115 e 20090102
-
Upload
abdou-chawi -
Category
Documents
-
view
216 -
download
0
description
Transcript of 115 e 20090102
Reservoir Characterization Study of an Iranian Carbonate Reservoir by SCAL Application Data
Hafizolah Kashani Birgani, Arvandan Oil&Gas Co.,
Abstract This paper contains the SCAL of carbonate reservoir field. It includes the main available data,
plots of relative permeability, capillary pressure, water saturation etc. The plots, analysis of Jfunction versus normalized water saturation and the relations applied for these analyses constitute another section. The main objectives of this paper are: Better understanding of the behavior and characteristics of the reservoir by integrating results using these results to characterize the carbonate reservoir properties. The main reservoir characteristics will include: Capillary pressure vs. water and gas saturation, Oilwater relative permeability vs. water saturation, Oilgas relative permeability vs. gas saturation The SCAL module in ECLIPSE is a tool to help engineers effectively use laboratory derived relative permeability and capillary pressure measurements in reservoir simulation. The program has facilities to: Import laboratory data, Perform quality control (such as curve smoothing), Group data according to litho logical parameters and endpoint values, Transform the laboratory data into rock curves suitable for input to simulators Such as ECLIPSE and automatically assign these curves to grid cells (according to a set of user defined rules, for example as a function of porosity, permeability or litho logical parameters). Key Words: SCAL relative permeability capillary pressure J, function.
Archive of SID
www.SID.ir
IntroductionSpecial core analysis (SCAL) is important to obtain more detailed information on reservoir parameters not available through
standard core analysis and usually includes capillary pressure, relative permeability and wettability measurements. SCAL is also some times needed to obtain more reliable input parameters for reservoir simulations and this paper concentrates on the latter for a mechanistic study of recovery mechanisms in carbonate reservoirs. Hydrocarbon recovery results from a competition between capillary and viscous forces and gravity. In most chalk reservoirs spontaneous imbibition is the major recovery mechanism. This dominance of capillary forces is due to narrow pore throats, more or less waterwet conditions and the low permeability of this rock. Essentially all reservoirs are affected by the interplay between capillary pressure and relative permeability at various wettability conditions but fractured chalk reservoirs with very low matrix permeability are in particular sensitive to these interactions. Questions arise about how and when fractures are crossed by the wetting fluid and if there is a component to oil recovery from viscous forces, and if so, when and how this occurs. The impacts of these parameters at the different wettability conditions seem to be the clue to the understanding of the oil recovery mechanisms in these reservoirs. (Ref.1, 2, 3, 4).
Experimental 1. Capillary Pressure & Pore Size Distribution by Mercury Injection
In determination of capillary pressure and pore size distribution by mercury injection, sample is cleaned and dried. It is evacuated inside the mercury pump and then mercury is injected at a series of increasing incremental pressures. There are two apparatus for mercury injection process, the mercury pump and Auto pore III. In mercury pump apparatus, the mercury is injected into rock sample at maximum 1500 psi and mercury penetrates pore down to 0.01 microns. Auto pore III used modern technology through automation and optional highpressure injection up to 60000 psi. This highpressure injection penetrates pore down to 0.003 microns diameter. As mercury is nonwetting, this is a drainage process and reverse process is imbibition. Pressure is then plotted as a function of Hg saturation to generate a capillary pressure curve. This can then be converted to an equivalent airbrine capillary pressure curve using standard conversion factors derived from constants for interfacial tension and contact angle. (Ref. 2, 5, 8, 13)
2. Relative Permeability Measurements, Unsteady state Method The measurements of absolute and relative permeability for oil and water are one of the most important tasks in core
laboratories. Generally, the tested sample plug is saturated initially with a wetting phase using vacuum pump and the absolute permeability for the Wetting phase is measured. Then the relative permeability measurements are conducted under twophase flow, steady or unsteady method. Based on the data collected in the two Measurements the absolute and relative permeabilities are calculated. The unsteady state method is also called Wedge’s method because the calculation is based on the theory of the improved Buckley and Leveret’s mechanism of fluid displacement in porous media. Our experiments will be carried out on Berea sandstone plugs with oil and water under unsteady state method, a constant pressure driving method. (Ref. 2, 3, 9, 10)
Results and Discussion 1. Relative Permeability Curves 11.OilWater
Table 1 shows the availability of sample data for this reservoir. Figure 1 demonstrates the oilwater relative permeability curves for samples 3, 4 and 5. These curves are normalized to remove the effect of different initial water saturation Swi) and residual oil saturation (Sor) and then denormalized (Figure 2). From the (normalized curves it is determined that samples 4 and 5 can be treated as one rock type, so an average was taken as a representative of oilwater permeability. The averaged curve shown in Figure 2 is an average of samples 4 and 5. To demonstrate the effect of including the sample from another rock type, in Figure 3 samples 3, 4 and 5 are all averaged, which shows the curve differs considerably from that of Figure 2.
