Li, Yuxiang (The University of Texas at Austin) | Lu, Jun (The University of Texas at Austin) | Churchwell, Lauren (The University of Texas at Austin) | Tagavifar, Mohsen (The University of Texas at Austin) | Weerasooriya, Upali (The University of Texas at Austin) | Pope, Gary A. (The University of Texas at Austin)
Primary and secondary oil recovery from naturally fractured carbonate reservoirs with an oil-wet matrix is very low. Enhanced oil recovery from these reservoirs using surfactants to alter the wettability and reduce the interfacial tension have been extensively studied for many years, but there are still many questions about the process mechanisms, surfactant selection and testing, experimental design and most importantly how to scale up the lab results to the field. We have conducted a series of imbibition experiments using cores with different vertical and horizontal dimensions to better understand how to scale up the process. There was a particular need to perform experiments with larger horizontal dimensions since almost all previous experiments have been done in cores with a small diameter, typically 3.8 cm. We adapted and modified the experimental method used for traditional static imbibition experiments by flushing out fluids surrounding the cores periodically to better estimate the oil recovery, including the significant amount of oil produced as an emulsion. We used microemulsion phase behavior tests to develop high performance surfactant formulations for the oils used in this study. These surfactants gave ultra-low IFT at optimum salinity and good aqueous stability. Although we used ultra-low IFT formulations for most of the experiments, we also performed tests at higher IFT for comparison. Even for the higher IFT experiments, the capillary pressure is very small compared to gravity and viscous pressure gradients. We also developed a simple analytical model to predict the oil recovery as a function of vertical and horizontal fracture spacing, rock properties and fluid properties. The model and experimental data are in good agreement considering the many simplifications made to derive the model. The scaling implied by the model is significantly different than traditional scaling groups in the literature.