Optimization of CO2-Enhanced Coalbed Methane Recovery Accommodating Swelling-Induced Permeability Evolution

The evolution of permeability in CO2-enhanced coalbedmethane (ECBM) recovery involves dynamic changes in coal shrinkage/swelling with the reduction/increase in gas sorption/desorption. Injection of CO2changes local pore pressures and induces related matrix volume strains. Modulated in part by the mechanical boundary conditions, changes in gas saturation and pressureinduce changes in permeability. Typically the recovery of methane induces shrinkage and the injection of CO2 induces swelling - concomitantly permeability decreases where net swelling results and increases where net shrinkage is present. However, permeabilities are also impacted by other important physical phenomena, including water saturations and the sequence of sweeps by different gases used in ECBM. These may be N2 or CO2. We use permeability evolution to investigate the performance of prototypical ECBM projects. Typical issues relate to premature breakthrough of CO2injectate at the recovery well and contamination of the methane output and also premature staunching of reservoir permeability through CO2 injection and in extreme cases the killing of the reservoir. We investigate these factors through lumped parameter and distributed parameter models to explore the evolution of permeability and productivity, respectively, where binary mixtures or CH4 and CO2 are present within coalbed methane reservoirs. Lumped parameters models incorporate the effects of sorption-induced strain and binary transport (CH4 and CO2) for rigid boundary conditions (no net strain).We use this model to explore the anticipated change in permeability with the dynamicallychanging stressregime.We fit observational data to models representing the evolution of permeability on coals subject to these prescribed mechanical boundary conditions to identify factors influencing permeability evolution in the swelling regime. Matrix swelling/shrinkage, matrix permeability, the ratio of fracture to matrix stiffness and fracture spacing and initial fracture permeabilityare the principal factors affecting the rate and magnitude of permeability change - and therefore in modulating the optimal sequencing of CH4 recovery and CO2 injection. We extend these lumped parameter characterizations to reservoir scale using a coupled flow-deformation model. We use this model to explore the intrinsic scaling at reservoir scale, including the spacing between recovery and withdrawal wells for typical recovery well patterns.