12.OilGas The only available oilgas relative permeability curves for samples 2, 4 and 5 were plotted (Figure 4). All the gasoil relative
permeability data have relatively good quality (Figure 4). By looking at the normalized curves it is specified that an average can be taken as a representative for oilgas relative permeability (Figure 5). Due to the various natures and characteristics of the relative permeability in the two fluid systems (oilwater and gasoil) the average relative permeability in two systems differs greatly. Figure 6 shows the averaged krog curve for samples 2, 4 and 5. Figure 7 demonstrates a comparison of averaged oilgas with oilwater relative permeability curves.
2. Capillary Pressure Curves There are two methods for calculating the capillary pressure. Due to the lack of the capillary pressure data from other
methods, we rely on the mercury injection method and the data resulting from this method were used. In this system mercury is
Archive of SID
www.SID.ir
considered the seen that all samples nonwetting phase. By generating capillary pressure curves for each sample, it can be except for sample 1 have a similar trend. Figure 8 shows the capillary pressure curves for the samples.
21.Jfunction Calculation and Curves As capillary pressure data are obtained on small core samples, which represent an extremely small part of the reservoir, it is
necessary to combine all the capillary data to classify a particular reservoir. So Jfunctions were generated for each capillary pressure. The J function curves were constructed using Microsoft Excel and then transferred into SCAL. Furthermore, in order to convert the Pc to a Jfunction some data including laboratory permeability and porosity and also mercury contact angle and interfacial tension were used as input. As for interface tension and contact angle values, regarding the variations in oilgas interfacial tension in different pressures, these parameters were extracted from the PVTi program accompanied by Parachor parameters. The interfacial tension estimation around the reservoir pressure is about 10 dyne/cm. The contact angle is also estimated to be 60 degrees. Figure 9 shows Jfunction plots for samples 3, 4 and 5. (Ref.12, 13)
3. Rock Type Determination To determine different rock types within the reservoir, primarily Jfunction curves were used. Capillary pressure and relative
permeability curves were also associated in this assessment. The other parameter that affects rock type determination is the depth of the samples, which was taken into account. From Figure 9 it can be seen that the slope of the curves of samples 3 and 5 are closer to one another than that of sample 4, so samples3 and 5 might lie within one single type. On the other hand oilwater relative permeability curves (Figure 1 and Figure 2) suggest that samples 4 and 5 follow the same trend and must be considered in one single rock type. Therefore, regarding these curves and also with regard to the depths of samples 3, 4 and 5 which are all within 1 m different from 2900 m, we categorize samples 3, 4 and 5 to be in one rock type. Since the trend of sample 3 in the krw curve is more different than the other two ones, we only average between samples 4 and 5 for creating the krw data output file. The result is given in Figure 10. However, in the Jfunction curves, since the curves for samples 3, 4 and 5 are all pretty much close to one another, we average all three samples for generating the data file. Figure 11 demonstrates this averaging. On the capillary pressure curve since the threshold and shape of the curves of samples 3 and 5 are really similar to each other, we only average between samples 3 and 5 curves (Figure 12). For comparison and better understanding, the averaged krg curves for samples 4 and 5 are shown in Figure 13. On the other hand capillary pressure curves (Figure 8) indicates that samples 1 and 2 are totally different and must belong to another rock types. The depths of these two samples are very close (2729 and 2730 m) and confirm the fact that they may belong to one single rock type. The normalized capillary pressure curves (Jfunction) of these two samples have the same trend. So we categorize samples 1 and 2 in one rock type. But since there aren’t sufficient data at the time of preparing this report, we cannot decide on the formation of each type at this point. Figure 14 shows SCAL main window with the two rock types specified in it.
