Numerous studies indicate that the pressure/volume/temperature (PVT) phase behavior of fluids in large pores (designated "unconfined" space) deviates from phase behavior in nanopores (designated "confined" space). The deviation in confined space has been attributed to the increase in capillary force, electrostatic interactions, van der Waals forces, and fluid structural changes.In this paper, conventional vapor/liquid equilibrium (VLE) calculations are modified to account for the capillary pressure and the critical-pressure and -temperature shifts in nanopores. The modified VLE is used to study the phase behavior of reservoir fluids in unconventional reservoirs. The multiple-mixing-cell (MMC) algorithm and the modified VLE procedure were used to determine the minimal miscibility pressure (MMP) of a synthetic oil and Bakken oil with carbon dioxide (CO 2 ) and mixtures of CO 2 and methane gas.We show that the bubblepoint pressure, gas/oil interfacial tension (IFT), and MMP are decreased with confinement (nanopores), whereas the upper dewpoint pressure increases and the lower dewpoint pressure decreases.

Summary Production of tight oil from shale reservoirs in North America reduces oil imports and has better economics than natural gas. Currently, there is a strong interest in oil production from Bakken, Eagle Ford, Niobrara, and other tight formations. However, oil-recovery fraction for Bakken remains low, which is approximately 4–6% of the oil in place. Even with this low oil-recovery fraction, a recent United States Geological Survey study stated that the Bakken and Three Forks recoverable reserves are estimated to be 7.4 billion bbl; thus, a large volume of oil will remain unrecovered, which was the motivation to investigate the feasibility of enhanced oil recovery (EOR) in liquid-rich shale reservoirs such as Bakken. In this paper, we will present both laboratory and numerical modeling of EOR in Bakken cores by use of carbon dioxide (CO2), methane/ethane-solvent mixture (C1/C2), and nitrogen (N2). The laboratory experiments were conducted at the Energy and Environmental Research Center (EERC). The experiments recovered 90+% oil from several Middle Bakken cores and nearly 40% from Lower Bakken cores. To decipher the oil-recovery mechanisms in the experiments, a numerical compositional model was constructed to match laboratory-oil-recovery results. We concluded that solvent injection mobilizes matrix oil by miscible mixing and solvent extraction in a narrow region near the fracture/matrix interface, thus promoting countercurrent flow of oil from the matrix instead of oil displacement through the matrix. Specifically, compositional-modeling results indicate that the main oil-recovery mechanism is miscible oil extraction at the matrix/fracture interface region. However, the controlling factors include repressurization, oil swelling, viscosity and interfacial-tension (IFT) reduction, diffusion/advection mass transfer, and wettability alteration. We scaled up laboratory results to field applications by means of a compositional numerical model. For field applications, we resorted to the huff ’n’ puff protocol to assess the EOR potential for a North Dakota Middle Bakken well. We concluded that long soak times yield only a small amount of additional oil compared with short soak times, and reinjecting wet gas, composed of C1, C2, C3, and C4+, produces nearly as much oil as CO2 injection.

Compositional modeling of hydraulically stimulated naturally fractured liquid-rich shale (LRS) reservoirs is a complex process that is yet to be understood. The flow and multiphase mass transfer in the nano-, meso-, and macro-scale pores, as in Eagle Ford, Woodford and Bakken is of great interest. Understanding the production mechanisms from such reservoirs is crucial in the overall effort to increase the ultimate hydrocarbon production. Thus, we focused on deciphering the physical fundamentals of various recovery mechanisms via reservoir modeling. The starting point was examining the phase behavior issues in unconventional reservoirs. Specifically, we constructed phase diagrams using a new correlation to shift the critical properties of components in the nano and meso-scale pores. The correlation was applied to three recently published Eagle Ford fluid samples. The new phase behavior correlation was used in a dual-permeability compositional model to determine the nature of pore-to-pore flow and, eventually, the hydrocarbon production from wells. In the simulation models we allowed for the phase behavior differences between fracture and matrix and included a multi-level flow hierarchy from matrix (nano, meso, and macropores) to fractures and finally to the well. To make computation accurate we resorted to a series of detailed logarithmic local grid refinement (LS-LGR) in various strategic subdomains in the matrix and fracture.As a result of this modeling study, we have concluded several reasons why hydrocarbon fluids can move in the shale reservoir nano, meso, and macro-scale pores and why we are able to produce from such low-permeability reservoirs. For instance, favorable phase envelope shift of hydrocarbon mixtures in the nano-and meso-scale pores is one of the contributing factors to economic production in gas-condensate and bubble-point systems. Also noted, when the phase envelope is crossed in gascondensate systems, a large gas-to-oil volume split in the nano, meso, and macro-pores plays a crucial role in hydrocarbon recovery during depletion. For the bubble-point oil region, the low viscosity of the liquid phase and the delay in gas bubble evolution appears as the main reason for favorable oil production. Furthermore, 'rubblizing' the reservoir in the vicinity of hydraulic fractures creates another favorable environment for improved drainage, which is why multi-stage hydraulic fracturing is so critical in successful development of shale reservoirs.

No abstract

Production of tight oil from shale reservoirs in North America reduces oil imports and has better economics than natural gas. Consequently, there is a strong interest in oil production from Bakken, Eagle Ford, and Niobrara. However, oil recovery factor for Bakken remains low, which is about four to six percent of the oil in place. Even with this low oil recovery factor, Bakken recoverable reserves are estimated to be 7.4 billion barrels; thus, a large volume of oil will remain unrecovered. This low level of oil recovery was the motivation to investigate the feasibility of enhanced oil recovery (EOR) in liquid-rich shale reservoirs such as Bakken. In this paper, we will present both laboratory and numerical modeling of EOR in Bakken cores using CO2, C1-C2 mixture, and N2. The laboratory experiments were conducted at the Energy and Environmental Research Center (EERC). The experiments recovered 90+ percent oil from several Middle Bakken cores and nearly 40 percent from Lower Bakken cores. To decipher the oil recovery mechanisms in the experiments, a numerical compositional model was constructed to match laboratory oil recovery results. We concluded that solvent injection mobilizes matrix oil by miscible mixing and solvent extraction in a narrow region near the fracture-matrix interface, thus promoting counter-current flow of oil from the matrix instead of oil displacement through the matrix. Specifically, compositional modeling results indicate that the main oil recovery mechanism is miscible oil extraction at the matrix-fracture interface region. However, the controlling factors include re-pressurization, oil swelling, viscosity and interfacial tension reduction, diffusion-advection mass transfer, and wettability alternation. We scaled up laboratory results to field applications via a compositional numerical model. For field applications, we resorted to the huff-and-puff protocol to assess the EOR potential for a North Dakota Middle Bakken well. We concluded that long soak times yields only a small additional oil compared to short soak times, and re-injecting wet gas, composed of C1,C2, C3, and C4+, produces nearly as much oil as CO2 injection.