Fluid Flow Patterns During Production from Gas Hydrates in the Laboratory compared to Field Settings: LARS vs. Mallik

Abstract:

The GFZ’s LArge Reservoir Simulator LARS allows for the simulation of the 2008 Mallik gas hydrate production test and the comparison of fluid flow patterns and their driving forces. Do we see the gas flow pattern described for Mallik [Uddin, M. et al., J. Can. Petrol Tech, 50, 70-89, 2011] in a pilot scale test? If so, what are the driving forces?

LARS has a network of temperature sensors and an electric resistivity tomography (ERT) enabling a good spatial resolution of gas hydrate occurrences, water and gas distribution, and changes in temperature in the sample. A gas flow meter and a water trap record fluid flow patterns and a backpressure valve has controlled the depressurization equivalent to the three pressure stages (7.0 – 5.0 – 4.2 MPa) applied in the Mallik field test. The environmental temperature (284 K) and confining pressure (13 MPa) have been constant.

The depressurization induced immediate endothermic gas hydrate dissociation until re-establishment of the stability conditions by a consequent temperature decrease. Slight gas hydrate dissociation continued at the top and upper lateral border due to the constant heat input from the environment. Here transport pathways were short and permeability higher due to lower gas hydrate saturation. At pressures of 7.0 and 5.0 MPa the LARS tests showed high water flow rates and short irregular spikes of gas production. The gas flow patterns at 4.2 MPa and 3.0MPa resembled those of the Mallik test. In LARS the initial gas surges overlap with times of hydrate instability while water content and lengths of pathways had increased. Water production was at a minimum. A rapidly formed continuous gas phase caused the initial gas surges and only after gas hydrate dissociation decreased to a minimum the single gas bubbles get trapped before slowly coalescing again. In LARS, where pathways were short and no additional water was added, a transport of microbubbles is unlikely to cause a gas surge as suggested for Mallik.