4. Simulation Rock Input Data 41.Single Porosity and Matrix Rock Properties From the ECLIPSE SCAL project described above and a SCAL report on the Sarvak Formation average curves were created for: Krw (Sw), Kro(Sw), Pcow(Sw), Krog(Sg) ,Krg(Sg), Pcgo(So+Sw). The given oilwater capillary pressure data contained multiple pressure entries for certain saturation values. These multiples had been removed. The grayshaded cells in Table 2 mark the entries that have been removed. Table 2: Removal of multiple pressure entries for the Pcow curve because the endpoints in the relative permeability curves were not consistent with the capillary pressure curve, a scaling of the relative permeability curves became necessary. The contradicting endpoints were Swc = 0.1105 for the capillary pressure Pcow and Swc=0.1633 for the relative permeability. Therefore the relative permeability curves have been scaled to the Swc entry of the Pc data. Figure 15 shows a plot with a comparison of the oilwater relative permeability curves before and after scaling. The yellow curves represent the original data and the green the new, scaled ones. Furthermore, the gasoil relative permeability data and the capillary pressure data have been scaled to match the Swc provided by the pcowd data. A comparative plot for pcogd before and after scaling can be found in Figure 16 and for the gasoil relative permeabilities in Figure 17. Again the yellow curves represent the original data and the green the new, scaled ones. The capillary pressure data provided in the SCAL report is valid at laboratory conditions where a mercury/air system was investigated. This made a conversion to reservoir conditions namely a water/oil and a gas/oil system necessary. Such a conversion can be done by using the following equation: PcR= σR cos θ/ σL cos θL PcL Equation 1, Where PcR is the capillary pressure under reservoir conditions, PcL is the capillary pressure measured under laboratory conditions, σR is the interfacial tension under reservoir conditions, σL is the interfacial tension under laboratory conditions, θR is the contact angle measured under reservoir conditions θL is the contact angle measured under laboratory conditions.
Archive of SID
www.SID.ir
Due to different interfacial tensions of wateroil and gasoil systems two different conversion factors had to be applied. For the wateroil system typical values were selected. These are σR =σw/o= 0.028 N/m and σL=σHg/a=0.48 N/m. The contact angles were neglected for the above equation. This decision was taken based on missing input data. Thus a conversion factor of 0.05833 was calculated for the wateroil system. For gasoil systems no typical values for interfacial tensions could be found in literature. Therefore the interfacial tension σgo was calculated by theWeinaug and Katz Equation, σg/o¼=∑ Pi (xiρo /Moyi ρg /Mg) Equation 2, Where σg/o is the interfacial tension between gas and oil phase [Dynes/cm], N is the number of components [ ], Pi is the parachor of component i [Dynes1/4cm11/4/moles ], Xi is the liquid mole fraction of component i [ ], Yi is the vapor mole fraction of component i [ ], ρo is the density of the oil phase [g/cm³], ρg is the density of the gas phase [g/cm³], Mo is the molecular weight of the oil phase [kg/kmol], Mg is the molecular weight of the gas phase [kg/kmol]. Based on the data of the PVT report listed in Table 4 and the oil and gas properties listed in Table 3 above Equation was evaluated to σg/o =21.5186 [Dynes/cm] =0.0215186 [N/m]. For σL=σHg/a=0.48 N/m was applied. Again the contact angles were neglected for the capillary pressure conversion equation (Eq. 1). This decision was taken based on missing input data. Thus a conversion factor of 0.04375 was calculated for the gasoil system. Table 3: Oil and gas properties used for σgo evaluation to take the oilwet characteristics of the reservoir into account it is necessary to use for initialization and production of the reservoir two different types of capillary pressure curves. These are a drainage curve for initialization and an imbibition curve for producing the reservoir. Since no imbibition curve was neither provided in the SCAL report nor in the ECLIPSE® SCAL project, an artificial imbibition curve had to be constructed. For the construction