Pore-scale modeling of seepage behaviors and permeability evolution in heterogeneous hydrate-bearing sediments and its implications for the permeability model

Abstract

Permeability is one of the key factors determining the recovery of gas resources from natural gas hydrate (NGH) reservoirs. The heterogeneous distribution of hydrates induced by random nucleation and growth in the pore space of hydrate-bearing sediments (HBS) inevitably results in pore space structure change, thereby significantly affecting fluid flow and permeability. In this paper, the seepage behaviors in a series of two-dimensional (2D) idealized HBS models were simulated based on the pore-scale computational fluid dynamics (CFD) modeling to obtain permeability and pore-scale seepage field. The 2D hydrate heterogeneity degree coefficient (DH) was proposed to quantitatively characterize the hydrate distribution heterogeneity. The pore-throat evolution characteristics were quantified by average pore-throat size and pore-throat size distribution. The results show that the occurrence and growth of hydrates within pores lead to the reconstruction and evolution of pore space structure and the formation of new fluid migration channels. The main reason for the reduction of HBS permeability is that with the increase of hydrate saturation, pores and throats with small sizes gradually become dominant, pore space connectivity is weakened, and the tortuosity for fluid flow increases. Additionally, the evolution of flow path and flow channel morphology caused by the growth of hydrates with different pore habits leads to the difference in pore-scale seepage behaviors and permeability variations in the homogeneous model. However, the effect of hydrate pore habits on permeability is no longer obvious due to the spatial heterogeneity of sediment grains and gas hydrates in the heterogeneous model. The effective flow path for pore fluid controlled by overall hydrate distribution and pore-throat characteristics would be critical. The comparison of calculated permeability results reveals that the accuracy of reported permeability reduction models is limited due to the ignorance of the HBS heterogeneity, the actual hydrate pore habits and morphology, and the pore structure geometry. It is hoped that the findings of this study can provide some insights into the pore-scale seepage characteristics of HBS, providing important implications for the development of the permeability model that can be easily implemented in the numerical simulators for field-scale gas production optimization of hydrate reservoir.