Executive Summary

This month, SPE Journal publishes 30 new papers, organized in five categories.

SPOTLIGHT: Data Analytics. Lee et al. use a long short-term-memory (LSTM) algorithm for predicting gas production of the Duvernay shale in Canada. The LSTM model is trained in 405 seconds by two features of production data and a shut-in (SI) period from 300 wells. The two-feature case can predict future production rates according to the SI period and provide a stable result for available time-series data.

Shale Rock and Fluid Characterization. Tandon and Heidari present a pore-scale simulation method for reliable modeling of nuclear-magnetic-resonance (NMR) response. The method quantifies the mechanisms that contribute to NMR surface relaxation of protons in kerogen pores of organic-rich mudrocks. The inputs to the simulator include pore geometries and the bulk and surface properties of different fluids present in the pore space. The outputs include T2 and T1 decay constants of the pore geometries.

Civan presents a correlation of the compressibility, porosity, and permeability of shale reservoirs by considering the effects of stress shock causing a slope discontinuity and loading/unloading hysteresis. Two approaches are used in the study. The first approach implements a kinetic model leading to a modified power-law equation, and the second approach applies an elastic cylindrical pore-shell model, leading to a semi analytical equation. Both approaches are shown to yield high-quality correlations.

Ratnakar and Dindoruk discuss flow and adsorptive characteristics of methane and other natural gases onto tight rock formations. Their method enables the simultaneous determination of nanodarcy permeability and adsorption isotherms of gas in such formations. The overall methodology can be applied to any type of adsorbing gases and shale/coal samples.

Li et al. conducted modeling and experimental work to investigate methane adsorption at high pressure. They found that the apparent excess isotherm determined by the He-based volume gradually becomes negative at high pressures, but the actual one determined by the adsorbate-accessible volume always remains positive. Also, the negative adsorption phenomenon in the apparent excess isotherm is a result of the overestimation in the adsorbate-accessible volume, and a larger overestimation leads to an earlier appearance of this negative adsorption. The positive amount in the actual excess isotherm indicates that the adsorbed phase is always denser than the bulk gas because of the molecule/pore-wall attraction aiding the compression of the adsorbed molecules.

To study the complex multiphase transitions of multicomponent fluids in multiscale volumes in shales, Zhao and Jin used density-functional theory (DFT) and explicitly considered fluid/surface interactions, inhomogeneous-density distributions in nanopores, volume partitioning in nanopores, and connected macropores/natural fractures. They found that vapor-like and liquid-like phases can coexist in nanopores when pressure is between the bubblepoint and dewpoint pressures of nanoconfined fluids, both of which are much lower than those of the originally injected hydrocarbon mixtures.

Bui et al. use molecular-dynamics (MD) simulations to investigate how the transport of hydrocarbons in kerogen and hydrocarbon recovery can be altered with the delivery of microemulsion and surfactant micelles into the pore network. The authors discuss how the maturation of kerogen during catagenesis relates to the qualities of the kerogen pore network, such as pore size, shape, and connectivity, and plays an important role in the recovery of hydrocarbons.

Baek and Akkutlu developed a computational method using molecular-simulation data to estimate the average mean-free-path length of multicomponent hydrocarbon molecules in an organic nanochannel. Grand-canonical Monte Carlo simulation is used first to construct the equilibrium distribution of the gas molecules in the channel. Their results show that the smaller the channel is, the denser the gas mixture becomes because of nanoconfinement effects.

Cao et al. propose a novel method to evaluate the gas content of shale through a new perspective—fractionation of carbon isotopes of methane.  Their method could provide a promising means for the identification of sweet spots in shale-gas reservoirs. Moreover, the method might have the potential to economically and rapidly evaluate the remaining resources in producing wells in future applications.

Xue et al. use local grid refinements and an embedded discrete fracture model for simulating hydraulically fractured shales using the fast-marching method (FMM). The authors indicate that the three main contributions of the proposed methodology are (i) unique mesh-generation schemes to link fracture and matrix flow domains, (ii) diffusive time of flight (DTOF) calculations in locally refined grids, and (iii) sensitivity studies to identify optimal discretization schemes for the FMM-based simulation.

Liu and Valkó discuss production-decline models in unconventional reservoirs using anomalous diffusion stemming from a complex multistage hydraulically fractured network performed in horizontal wells. The results of all case studies display good matches between their model and the production data. This highlights the model’s capability to accurately describe the transient regime of the flow in the extremely heterogeneous fracture networks on the basis of average values of the formation properties.

Almubarak et al. discuss how to enhance deep-reservoirs productivity with the use of slickwater fracturing fluids in unconventional shale reservoirs and crosslinked fracturing fluids in conventional reservoirs. The double challenge is the combination of low permeability and higher temperatures. The authors developed a new hybrid dual-polymer hydraulic-fracturing consisting of a guar derivative and a polyacrylamide-based synthetic polymer. The major benefit of using a mixed-polymer system is reduced polymer loading.