process Swc = 0.1105 and Sor = 0.4981 were taken as interval limits on the Sw axis and on the Pc axis the maximum wateroil capillary Pressure was taken as interval limits, once positive at Swc and negative at Sor for an oil wet system. For the water wet system Pc equals zero at Sor. Between these endpoints the imbibition capillary pressure curve was constructed. The drainage and imbibition water oil capillary pressure curves are displayed in Figure 18 and Table 5 contains the data of the oil wet and water wet wateroil imbibition curves. For oilgas displacement, no distinction between drainage and imbibition processes has been made. The used capillary pressure curve, valid for drainage and imbibition, is shown in Figure 19. The rock data received does not contain information about the rock compressibility factor. Therefore, a typical value of 4.3000E05 1/bar was chosen as input for the simulation models. (Ref.12, 13, 14)
42.Fracture Rock Properties For the fracture properties, constant parameters were used. For fracture permeability a value of 10 mD was assumed. For the
shape factor a constant value of 0.1 m2 was assigned. For gravity drainage calculation purposes a fracture block height of 10 m was selected for the single well model simulation runs group. This group consists of the column models, the radial well models and the horizontal well models. For the crosssection models a matrix block height value of 100 m was selected. For fracture porosity a constant value of 0.5% was assumed. It is common practice to use a capillary pressure equal to zero for the fractures and a linear relationship between saturation and relative permeability. Therefore, these settings were also selected for the fracture rock properties of the Sarvak formation. In Figure 22 and Figure 23, the relative permeability curves for the fractures are displayed. To calculate the relative permeability curves for the two cases, Corey’s equation is used.
OilWater Data: The corresponding values for the parameters in Corey’s equation for both wettability cases are given in Table 7. Figure 14 to
Figure 17 show the corresponding curves of the relative permeability data for the two cases used in the simulation. Based on the endpoints defined for Corey’s equation artificial capillary pressure curves had to be constructed. Capillary pressure curves are not solely functions of the saturation, but depend on the direction of the saturation change, too. To distinguish between drainage and imbibition is called hysteresis in the capillary pressure data and is common practice in reservoir simulation. These are a drainage curve for initialization and an imbibition curve for producing the reservoir. The capillary pressure curves for the oilwet case are shown in Figure 12. To construct the oilwet imbibition curve Siw = 0.2 and Sor,w = 0.4 were taken as interval limits on the Sw axis and on the Pc axis the maximum wateroil capillary pressure were taken as interval limits, once positive at Siw and negative at Sor,w for an oil wet system. For the water wet system Pc equals zero at Sor, w. between these endpoints the imbibition capillary pressure curve was constructed. For oilgas displacement, no distinction between drainage and imbibition processes has been made. The used artificial capillary pressure curve, valid for drainage and imbibition. Therefore, a typical value of 4.3000E 05 1/bar was chosen as input for the simulation models.
Conclusions 1. Technique is reservoir specific and yields detailed categorization of changing rock quality.
Archive of SID
www.SID.ir
2. Allows detailed assessment of reservoir quality before the well is completed. 3. Jfunction is useful for averaging capillary pressure data from a given rock type from a given reservoir. 4. Jfunction can sometimes be extended to different reservoirs having same lithology 5. Jfunction usually not accurate correlation for different lithologies. 6. Gas relative permeability measurements were carried out using a ‘depletion’ technique on the same core at low flow rates. The results showed no rate dependency, but there was evidence for an increase in relative permeability due to low IFT. 7. Leveret’s Jfunction vs. water saturation plots reveal big scatter, and no definite trend could be established.
