Analysis of spontaneous imbibition in fractured, heterogeneous sandstone

Multiphase flow studies in naturally fractured reservoirs have numerous applications in hydrocarbon recovery, hydrogeology and environmental remediation of subsurface contamination. The presence of natural fractures has significant effects on recovery from oil and gas reservoirs. In these reservoirs, fracture networks serve as better flow paths for fluids, while the porous rock provides storage space. The efficiency of hydrocarbon recovery as well as pollution and contaminant removal from soil and groundwater are mainly controlled by our ability to understand and define fluid transport mechanisms in naturally fractured formations. Appropriate representation of transport properties, such as relative permeability and capillary pressure, is essential for the success of any predictive flow process in permeable media. In addition, the presence of heterogeneities such as fractures, faults, and other geological features adds to the difficulty of assigning flow properties that are representative of the system. The present study focuses on the development and implementation of a numerical model of two-phase flow in fractured rocks showing contrasting rock properties in the form of bedding planes. A unique experimental data set of spontaneous imbibition in a fractured sandstone core is used for the development and verification of the model. This data set consists of a series of high-resolution X-ray computed tomography scans of a rock sample showing local rock heterogeneities, fracture orientation, bedding planes, and fluid saturation as a function of time. An automated history matching approach is proposed to determine relative permeability and capillary pressure curves. A commercial reservoir simulator is used in coordination with an optimization protocol in the proposed history matching iv method. Four different synthetic and a semi-synthetic data sets were used to test the automated history matching approach. Absolute permeability and oil relative permeability curves were predicted in those synthetic cases, using different relative permeability representations. In the semi-synthetic case, known relative permeability for the matrix and capillary pressure for both matrix and the fracture were determined simultaneously using experimental spontaneous imbibition data of a heterogeneous fractured core sample. Results of this study indicate that the automated history matching approach was successful in predicting the absolute permeability, relative permeability and capillary pressure. The effects of transport properties on the imbibition process were also investigated by sensitivity analysis. Results from this work improve understanding of multiphase flow in fractured and heterogeneous porous media and ability to predict fluid migration in fractured reservoirs.