Luo, Haishan (The University of Texas at Austin) | Mohanty, Kishore K. (The University of Texas at Austin) | Delshad, Mojdeh (The University of Texas at Austin) | Pope, Gary A. (The University of Texas at Austin)
Upscaling of unstable immiscible flow remains an unsolved challenge for the oil industry. The absence of a reliable upscaling approach greatly hinders the effective reservoir simulation and optimization of heavy oil recoveries using waterflood, polymer flood and other chemical floods, which are inherently unstable processes. The difficulty in upscaling unstable flow lies in estimating the propagation of fingers smaller than the gridblock size. Using classical relative permeabilities obtained from stable flow analysis can lead to incorrect oil recovery and pressure drop in reservoir simulations.
In a recent study based on abundant experimental data, it is found that the heavy-oil recovery by waterfloods and polymer floods has a power-law correlation with a dimensionless number (named viscous finger number in this paper), which is a combination of viscosity ratio, capillary number, permeability, and the cross-section area of the core. Based upon this important finding as well as the features of unstable immiscible floods, an effective-finger model is developed in this paper. A porous medium domain is dynamically identified as three effective zones, which are two-phase flow zone, oil single-phase flow zone, and bypassed oil (isolated oil island) zone, respectively. Flow functions are derived according to effective flows in these zones. This new model is capable of history-matching a set of heavy-oil waterflood corefloods under different viscosity ratios and injection rates. Model parameters obtained from the history match also have a power-law correlation with the viscous finger number.
The build-up of this correlation contains reasonable physical meanings to quantitatively characterize the upscaled behavior of viscous fingering effects. Having such a correlation enables the estimation of model parameters in any gridblock of the reservoir by knowing the local viscous finger number in reservoir simulations. The model is applied to several heavy-oil field cases with waterfloods and polymer floods with different heterogeneities. Oil recovery in water flooding of viscous oils is overpredicted by classical simulation methods which do not incorporate viscous fingering properly. Simulation results indicate that the new model reasonably differentiates the oil recoveries at different viscous finger numbers, e.g., lower injection rate leads to higher oil recovery. In contrast, classical simulations obtain close oil recoveries under different injection rates or degrees of polymer shear-thinning, which is apparently incorrect for unstable floods. Moreover, coarse-grid simulations using the new model are able to obtain consistent saturation and pressure maps with fine-grid simulations when the correlation lengths are not smaller than the coarse gridblock size. Furthermore, it is well captured by the model that the shear-shinning polymer solution can strengthen the fingering in high-permeability regions due to increased capillary number and viscosity ratio, which is not observed in waterflood. As a whole, the new model shows encouraging capability to simulate unstable water and polymer floods in heavy oil reservoirs, and hence can facilitate the optimization of heavy-oil EOR projects.
Okwen, Roland T. (Illinois State Geological Survey, Prairie Research Institute, University of Illinois at Urbana-Champaign) | Frailey, Scott M. (Illinois State Geological Survey, Prairie Research Institute, University of Illinois at Urbana-Champaign)
Historically, deep oil reservoirs with temperatures and pressures above the critical point of carbon dioxide (CO2) are generally preferred over shallower reservoirs in enhanced oil recovery (EOR) and CO2 storage operations because of high recovery and storage efficiencies associated with miscible floods. As a result, shallower reservoirs containing significant volumes of recoverable resource are generally overlooked. However, basins with relatively low geothermal gradients and high fracture gradients, such as the Illinois Basin, can sustain pressures above the vapor pressure of CO2 where CO2 changes from a gas to liquid. Liquid CO2 has fluid properties similar to that of supercritical CO2 and is more readily miscible with oil.
This study evaluates the EOR potential of low-temperature reservoirs based on the performance of a miscible liquid CO2 flood pilot at the Mumford Hills oil field in Posey County, Indiana. About 7,000 tons (6,350 tonnes) of CO2 were injected into a Mississippian sandstone reservoir over a period of 1 year to demonstrate miscible CO2 EOR in low-temperature oil reservoirs. The reservoir model was calibrated with available historical primary waterflood, and CO2 flood pilot data. The calibrated reservoir model was used to simulate different full-field CO2 EOR development scenarios. The projected oil recovery factors range between 10% and 14%, which compares well to the Permian Basin supercritical CO2 flood recovery range of 8% to 16%.