References 1. Anderson, W.G.: “Wettability Literature Survey – Part 2: Wettability Measurement”, JPT (Nov. 1986) 12461262. 2. Anderson, W.G.: “Wettability Literature Survey – Part 4: Effects of Wettability on Capillary Pressure”, JPT (Oct. 1987) 1283 1300. 3. Anderson, W.G.: “Wettability Literature Survey – Part 5: Effects of Wettability on Relative Permeability”, JPT (Nov. 1987) 14531468. 4. Honarpour, M., Koederitz, L. And Harvey, A.H.:”Relative Permeabilities of Petroleum Reservoirs”, CRC Press, Boca Raton, FL (1986) 45. 5. Negative Capillary Pressure Curves for Reservoir Rock using the Centrifuge”, 4th International Reservoir Characterization Technical Conference Proceedings (1997). 6. Anderson, D.M., McFadden, G.B., and Wheeler, A.A.: “Diffuseinterface methods in fluid mechanics,” Ann. Rev. Fluid Mech. (1998) 30, 139–165. 7. Seppecher, P.: “Moving contact lines in the CahnHilliard theory,” Int. J. Engng. Sci. (1996) 34, 977–992. 8. Li, K. and Horne, R.N.: “An Experimental and Analytical Study of Steam/Water Capillary Pressure,” SPEREE (Dec. 2001) 477–482. 9. Li, K. and Horne, R.N.: “SteamWater Relative Permeability by the Capillary Pressure Method,” presented at the 2001 International Symposium of the Society of Core Analysts, Edinburgh, Sep. 19–21. 10. Kjosavik, A., Ringen, J.K., and Skjaeveland, S.M.: “Relative Permeability Correlation for MixedWet Reservoirs,” SPEJ (March 2002) 49–58. 11. Dullien, F.A.L.: Porous media: fluid transport and pore structure, Academic Press, san Diego (1992). 12. Firoozabadi, A.: Thermodynamics of hydrocarbon reservoirs, McGrawHill, New York (1999). 13. Skjaeveland, S.M., Siqveland, L.M., Hammervold, W.L., and Virnovsky, G.A.: “Capillary Pressure Correlation for Mixed Wet Reservoirs,” SPEREE (Feb. 2000) 60–67. 14. Barrett, J.W., Blowey, J.F., and Garcke, H.: “Finite element approximation of the Cahn Hilliard equation with degenerate mobility,” SIAM J. Numer. Anal. (1999) 37, 286–318.
Table 1: Availability of sample data
SAMPLES 1 2 3 4 5 40H 41H 48H
Depth (m) 2729 2730 2899.8 2900.3 2900.7
Pc Data Yes Yes Yes Yes Yes
Kr Data N/A Krg Krw Krw &Krg
Krw &Krg
Krw &Krg Krw Krw
Porosity 0.0641 0.1107 0.2029 0.1415 0.1034 0.095 0.069 0.1689
KO(Swi) N/A N/A 0.139 0.768 0.176 0.1797 0.1258 0.427
Swirr N/A N/A 0.22 0.245 0.12 0.2786 0.2861 0.1415
Swcrit 0.4103 0.0656 0.1105 0.0229 0.0833 0.2786 0.2881 0.1415
Kair (mD) N/A 0.332 4.729 3.091 2.089 5.423 2.478 2.045
Archive of SID
www.SID.ir
Table 2: Removal of multiple pressure entries for the Pcow curve
Sw Pc [bar]
0.110500 20.28 0.113400 18.93 0.116300 17.57 0.122100 16.22 0.130800 14.87 0.133700 13.52 0.133700 12.17 0.145300 10.82 0.168600 9.46 0.200600 8.11 0.255800 6.76 0.360500 5.41 0.494200 4.06 0.668600 2.70 0.930200 1.35 0.988400 0.68 0.994200 0.18 0.997100 0.11 0.997100 0.08 1.00000 0.05 1.00000 0.02 1.00000 0.00
Table 3: Oil and gas properties used for go evaluation
Insitu oil density 0.809 g/cm³ Insitu gas density 0.372 g/cm³ Oil molecular weight 168kg/kmol Gas molecular weight 30.86kg/kmol
Table 4: Evaluation of g/ o
Components Parachors xi [%] xi yi [%] yi IFT
H2S 80 0.00 0.00 0.41 0.00 0.0039
N2 41 0.00 0.00 0.62 0.01 0.0031
CO2 48 0.00 0.00 6.54 0.07 0.0615
C1 77 0.00 0.00 50.54 0.51 0.4691
C2 108 0.00 0.00 16.73 0.17 0.2178
