Unconventional gas reservoirs have become the focus of considerable attention as primary energy resource over the past decades worldwide. Numerical modelling technique plays a critical role in providing the essential tools for evaluating, optimizing and managing the development of such complex systems. In this work, we develop a generic simulation platform which allows investigators to rapidly implement and experiment with a wide array of alternate physical and constitutive models. The simulation platform is designed to incorporate the spectrum of known physics inherent in unconventional gas reservoirs, such as the non-Darcy effect covering various flow regimes, multi-phase behavior, adsorption/desorption, high-velocity turbulent flow, as well as the rock un-consolidation of natural fractures network.

In addition, the platform provides maximum flexibility of representation for the complex fractured network with irregular and non-ideal fracture geometries in unconventional formations. Two types of hybrid fractures models which integrate discrete fracture models (DFMs) with continuum-type approaches were developed for describing the multi-scaled multi-continuum nature of the stimulated fractured system. The hybrid modelling techniques could be utilized for applications with different requirements for efficiency and accuracy considerations, such as long-term gas recovery evaluation, multi-well interaction, completion optimization and transient behavior characterization, etc.

The simulation platform is designed and applied using a general abstraction that is built on top of the Automatically Differentiable Expression Templates Library (ADETL). In this paper, we conduct preliminary sensitivity studies to determine the key factors of reservoir and fractures that affect the production performance of unconventional gas wells. We present preliminary simulation results to demonstrate the model applications, and show the results of our model validation effort.

You can access this article if you purchase or spend a download.