Multiphase flow of CO2 and water in reservoir rocks at ... · reservoir rocks at reservoir conditions ... representing a range of reservoir properties and rock types. Table 1 summarizes

Multiphase flow of CO2 and water in reservoir rocks at reservoir conditions

Ronny Pini, Sam Krevor, Lin Zuo, Sally Benson Department of Energy Resources Engineering

Stanford University

2

Multiphase flow properties

Ronny Pini – [email protected] - Stanford University

Core flooding experiments

sample collection

relative permeability residual trapping

capillary pressure

simulations

X - 16 BALES ET AL.: MULTIPHASE FLOW OF CO2 AND WATER

Berea

Mt. Simon

Water Saturation

Rel

ativ

e pe

rmea

bilit

y

Paaratte

Tuscaloosa1

0.4

Figure 11. Drainage CO2 and water relative permeabil-

ity curves for the four samples. Dashed lines are Brooks-

Corey or Purcell-Burdine relative permeability curves de-

rived from the mercury porosimetry measurements. Solid

lines are best fit Brooks-Corey curves. For the Berea, the

best fit curve and mercury porosimetry curve were nearly

the same and only the best-fit curve is shown.

Rel

ativ

e pe

rmea

bilit

y

Water saturation

CO2

water

0 . 4 0 . 6 0 . 8 1

24 Samuel C.M. Krevor et al.

Berea Paaratte

Mt. Simon Tuscaloosa

00.8

00.8

S CO

2 ,r [-]

SCO2 ,i [-]

Fig. 13 Plots of maximal CO2 saturation vs. the residual CO2 saturation achieved with100% water flooding. Black dots show measurements, solid lines show the best-fit Spiterihysteresis model and dashed lines show the Land hysteresis model using the average Landcoefficient. All parameters are provided in table 5.

carbonate, shale, and anhydrite rocks. SPE Reservoir Evaluation and Engineering,June:487–496, 2008.

3. D. Brant Bennion and Stefan Bachu. Drainage and imbibition co2/brine relative perme-ability curves at reservoir conditions for carbonate formations. In SPE Annual TechnicalConference and Exhibition, number SPE 134028, 2010.

4. McMillan Burton, Navanit Kumar, and Steven L. Bryant. Co2 injectivity into brineaquifers: Why relative permeability matters as much as absolute permeability. EnergyProcedia, 1(1):3091–3098, 2009.

5. P. Chiquet, D. Broseta, and S. Thibeau. Wettability alteration of caprock minerals bycarbon dioxide. Geofluids, 7:112–122, 2007.

6. Y. Cinar, A. Riaz, and H.A. Tchelepi. Experimental study of co2 injection into salineformations. In SPE Annual Technical Conference and Exhibition, number 110628-MS,2007.

CO

2 max

. sat

urat

ion

CO2 initial saturation 0 0 . 2 0 . 4 0 . 6 0 . 8

Land

Spiteri

0 0.2 0.4 0.6 0.8 10

5

10

15

20

25

30

Water saturation

Cap

illary

Pre

ssur

e [k

Pa]

298K

323K

CO2 - water

CO2

!!!

3

Core-flooding experiments • Replicate reservoir conditions

• Ppore : 9 MPa

• Pconf: 11.8 MPa

• T : 50C

• Continuous circulation

• Immiscible displacement

• Experimental variables:

• Flow rates

• Pressure drop

• Saturation (CT scanner) !"#

Perrin, J.C., Benson, S.M.: An Experimental Study on the Influence of Sub-Core Scale

Heterogeneities on CO2 Distribution in Reservoir Rocks. Transport in Porous Media, DOI

10.1007/s11242-009-9426-x, 82:93–109 (2010)

Riaz, A., Hesse, M.A., Tchelepi, H.A., Orr, F. M.: Onset of Convection in a Gravitationally

Unstable Diffusive Boundary Layer in Porous Media. Journal of Fluid Mechanics, 548: 87-

111 (2006)

Riaz, A., Tchelepi, H.A.: Unstable Dynamics of Vertical Displacement in Porous Media

Associated With CO2 Sequestration. SPE 103169 (2006)

Saadatpoor, E., Bryant, S.L., Sepehrnoori, K.: Effect of Heterogeneous Capillary Pressure on

Buoyancy-Driven CO2 Migration. SPE 113984-MS (2008)

Silin, D., Patzek, T. W., Benson, S. M.: A one-dimensional model of vertical gas plume migration

through a heterogeneous porous medium. International Journal of Greenhouse Gas Control

(2008). ISSN 1750-5836, DOI: 10.1016/j.ijggc.2008.09.003

Smith, G.E.: Fluid Flow and Sand Production in Heavy Oil Reservoirs Under Solution Gas Drive.

