No abstract

No abstract

Summary Unconventional reservoirs are the focus of considerable attention as a primary energy source. Numerical simulation is a core kernel of reservoir-engineering work flows for reservoir evaluation, optimization, and management. Accurate and efficient numerical simulation of unconventional reservoirs is challenging. There is substantial physical complexity involving a number of tightly coupled mechanisms in the modeling of these reservoirs. The complexity is further amplified by the multicontinuum nature of the stimulated formation, and the complex fracture networks with a wide range of fracture-length scales and topologies. To adequately capture the effects of the multiscaled fracture system, we develop two alternative hybrid approaches that are aimed at combining the advantages of multicontinuum and discrete-fracture/matrix (DFM) representations. During the development of unconventional resources, geological and geophysical information may be available in some cases to suggest a prior characterization, whereas in many other cases, this prior model may be incomplete and limited to hydraulic fractures. The two hybrid approaches could be used for different applications depending on the available characterization data and different requirements for efficiency and accuracy considerations. The first hybrid model couples an embedded-discrete-fracture model (EDFM) with multiple interacting continua (MINC) into EDFM/MINC, which simulates the fracture network characterized by stimulated-reservoir-volume (SRV) concept. This optimized model can reduce the computational cost that is associated with the widely applied logarithmically spaced/locally refined (LS/LR) DFM technique, while improving the flexibility to model the complex geometry of hydraulic fractures. The MINC concept allows the hybrid model to handle the extreme contrast in conductivity between the small-scale fracture network and the ultratight matrix that results in steep potential gradients. For the second class of hybrid model (unstructured DFM/continuum), the primary fractures are described by use of DFM with unstructured gridding, and the small-scale fractures are simulated by continuum-type approaches in a fully coupled manner. Optimized local-grid refinement is used to accurately handle the transient-flow regime around primary fractures. An upscaling technique that applies EDFM on the detailed realization of the discrete-fracture network by use of the target unstructured grid to generate an appropriate dual-permeability model is also developed. The upscaling technique is suitable for cases where a detailed prior model for the complete fracture network is available. Simulation studies demonstrate the applicability of the developed hybrid-fracture models. Model verification is conducted against several reference solutions.

Shale gas reservoirs have been proposed as feasible choices of location for injection of CO2 and/or N2 because this method could enhance recovery of natural gas resources, while at the same time sequester CO2 underground. In this paper, a fully coupled multi-continuum multi-component simulator which incorporates several transport/storage mechanisms is developed. To accurately capture physics behind the transport process in shale nanopores, Kundsen diffusion and gas slippage are included in the flow model. An extended Langmuir isotherm is used to describe the adsorption/desorption behavior of different gas components and the displacement process of methane as free gas. Pressure-dependent permeability (due to rock deformation) of natural fractures induced by hydraulic fracturing is also considered in the simulator. In addition, modeling of complex fracture networks is very crucial for simulating production of shale gas reservoirs because there exists various scales of fractures with multiple orientations after the fracturing treatment for horizontal well. In this work, a hierarchical approach which integrates EDFM with dual-continuum concept is adopted. The hybrid model includes three domains: matrix, major hydraulic fractures and large-scale natural fractures (described by EDFM), and micro-fractures in SRV region which are modeled by dual-continuum approach. Embedded Discrete Fractures Model (EDFM) is an efficient approach for explicitly simulating large-scale fractures in Cartesian grid instead of complicated unstructured grid. Moreover, a nested-grid refinement method is used to capture the fluids transfer from matrix to fractures. Fully implicit scheme is applied for discretizing fluid equations, and the corresponding Jacobian matrix is evaluated by Automatic Differentiation with Expression Templates Library (ADETL). The AD-Library framework allows wide flexibility in the choice of variable sets and provides generic representations of discretized expressions for gridblocks. Several simulations and sensitivity analysis are performed with the developed research code for determining the key factors affecting shale gas recovery. Modeling studies indicate that the properties of fracture networks could greatly influence methane production. Different injection strategies including huff-n-puff scenario are also evaluated to provide insights for optimizing production of multi-fractured horizontal well. Results show that CO2/N2 injection can be an effective approach with great application potential for enhancing shale gas recovery.

No abstract