C3 150.3 0.92 0.01 11.49 0.11 0.2014 IC4 181.5 0.39 0.00 1.93 0.02 0.6389
NC4 189.9 1.83 0.02 5.46 0.05 0.1082
IC5 225 1.64 0.02 1.66 0.02 0.0271
NC5 231.5 2.28 0.02 1.87 0.02 0.0268
C6 271 6.33 0.06 1.75 0.02 0.0255 C7 312.5 4.81 0.05 0.78 0.01 0.0428 C8 351.5 5.34 0.05 0.21 0.00 0.0814 C9 380 5.06 0.05 0.02 0.00 0.0919
Archive of SID
www.SID.ir
C10 404.9 4.92 0.05 0.00 0.00 0.0959 C11 429.3 4.06 0.04 0.00 0.00 0.0839 C12+ 961.53 62.42 0.62 0.00 0.00 2.8902
21.5186
Table 5: Conversion of laboratory capillary pressure data to reservoir conditions
SW PCOWD_LAB(bar) PCOWD_Res(bar) Sf Pcgfd_Lab (bar) Pcgfd_Res (bar)
0.110500 20.280001 1.183000 0.110500 20.280001 0.867250 0.113400 18.930000 1.104250 0.118306 17.570000 0.768688 0.116300 17.570000 1.024917 0.145627 16.219999 0.709625 0.122100 16.219999 0.946167 0.172853 14.870000 0.650563 0.130800 14.870000 0.867417 0.211882 13.520000 0.591500 0.133700 13.520000 0.788667 0.243106 12.170000 0.532438 0.145300 10.820000 0.631167 0.289942 10.820000 0.473375 0.168600 9.460000 0.551833 0.360196 9.490000 0.415188 0.200600 8.110000 0.473083 0.453772 8.110000 0.354813 0.255800 6.760000 0.394333 0.617698 6.760000 0.295750 0.360500 5.410000 0.315583 0.820558 5.410000 0.236688 0.494200 4.060000 0.236833 0.937552 4.060000 0.177625 0.668600 2.700000 0.157500 0.976582 2.700000 0.118125 0.930200 1.350000 0.078750 1.000000 0.000000 0.000000 0.988400 0.680000 0.039667 1.000000 0.000000 0.000000
Table 6: Water and oil wet capillary pressures for imbibition
Sw PCOWi (Water Wet) (bar)
PCOWi (Oil Wet) (bar)
0.11050 1.18300 1.18300
0.11300 1.00000 0.70000
0.13000 0.70000 0.40000
0.15000 0.50000 0.30000
0.17000 0.45000 0.20000
0.19000 0.40000 0.10000
0.22000 0.33000 0.00000
0.25000 0.28000 0.10000 0.28000 0.22000 0.20000 0.32000 0.18000 0.30000 0.35000 0.12000 0.40000 0.39000 0.09000 0.50000 0.42000 0.07500 0.60000 0.45000 0.05000 0.70000 0.48000 0.02500 0.80000 0.49810 0.00000 1.18300
Archive of SID
www.SID.ir
Table 7: Used values for Corey’s equation parameters
waterwet case oilwet case
OilWater
Sor,w 0.3 0.4 Siw 0.3 0.2 Kro,iw 1 1 Krw,i 1.07 1.6 No 2 3 nw 3 2
GasWater
Sor,g 0.2 0.2 Sgc 0.1 0.1 Kro,ig 1 1 Krg,i 1.53 0.98 No 2 2 ng 2 2
Figure 1: Krwo curves
Figure 2: Denormalized krw curves (samples 4 and 5 averaged)
Archive of SID
www.SID.ir
Figure 3: Denormalized krw curves (samples 3, 4 and 5 averaged)
Figure 4: krog curves
Figure 5: Normalized krog curves
Archive of SID
www.SID.ir
Figure 6: Averaged krog curve
Figure 7: Mercury capillary pressure for samples 1, 2, 3, 4 and 5
Figure 8: Jfunction curves for sample 3, 4 and 5.
Archive of SID
www.SID.ir
Figure 9: Samples 4 and 5 averaged for creating the krw data output file
Figure 10: Samples 3, 4 and 5 averaged for creating the output file
Figure 11: Samples 3 and 5 averaged for creating the output Pc file
Archive of SID
www.SID.ir
Figure 12: Samples 4 and 5 averaged for creating krg output file
Figure 13: shows SCAL main window with the two rock types specified in it.
Figure 14: Wateroil relative permeability curves used for Sarvak simulation Models
Archive of SID
www.SID.ir
Figure 15: Oilgas relative permeability curves used for Sarvak simulation Model
Figure 16: Fracture wateroil relative permeabilities
Figure 17: Fracture oilgas relative permeabilities
Archive of SID
www.SID.ir