SPE Production Engineering, 3(2):169-180 (1988)

Tang, G.Q., Firoozabadi, A.: Gas and Liquid-Phase Relative Permeabilities for Cold Production

from Heavy Oil Reservoir. SPE 56540 (2003)

Tang, G.Q., et al.: Heavy-oil Solution Gas Drive in Consolidated and Unconsolidated Rock. SPE

87226 (2006)

Wilt, P.M.: Nucleation Rates and Bubble Stability in Water Carbon Dioyde Solutions. J. Coll. Int.

Sci., 112, 530-538 (1989)

Zhang, Y.Q., Oldenburg, C.M.,Benson, S.M.: Vadose Zone Remediation of Carbon Dioxide

Leakage from Geologic Carbon Dioxide Sequestration Sites. Vadose Zone Journal, 3:858-866

(2004)

Figure Legend

two-phase separator

R

RC

O2

Cyl

ind

er

CO2 water

back pressure pump

core holder confining pressure pump

heater

R

pressure transducer

manual valve

electric valve

relief valve

check valve

two-phase separator

RR

RRC

O2

Cyl

ind

er

CO2 water

back pressure pump

core holder confining pressure pump

heater

RR

pressure transducer

manual valve

electric valve

relief valve

check valve

X-ray CT scanner Core-holder

8

!"#$%&'()*+,-#./,0)1.2)$.23

4+5#%6)$+,%7,*%$" 6.78%$%.7*& '2+**"2+& 9+:5+2)$"2+& ;2%7+,6.:5.*%$%.7

CO2/Water ISCO pumps

Perrin J-C. and Benson S., Trans Porous Media. 2010, 82, 93-109


4



Core flooding experiments relative permeability

residual trapping capillary pressure


Berea

Mt. Simon

Water Saturation

Rel

ativ

e pe

rmea

bilit

y

Paaratte

Tuscaloosa1

0.4








Rel

ativ

e pe

rmea

bilit

y

Water saturation

CO2

water

0 . 4 0 . 6 0 . 8 1


Berea Paaratte


00.8

00.8

S CO

2 ,r [-]

SCO2 ,i [-]







CO

2 max

. sat

urat

ion


Land

Spiteri

0 0.2 0.4 0.6 0.8 10

5

10

15

20

25

30

Water saturation

Cap

illary

Pre

ssur

e [k

Pa]

298K

323K

CO2 - water

CO2

5

Relative permeability


u

P1

P2

BP PCO2

Pw

ui = !kkri (Si )µi

dPidz

CO2 / water injection

z = 0 z = L

steady state

PC

fi

Si = constant ! dPc

dz= 0

ui = !kkri (Si )µi

"PL

Steady state method

rock sample

6

Rock samples

Porosity

0.18 0.24 0.30


•  Sandstones •  Berea: “model” rock

•  Others: target CO2 storage reservoirs


flow properties of the rocks. In addition, flow experiments

were performed using the steady state relative permeability

method with x-ray CT imaging. With this method, the re-

sults are independent of the assumptions necessary to justify

the use of Buckley-Leverett theory. The absence of common

problems in core-flooding experiments such as capillary end

effects and gravity fluid segregation can also be confirmed

through x-ray imaging. The main goal of this work is to de-

rive important petrophysical flow parameters for a range of

sandstone rock types at reservoir conditions including suffi-

cient observation and analysis to address major uncertain-

ties common in these types of characterizations.

2. Materials2.1. Rocks

Four sandstone rock cores, 5.08cm in diameter, were used

representing a range of reservoir properties and rock types.

Table 1 summarizes various properties of the rocks. Fig-

ure 1 shows the variation in porosity of each rock along the

length of the cores. A Berea was used because of its utility in

making comparisons with other studies. Three other rocks,

one each from the Paaratte formation in Southern Australia,

the Mt. Simon in Illinois, and the Tuscaloosa massive sand

from Alabama are from target reservoirs where large volume

CO2 injection pilot projects are either underway or under

development.

Berea sandstone: A high permeability homogenous

Berea sandstone was obtained from the Berea sandstone core

company. A thin section micrograph is shown in figure 2.

