While there has been a steady increase in production from shale plays in recent years, recovery factors are still relatively low when compared to conventional formations. To improve on the existing technology and production from unconventional plays, a better understanding of reservoir fluid properties and phase behavior are vital. Understanding the dynamic behavior of reservoir fluids and the reliable prediction of reservoir performance in unconventional formations like shale are quite difficult. Reservoir simulation is an important tool that makes this easier and the type of reservoir simulation model used is a significant factor in this process.
This paper focuses on performance analyses of shale volatile oil reservoirs using black-oil and compositional simulation models. The performance analyses were done using single-phase (oil) and two-phase (oil and gas) black-oil simulations, as well as compositional simulations. The three approaches were compared in this paper. While black-oil simulations are easier, more accessible to the user and less time-consuming compared to compositional simulations, we assume that results are not as accurate as with compositional simulations. Nonetheless, how accurate are black-oil simulation results compared to compositional simulation results? Can the results be trusted to a reasonable extent? We have attempted to answer these questions and many others in this paper. The answers are vital because we cannot afford to use easier and less time-consuming methods at the expense of jeopardizing the accuracy of results of production forecasts.
Results show that the two-phase black-oil simulations are different and probably more accurate than single-phase black-oil simulations. As we have no field data to support our assumption for now, our opinion is based solely on the impact of the gas phase (for two-phase flow) on production performance. Also, the effects of fluid composition on cumulative oil production and oil rates were analyzed using compositional and two-phase black-oil simulations. Results from compositional simulations were different and presumably more accurate than two-phase black-oil simulations. This hypothesis is based on the fact that compositional simulation includes more of the physics that we assume are important in modeling reservoir fluids. Therefore, for thorough analysis of fluid composition effects and improved production forecasts (especially for reservoir fluids like volatile oils in shale formations), compositional simulations are necessary in most cases. We believe the compositional simulation will provide the most accurate forecasts, two-phase black-oil simulation will be the next most accurate, followed by single-phase black-oil simulation. However, this list is in inverse order of ease-of-use, so the simpler approaches are well worth considering to see whether they might be adequate. Finally, in this paper, the impacts of fluid sampling errors on cumulative oil production and oil rates were examined and found to have substantial effects on production forecasts.