RESUMO
This paper presents a simple method to estimate ultimate recovery factors (URF) of ultratight reservoirs based on equilibrium by diffusion in which URF is only a function of changes in hydrocarbon density between initial and final states. URF is defined at infinite time and therefore does not depend on the transient behavior. Although URF may not be achievable during the life-cycle of the field development and production, it provides valuable insights on the role of phase behavior. We show that equilibrium phase behavior defines the absolute upper-bound for URF during primary production and explains the poor recovery from shale oil reservoirs compared to the high recovery factor in shale gas reservoirs in a unifying way. Further, we quantify how injected solvent compositions (CH4, CO2, N2, and C2H6) during huff'n'puff enhanced oil recovery (EOR) improve recovery based on density reduction and compositional dilution, and show that the largest percentage increase in recovery occurs for heavier oils. Our calculations provide a practical means to define the URF from primary production as a function of reservoir fluid composition, temperature, and pressure drawdown. In addition, our calculations articulate incremental URF (IURF) of solvent huff'n'puff based on net solvent transfer into ultratight rock, which is a key design consideration. The results illustrate that solvent transfer dilutes the hydrocarbons in place, thus maximizing long-term hydrocarbon recovery. Net mass transfer can be improved by enhancing the diffusion of solvent into the matrix based on the huff'n'puff design parameters including solvent composition, drawdown pressure, and the net amount of solvent injected based on optimal frequency and cycle duration.
RESUMO
Prediction of microemulsion phase behavior for changing state variables is critical to formulation design of surfactant-oil-brine (SOB) systems. SOB systems find applications in various chemical and petroleum processes, including enhanced oil recovery. A dimensional equation-of-state (EoS) was recently presented by Ghosh and Johns1 that relied on estimation of the surfactant tail length and surface area. We give an algorithm for flash calculations for estimation of three-phase Winsor regions that is more robust, simpler, and noniterative by making the equations dimensionless so that estimates of tail length and surface area are no longer needed. We predict phase behavior as a function temperature, pressure, volume, salinity, oil type, oil-water ratio, and surfactant/alcohol concentration. The dimensionless EoS is based on coupling the HLD-NAC (Hydrophilic Lipophilic Difference-Net Average Curvature) equations with new relationships between optimum salinity and solubility. An updated HLD expression that includes pressure is also used to complete the state description. A significant advantage of the dimensionless form of the EoS over the dimensional version is that salinity scans are tuned based only on one parameter, the interfacial volume ratio. Further, stability conditions are developed in a simplified way to predict whether an overall compositions lies within the single, two-, or three-phase regions. Important new microemulsion relationships are also found, the most important of which is that optimum solubilization ratio is equal to the harmonic mean of the oil and water solubilization ratios in the type III region. Thus, only one experimental measurement is needed in the three-phase zone to estimate the optimum solubilization ratio, a result which can aid experimental design and improve estimates of optimum from noisy data. Predictions with changing state variables are illustrated by comparison to experimental data using standard diagrams including a new type of dimensionless fish plot. The results show that the optimum solubilization ratio and salinity using the tuned dimensionless EoS are within average errors of 2.44% and 1.17% of experimental values for the fluids examined. We then use the dimensionless equations and thermodynamic first-principles to derive the constant in Huh's equation for interfacial tension prediction.