The rock is typical of Berea sandstones; well sorted with sub-

mature granular features and around 10% clay. The rock is

homogeneous with no apparent bedding planes and minimal

porosity variation along the length of the core. The rock was

fired at 700◦C to stabilize swelling clays which are known to

be present in Berea sandstones. Firing was not performed

on the other cores so as to best preserve their original flow

properties.

Paaratte sandstone: A high permeability core from

the Paaratte sandstone formation was obtained from the

CO2CRC’s Otway pilot sequestration project in Southern

Australia [Sharma and Cook , 2007]. The core was taken

from 1400m depth where the reservoir is at 13.8 MPa and

55◦C. A thin section micrograph is shown in figure 3. The

sandstone is well sorted but otherwise texturally subma-

ture. There is no apparent cementing material and little

clay. There are several low-porosity bedding planes perpen-

dicular to the direction of flow (Fig. 1). Outside of these

zones, porosity ranges between .28-.29. The rock is simi-

lar in morphology and component composition to the Berea

sandstone.

Mt. Simon: A low permeability sandstone from the

Mt. Simon sandstone formation was obtained from a well

in Macon County, Illinois. The Mt. Simon underlies one

of the largest concentrations of coal-fied power stations in

the world and is the target formation of actively develop-

ing CO2 storage projects [Leetaru et al., 2009]. The core

was taken from 1650m depth. A thin section micrograph

Table 1. Rock properties and experimental conditions

Name Porosity Absolute Length[-] Permeability [mD] [cm]

Berea 22.1 914 10Paaratte 28.3 1156 9.5Mt. Simon 24.4 7.5 9.6Tuscaloosa 23.6 220 10.8

is shown in figure 4. The rock grains are well sorted and

have mature textural features. The grain-size is generally

smaller than the Berea and Otway cores and grain-coating

and pore-filling clays (fibrous illite) can be seen clearly in

the thin-section and SEM micrographs (Fig. 4). The rock

has sloping changes in the porosity along the core length

which range from .23-.26.

Tuscaloosa: A conglomerate sandstone from the

Tuscaloosa Massive Sand formation was obtained from the

Cranfield CO2 injection site in Alabama [Lu et al., 2011].

The core was taken from 3200m depth. Thin-section and

SEM micrographs are shown in Fig. 5. The rocks grains are

compositionally and texturally immature, with significant

rock fragment components, a wide range in grain sizes, and a

high angularity and low sphericity. There is significant pore-

filling and grain-lining quartz cementation, as can be seen

in both the thin-section and SEM micrographs. Clay com-

ponents are platy and appear to be the serpentine-chlorite

clay that has been identified in Tuscaloosa sands [Ryan and

Reynolds, 1996; Hosseini and Hayatdayoudi , 1986]. There is

significant heterogeneity in the porosity of the core, which is

randomly distributed along its length ranging from .21-.26.

2.2. Fluids

CO2 and tap water (salinity less than 50 mg/kg) were

used for the flow experiments. At 9 MPa and 50◦C, the vis-

cosities of the fluids are taken to be µCO2 = 2.3×10−5

[Pa

s] and µwater = 5.5×10−4

[Pa s] [Lemmon et al., 2011]. The

viscosity ratio of CO2:water, M = .042, is midrange for tem-

perature and pressure conditions seen in sedimentary basins

of North America [Nordbotten et al., 2005].