Ding studied the multiple-interacting-continua (MINC) method for flow modeling in fractured reservoirs because it is commonly considered as an improvement over the dual-porosity model. To overcome the shortcomings in the standard MINC approach in dealing with matrix/fracture transfers under variable matrix-block sizes, he introduced new approaches for the MINC subdivision and the transmissibility computations.

Improved Oil Recovery (IOR)/Enhanced Oil Recovery (EOR). Singh et al. discuss the use of nanoparticle-microencapsulated acids (MEA) for treatment of calcite-rich shales. Fracture closure was simulated by increasing the overburden pressure. The authors measured conductivity of the cores before and after the MEA treatment. The result was outstanding as it led to the conclusion that the MEA could dramatically improve (up to 40 times) the permeability of the unpropped fractures by creating concentration-dependent, nonuniform localized surface etching.

Ghalamizade Elyaderani et al. describe an experimental investigation of mechanisms for enhancing recovery from heavy-oil reservoirs. The authors use functionalized multiwalled carbon nanotubes (MWCNT) in three concentrations—0.01, 0.05, and 0.1 wt%—to conduct micromodel, interfacial-tension (IFT), wettability, viscosity, phase-behavior, and static-adsorption tests. They conclude that the functionalized MWCNT might help increase heavy-oil recovery through mechanisms of reducing IFT, changing wettability from oil-wet to water-wet in oil-wet reservoirs, increasing viscosity (slightly), and developing a water-in-oil emulsion.

Ma et al. develop an analytical material balance model (MBM) for primary production and cyclic solvent injection in a heavy-oil reservoir. The MBM predicts cumulative heavy-oil and gas-production data, as well as the average reservoir pressures, during the primary production and subsequent cyclic solvent injection (CSI) in a heavy-oil reservoir. The theoretical MBM considers the nonequilibrium foamy-oil phase behavior and foamy-oil flow by invoking two kinetic equations with nucleation and decay coefficients. Laboratory sandpack tests of the primary production and subsequent CSI validate the MBM.

Perez-Perez et al. perform simulations of the in-situ upgrading (IU) process that include interpretation of laboratory experiments and study of field-scale tests. A general IU numerical model for the different experimental setups is developed and compared with experimental data, using a commercial reservoir-simulator framework. This model is capable of representing the phase distribution of pseudocomponents, the thermal decomposition reactions of bitumen fractions, and the generation of gases and residue (solid) under thermal cracking conditions. Simulation results for the cores exposed to a temperature of 380°C and production pressure of 15 showed that oil production and oil-sample quality were well-predicted by the model. 

Azad and Trivedi present a critical review of quantification of the viscoelastic effects during polymer flooding. Discussions on the shortcomings of the existing viscoelastic models caution the current chemical-enhanced-oil-recovery researchers about their applications and potential consequences. Most of the earlier models rely on the exclusive use of the Deborah number to quantify the viscoelastic effects, but the Deborah number overlooks mechanical-degradation effects. The main limitation with all the existing continuum viscoelastic models is the empirical reliance on coreflood data to predict the shear-thickening effects in porous media. The conventional capillary number (Nc) fails to explain the reduction in residual oil saturation (Sor) during viscoelastic polymer flooding.

Janssen et al. perform experimental work with a view to increment oil recovery by alkaline/surfactant/foam flooding and investigate the effect of drive-foam quality on oil-bank propagation. Foam qualities range between 57 and 97%. Their findings suggest that dispersive characteristics at the leading edge of the generated oil bank are strongly related to the surfactant slug size used, the lowest drive-foam quality assessed that yielded the highest apparent foam viscosity (and, thus, the most stable oil-bank displacement), and drive-foam strength increased upon touching the oil bank when using drive-foam qualities of 57 and 77%.

Lenchenkov et al. discuss laboratory propagation of polymer nanospheres in outcrop cores. When nanospheres are used for in-depth diversion in heterogeneous reservoirs, it is desired that spheres propagate deep into the reservoir along highly permeable zones with a resistance-factor (RF) buildup over time. This results in the reduced permeability of these reservoir zones and the diversion of subsequently injected water into unswept areas with higher oil saturation. Laboratory work described in this article shows that the propagation of nanospheres in porous media is highly dependent on the brine salinity in cores with single- and multiphase saturations. For the same experimental conditions, the RF of nanospheres in porous media depends on the flow rate.

