Khodaparast, Pooya (Department of Energy and Mineral Engineering and EMS Energy Institute, The Pennsylvania State University, University Park) | Johns, Russell T. (Department of Energy and Mineral Engineering and EMS Energy Institute, The Pennsylvania State University, University Park)
Surfactant floods can attain high oil recovery if optimum conditions with ultra-low interfacial tensions (IFT) are achieved in the reservoir. A new equation-of-state (EoS) phase behavior model based on the hydrophilic-lipophilic difference (HLD-NAC) has been shown to fit and predict phase behavior data continuously throughout the Winsor I, II, III, and IV regions. The state-of-the-art for viscosity estimation, however, uses empirical non-predictive models based on fits to salinity scans, even though other parameters change, such as the phase number and compositions. In this paper, we develop the first-of-its-kind microemulsion viscosity model that gives continuous viscosity estimates in composition space. This model is coupled to our existing HLD-NAC phase behavior EoS.
The results show that experimentally measured viscosities in all Winsor regions (two and three-phase) are a function of phase composition, temperature, pressure, salinity, and
The objective of this research was to develop a model to predict the optimum phase behavior of chemical formulations for a given oil based on the molecular structure of the surfactants and co-solvents. The model is sufficiently accurate to provide a useful guide to an experimental testing program for the development of chemical EOR formulations. There are thousands of combinations of surfactants and co-solvents that could be tested for each oil, so even approximate predictions are very useful in terms of reducing the time and effort required for testing and for prioritizing the chemical combinations to test that are most likely to yield ultra-low IFT at reservoir conditions. The effects of changing molecular structures (e.g. swapping head groups, swapping hydrophobes, increasing the length of hydrophobes, increasing the number of PO and EO groups, adjusting the ratios of surfactants) are shown. The variables with the greatest impact on the optimum salinity and solubilization ratio were identified, and methods are proposed to shift the optimum salinity and the optimum solubilization ratios in any desired direction. The structure-property model was developed and tested using a large dataset consisting of 684 microemulsion phase behavior experiments using 24 oils. The chemical formulations used 85 surfactants and 18 co-solvents in various combinations. Both optimum salinity and optimum solubilization ratio (and thus IFT) are modeled whereas other models have focused almost exclusively on the optimum salinity. Predicting the optimum solubilization ratio is actually of more value because of its relationship to IFT. The models include the effects of co-solvent partitioning, soap formation and the molecular structure of both the surfactants and co-solvents.
Alboudwarej, Hussein (Chevron Energy Technology Company) | Sheffield, Jonathan M. (Chevron Energy Technology Company) | Srivastava, Mayank (Chevron Energy Technology Company) | Wu, Stanley S. (Chevron Energy Technology Company) | Zuo, Lin (Chevron Energy Technology Company) | Inouye, Arthur (Chevron Energy Technology Company) | Zhou, Dengen (Chevron Energy Technology Company) | Oghena, Andrew (Chevron Energy Technology Company)
Standard gas Enhanced Oil Recovery (EOR) Pressure-Volume-Temperature (PVT) program includes experiments such as solubility/swelling, multi-contact, slim tube, vapor-liquid equilibria (VLE) tests, and fluid property measurements. These tests are designed to determine the extent of gas miscibility and mixture phase behavior during gas injection in hydrocarbon reservoirs. These experimental programs are known to be expensive and time consuming. The degree of complexity increases as the industry move into conducting gas EOR PVT for high/ultra-high-pressure reservoirs. The focus of this paper is to demonstrate the challenges associated with these measurements and evaluate the merit, applicability, and usage of such data for fluid model development for high pressure gas EOR studies.
