Laboratory Investigations of CO2 Nearmiscible Application in Arbuckle Reservoir

SPE 129710 Laboratory Investigations of CO2 Near-miscible Application in Arbuckle Reservoir L H Bui, J S Tsau, and G P Willhite, SPE, University of Kansas Copyright 2010, Society of Petroleum Engineers This paper was prepared for presentation at the 2010 SPE Improved Oil Recovery Symposium held in Tulsa, Oklahoma, USA, 24–28 April 2010 This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s) Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s) The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied The abstract must contain conspicuous acknowledgment of SPE copyright Abstract Carbon dioxide (CO2) is a proven enhanced oil recovery technology However, many reservoirs are located at shallow depths or geologic conditions such that CO2 can not be injected at pressures above the MMP CO2 injection is usually not considered as an enhanced oil recovery process in these reservoirs When CO2 is injected below the MMP, displacement efficiency decreases as a result of the loss of miscibility Near miscible displacement has sometimes been referred to as the process occurring between immiscible and miscible pressures, but has never been clearly defined This paper describes laboratory study of CO2 near miscible displacement in an Arbuckle reservoir in Kansas Phase behavior studies between CO2 and Arbuckle crude oil were carried out to define near miscible conditions at reservoir conditions Swelling/extraction tests combined with slim-tube experiments were interpreted to identify the mass transfer mechanisms at near miscible condition A phase behavior model was developed to match PVT data and MMP in the slim-tube experiment Good agreement was obtained between simulated and observed data from slim-tube experiments Core flooding tests were conducted to evaluate oil recovery at near miscible condition at which pressure varies from 1350 psi (MMP) to 1150 psi Recovery of over 50% of the waterflood residual oil saturation was observed when CO2 was used to displace Arbuckle oil from Berea, Baker dolomite and Arbuckle dolomite cores At near miscible conditions, extraction appears to be the primary mechanism for mass transfer between hydrocarbon components and CO2 However, the reduction of oil viscosity by a factor of five occurred when CO2 dissolved in the oil This suggests that some of the additional oil recovery may be attributed to reduction of the mobility ratio between CO2 and resident oil Introduction Arbuckle reservoirs are a significant resource in Kansas for improved oil recovery These reservoirs have produced an estimated 2.2 billion barrels of oil representing 35% of the 6.1 billion barrels of oil of total Kansas oil production (Franseen et al., 2004) Most Arbuckle reservoirs have active water drives which have maintained reservoir pressure at 1000-1100 psig for nearly 50 years eventhough millions of barrels of fluid have been produced Initial studies of CO2 miscible flooding indicated that miscibility is not achievable at the reservoir operating pressure in most Arbuckle reservoirs For example, the Arbuckle reservoir oil in the Bemis-Shutts field has a MMP of 1400 psi while the current operating pressure is 1100 psi in a large portion of the field The possibility of operating at pressures below the MMP means that many of these fields in central Kansas, which might be otherwise abandoned with substantial remaining oil left in place could be considered for CO2 injection Carbon dioxide (CO2) injection is normally operated at pressures above minimum miscibility pressure (MMP), which is determined by crude oil composition and reservoir conditions When carbon dioxide is injected at pressures slightly below MMP, the process is commonly referred as a near miscible displacement The effectiveness of near miscible displacement within a certain range of pressure also depends on the oil composition and reservoir condition Within this pressure range, significant oil recovery efficiency has been observed in slim-tube experiments and to a lesser extent in core tests Although recovery efficiency is less than that from miscible displacement, the recovery efficiency is still much greater SPE 129710 than for waterflood recovery This effect of pressure on the oil recovery has been investigated with different conclusions Shyeh-Yung (1991) suggests that recovery of residual oil by injection of CO2 at pressure below MMP benefits from a possible improvement of mobility ratio