The oil recovery factors from the simulated scenarios suggest that liquid CO2 floods in low-temperature oil reservoirs can achieve an incremental oil recovery similar to deeper, supercritical CO2 floods. Re-evaluating previously overlooked shallow depleted reservoirs as potential candidates for liquid CO2 EOR provides the opportunity to increase the development of these shallow oil reservoirs available for miscible CO2 flooding
San, Jingshan (New Mexico Institute of Mining and Technology) | Wang, Sai (New Mexico Institute of Mining and Technology) | Yu, Jianjia (New Mexico Institute of Mining and Technology) | Lee, Robert (New Mexico Institute of Mining and Technology) | Liu, Ning (New Mexico Institute of Mining and Technology)
This paper reports the study of the effect of different ions (monovalent, bivalent, and multiple ions) on nanosilica-stabilized CO2 foam generation. CO2 foam was generated by co-injecting CO2/5,000 ppm nanosilica dispersion (dispersed in different concentrations of brine) into a sandstone core under 1,500 psi and room temperature. A sapphire observation cell was used to determine the foam texture and foam stability. Pressure drop across the core was measured to estimate the foam mobility. The results indicated that more CO2 foam was generated as the NaCl concentration increased from 1.0% to 10%. Also the foam texture became denser and foam stability improved with the NaCl concentration increase. The CO2 foam mobility decreased from 13.1 md/cp to 2.6 md/cp when the NaCl concentration increased from 1% to 10%. For the bivalent ions, the generated CO2 foam mobility decreased from 19.7 md/cp to 4.8 md/cp when CaCl2 concentration increased from 0.1% to 1.0%. Synthetic produced water with total dissolved solids of 17,835 ppm was prepared to investigate the effect of multiple ions on foam generation. The results showed that dense, stable CO2 foam was generated as the synthetic produced water and nanosilica dispersion/CO2 flowed through a porous medium. The lifetime of the foam was observed to be more than two days as the foam stood at room temperature. Mobility of the foam was calculated as 5.2 md/cp.
In this paper, we estimate foam parameters and investigate foam behavior for a given range of water saturation using two local equilibrium foam models: the population balance and the Pc*. Our method uses an optimization algorithm to estimate foam model parameters by matching foam measured pressure gradient from steady-state coreflood experiments. We calculate the effective foam viscosity and the water fractional flow using experimental data and we then compare lab data against results obtained with the matched foam models to verify the foam parameters. Other variables, such as the foam texture and foam relative permeability are used to further investigate the behavior of the foam during each experiment. We propose an improvement to the Pc* model with a better match in high quality regime by assuming resistance factor and critical water saturation is a linear function of pressure gradient. Results show that the parameter estimation method coupled with an optimization algorithm successfully matches the experimental data using both foam models. In the population balance, we observe different values of the foam effective viscosity for each pressure gradient due to variations of the foam texture and shear thinning viscosity effect. The Pc* model presents a constant effective viscosity for each pressure gradient; we propose the use of resistance factor and critical water saturation as a linear function of pressure to improve the match in the high quality regime, when applicable. Foam has been successfully used in the oil industry for conformance and mobility control in gas injection processes. The efficiency of a foam injection project must be assessed by means of numerical models. Although there are several foam flow models in the literature, the prediction of foam behavior is an important issue that needs further investigation.
Al Ayesh, A. H. (Department of Geoscience and Engineering, Delft University of Technology) | Salazar, R. (Department of Geoscience and Engineering, Delft University of Technology) | Farajzadeh, R. | Vincent-Bonnieu, S. | Rossen, W. R.
