A spectrum of behavior is encountered in methane hydrate provinces, especially ocean sediments, ranging from essentially static accumulations where the pore space is filled with hydrate and brine, to active seeps where hydrate and methane gas phase co-exist in the hydrate stability zone (HSZ).
We present grain-scale models of drainage and fracturing, key processes involved in pressure-driven gas phase invasion of a sediment. A novel extension of invasion percolation to infinite-acting, physically representative networks is used to evaluate the connectivity of water in a gas-drained sediment. A novel implementation of the level set method (LSM) is used to determine the capillarity-controlled displacement of brine by gas from sediment and from fractures within the sediment. The discrete element method (DEM) is extended to model the coupling between the pore fluids and the solid and thereby predict the onset of sediment fracturing by gas phase pressure under in situ loading conditions. The DEM grain mechanics model accounts for the different pressure of brine and methane gas, in a “membrane” two-fluid model. The fluid-fluid configuration from LSM can be mapped directly to the pore space in DEM, thereby coupling the drainage and mechanics models. The type of behavior that can emerge from the coupled processes is illustrated with an extended LSM model. The extension computes grain displacement by the gas phase with a simple kinematic rule.
The LSM predicts gas/brine interface movement in sediment/fractures, and the DEM shows that the interface location affects the strength of sediment. Network simulations indicate that brine in drained sediment is better connected than previously believed. This increases the availability of water, water-gas interface area and consequently the rate of counter-diffusion of salinity ions, thus relaxing the limit on hydrate build-up within a gas-invaded region. The DEM model reproduces the key processes in the fluid-solid interaction. It was validated by means of isochoric pressurization tests, multistep inflow/outflow experiments, and drained/undrained consolidation tests. DEM simulations reproduce the sediment stress-strain behavior in laboratory triaxial experiments. The DEM model shows that vertical fracturing of the sediment is favored at the base of the hydrate stability zone. The kinematic extension of LSM yields much broader capillary pressure curves and self-reinforcing invasion paths, confirming the utility of ongoing efforts to integrate the DEM and network/LSM models.