Development of advanced reservoir characterisation and simulation tools for improved coalbed methane recovery (ICBM)
Bidisperse pore-diffusion model seeking to linearise the extended Langmuir equation in a piece-wise manner in the temporal dimension
The pore structure of coal matrix is highly heterogeneous, with the pore size varying from a few Angstroms to frequently over a micrometer in size. Many coal exhibit a bidisperse pore structure, with significant fractions of the total pore volume being found in size greater than 30nm and less than 1.2nm. Gas diffusion in micropores (<2nm) is controlled by a distinctively different mechanism from those identified in macropores (molecular diffusion, Knudsen diffusion and surface diffusion). A bidisperse pore-diffusion model that accounts for competitive gas adsorption/desorption is developed. A solution procedure that is more suitable for coal due to its low micropore diffusivity is proposed. This procedure seeks to linearise the extended Langmuir equation in a piece-wise manner in the temporal dimension. It dose not directly enforce the parabolic concentration profile within a particle. In other words, the parabolic concentration profile assumption is used to evaluate the concentration gradient at the particle surface only and no explicit assumption is made regarding to the concentration profile within the particle. When applied to a laboratory core flush test, an excellent match to the test data was achieved using sorbate concentration dependent apparent micropore diffusivity.
The cleat permeability of coalbeds has been shown to vary exponentially with changes in the effective horizontal stress acting across the cleats, through the cleat volume compressibility, which is analogous to pore compressibility in conventional hydrocarbon reservoir rocks. A new formulation for changes in the effective horizontal stress (under uniaxial strain conditions) accounting for matrix swelling as well as shrinkage is developed. The model has been validated using field primary recovery permeability data from coalbed methane wells at two different sites in the San Juan Basin. Using the reservoir parameters calibrated from the primary recovery field data, permeability changes during enhanced recovery by injecting CO2, N2 or flue gas may be estimated. In particular, it is shown that matrix swelling caused by CO2 adsorption could lead to a two-order of magnitude reduction in coalbed permeability.
Laboratory core scale CO2 flooding test capability enabling investigation into the influence of pressure, temperature and flow constraints
An improved laboratory core scale CO2 flooding test capability, including a 1000cc sample bearing high pressure (50MPa annular pressure, 480bar pore pressure) and temperature device, which enables investigation into the influence of pressure, temperature, and flow constraints on sorption characteristics of CO2 and CH4 in dry and water saturated coal samples under simulated reservoir conditions. The flushing experiments are designed to study the behaviour of water and methane-bearing coal on its transition to a maximum CO2 saturated coal. The outcomes can be used to define sorption, diffusion, permeability and gas flow parameters for field simulations. Coal cores are placed in a lead and rubber sleeve and built into a high-pressure metal tube. The metal tube is filled with hydraulic oil to apply a hydrostatic pressure on the coal core. Both ends are connected to permeable metal plates to permit a gas flow through the core. The amount of gas transferred into and out the core is monitored by flow meters on the injection and production side. Pressure, displacement and temperature transducers are used to monitor the experiment permanently and the data are transmitted to a data acquisition system. The produced gas composition is analysed in regular intervals.
Methodology for cleat angle characterisation from drilling cuttings using quantitative image analysis
This methodology characterises the preferable cleat angles and cleat spacing form frilling cuttings of coal seams. This angle distribution is of use for modelling sub-surface cleat orientations. The measurements are performed using the Image analysis program Qwin form Leica Microsystem. Microsoft Excel is used for the processing of the measured data. The procedure involves: - Preparation of the sample: Crushing and sieving. - Scanning and acquiring multigrain images. - Separation of multi grain images into single particle images. - Gathering shape information using the rotation program. - Processing image analysis results.
Methodology for stochastic realisations of matrix-fracture systems and upscaling of natural fractures and reservoir properties for coals
This is an operational methodology of characterizing static and dynamic reservoir properties as they apply to ECBM recovery. The particular structure of coal (fractured media) made of macro and micro fracture features obliges us to consider the microscopic and macroscopic scale in order to understand (and thus simulate) methane production and furthermore CO2 adsoption. More classical methods could be applied to CBM production. Yet, a true understanding of the adsorption/diffusion mechanism of CH4 and CO2 as well as the diffusional effects taking place within the matrix (at small scale) needs an accurate understanding of the cleat system at the small scale. It is our ambition to show that from this scale one can proceed through a series of steps leading to an accurate petrophysical numerical (reservoir size cells) assignation at the inner-well scale. By accurate we mean that values assigned minimize the inherent errors due to up-scaling but above all are representative. The basis of the approach is the assumption that the geometry of the cleat system (described statistically) at the small scale can serve to generate a realistic network of cleat features at the numerical size scale (ex. 100m). Once that done, one can generate within this unitary volume a series of petrophysical values (typically K and PHI) at will, through up-scaling, thus creating a petrophysical database. Once this database used in conjunction with logs, more classical characterization methods could then be used.
The initial CBM simulator METSIM developed at Imperial College was designed for primary methane recovery. The simulator solves a set of Darcy equations for the water and gas (methane) phases in the cleats. In order to simulate CO2 sequestration/enhanced recovery, three new equations needed to be solved simultaneously for the methane, CO2, and water components in the upgraded simulator (METSIM2). - METSIM2 has been verified against other ECBM simulators in a numerical simulator comparison study. The results are in good agreement with those from the other ECBM simulators participating in the study. - Using the laboratory data obtained by TUD, a bidisperse pore diffusion model was developed and implemented. Further development of the stress-permeability theory led to the pore pressure dependent absolute permeability relationship for the coalbed reservoir during both primary and CO2-enhanced recovery/CO2 storage. - METSIM2 has been successfully used to history match the Allison Unit ECBM pilot in the San Juan basin.