煤层气
地质学
煤
甲烷
煤矿开采
石油工程
拉伤
解吸
采矿工程
地球化学
废物管理
医学
化学
工程类
有机化学
吸附
内科学
作者
Xiaojun Cui,R.M. Bustin
出处
期刊:AAPG Bulletin
[AAPG/Datapages]
日期:2005-07-29
卷期号:89 (9): 1181-1202
被引量:711
摘要
The permeability of deep (1000 m; 3300 ft) coal seams is commonly low. For deep coal seams, significant reservoir pressure drawdown is required to promote gas desorption because of the Langmuir-type isotherm that typifies coals. Hence, a large permeability decline may occur because of pressure drawdown and the resulting increase in effective stress, depending on coal properties and the stress field during production. However, the permeability decline can potentially be offset by the permeability enhancement caused by the matrix shrinkage associated with methane desorption. The predictability of varying permeability is critical for coalbed gas exploration and production-well management. We have investigated quantitatively the effects of reservoir pressure and sorption-induced volumetric strain on coal-seam permeability with constraints from the adsorption isotherm and associated volumetric strain measured on a Cretaceous Mesaverde Group coal (Piceance basin) and derived a stress-dependent permeability model. Our results suggest that the favorable coal properties that can result in less permeability reduction during earlier production and an earlier strong permeability rebound (increase in permeability caused by coal shrinkage) with methane desorption include (1) large bulk or Young's modulus; (2) large adsorption or Langmuir volume; (3) high Langmuir pressure; (4) high initial permeability and dense cleat spacing; and (5) low initial reservoir pressure and high in-situ gas content. Permeability variation with gas production is further dependent on the orientation of the coal seam, the reservoir stress field, and the cleat structure. Well completion with injection of N2 and displacement of CH4 only results in short-term enhancement of permeability and does not promote the overall gas production for the coal studied.
科研通智能强力驱动
Strongly Powered by AbleSci AI