Associated challenges include utilizing gas concentrations up to 90% mole during swelling tests to determine critical mixture composition. Determination of dew point pressures by visual inspection or liquid build-up method proved inefficient. An interpretation of pressure-volume data showed good promise for determining both saturation pressure and liquid build-up curve for opaque dew point systems. VLE tests were designed at gas concentrations close to critical mixture composition to generate phases with increased affinity for mass transfer. Measured Minimum Miscibility Pressure (MMP) for all studied oil and gas systems were less than 6000 psia, except for N2 gas. Such relatively low MMP values suggest that development of full miscibility is not a concern for these high-pressure fluid systems. Such relatively low MMP values suggest that generated miscibility is not a concern for these high-pressure fluid systems. Instead the focus shifted to determine and effectively model the first contact miscibility pressures for these fluid systems (if it existed at pressures less than initial reservoir pressure). Measured MMPs were assigned a low weight factor in EoS model optimization process. Swelling test data for gas concentrations lower than 50% mole was of little value for EoS model optimization. Presence of precipitated asphaltenes challenged accurate measurement of liquid phase density and viscosity, as capture and analysis of a representative sample was very difficult. A knowledge of asphaltenes phase envelope for mixtures of reservoir fluid and injection gas proved to be invaluable.
A robust gas EOR PVT database was generated for mixtures of 6 injection gases and 7 deep water Gulf of Mexico Wilcox Trend reservoir fluids. Tests were carried out at pressures as high as 20000 psia and temperatures ranging between 230-265 °F. New high-pressure testing capabilities were developed, and modified data interpretation techniques were implemented. Lessons learned during the measurement, interpretation and application of these high-pressure gas EOR PVT data helped us design an effective measurement program. The developed fit-for-purpose experimental program leads to reduced cost (elimination of unnecessary tests) and increased reliability of high pressure phase equilibria and fluid property data for gas EOR reservoir simulation studies.
Large scale polymer flooding projects in heavy oil are now ongoing in several countries and numerous other projects are at the pilot or design stages. However, there is currently no guideline for the maximum acceptable oil viscosity, one of the important parameters in the screening of new projects. Standard screening criteria do not take the latest field results into account and more recent guidelines rely mostly on viscosity averages whereas they should focus on the extreme values instead.
Since the laboratory can only provide little help to settle this issue we propose to examine current field projects for guidance.
To the best of the author's knowledge, the Pelican Lake and the Seal polymer floods, both in Canada, are operating in the highest oil viscosity ranges; moreover, the data is public and can easily be accessed. We have therefore examined the performances of polymer injection in the highest ranges of oil viscosity in both fields to get an understanding of the limits. This involved first the identification of the highest oil viscosity patterns, then the estimation of the live oil viscosity during the polymer flood in these patterns and finally the performances of the polymer flood.
Viscosity measurements are notoriously difficult and not always very reliable in heavy oil and the evaluation of in-situ viscosity is even more difficult; therefore, we used ranges of viscosity rather than definite values. The observations from Pelican Lake and Seal seem in good agreement and suggest that polymer flood is still feasible and can provide an acceleration in production for live oil viscosities up to 10,000-12,000 cp. There is little experience beyond these values, but it appears that for higher ranges of viscosity polymer injection becomes much more difficult; in Seal polymer flood does not appear to be working satisfactorily in oil viscosities above 14,000 cp.
To the best of the author's knowledge, this is the first time that the issue of maximum oil viscosity is investigated in such a manner. Although these results are preliminary and would require further confirmation from other field cases, this paper will provide guidance to engineers screening heavy oil reservoirs for potential application of polymer flood.
The improved oil recovery of unconventional shale reservoirs has attracted much interest in recent years. Gas injection, such as CO2 and natural gas, is one of the most considered techniques for its sweep efficiency and effectiveness in low permeability reservoirs. However, the uncertainties of fluid phase behavior in shale reservoirs pose a great challenge in evaluating the performance of gas injection operation. Shale reservoirs are featured with macro-scale to nano-scale pore size distribution in the porous space. In fractures and macropores, the fluid shows bulk behavior, but in nanopores the phase behavior is significantly altered by the confinement effect. The integrated behavior of reservoir fluids in this complex environment remains uncertain.
In this study, we investigate the nano-scale pore size distribution effect on the phase behavior of reservoir fluids in gas injection for shale reservoirs using a multi-scale equation of state modeling. A case of Anadarko Basin shale oil is used. The pore size distribution is discretized as a multi-scale system with pores of specific diameters. The phase equilibria of methane injection into the multi-scale system are calculated. The constant composition expansions are simulated for oil mixed with various fractions of injected gas. Bubble point, swelling factor, criticality and fluid volumetrics are studied in comparison to the behavior of the bulk fluid. It is found that fluid in nanopores becomes supercritical with injected gas, but lowering the pressure below bubble point will turn it into the subcritical state. The swelling factor is slightly higher with nanopores, and bubble point is lower than the bulk. The degree of deviation depends on the amount of injected gas.
