THE EFFECT OF RESERVOIR HETEROGENEITY ON GAS PRODUCTION FROM HYDRATE ACCUMULATIONS IN THE PERMAFROST

The quantity of hydrocarbon gases trapped in natural hydrate accumulations is enormous, leading to significant interest in the evaluation of their potential as an energy source. Large volumes of gas can be readily produced at high rates for long times from methane hydrate accumulations in the permafrost by means of depressurization‑induced dissociation, and using conventional technologies and horizontal or vertical well configurations. Initial studies on the possibility of natural gas production from permafrost hydrates assumed homogeneity in intrinsic reservoir properties and in the initial condition of the hydrate‑bearing layers (either due to the coarseness of the model or due to simplifications in the definition of the system). These results showed great promise for gas recovery from Class 1, 2, and 3 systems in the permafrost. This work examines the consequences of inevitable heterogeneity in intrinsic properties, such and the porosity and permeability of the hydrate‑bearing formation, as well as heterogeneity in the initial state of hydrate saturation. Heterogeneous configurations are generated through multiple methods: 1) through defining heterogeneous layers via existing well‑log data, 2) through randomized initialization of reservoir properties and initial conditions, and 3) through the use of geostatistical methods to create hetereogenous 2‑D and 3‑D fields that extrapolate from localized information. These extrapolations use all available information and established geophysical methods to capture a range of deposit properties and hydrate configurations. The results show that some forms of heterogeneity, such as horizontal stratification, can assist in production of hydrate‑derived gas. However, more heterogeneous structures can lead to complex physical behavior within the deposit and near the wellbore that may obstruct the flow of fluids to the well, necessitating revised production strategies. The need for fine discretization is crucial in all cases to capture dynamic behavior during production.