Foam can divert flow from higherto lower-permeability layers and thereby improve vertical conformance in gas-injection enhanced oil recovery. Recently,
The effectiveness of diversion varies greatly with injection method. In a SAG (surfactant-alternating-gas) process, diversion of the first slug of gas depends on foam behavior at very high foam quality. Mobility in the foam bank during gas injection depends on the nature of a shock front that bypasses most foam qualities usually studied in the laboratory. The foam with the lowest mobility at fixed foam quality does not necessarily give the lowest mobility in a SAG process. In particular, diversion in SAG depends on how and whether foam collapses at low water saturation; this property varies greatly among the foams reported by Kapetas et al. Moreover, diversion depends on the size of the surfactant slug received by each layer before gas injection. This of course favors diversion away from high-permeability layers that receive a large surfactant slug, but there is an optimum surfactant slug size: too little surfactant and diversion from high-permeability layers is not effective; too much and mobility is reduced in low-permeability layers, too. For a SAG process, it is very important to determine if foam collapses completely at irreducible water saturation.
In addition, we show the diversion expected in a foam-injection process as a function of foam quality. The faster propagation of surfactant and foam in the higher-permeability layers aids in diversion, as expected. This depends on foam quality and non-Newtonian foam mobility and varies with time of injection. Injectivity is extremely poor with foam injection, but is not necessarily worse than waterflood in some effective SAG foam processes
Jong, Stephen (University of Texas at Austin) | Nguyen, Nhut M. (University of Texas at Austin) | Eberle, Calvin M. (University of Texas at Austin) | Nghiem, Long X. (Computer Modelling Group Ltd.) | Nguyen, Quoc P. (University of Texas at Austin)
Low Tension Gas (LTG) flooding is a novel EOR process which can address challenging reservoir conditions such as high salinity, high temperature, and tight rock. Current process understanding is limited, and a joint experimental and modeling approach allows for both interpretation and insight into the complex interactions between the key process parameters of salinity gradient, foam strength, microemulsion phase behavior, and phase desaturation in order to achieve a physically correct and predictive process model.
We performed a series of corefloods in high permeability Berea sandstones (~500 mD) to demonstrate the impact of salinity gradient on the LTG process and interactions between key mechanisms such as microemulsion phase behavior and foam stability. In order to provide additional insight into the experimental study and improve understanding of the LTG process, we used our newly developed LTG simulator which we built within CMG GEM.
The results demonstrate that decreasing slug injection salinity can lead to a 15% increase in residual oil in place (ROIP) recovery over a slug injected at optimum salinity, with earlier breakthrough and steeper recovery slope. In addition, there is evidence of a late time pressure buildup as salinity is decreased through mixing with drive salinity which is indicative of increasing foam stability. This may be due to an inverse relationship between oil-water IFT and foam stability and thus designing an optimal salinity gradient for an LTG process requires balancing oil mobilization due to ultralow IFT and effectively displacing mobilized oil with adequate foam mobility control.
We introduce and show the strength our compositional LTG simulator in a pioneering laboratory and simulation study that sheds light on the interaction between salinity, microemulsion phase behavior, and foam strength. Our conclusions indicate a significant departure from traditional ASP understanding and methodology when designing an LTG salinity gradient and serve as a foundation for future investigation.
Enhanced-oil-recovery techniques by gas injection in shale reservoirs have been introduced and investigated. Laboratory and simulation works have shown good results for enhanced shale oil recovery, but one problem with gas injection is asphaltene precipitation and deposition. Damage due to asphaltene precipitation and deposition in conventional reservoirs has been reported in the literature. In shale reservoirs, pore and throat sizes are much smaller than in conventional reservoirs. Thus, large asphaltene aggregates may cause more serious problems in shale reservoirs.
This experimental study used a nanofiltration technique to investigate the size of asphaltene aggregates precipitated during CO2 and CH4 injection in a shale oil sample. Nano membranes of 200nm, 100nm and 30nm were used to filtrate oil samples injected with different mole fractions of CO2 and CH4 gas. The distribution of asphaltene aggregates’ size at different injected CO2 and CH4 concentrations were obtained and compared with the pore size distribution data of shale cores measured by mercury intrusion porosimeters. Results showed that a higher injected CO2 and CH4 concentration caused more asphaltene precipitation and growth in asphaltene aggregates’ size. The precipitated asphaltene particle size was large enough to cause a pore-blocking problem in tested shale cores.