Rognmo et al. investigate pore-to-core EOR upscaling for CO2 foam, for carbon capture, utilization, and storage (CCUS). At pore-scale, high-pressure silicon-wafer micromodels showed in-situ foam generation and stable liquid films over time during no-flow conditions. Intrapore foam bubbles corroborated high apparent foam viscosities measured at core scale. CO2-foam apparent viscosity was measured at different rates (foam-rate scans) and different gas fractions (foam-quality scans) at core scale. Strong foam was generated in EOR corefloods at a gas fraction of 0.70, resulting in an apparent viscosity of 39.1 mPa·s. Foam parameters derived from core-scale foam floods were used for numerical upscaling and field-pilot performance assessment.

Othman et al. discuss the effect of fines migration during CO2 injection using pore-scale characterization. Results show the dissolution of dolomite and other high-density minerals. Mineral dissolution dislodges fines particles, which migrate during water-saturated-scCO2 injection. During CO2-saturated-water injection, the permeability of Berea 1 and Berea 2 increase by 29 and 13%, respectively. After water-saturated-scCO2 injection, the permeability of Berea 1 and Berea 2 decrease by 60%. The permeability damage of the sample can be explained by fines migration and subsequent blockage. SEM-EDS images also show instances of pore blockage.

Dong et al. investigate how to increase oil recovery from high-temperature ultrahigh-salinity fractured oil-wet carbonate reservoirs. In their experimental work, they inject ultralow-IFT foam in oil-wet homogeneous and fractured cores. The ultralow-IFT formulation helps to mobilize the residual oil for better displacement efficiency and reduce the unfavorable capillary entry pressure for better sweep efficiency. The selective diversion of foam makes it a good candidate for a mobility-control agent in a fractured system for better sweep efficiency.

Low-Salinity Waterflood. Du et al. discuss a microfluidic investigation of low-salinity effects during oil recovery. This is a no-clay and time-dependent mechanism. They conclude that low-salinity tertiary waterflooding can improve oil recovery by an improvement of sweep efficiency, which is a consequence of residual-oil dewetting and swelling; the low-salinity effect can occur without the existence of clay; and the wettability alteration and oil swelling are time-dependent processes and should be expressed as a function of oil/water contact time rather than dimensionless time (pore volume), which explains some observations from previous coreflood experiments.

Mahzari et al. evaluate the decisive role of microdispersion formation in improved oil recovery by low-salinity-water injection in sandstone formations. The results demonstrate that the ratio of the microdispersion quantity to bond water is the main factor controlling the effectiveness of low-salinity-water injection. In general, a monotonic trend was observed between incremental oil recovery and the microdispersion ratio of the different crude-oil samples. In addition, it can be inferred from the results that geochemical interactions (pH and ionic interactions) would be mainly controlled by the rock’s initial wettability, and also that these processes could not affect the additional oil recovery by low-salinity-water injection.

Al-Ibadi et al. present an extended fractional-flow model of low-salinity waterflooding accounting for dispersion and effective salinity range. The authors observed that dispersion of the salinity profile affects the fractional-flow behavior depending on the effective salinity range. They derived an extended form of the fractional-flow analysis to include the impact of salinity dispersion. The simulator solution is equal to analytical predictions from fractional-flow analysis when the midpoint of the effective salinity range lies between the formation and injected salinities. This improves predictive ability and also reduces the requirement for full simulation.

Chemically-reactive flow-enhancing chalk compaction is of significant importance for EOR, compaction, and subsidence in North Sea chalk reservoirs. Andersen and Berawala focus on Ca, Mg, and NaCl brines that interact with the chalk by the dissolution of calcite and the precipitation of magnesite. The model was validated and parameterized against data from 22 core samples from two chalk types at ambient to Ekofisk-reservoir conditions (130°C). This model is the first to link a vast set of data on this subject and predict performance under new experimental conditions. This also represents a first step in upscaling such results from the laboratory toward the field.

Wellbore Modeling. Tang et al. develop a new drift-flux (DF) multiphase-flow model that exhibits equivalent or better performance compared with the existing models. More significantly, the model is shown to be numerically smooth, continuous, and stable for cocurrent flow when implemented in a fully implicit and coupled wellbore/reservoir simulator.

When it comes to fully understanding the mechanism of mudcake buildup, Jaffal et al. fill a gap in the literature by extending a previously developed model, along with experimental validation, for mudcake buildup at the wellbore wall to capture the basic physics of mudcake buildup inside fractures and the resulting fracture sealing.

Jayasinghe et al. present a new well model for reservoir simulation that relates the volumetric flow rate and the bottomhole pressure of the well to the reservoir pressure through a spatially distributed source term that is independent of the numerical method and the discrete mesh used to solve the flow problem. This is in contrast to the widely used Peaceman-type well models. They show that for high-order discretizations and mesh-adaptation schemes, their model is more adaptable. 

Reza Fassihi, SPE J. Executive Editor;
BHP Petroleum, Houston