(353f) Considerations in Production of Liquids in Shales

Surrogate reservoir models were generated through compiling eight main factors that most affect oil and gas recovery in ultra low-permeability reservoirs. Experiments are designed using partial factorial method.  Gas oil ratio and two different recovery outcomes were sought – one based on a terminal time and the other based on economic rate.  Reservoir simulations were performed to obtain the necessary outcomes. Reservoir permeability, fracture spacing, initial reservoir pressure and the initial gas oil ratio were the most important factors.  Multivariate regression was used to obtain coefficients for the second-order response surface models. Surrogate models of recoveries were created using the important factors and validated. Probability density functions (PDF) of recoveries and gas oil ratio were generated using uncertainties in the primary input parameters. The heavy hitters are identified from tornado plot. The order of importance of these factors changed slightly as the type of outcome sought changed.  Average oil recoveries in the PDFs generated were about 15% after 10 years of production. The study also provided a method of optimizing operational parameters given the known uncertainty in the geologic input parameters.