Mukherjee, Joydeep (The Dow Chemical Company) | Nguyen, Quoc P. (The University of Texas at Austin) | Scherlin, John (Fleurde Lis Energy) | Vanderwal, Paul (The Dow Chemical Company) | Rozowski, Peter (The Dow Chemical Company)
A supercritical CO2 foam pilot, comprised of a central injection well in an inverted 5-spot pattern, was implemented in September 2013 in Salt Creek field, Natrona County WY. In this paper we present a thorough analysis of the pilot performance data that has been collected to date from the field. A monitoring plan was developed to analyze the performance of the pilot area wells before and after the start of the foam pilot. The injection well tubing head pressure was controlled to maintain a constant bottom hole pressure and the fluid injection rates were monitored to capture the effect of foam generation on injectivity. Inter-well tracer studies were performed to analyze the change in CO2 flow patterns in the reservoir. Production response was monitored by performing frequent well tests. The CO2 injection rate profile monitored over several WAG cycles during the course of the implementation clearly indicates the formation and propagation of foam deep into the reservoir. CO2 soluble tracer studies performed before and after the start of the foam pilot indicate significant areal diversion of CO2. The production characteristics of the four producing wells in the pilot area indicate significant mobilization of reservoir fluids attributable to CO2 diversion in the pattern. The produced gas-liquid ratio has decreased in all four of the producing wells in the pattern. Analysis of the oil production rates shows a favorable slope change with respect to pore volumes of CO2 injected. Segregation of CO2 and water close to the injection well seems to be the primary factor adversely affecting CO2 sweep efficiency in the pilot area. Foam generation leads to a gradual expansion of the gas override zone. The gradual expansion of the gas override zone seems to be the principal mechanism behind the production responses observed from the pilot area wells.
Given limited CO2 supply, operational constraints, and pattern specific reservoir performance, WAG schedule can be customized such that NPV or other metrics are optimized. Depending on the WAG schedule, recovery can fluctuate between 5–15% at the pattern scale due to reservoir heterogeneity causing variations in sweep efficiency. An analytical method was developed to optimize WAG schedules that couples traditional reservoir modeling and simulation with machine learning, enabling the discovery of optimal WAG schedules that increase recovery at the pattern level. A history-matched reservoir model of Chaparral Energy's Farnsworth Field, Ochiltree County, TX was sampled intelligently to perform predictive reservoir flow simulations and artificially build an intelligent reservoir model that samples a broad range of possible WAG scenarios for optimization. The intelligent model generates the next "best" sample to investigate in the numerical simulator and converges on the optima, quickly reducing the number of runs investigated. Results in this paper demonstrate that there can be significant improvements in net present value as well as net utilization rates of CO2 using this analytical technique. The WAG design generated by the intelligent reservoir model should be deployed in the field in early 2016 for validation. It is intended that the intelligent reservoir model will be updated on a regular basis as injection and production data is obtained. This effort represents the beginning of a paradigm shift in the application of modeling and simulation tools for significant improvements in field production operations.
It is generally assumed that while the presence of foam reduces the mobility of the gas phase, it does not alter the mobility of the liquid phase. Here, the effect of surfactant type and concentration on the behavior of nitrogen foam flow in porous media is investigated by simultaneous injection of gas and surfactant into Bentheimer sandstone cores. Different surfactant types, viz., anionic alpha-olefin-sulfonate (AOS) and zwitterionic Betaine with different surfactant concentrations from critical-micelle-concentration (CMC) to higher concentration are used in this study. The foam strength is quantified by measuring the pressure drop in different sections of the core. The liquid saturation is measured by analyzing the X-ray images obtained in a medical CT-scanner.
It is shown that the connate water saturation is reduced by increasing the surfactant concentration, and therefore the relative permeability relation for the aqueous phase should be modified when fitting the data to the foam models. It is observed that it is not possible to fit one monotonic liquid relative permeability curve to all the data points, obtained with different surfactant type and concentration in one rock type. Moreover, increasing AOS concentration above a certain value does not have a significant effect on the mobility reduction of the gas phase; however it modifies the liquid relative permeability. These results indicate that the water relative permeability measured in absence of surfactant should not be used to model the flow of foam in porous media, as it can lead to erroneous calculations of the liquid saturation.