Jin, Luchao (University of Oklahoma) | Budhathoki, Mahesh (University of Oklahoma) | Jamili, Ahmad (University of Oklahoma) | Li, Zhitao (The University of Texas at Austin) | Luo, Haishan (The University of Texas at Austin) | Delshad, Mojdeh (The University of Texas at Austin) | Shiau, Ben (University of Oklahoma) | Harwell, Jeffrey H. (University of Oklahoma)
The surfactant screening process to develop an optimum formulation under reservoir conditions is typically time consuming and expensive. Theories and correlations like HLB, R-ratio and packing parameters have been developed. But none of them can quantitatively consider both the effect of oil type, salinity, hardness and temperature, and model microemulsion phase behavior.
This paper uses the physics based Hydrophilic Lipophilic Difference (HLD) Net Average Curvature (NAC) model, and comprehensively demonstrated its capabilities in predicting the optimum formulation and microemulsion phase behavior based on the ambient conditions and surfactant structures. By using HLD equation and quantitatively characterized parameters, four optimum surfactant formulations are designed for target reservoir with high accuracy compared to experimental results. The microemulsion phase behavior is further predicted, and well matched the measured equilibrium interfacial tension. Its predictability is then reinforced by comparing to the empirical Hand's rule phase behavior model. Surfactant flooding sandpack laboratory tests are also interpreted by UTCHEM chemical flooding simulator coupled with the HLD-NAC phase behavior model.
The results indicate the significance of HLD-NAC equation of state in not only shorten the surfactant screening processes for formulators, but also predicting microemulsion phase behavior based on surfactant structure. A compositional reservoir simulator with such an equation of state will increase its predictability and hence help with the design of surfactant formulation.
Ghasemi, M. (Petrostreamz AS) | Astutik, W. (Petrostreamz AS) | Alavian, S. A. (PERA AS) | Whitson, C. H. (PERA AS/NTNU) | Sigalas, L. (Geological Survey of Denmark and Greenland) | Olsen, D. (Geological Survey of Denmark and Greenland) | Suicmez, V. S. (Maersk Oil & Gas A/S)
This paper presents a novel technique to determine multi-component diffusion coefficients for CO2 injection in a North Sea Chalk Field (NSCF) at reservoir conditions. Constant volume diffusion (CVD) method is used, consisting of an oil-saturated chalk core in contact with an overlying free-space, which is filled with the CO2. The experimental data are matched with an EOS-based compositional model.
Transport by diffusion controls the dynamics of the constant-volume system, together with phase equilibria, allowing a consistent estimation of diffusion coefficients needed to describe the observed changes in system pressure.
We conduct two experiments at reservoir condition: one utilizes a core plug saturated with live-oil, and the other with stock tank oil (STO). Once the experiments are completed, EOS-based compositional simulation is performed to match the experimental data using the oil and gas diffusion coefficients as history matching parameters. The modeling work is conducted with a commercial reservoir simulator using a two dimensional radial grid model to describe the experimental setup.
The experiment utilizes a vertically-oriented core holder with a height of 92 mm and 37.6 mm in diameter. An outcrop chalk core with a sealing sleeve is mounted in the core holder, which has the same diameter and a height of 64.6 mm, thus resulting in an overlying void space. The system is initially saturated with oil at reservoir condition. CO2 is then injected from the top, forming an overlying CO2 chamber, and displacing oil towards the bottom of the core holder. Once CO2 fills the overlying bulk space, the system is isolated with no further injection or production.
The CO2 and oil reach and remain in equilibrium locally at the gas-oil interface throughout the test, initiating and maintaining the diffusion mechanism. Diffusion of CO2 into the oil results in a decreasing pressure, which is the main history matching parameter.
The multi-component diffusion coefficients are found to match the model pressure-time prediction to the experimental data. This suggests the modelling workflow incorporates a representative EOS model and the main transport dynamics controlled by diffusion are being treated properly.