2.3. Core flooding experimental setup

The core flooding apparatus is a modified version of a

setup described in a previous study [Perrin et al., 2009]. A

schematic is shown in figure 6. The rock sample is wrapped

in a sleeve with layering from the core outwards of heat-

shrinkable teflon, nickel foil, another layer of heat-shrinkable

teflon, and a viton rubber sleeve and placed in an aluminum

core holder. Two high accuracy pressure transducers (Oil

filled Digiquartz Intelligent Transmitter, Model 9000-3K-

101) are tapped into the core holder with pressure measured

at the inlet and outlet faces of the core. A displacement

pump (Teledyne Isco, Model 260D) injects water around

the sleeve to create the overburden pressure. Two electric

heaters heat the water inside the confining fluid to maintain

the core at the experimental temperature. Two dual-pump

systems are used to inject water and CO2 in the core sam-

ple (Teledyne Isco, Model 500D). The systems are composed

of two pumps connected with a set of electric valves. The

dual pump configuration provides continuous fluid delivery

by synchronizing the pump and refill strokes so that one

pump is always delivering fluid. The refill and delivery of

each pump is automated through a controller. The bodies of

the CO2 pumps are surrounded by water-regulated temper-

ature jackets and CO2 was kept at a constant temperature

of 50◦C during the experiments so that delivery rates from

the CO2 pumps corresponded to flow rates in the core. Be-

fore entering the core, CO2 and water pass through a heat

exchanger heated to the experimental temperature. The flu-

ids combine in the lines prior to the core holder and then

enter the core. After flowing through the core, the CO2 and

water are separated in a TEMCO AMS-900 high pressure

two-phase separator. The separator includes an electronic

pulser receiver and uses acoustic signals to constantly mon-

itor the volume of fluids in the separator. CO2 from the top

of the separator returns to the CO2 pumps as they refill,

whereas water passes through a pressure regulating pump

(Teledyne Isco, Model 1000D) that is set to deliver or re-

ceive as is necessary to maintain the fluid pressure at the

Berea Paaratte (Australia)

Mt. Simon (Illinois) Tuscaloosa (Alabama) 10 cm

7

Relative permeability - Results

•  Flow rate:

•  Steady-state: 5 PVI •  100% CO2 injection

alternative technique* à Flat saturation profiles à Core heterogeneity

Mt. Simon

Distance from the inlet [mm]

CO2

satu

ratio

n [-]

Paaratte

Tuscaloosa

Berea

9010 20 30 40 50 60 70 800

.2

.4

.6

.8

9010 20 30 40 50 60 70 80

0

.2

.4

.6

.8

fCO2 =qCO2qt

= 0.1!1

qt =10!15ml/min100% CO2

0.9

0.1

fCO2

100% CO2

0.9

0.1

fCO2

100% CO2

0.9

0.1

fCO2

100% CO2

0.9

0.1

fCO2

CO2 saturation profiles

Berea Paaratte



*Ramakrishnan T.S. and A. Cappiello, Chem. Eng. Sci. 1991, 46(4), 1157-1163

8

•  Features are qualitatively predicted from MICP measurements

•  Typical behavior of a strongly water-wet gas/water system

•  Viscosity ratio controls end-point saturation ( fCO2=1 )

Berea

Mt. Simon

Water Saturation

Rel

ativ

e pe

rmea

bilit

y

Paaratte

Tuscaloosa1

0.4

CO2 / water rel. perm. curves Berea Paaratte


Relative permeability curves

Rel

ativ

e pe

rmea

bilit

y

Water saturation

Krevor S. et al., Water Resources Research 2011, submitted

prediction from Pc curve

prediction

fit


CO2

water

kri (Si ) = !Lµiui"Pk

9



Core flooding experiments relative permeability

residual trapping capillary pressure


Berea

Mt. Simon

Water Saturation

Rel

ativ

e pe

rmea

bilit

y

Paaratte

Tuscaloosa1

0.4








Rel

ativ

e pe

rmea

bilit

y

Water saturation

CO2

water

0 . 4 0 . 6 0 . 8 1


Berea Paaratte


00.8

00.8

S CO

2 ,r [-]

SCO2 ,i [-]







CO

2 max

. sat

urat

ion


Land

Spiteri

0 0.2 0.4 0.6 0.8 10

5

10

15

20

25

30

Water saturation

Cap

illary

Pre

ssur

e [k

Pa]

298K

323K

CO2 - water

CO2

10

The method


P1

P2

BP

u = ! kµ"PL

u

100% water (single phase)

Capillary pressure measurement during a core-flooding experiment

Darcy’s law:

11

The method


u

P1

P2

BP PCO2

Pw

uw = 0 ! dPdz

=dPc

dz ! "P = Pc z=0

100% CO2 (drainage)

Capillary pressure measurement during a core-flooding experiment

z = 0 z = L

Steady state:

12

Experiment - Pressure drop


T = 323K 2 4

7 10

15

25 35

50

1

Flow rate [ml/min]

T = 298K

•  Berea (280 mD) •  Flow rates:

•  1 – 50 ml/min •  Injection of 5 PVI

for each step •  Average over the

last 1 PVI •  Viscosity

•  298 K: 7.1 10-5 Pa s •  323 K: 2.3 10-5 Pa s

2 4

7 10 15

25

35

50

1

Flow rate [ml/min]

