Molecular Modeling of Shale Gas Transport Mechanisms in Shale Nanopores

Shale reservoir contains different rock components that exhibit multi-scale hierarchical pore structure and complex mineralogical compositions. Experimental evidence has indicated that the majority of the natural gas in shale is adsorbed in clay and organic matter nanopores. The impact of different geological features and adsorption effect on fluid transport mechanism is yet poorly understood. In this dissertation, we present a molecular modeling on fluid flow at nanoscale and investigate the impact of different features on gas transport. It further facilitates large-scale shale reservoir property characterization.

We first apply boundary-driven non-equilibrium MD simulation (BD-NEMD) to estimate the gas transport diffusivity in slit-like clay and kerogen nanopores. We scrutinize the performance of six temperature control schemes in BD-NEMD simulation. Then, we examine the validity of the Knudsen model for predicting transport diffusivity under the effective temperature control schemes. Results indicate that the reservoir models based on Knudsen theory fails to predict shale gas production accurately because it neglects the effect of gas adsorption.

Second, we apply external field NEMD (EF-NEMD) to estimate gas transport diffusivity in nano-scale digital rocks with complex pore structures, which is reconstructed by Markov Chain Monte Carlo (MCMC) simulation or Focused Ion Beam Scanning Electron Microscope (FIB-SEM). On top of its dependency on Knudsen number, gas transport diffusivity is found to be sensitive to pore geometry factors (e.g., surface area, pore tortuosity, etc.), for which an effectiveness factor has been proposed.

Third, we apply fractional partial differential equations (F-PDEs) to model the shale gas sub-diffusion process in micron-scale digital rocks with nanometer resolutions. Time-fractional PDE is capable of capturing the transient state of gas sub-diffusion process, especially when the pore structure is heterogeneous with strong confinement. An integrated workflow has been developed to efficiently model such process by coupling results from MD simulations. Slow rate of transport has been identified and the micron-scale effective transport diffusivity has been estimated. Such reservoir properties are then used in upscaling workflow to recover core-plug-scale (~centimeter) rock properties for practical applications.