between CO2 and oil thereby increasing oil recovery Grigg et al (1987) report that oil recovery below the MMP in the near miscible region is not effective due to an inefficient extraction process This paper summarizes a study of the injection of CO2 at near miscible condition to improve oil recovery in an Arbuckle reservoir where the current operating pressure is approximately 1150 psig Phase behavior studies and core flood tests were carried out to improve understanding of the mechanisms affecting the process of near-miscible CO2 flooding and evaluate the feasibility of near miscible carbon dioxide injection in Arbuckle reservoirs All experiments were performed with centrifuged and filtered Ogallah stock tank oil obtained from Ogallah Unit, Trego Country, Kansas The location of the unit is shown in Figure Phase Behavior Studies Phase behavior studies consisted of slim-tube experiments to determine the minimum miscibility pressure (MMP) of the Ogallah oil, swelling/extraction tests to determine the effect of pressure on the solubility of CO2 in oil and measurement of the viscosity and density of Ogallah crude oil saturated with CO2 at various pressures A phase behavior model was constructed by tuning Peng-Robinson Equation of State (PREOS) with experimental data The composition of Ogallah oil was determined by gas chromatography analysis and shown in Figure Physical properties of the oil and the lumped heavy component are presented in Table Commercial CO2 of 99.99 % purity was used Slim-Tube Experiments The slim-tube setup is shown in Figure The slim-tube consists of a coiled 40 ft-long stainless steel tube with an ID of 0.24 in packed with glass beads The slim tube has a pore volume (PV) of 128 cm3 and a permeability of 4900 md The pressure of the system was maintained by the back-pressure regulator at the outlet CO2 was injected at a constant rate of 0.1cc/min to displace the oil Density of the effluent was measured continuously using an inline densitometer The effluent was continuously flashed to atmospheric conditions The separator gas was connected to a flow meter for measurement of the gas flow The amount of fluid produced was collected in a graduate cylinder and determined by weight The composition of the produced fluid was not determined An oil recovery factor of at least 90% at 1.2 HCPV of CO2 injected was used to define MMP in the slim-tube experiments for this oil Experiments were done at two temperatures (110˚F and 125˚F) representing the range of temperatures obtained from field data Oil recovery at 1.2 HCPV CO2 injected was plotted in Figure against average pressure The MMP was estimated to be 1350 psig at 110 oF and 1650 psig at 125 oF The measurement of MMP indicates that miscibility would not be reached during carbon dioxide injection at the current reservoir pressure of 1150 psig However, the recovery efficiency from the slim-tube experiments was about 80% at the current reservoir pressure These data indicate that substantial amounts of hydrocarbon constituents were extracted by the CO2 as it was displaced through the slim-tube Figure shows the density profile of the effluent at pressures below MMP Prior to the breakthrough of CO2 the effluent density was equal to the oil phase density (0.834 g/cc) at reservoir temperature and slim-tube average pressure The abrupt change in density of the effluent corresponds to the breakthrough of CO2 Significant reduction of effluent density occurred at pressures below MMP following CO2 breakthrough After breakthrough of CO2, average densities of the effluent were 0.434, 0.535 g/cc at average pressures of 1100, 1200 psig At the same pressure and temperature, the densities of pure CO2 are 0.221 g/cc and 0.275 g/cc The increase of density in the effluent profile is evidence that light hydrocarbon components from the oil continued to be vaporized or extracted by CO2 contributing to relatively high recovery efficiencies for near miscible CO2 displacement Swelling/Extraction Tests Swelling/extraction experiments were conducted in a visual PVT cell with a total volume of 26 cc Details of the experimental setup are described in a companion paper (Tsau et al., 2010) Typically, a pre-determined volume of crude oil is injected into the view cell The cell pressure is increased in discrete steps by CO2 injection from the top of the view cell until a desired pressure is achieved A stir bar inside the view cell is used to accelerate the mass transfer between oil and CO2 When the equilibrium