: Behrooz Koohmareh Hosseini
: 31.61 MB
The increasing demand for hydrocarbons and decreasing reserves have created the necessity to produce oil and gas more efficiently and economically. Increasingly, oil and gas companies are focusing on unconventional hydrocarbons; oil sands, shales and CBM. For this class of reservoir materials, the geomechanical response of the reservoir can play an important role in the recovery process. For naturally fractured, stress sensitive reservoirs or thermal recovery processes, geomechanical processes play an even greater role in efficient, economic recovery. For simulations of these processes, most research efforts have been focused on reservoir geomechanical simulations using conventional reservoir simulators coupled to geomechanical codes. While coupled reservoir-geomechanics modeling has been recently widely studied in the literature, there is no applicable methodology implemented or proposed to mitigate the challenging computational cost involved with the inclusion of geomechanics in large multimillion-cell reservoirs. Past studies so far have focused on different coupling schemes, but not on the efficient and robust simulation workflows. This research was conducted with the aim of development and application of various different strategies to include geomechanics into reservoir simulation workflows in large scale reservoirs and in a timely fashion process. The research was performed to allow the future simulators to perform high resolution reservoir-geomechanical simulations in a large scale (near field and far field) with long simulation time windows and lowest computational cost. Initially, analytical proxies were developed and recommending for implementation in lieu of complex reservoir simulations. The analytical model was for prediction of heavy oil geomechanical responses everywhere in the reservoir. The model adopted the use of the mathematical domain decomposition technique and a novel temperature front tracking developed in the very early stage of the research. As opposed to classical analytical models, the proxy predicted reservoir flow and mechanical behavior (on a synthetic case geometry with real hydraulic data) everywhere in the reservoir and in dynamic and transient flow regimes. Subsequent research was aimed at reservoir-geomechanics coupled model order reduction by use of a numerical proxy. The proxy took advantage of streamline linear space behavior and power in decomposition of the reservoir domain into sub-systems (delineation/drainage areas). The combination of localization and linearization allowed predicting both mechanical and fluid flow responses of the reservoir with only solving the pressure equation in Cartesian underlying 3D grids and the solution of saturation transport equation along only one streamline. Following this, a streamline-based reservoir-geomechanics coupling was proposed and was implemented within a Fortran-C++ based platform. The new developed technique was compared in terms of computational cost and results accuracy with the conventional hydromechanical coupling strategy that was developed on a C++ based platform by use of collocated FV-FEM discretization scheme. One of the final stages of the research explored different streamline-based reservoir-geomechanics coupling strategies for full-field reservoir simulations. Various coupling strategies including sequential coupling schemes and a semi-fully coupling scheme to embed geomechanics into streamline simulation workflow was developed and performed. Numerical software with advanced GUI was coded on QT programming language (C++ based) developed to couple mechanical simulator to streamline simulation engine. While streamline simulations were the center of the research, the last stage of research was conducted on numerical and physical stability, convergence and material balance errors of SL-based reservoir-geomechanics class of couplings. The results provided a solid foundation for proper selection of time-steps in SL-based coupling to ensure a numerically stable and physically robust hydromechanical simulation. As a result we showed that use of streamline simulation in both proxy forms and simulator forms have significant added value in full-field reservoir-geomechanics simulations.