CO2 injection has been commonly applied in naturally fractured reservoirs (NFR) for EOR purpose. The matrix part of those types of reservoirs could be potentially a good storage medium as well. Understanding the matrix-fracture interaction during this process and the dynamics of the flow in this dual porosity system requires visual analyses. We mimicked fully miscible CO2 injection in NFRs using 2-D models with a single fracture and oil (solute) - hydrocarbon solvent pairs. The focus was on the visual pore scale analysis of miscibility interaction, breakthrough of solvent through fracture, transfer between matrix and fracture, and the dynamics of miscible displacement inside the matrix.
First, matrix-fracture interaction was intensively studied using 2D glass bead models experimentally. The model was prepared using acrylic sheets and glass beads saturated with oil as a porous media while a narrow gap of 1 mm size containing filter paper serving as a fracture. The first contact miscible solvent (pentane) was injected into the fracture and the flow distribution was monitored using an image acquisition and processing system. The produced solvent and solute were continuously analyzed for compositional study. The effects of several parameters such as flow rate, viscosity ratio (oil/solvent), and gravity were studied. Next, the process was modeled numerically using a commercial compositional simulator and the saturation distribution in the matrix was matched to experimental data. The key parameters in the matching process were effective diffusion and longitudinal and transverse dispersion coefficients. These coefficients were assigned into grid blocks in the model according to their position to the fracture.
Introduction
Miscible displacement in porous media is an important subject in many different fields including hydrogeology, environmental and petroleum engineering. What is critical for petroleum engineers is the efficiency of this process during enhanced oil recovery applications. Over the last four decades, scientists have investigated the factors and mechanisms controlling miscible displacement. Due to high influence on the process, primary interest was on diffusion and dispersion in the oil industry (Perkins and Johnston, 1963). Many investigators developed experimental and numerical methods to calculate diffusion and dispersion coefficients for the displacing and displaced fluids (Leayh-Dios and Firoozabadi, 2004; Goss, 1971; Islaz-Juarez et al; 2004 and Garder et al, 1963). Reason behind those studies is that the complex nature of multicomponent reservoir fluids limits the usage of available analytical methods.
The efficiency problem is more crucial in fractured medium as highly fractured system creates complexities causing irregular distribution of the injected phase. Factors affecting the efficiency and mechanisms in the presence of fractures were also studied (Silvia and Belery, 1989; Thompson and Mungan, 1969; Er and Babadagli, 2007; Trivedi and Babadagli, 2006, and Firoozabadi and Markset 1994). Mass transfer and convective dispersion between matrix and fracture lead to the higher ultimate recoveries compared to immiscible displacements. Early breakthrough, however, is the major issue for miscible displacement in fractured reservoirs. Geometrical properties of fracture are important as well as matrix and fluid properties in controlling the breakthrough of the injected fluid. Thompson and Mungan (1969) analyzed effect of fracture length, orientation and density in their experimental study with core samples. As similar, Firoozabadi (1994) investigated the effects of fracture aperture on the process. His study showed that the miscible displacement is still efficient even in the case of high fracture apertures.