is achieved, the volume of CO2-saturated oil is measured with a cathetometer Solubility of CO2 in oil, relative volume change of oil and density of CO2-saturated oil are determined as a function of pressure Figure shows the swelling/extraction curve for oil/CO2 system at 110 ºF with cc of sample (12 % volume of cell volume) CO2 solubility is also plotted in the same figure as a function of pressure The swelling factor (SF) of oil is the ratio of volume at reservoir conditions to volume at stock tank conditions This value was determined by measuring the change of interface level as a result of CO2 dissolution in the oil CO2 solubility was calculated based on the assumption that negligible hydrocarbon component in the vapor phase Maximum swelling occurred at 1150 psig, when its volume was 1.21 of its SPE 129710 original volume with 0.728 mole fraction of CO2 dissolves in the liquid phase Extraction appears to have started at approximately the same pressure As the pressure increased, the oil volume continued to shrink and CO2 extracted more hydrocarbon components from liquid phase The rate of oil volume shrinking by extraction was faster than the rate of swelling by continued dissolution of CO2 At 2035 psig, the oil volume shrank as much as 39.2 % of its original volume Viscosity Measurement The viscosity of CO2-saturated oil was measured using a Cambridge Applied Systems high pressure viscometer (ViscoPro 2000 System 4-SPL-440 with Viscolab software) The schematic drawing of this setup is shown in Figure Detail of the experimental setup is described elsewhere by Ahosseini, et al (2008) Figure shows the effect of solubility of CO2 on viscosity of CO2-saturated oil at 110 oF Dissolution of CO2 into crude oil reduces the viscosity of crude oil to as much as a factor of five The reduction of oil viscosity observed in the near miscible pressure range reduces the mobility ratio between CO2 and oil in the displacement process and consequently viscous fingering Phase Behavior Model A phase behavior model was constructed using the PREOS Parameters of the EOS were adjusted to match the laboratory-determined PVT data The fluid system consists of CO2 and hydrocarbon with four pseudocomponents Molecular weight of the plus fraction was adjusted to match the oil density Coefficients of Pedersen viscosity correlation were adjusted to match the oil viscosity Binary interaction coefficients between CO2 and hydrocarbon components as well as CO2 volume shift factor were adjusted to match saturation pressure and swelling data Figure presents the match of density and viscosity of oil Figure 10 shows the match of swelling factor and saturation pressure while Figure 11 shows the match of crude oil saturated with CO2 as a function of saturation pressure Core Flow Tests Cores from Arbuckle reservoirs are limited Core tests were made using Berea sandstone, Baker dolomite and Arbuckle dolomite Berea sandstone and Baker dolomite were quarried rock samples whereas Arbuckle dolomite was cored sample from Hadley well, Bemis-Shutts Field at Ellis County, Kansas Cores were epoxy encased and cast inside an aluminum cylinder with high strength epoxy The core properties are tabulated in Table The pore volumes of the cores were determined by measuring the volume of brine imbibed by the evacuated core and confirmed by tracer test with wt% MgNO3 as the tracer The cores were cleaned and reused after completion of each CO2 flood During core cleaning, the core was flushed with 10 PV of methylene chloride followed by 10 PV of methanol The sequence was repeated at least three times and finally the core was flushed with 10 PV of brine prior to be used for the flow test All flow tests were conducted with 1wt% total dissolved solids (TDS) brine consisting of 0.5 wt% MgCl2 and 0.5 wt% CaCl2 in deionized water Density and viscosity of brine measured at 110 ºF and atmospheric condition were 0.9959 g/cc, 0.7250 cp, respectively A schematic diagram of the core displacement apparatus is shown in Figure 12 The core displacement apparatus consists of a core holder, injection system, a production system and a data acquisition system The injection system consists of three pumps (for fluid transfer and injection at a desired rate) and a transfer cylinder (for crude oil storage) The production system utilizes a back pressure regulator to control the core outlet pressure at a set level During the experiment, the core effluent was flashed to atmospheric conditions The separator gas