1 Saleh AlMansoori Dr. Stefan Iglauer Christopher Pentland Professor Martin J. Blunt Measurements of Residual Saturation Implications for CO 2 Trapping.

1

Saleh AlMansooriDr. Stefan Iglauer

Christopher PentlandProfessor Martin J. Blunt

Measurements of Residual Saturation

Implications for CO2 Trapping

2

Measurements of Residual Saturation - Implications for CO2 Trapping

• Introduction - Residual Saturation & Implications for CO2 Trapping

• Objectives – The Big Picture & Experimental Design

• Experimental Procedure & Results– Horizontal oil water core floods– Vertical Sgr Experiment– Vertical oil water Sor vs. Soi core floods– Displacement Experiment

3

Introduction - Residual Saturation & Implications for CO2 Trapping

• Carbon Capture and Storage (CCS) is one technology to prevent atmospheric emissions of CO2 (Benson, 2006).

• Fossil fuel burring during the past 20 years has accounted for around three-quarters of human-made CO2 emissions (Marland et al., 2003).

• The volume of oil currently produced worldwide would be similar to the volume of CO2 injected if 109 tonnes (1 Gt) of carbon per year were captured and stored underground (Pacala and Socolow, 2004).

• Global emissions are 7 GT carbon a year.

4

Introduction - Residual Saturation & Implications for CO2 Trapping

• CO2 critical point: T = 31.04 ºC, P = 7.38 MPa, and D ≈ 800-850 m (Pruess et al., 2001).

• CO2 at supercritical conditions has a high density, similar to oil and water, but a gas-like viscosity (Holloway and Savage, 1993).

• CO2 at supercritical conditions is non-polar, slightly soluble in water, and a very good solvent for organic compounds (Holloway and Savage, 1993).

• Therefore, it is widely used worldwide to increase oil mobility and may displace more than 40% of the residual oil in EOR operations (Blunt et al., 1993).

• However, supercritical CO2 will not be used in our experiments (high pressure).

• Experiments conducted with oil/water and gas/water systems initially.

5

Objectives – The Big Picture

• Investigate pore-scale displacement processes.

• Measure the amount of oil trapped under different conditions. Trapping at low to intermediate saturations is important (implications for CCS).

• Experimental results test and validate the results of pore-scale modelling.

• Properties can then be used in field scale modelling.

• Overall aim of being able to design CO2 injection projects such that the vast majority of CO2 is trapped as a residual phase before it reaches any stratigraphic or structural seal (Ran Qi’s talk about CO2 storage design)

6

Objectives – Experimental Design

• Start with 2 phase systems (oil/water or gas/water).

• Theoretical-empirical prediction of two and three-phase relative permeability requires experimental validation.

• Models (e.g. Land) can not predict the residual oil and gas saturations reliably at low residual saturations.

7

Experimental Procedure – Horizontal Core Floods

• Mass balance before and after packing to measure porosity (grain size route).

• Flush with CO2.

• Sand-packed column was oriented horizontally.

• 5 pore volumes of de-aired brine was injected to reach full saturation.

• Measure absolute permeability for 3 rates.

• Volume balance before and after saturation to measure porosity (pore volume route).

• 5 pore volumes of n-octane were injected.

• Soi was measured via mass balance and volume balance.

• De-aired brine was injected, then Sor was measured via mass balance and volume balance.

8

Experimental Procedure – Horizontal Core Floods

• Several sands have been screened and based on preliminary residual oil saturation results, Levenseat 60 was chosen.

0

20

40

60

80

100

0 200 400 600 800 1000Sieve opening (mm)

Acc

umul

ativ

e w

eigh

t per

cent

age

Imperial College

WBB Minerals

Photo micrographs of LV 60 sand. Grain size distribution of LV 60

9

Experimental Procedure – Horizontal Core Floods

10

Experimental Results – Horizontal Core Floods

Static Rock Properties• Three replicates performed (columns broken down & re-packed) to ensure quality,

precision, and reproducibility.

• Porosity = 37% (± 0.21%) – Porosity Lit. ranges from 33% to 42% for unconsolidated sands – (Willhite, 1986; Kantzas et al., 1988; Zhou and Blunt, 1997; Fanchi, 2000; Schaefer et al., 2000; Irle and Bryant, 2005)

• Permeability = 3.18 x 10-11 m2 (± 3.0 x 10-13 m2)– Permeability Lit. values (4.85 x 10-11 m2 to 3.22 x 10-10 m2) – (Zhou and Blunt, 1997)

Porosity Replicate 1 Replicate 2 Replicate 3

1 37.21% 36.99% 36.99%

2 37.31% 37.08% 36.81%

3 --- 35.82% 36.72%

4 --- 37.02% 37.02%

Mean 37.26% 36.73% 36.88%

Perm Replicate 1 Replicate 2 Replicate 3

qm3/s

Δp,Pa

k,m2

Δp,Pa

k,m2

Δp,Pa

k,m2

8.33E-08

9.38E+03

3.28E-11

9.72E+03

3.16E-11

9.86E+03

3.12E-11

5.00E-08

5.65E+03

3.26E-11

5.93E+03

3.11E-11

5.93E+03

3.11E-11

1.67E-08

1.86E+03

3.30E-11

1.93E+03

3.19E-11

2.00E+03

3.08E-11

11

Experimental Results – Horizontal Core Floods

Dynamic Properties• Three replicates performed (columns broken down & re-packed) to ensure quality,

precision, and reproducibility.

