coal samples were characterized by nuclear magnetic resonance (NMR), X-ray computed tomography (CT Scan), X-ray diffraction (XRD), and Scanning Electron Microscopy (SEM) analyses to study qualitatively the coal microstructure, size and distribution of pores, and their cleat system. They showed that the microstructure of coal was made up of a range of pore shapes and sizes, with clay minerals filling the pore space. These minerals reduced the cross-sectional areas of pore throats and consequently the coal permeability. Capillary pressure of the coal samples was measured by mercury injection. The measurement showed a bimodal pore size distribution representing micropores and cleats. These data were then used to generate relative permeability of coal curves based on Naar-Wygal-Henderson equations. The resulting curves were smooth because the equations were based on a bundle of capillary tubes, which was not representative of coal. Nevertheless, it showed that theoretically relative permeability of coal curves could be derived from capillary pressure. An experimental apparatus was designed, built, tested, and commissioned for various two-phase gas-water relative permeability measurements. The gas phases were helium, methane, and carbon dioxide in this order of adsorbing strength on coal and the water phase was formation brine. Carried out 48 drainage and imbibition displacement runs on the two coal samples using helium-brine, methane-brine, and carbon dioxide-brine at three operating pressures and the results were analyzed on a spreadsheet developed for this work. lmbibition relative permeability showed varying irreducible water saturation (Swir) correlated with pressure. The absolute permeability of the coal samples changed after each experiment. The Goldsource coal sample showed a higher two-phase flow saturation range than the Grande Prairie coal sample. In both cases, the relative permeability to gas was low (of the order of 0.10 at Swir) and high (1 .0) for water. The coal samples tended to become more water-wet at higher pressures in the case of adsorbing gases, namely methane and carbon dioxide. In the case of non-adsorbing gas helium, higher pressure prevented the inflow of water into the smaller pores leading to a decrease in Swir with an increase in pressure. The drainage experiments were highly unstable based on the Instability Number calculated. The resulting relative permeability curves showed oscillation. A simulator was used to simulate two laboratory experiments of drainage and imbibition using the measured relative permeability of coal samples to gas and water. The agreement between the calculated and experimentally observed gas and water production was fair. Carried out several coalbed methane full-field simulations using 40 and 160 acres per well spacing. Given the low permeability (< 0.1 mD), the higher recovery factor was 6% of the gas in place. This decreased to 1 % when bottom water was present. The production rate of a well was of the order of 10 Mscf/day over a 30 year period. To develop such a reservoir, a smaller well spacing would have to be used.
Bibliography: p. 348-364