was connected to a flow meter for flow rate measurement The separator fluid was collected in glassware designed for a specific flooding The amount of fluids produced was determined gravimetrically and/or volumetrically Secondary and tertiary CO2 flooding experiments were conducted to evaluate the recovery efficiency at operating pressure in the near miscible condition Injection of CO2 was controlled at 0.1 cc/min at operating pressure The amount of fluid recovered by CO2 displacement was compared at PV of injection Secondary CO2 Flooding Berea sandstone was used in this series of experiments with the core saturated with oil prior to injection of CO2 The recovery efficiency was determined by the amount of oil recovered at PV of CO2 injections The recovery efficiency is presented in Figure 13 where the recovery efficiency from slim-tube experiment is also plotted for comparison The recovery efficiency in a short core was much less than that from slim-tube displacements The lower recovery at pressure above MMP is probably due to lack of development of multiple-contact miscibility in a short core At pressure below MMP, the extraction was also less effective as the dispersion is dominated for flow in the core plug as compared to that in a slim-tube Nevertheless, the density profiles of core flooding effluents showed similarity in density profiles of slim tube effluents at pressure below MMP Density of effluent during the displacement was higher than density of pure CO2 at near miscible pressure The density behavior of the effluent suggested that the vaporization process took place during core flooding process despite the length of core is short Tertiary CO2 Flooding Core plugs of Arbuckle dolomite, Baker dolomite and Berea sandstone, were used in this series of experiments Each core sample was saturated with brine at the test pressure and permeability was measured The core was then flooded with oil to connate water saturation at flow rate of 0.1cc/min After connate water saturation was established, the core was water flooded at same rate to residual oil saturation At least 10 PV of brine and crude oil were used in each SPE 129710 sequence of displacement to establish a steady state residual fluid saturation Carbon dioxide was finally injected to displace the remaining oil in the core The amount of oil recovered by CO2 flooding was determined volumetrically A typical result of CO2 flooding is presented in Figure 14 where the recovery history of fluid is plotted Most of recovery occurred before PV of CO2 injection No significant fluid recovery was observed after PV of CO2 were injected Figure 15 gives a comparison of recovery efficiency between secondary and tertiary CO2 flooding with Berea sandstone Higher recovery efficiency of remaining oil in place (ROIP) in tertiary CO2 flooding indicates the existence of water phase is not necessarily detrimental to CO2 displacement efficiency due to its blocking effect Instead, the relative permeability of CO2 at presence of water might be reduced Coupled with the reduction of the oil viscosity, the mobility ratio between the oil and CO2 is reduced and therefore the recovery efficiency is improved The results of tertiary CO2 flood in different cores are summarized in Table to Relatively high values of SorCO2, 0.21 to 0.29 were found in Berea sandstone as it had an unusual high Sorw, 0.48 to 0.50 prior to CO2 injections On the other hand, the Sorw of the dolomite core was found to vary from 0.32 to 0.41 with the SorCO2 from 0.07 to 0.17 at the near miscible condition Figure 16 presents the comparison of recovery efficiency among the cores tested The recovery efficiency of ROIP varied from 60% to 80% for dolomite cores while it varied from 35% to 58 % for sandstone core as pressure increased from 900 psig to 1400 psig Although the recovery efficiency differed among the rock types, substantial recovery was observed for Arbuckle rock at current reservoir operating pressure of 1150 psig The recovery efficiency was similar between two dolomite cores and was substantially higher than that in Berea core Wylie and Mohanty (1998) in their study of effect of wettability on oil recovery by gas injection concluded that the mass transfer from the bypassed region to the flowing gas inside a core is enhanced under oil-wet conditions over water-wet conditions Although the wettability of core was not determined in this study, it is generally believed that Berea sandstone is strongly water wet whereas the dolomite