Soi Replicate 1 Replicate 2 Replicate 3

Mass Balance73.04% 71.62% 71.49%

Volume Balance73.74% 74.82% 74.43%

Mean, Soi 73.39% 73.10% 72.96%

Sor Replicate 1 Replicate 2 Replicate 3

Mass Balance12.793% 13.518% 12.616%

Volume Balance13.159% 13.641% 12.570%

Mean, Sor 12.976% 13.579% 12.593%

• Soi = 73% (± 0.2%)• Lit. values (70-85%) (Blunt et al., 1994b; Skauge et al., 1994; Schaefer et al., 2000)

• Sor = 13% (± 0.4%)• Lit. values ( in the range of 13%) (Firoozabadi et al., 1987; Sahni et al., 1998).

12

Experimental Procedure – Vertical Sgr Experiment

lower-end cap

upper- end cap

metal meshfilter paper

lower stream pressure transducer

P2

P1

fluidreservoir

effluent collector

upper stream pressure transducer

lower 3-way metal valve

1/8 inch tube

Flo

w p

ath

of b

rine

into

sat

urat

ed s

and-

pack

ed c

olum

n

500D Pump

Cyl

inde

r

saturated sand-packed colum

n

13

Experimental Procedure – Vertical Sor/Soi Experiment

• Mass balance before and after packing to measure porosity (grain size route).

• Flush with CO2.

• Sand-packed column was oriented vertically.

• 5 pore volumes of de-aired brine was injected to reach full saturation.

• Volume balance before and after saturation to measure porosity (pore volume route).

• Bottom valve was opened to drain brine and top valve was opened to allow air to enter into the column.

• The column was reversed to remove end effects.

• Swi was measured via mass balance.

• De-aired brine was injected, then Sgr was measured via mass balance.

14

Experimental Results – Vertical Sgr Experiment

Dynamic Properties• Three replicates performed (columns broken down & re-packed) to ensure quality,

precision, and reproducibility.

SwiReplicate

1Replicate

2Replicate

3Mean

Mass Balance

42.06% 45.89% 40.00% 42.65%

SgrReplicate 1 Replicate 2 Replicate 3 Mean

Mass Balance

13.44% 13.65% 14.06% 13.72%

Sgr Lit. values (10-50%) (Geffen et al., 1952; Chierici et al., 1963; Mulyadi et al., 2000; Schaefer et al., 2000)

Swi = 43% (± 2.3%)

Sgr = 14% (± 0.2%)

15

Experimental procedure – Sand Packed Columns

Flow path of oilSand-packed column injected with non-wettingfluid (oil dyed red).

16

Experimental Procedure - Vertical oil water Sor vs. Soi core floods

• 2 IDENTICAL COLUMNS

• Mass balance before and after packing to measure porosity (grain size route).

• Flush both columns with CO2.

• Sand-packed columns were oriented vertically.

• 5 pore volumes of de-aired brine were injected to reach full saturation.

• Decane reservoir connected to top of columns and brine allowed to drain under gravity from the base. Decane enters the top of the column. No pumping.

• Equilibrium reached where both columns have a (theoretically) identical oil saturation profile versus height.

• One column removed for slicing and sampling – Soi.

• Second column has brine injected from the base, Brine sweeps oil leaving an Sor. Coulmn sliced and sampled.

17

Experimental Procedure - Vertical oil water Sor vs. Soi core floods

brine flow

Oil

flow

Oil

flow

brine flow

COLUMN A - Soi COLUMN B - Sor

18

Experimental Results - Vertical oil water Sor vs. Soi core floods

Saturation versus Height

0

25

50

75

100

0 20 40 60 80 100

Oil Saturation (%)

He

igh

t fr

om

co

lum

n b

as

e (

cm

)

Experiment 2 - SoiExperiment 2 - Sor

Experiment 1 - SoiExperiment 1 - Sor

19

Experimental Results - Vertical oil water Sor vs. Soi core floods

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.0 0.2 0.4 0.6 0.8 1.0

Soi

So

r

Land - Experiment 2Experimental Data - Exp. 2Core Flood ResultsLand - Core Flood

20

Experimental Procedure – Displacement Experiment

• Generic behaviour

• No clear evidence of upward movements and how it forms (Glass et al., 2003).

• Phase distribution (uniform or fingered)

• If fingering, less CO2 trapped

• If uniform, more CO2 trapped

(Uniform front versus fingering flow (Irle and Bryant, 2005).

21

Experimental Procedure – Displacement Experiment

• Inject n-octane into water-saturated column.

• Once it reaches half-way up, cease injection.

• Rotate the column through 180 degrees so that the n-octane is now at the base, n-octane will rise under buoyancy forces.

• Experiment ends when injected n-octane reaches the top.

• Then to GC (Gas Chromatography). bottom

top

brin

eb

rin

e-o

il

bottom

top

brin

eb

rin

e-o

il

22

Measurements of Residual Saturation – Future Work

• Sor versus Soi Experiment:– Improve experimental accuracy (e.g. reduce evaporation)– Eliminate air from the system (modify column design)– Expand to water / air systems– In the long term look to work with super-critical CO2

• Use GC to measured Sor in three-phase experiments as a function of distance (down to 0.1%).

• Link in to CT experimental work (Dr. Stefan Iglauer)

• Wettability studies through ageing of columns

• Use the network models to predict observed behaviour in the lab– Experimental measurements of contact angle to feed in to the network models

• Use of consolidated rock (limestone and sandstone)

23

Measurements of Residual Saturation – Future Work

a) Column experiments of buoyancy-driven flow are performed in the labb) Small samples are imaged using micro-CT scanningc) A network model representing the system is used to predict the pore-

scale fluid configurations and compared to the image. Average flow properties are predicted

d) Properties are used in field-scale modelling.

24

Questions