13 Ronny Pini – [email protected] - Stanford University

Experiment – CT scan (323 K) Inlet slice (20x) Outlet slice

CO

2 saturation CO2

water

aluminum

4 ml/min

15 ml/min

50 ml/min

2 ml/min

10 ml/min

35 ml/min

1 ml/min

7 ml/min

25 ml/min

4 ml/min

15 ml/min

50 ml/min

14

Inlet slice (20x)


Experiment – CT scan (298 K)

4 ml/min

15 ml/min

50 ml/min

2 ml/min

10 ml/min

35 ml/min

1 ml/min

7 ml/min

25 ml/min

Outlet slice

CO2

water

aluminum

4 ml/min

15 ml/min

50 ml/min

CO

2 saturation

15

CA: IFT:

Capillary pressure curve


0 0.2 0.4 0.6 0.8 10

5

10

15

20

25

30

Water saturation

Cap

illary

Pre

ssur

e [k

Pa]

T = 323K

T = 298K

CO2 / water

MICP Pc,CO2/w = Pc,m/a

! CO2/w cos"CO2/w!m/a cos"m/a

!CO2/w =180o

!m/a = 485 mN/m

298 K 323 K

Fit (exps.) 28.1 38.7

Literature* 29.5 35.5

! CO2/w[mN/m]

* Georgiadis A. et al, J. Chem. Eng. Data 2010, 55, 4168–4175 Chiquet P. et al., Energy Convers. Manage. 2007, 48, 736–744

!m/a =140o

16

Capillary pressure - heterogeneity


CO

2 saturation

Porosity

Porosity Saturation*

At the sub-core scale, a saturation distribution can be associated to a given capillary pressure

* 323 K, 10 ml/min

17

CT scan precision - assessment

•  Normal distribution

•  Random error averaging helps! •  Error propagation


100 50 0 50 1000

0.05

0.1

0.15

0.2

CT units

Prob

abilit

y

observations normal distribution

µ  = -0.35 σ = 25.6

Subtracting two scans

*120 kV, 200 mA, 25 DFOV

c = f (a,b) ! ! c2 =! a

2 !c!a!

"#

$

%&

2

+! b2 'c'b!

"#

$

%&

2

N µ,! 2( )

18

CT scan precision - assessment


*120 kV, 200 mA, 25 DFOV

CO2 saturation

Uncertainty σS,1 σS,20

1×1 0.22 0.049

3×3 0.13 0.03

5×5 0.077 0.017

g/rws/r

wsg/rws/r

CTCTCTCT

S−

−=

2

g/rws/r

wsg/rws/r

g/rws/r

pix 12

⎟⎟⎠

⎞⎜⎜⎝

⎛

−

−+

−=

CTCTCTCT

CTCTS

σσ

with CTi/r affected by σpix

≈ 1

5 10 15 200.01

0.1

0.3

1x1

3x3

5x5

S

Repeated scans

19 Ronny Pini – [email protected] - Stanford University

Experiment – CT scan (323 K) Inlet slice (20x) – CO2 saturation

CO

2 saturation 4 ml/min

15 ml/min

50 ml/min

2 ml/min

10 ml/min

35 ml/min

1 ml/min

7 ml/min

25 ml/min

20

Experiment – CT scan (323 K)


Inlet slice (20x + 5x5) – CO2 saturation C

O2 saturation

4 ml/min

15 ml/min

50 ml/min

2 ml/min

10 ml/min

35 ml/min

1 ml/min

7 ml/min

25 ml/min

21

0 0.2 0.4 0.6 0.8 10

5

10

15

20

25

Water saturation

Cap

illary

Pre

ssur

e [k

Pa]

Capillary pressure - heterogeneity


4 ml/min

10 ml/min

•  Coarsening •  5 x 5

•  Pixel size: •  2.5 x 2.5 mm

•  Uncertainty S •  σS = 1.7% (abs.)

Each pixel possesses a unique capillary pressure curve!

22

Concluding remarks •  CO2/water relative permeability and capillary pressure curves

have been measured on reservoir rocks at reservoir conditions

•  Generally, results are typical for a strongly water-wet system

•  Relative permeability: •  Low CO2:water viscosity ratio results in low CO2 saturations and

accordingly low relative permeability

•  Capillary pressure: •  Results are consistent with MICP and expectations from changes in

temperature •  The technique allows to assess and quantify the heterogeneity of the

capillary pressure at the sub-core scale


Multiphase flow of CO2 and water in reservoir rocks at ... · reservoir rocks at reservoir conditions ... representing a range of reservoir properties and rock types. Table 1 summarizes

Documents