is less water wet After CO2 breakthrough from the core, the extraction or the mass transfer between the bypassed region and flowing CO2 becomes more important to extract the remaining oil inside the core The findings from Wylie’s study may explain why the recovery efficiency is higher in dolomite than that in sandstone tested in this study Simulation of Slim-Tube Experiments Slim-tube displacements were simulated using a 1D compositional simulator (GEM, CMG) with the tuned EOS A series of simulations were run over a range of pressures Figure 17 compares the recovery efficiencies from simulation and experiment at 1.2 HCPV of injection Figure 18 compares the density of the effluent after CO2 breakthrough calculated from the model with measured effluent densities at 1100 psig The calculated density is consistent with two phase flow with liquid dispersed in vapor phase as observed from the experiment This supports that extraction/vaporization is a primary mechanism at near miscible condition to result in a relatively high recovery efficiency of this particular oil The phase behavior model predicts the MMP and the oil recovery reasonably well and will be used in future work to simulate oil recovery from CO2 injection in an Arbuckle reservoir Field Applications Results of this study indicate that CO2 injection at current reservoir pressure in an Arbuckle reservoir could mobilize more than 50% of the waterflood residual oil eventhough the reservoir pressure is substantially less than the MMP Principal oil recovery mechanism in near miscible flooding appears to be extraction/evaporation of light hydrocarbon constituents into the carbon dioxide rich vapor phase coupled with enhanced mobility control due to the reduction of oil viscosity due to dissolution of CO2 This suggests that application of carbon dioxide in the field would require injection of and recycling of large volumes of carbon dioxide Further study is needed to determine if such a process is economically feasible However, the potential of recovering up to billion barrels of oil from Arbuckle reservoirs offers significant economic potential Conclusions Properties of Ogallah unit oil produced from an Arbuckle reservoir in Kansas were determined at reservoir temperature from a series of phase behavior and slim tube experiments where CO2 was dissolved in or used to displace the oil The MMP at 110 ºF was 1350 psig The MMP increased to 1650 psig when the temperature increased to 125 ºF At near miscible conditions (p>1100 psig), the oil viscosity was reduced by a factor of five due to the dissolution of carbon dioxide Phase behavior data were used to develop an equation of state that correlated properties of carbon dioxide saturated crude oil as a function of pressure at reservoir temperature Recovery of more than 50% of the waterflood residual oil from Berea, Baker dolomite and Arbuckle reservoir rock was obtained when CO2 was injected at the current average reservoir pressure of 1150 psig, substantially less than the SPE 129710 MMP(1350 psig) Good agreement was observed between simulated and measured oil recovery from slim-tube tests for CO2 injection over pressures ranging from 1000 psig to 1500 psig Simulated effluent densities from the slim-tube experiments were in good agreement with measured effluent densities At near miscible conditions, relatively high recovery efficiency in the slim-tube experiment supports extraction/vaporization as a principle displacement mechanism NOMENCLATURES Swr Swf Sorw Sorco2 Oil flood residual water saturation CO2 flood residual water saturation Waterflood residual oil saturation CO2 flood residual oil saturation ACKNOWLEDGEMENTS The authors wish to acknowledge the funding support by Research Partnership to Secure Energy for America (RPSEA) small producer program, RPSEA Contract DE-AC26-07NT42677/Subcontract 07123-03, Scott Ramskill of TORP for help in laboratory work, and Computer Modeling Group Inc for the software package used in the simulation RERERENCES Ahosseini, A and Scurto, A.: “Viscosity of Imidazolium-Based Ionic Liquids at Elevated Pressures: Cation and Anion Effects,” International Journal of Thermophysics, 2008 29 (4), 1222-1243 Franseen, E K., Byrnes, A P Cansler, J R Steinhauff, D M and Carr, T R.: “The Geology of Kansas ARBUCKLE GROUP,” Current Research in Earth Sciences, Bulletin 250, part 2, 2004 Grigg, R.B and Gregory, M.D., and Purkaple, J.D.: “The Effect of Pressure on Improved Oil flood Recovery from Tertiary Gas