The two main challenges in the modeling are (1) the limitation on setting an appropriately-high permeability for the CO2 chamber, and (2) the reservoir simulator neglects compositional dependency of diffusion coefficients.
Proper simulation of CO2 injection in fractured chalk reservoirs requires the ability to model multi-component diffusion accurately. The proposed CVD-method provides such modeling capabilities. Our modeling and experimental work indicate the novelty of the CVD method to determine the diffusion coefficients of a system where diffusion is the dominant displacement mechanism. The fact that the oil is contained within a low-permeability chalk sample reduces density-driven convection that could result due to non-monotonic oil density changes as CO2 dissolves into the oil.
Compositional reservoir simulation plays a vital role in the development of conventional and unconventional reservoirs. Two major building blocks of compositional simulation are phase behavior and fluid transport computations. The oil and gas reserves and flow of reservoir fluids are strongly dependent on phase behavior. In conventional reservoirs, capillary pressure is relatively small and is typically ignored in phase behavior calculations. However, large capillary pressure values are encountered in tight formations such as shales; and therefore, its effects should not be ignored in phase equilibria calculations. Neglecting the effects of capillary pressure on phase behavior can lead to an inaccurate estimation of original oil and gas in place as well as recovery performance. In spite of this, the effect of capillary pressure on phase behavior in tight reservoirs has not been well studied using compositional simulation, especially for hydraulically-fractured reservoirs.
In this paper, we develop a new compositional reservoir simulator capable of modeling discrete fractures and incorporating the effect of capillary pressure on phase behavior. Large-scale natural and hydraulic fractures in tight rocks and shales are modeled with a technique called Embedded Discrete Fracture Model (EDFM) where fractures are modeled explicitly without using local grid refinement or an unstructured grid. Flow of hydrocarbons occurs simultaneously within similar and different porosity types. Capillary pressure is considered in both flow and flash calculations, where simulations also include variable pore size as a function of gas saturation in each grid block. We examine the impact of capillary pressure on the original oil in place and cumulative oil production for different initial reservoir pressures (above and below the bubble-point pressure) on Bakken and Eagle Ford fluids. The importance of capillary pressure on both flow and flash calculations from hydraulically fractured horizontal wells during primary depletion in fractured tight reservoirs using Bakken fluid composition is demonstrated.
Phase behavior calculations show that bubble-point pressure is suppressed allowing the production to remain in the single-phase region for a longer period of time and altering phase compositions and fluid properties such as density and viscosity of equilibrium liquid and vapor. The results show that bubble-point suppression is larger in the Eagle Ford shale than for Bakken. When capillary pressure is considered, we found an increase in original oil in place up to 4.1% for Bakken and 46.33% for the Eagle Ford crude. Depending on the initial reservoir pressure, cumulative primary production after one year increases owing to capillary pressure by approximately 9.0 – 38.2% for Bakken oil and 7.2 – 154% for Eagle Ford oil. The recovery increase caused by capillary pressure becomes more significant when reservoir pressure is far below bubble-point pressure. The simulation results with hydraulically fractured wells give similar recovery differences; cumulative oil production after 1 year is 3.5 – 5.2% greater when capillary pressure is considered in phase behavior calculations for Bakken.
The use of isenthalpic flash has become of interest for the simulation of some heavy oil recovery processes where large temperature changes are experienced. For these thermal simulations energy can be used as a primary variable. This leads to thousands or millions of individual multiphase isenthalpic flash calculations. Robust and efficient algorithms for multiple-phase isenthalpic flash are required to improve the efficiency of compositional simulations for thermal recovery.
The general framework on state function based flash specifications proposed by
Narrow boiling mixtures can be dealt with in the majority of cases without any significant difficulty. This is true of the direct substitution algorithm and the proposed solution procedure. The vast majority of examples can be solved without using Q function maximisation. The challenges associated with multiphase calculations in the Newton steps are investigated. In particular, inadequate initial estimate of the equilibrium type may lead to non-convergent iteration. This can usually be solved by introduction of a new phase and/or elimination of an existing phase. The speed of the method is analysed for a large number of specifications and is found to be only slightly more expensive than isothermal flash in the majority of cases.