Injection,” SPERE, August 1997, 179-187 Shyeh-Yung, J-G, J: “Mechanisms of Miscible Oil Recovery: Effects of Pressure on Miscible and Near-Miscible Displacements of Oil by Carbon Dioxide,” paper SPE 22651 presented at 1991 Annual Technical Conference at Dallas, Texas, October 6-9 Tsau, J S., Bui, L H., and Willhite, G P.: “Swelling/Extraction Test of a Small Sample Size for Phase Behavior Study,” paper SPE 129728 to be presented at the Improved Oil Recovery Symposium, Tulsa, OK April 24-28, 2010 Wylie, P and Mohanty, K K.: “Effect of Wettability on Oil Recovery by Near-miscible Gas Injection,” paper SPE 39620 presented at the Improved Oil Recovery Symposium, Tulsa, OK April 19-22, 1998 6 SPE 129710 Table Physical properties of Ogallah stock tank oil Molecular Weigh, g/mol API o Density @ 14.7 psi & 60 F, g/cc o Viscosity @ 14.7 psi & 60 F, cp C36+ molecular weight, g/cc o C36+ density @ 14.7 psi & 60 F, g/cc 228.71 33.34 0.8584 13.4 873.24 0.9978 Table Core properties Type Length (cm) Cross section Area (cm ) Pore volume (cc) Porosity Permeability (mD) Berea sandstone 5.861 2.53 5.007 5.796 19.7 % 238.50 Arbuckle dolomite 5.967 2.46 4.750 6.046 21.3% 2.5 Baker dolomite 8.068 2.34 4.301 7.195 20.7% 89.7 Table Tertiary CO2 flood results of Berea sandstone Pressure Swr Sorw Sorco2 Swf Recovery 1-(Sorco2/Sorw) (psig) 905 1104 1198 1317 1413 0.318 0.318 0.318 0.318 0.318 0.483 0.500 0.483 0.500 0.483 0.311 0.293 0.259 0.207 0.207 0.370 0.388 0.405 0.336 0.336 35.71 41.38 46.43 58.62 57.14 Table Tertiary CO2 flood results of Arbuckle dolomite Pressure Swr Sorw Sorco2 Swf Recovery 1-(Sorco2/Sorw) (psig) 901 1100 1200 1305 1407 0.380 0.380 0.446 0.446 0.380 0.414 0.414 0.331 0.331 0.380 0.165 0.165 0.083 0.066 0.099 0.512 0.553 0.636 0.636 0.529 60.00 60.00 75.00 80.00 73.91 Table Tertiary CO2 flood results of Baker dolomite Pressure Swr Sorw Sorco2 Swf Recovery 1-(Sorco2/Sorw) (psig) 905 1109 1201 1303 1402 0.284 0.312 0.340 0.368 0.368 0.389 0.375 0.347 0.347 0.320 0.153 0.125 0.097 0.069 0.069 0.437 0.409 0.451 0.534 0.465 60.71 66.67 72.00 80.00 78.26 SPE 129710 Figure Ogallah unit, Trego County, Kansas Figure Minimum miscibility pressure (MMP) Figure Carbon number distributions of Ogallah crude oil Figure Densities of slim-tube effluents during CO2 injection at pressures below MMP and a reservoir temperature of 110 °F DPT 1.0 1.4 PT 1.2 CO2 BPR Gas Swelling Factor Test Oil Densitometer 0.6 0.8 0.6 0.4 0.4 Oil 0.2 Swelling Factor CO2 solubility 0.2 0.0 ISCO Pump Electronic Balance Figure Slim-tube experimental setup ISCO Pump 500 1000 1500 2000 0.0 2500 Pressure, psi Figure Swelling/extraction curve of Ogallah crude oil with carbon dioxide at 110 °F CO solubility 0.8 1.0 SPE 129710 PT CO2 Inlet RD Liquid Outlet Liquid Inlet High pressure sensor RTD Circulation Pump Oven Liquid Outlet High Pressure Generator Figure 10 Match of saturation pressure and swelling factor at 110 °F Figure 7: Experimental setup for high pressure viscosity measurement 5.0 Run Run Run Viscosity (cp) 4.0 3.0 2.0 1.0 0.0 500 1000 1500 2000 Pressure (psi) Figure 11 Match of viscosity of oil saturated with carbon dioxide at 110 °F Figure Viscosity of crude oil saturated with carbon dioxide at 110 °F DPT Brine Quizix Pump CO2 ISCO Pump BPR Gas Test Oil Oil/Brine ISCO Pump Figure Match of density and viscosity of oil at 110 °F Figure 12 Coreflood experimental setup SPE 129710 Figure 13 Comparison of recovery efficiency between slim-tube and coreflood experiment at 110 °F Figure 16 Effect of rock type on recovery efficiency at 110 °F Figure 14 Effluent profile of production fluid during CO2 flooding at 1317 psig and 110 °F Figure 17 Match of oil recovery efficiency and MMP in slim-tube test at 110 °F Figure 15 Effect of water saturation on recovery efficiency at 110 °F Figure 18: Comparision of simulated and measured effluent density for slim-tube experiment at 1100 psig ... improve understanding of the mechanisms affecting the process of near-miscible CO2 flooding and evaluate the feasibility of near miscible carbon dioxide injection in Arbuckle reservoirs All experiments... was maintained by the back-pressure regulator at the outlet CO2 was injected at a constant rate of 0.1cc/min to displace the oil Density of the effluent was measured continuously using an inline... operating pressure in the near miscible condition Injection of CO2 was controlled at 0.1 cc/min at operating pressure The amount of fluid recovered by CO2 displacement was compared at PV of injection