9.

CAPILLARY PRESSURE

9.1

Definitions

When two immiscble fluids are in contact in the interstices of a porous medium, a discontinuity in pressure exists across the interface separating them. The difference in pressure Pc is called capillary pressure, which is pressure in the non-wetting phase minus the pressure in the wetting phase
Pc = Pnon wetting Pwetting

(9.1)

Thus, the capillary pressure may have either positive or negative values. For an oil-water, gas-water or gas-oil system, capillary pressure is defined as Pc = Po Pw
Pc = Pg Pw Pc = Pg Po

(9.2)
(9.3) (9.4 )

The hydrostatic pressure of a liquid with density is dependent on the elevation z as follows
dP = g dz

(9.5)

For an oil-water system, the capillary pressure becomes

(9.6)

If fluid columns are continuous in the reservoir, the following relationship holds Pc ( z ) = Pc ( z 0 ) + g ( w o )dz
z0 z

(9.7 )

where Pc(z) is capillary pressure at height z above z0 and w, o are densities of water and oil, respectively. The capillary pressure is a result of the curvature of fluid interfaces, according to the well-known Laplace equation

1 1 Pc = + r r 2 1

(9.8)

54

where is interfacial tension between the two fluids and r1 and r2 are principle radii of curvature. The condition for capillary forces to exist is a certain curvature of the fluid-fluid interface. The relation between fluid saturation and capillary pressure in the reservoir is a function of pore sizes, wettability, interfacial tension and fluid saturation history (drainage and imbibition). Based on laboratory measurements of capillary pressure, it is possible to convert those into reservoir capillary pressure. From these values fluid saturations in the reservoir can be evaluated. An example of the capillary pressure versus saturation relationship (capillary pressure function) is shown in Fig. 9.1. It is apparent that the relationship between capillary pressure and saturation is not unique, but depends on the saturation history of the system. Definitions of the main terms are as follows (oil-water system):
SWI
Primary DRAINAGE (Forced)

PC (+)
IMBIBITION (Spontaneous) Secondary DRAINAGE (Forced)

A1
SWS SOS Swt
IMBIBITION (Forced)

S ws S wt S os ro = S wt (-) rw =

A2

(Residual Oil)

WIAmott = rw-ro
W I USBM = log
0

A1 A2
SW %

Secondary DRAINAGE (Spontaneous)

100

Fig. 9.1: Typical capillary pressure curves and the relationships of wettability measurements by Amott and USBM tests to Pc.

Irreducible water saturation Swi: The reduced volume of the wetting phase retained at the highest capillary pressure where wetting phase saturation appears to be independent of further increases in the externally measured capillary pressure. Residual oil saturation Sor: The reduced volume of the non-wetting phase which is entrapped when externally measured capillary pressure is decreased from a high value to a big negative value. Primary drainage curve: The relationship characteristic of the displacement of the wetting phase from 100% saturation to the irreducible saturation.
55

Imbibition curve: The relationship characteristic of the displacement of the non-wetting phase from the irreducible saturation to the residual saturation. Secondary drainage curve: The relationship characteristic of the displacement of the wetting phase from residual saturation to the irreducible saturation. Most experimental evidences indicate that the irreducible saturation obtained by primary drainage is the same as that obtained by secondary drainage. When the residual saturation is the same, the imbibition after secondary drainage will follow exactly the imbibition curve obtained after primary drainage. Thus, the secondary drainage curve and the imbibition curve constitute a closed and reproducible hysteresis loop. The capillary pressure hysteresis as can be seen in Fig. 9.1 is a result of the different mechanisms governing filling/ emptying of pores with a non-wetting or a wetting fluid respectively. Contact angle hysteresis is one factor also contributing to hysteresis.

9.2
9.2.1

Capillary Pressure Measurement Methods

Water saturated samples for air-water or oil-water tests and oil saturated cores for air-oil tests are placed on a semi-permeable diaphragm, and a portion of the contained liquid is displaced with the appropriate fluid of air or oil. A schematic diagram of an apparatus for performing such tests is seen in Fig. 9.2. It consists of a cell for imposing pressure, a semi-permeable diaphragm C, manometer for recording pressure M, and a measuring burette for measuring produced volumes.

Fig. 9.2: The porous plate method assembly.

During measurement, the pressure is increased in steps and final equilibrium produced volumes of the wetting phase are recorded for each step. The porous plate method is slow and one full curve may take up to 40 days or more to obtain. However, equipment needed for this method is simple and inexpensive and the
56

work needed is limited to some volume reading during the process. Several samples may be run in one chamber. Then the samples have to be removed in order to weigh them separately between each pressure increase. Preferably, one and one sample should be run in an assembly of one-sample cells. Then it is not necessary to decrease pressure between each reading. This method is regarded as the standard method against which all other methods are compared. Routinely only the drainage curve is measured, but with appropriate modifications the imbibition curve may be determined in the same manner. The weakness, as with all the other methods, is the transformation of data to reservoir conditions.

9.2.2

Centrifuge Method

Hassler and Brunner (1945) presented the basic concepts involved in the use of the centrifuge by relating the performance of a small core in a field of high acceleration. If the cylindrical core of length L is subjected to an acceleration ac = -2r where is angular velocity of the centrifuge and r is the distance from the axis of rotation, then from Eq. (9.6) we have
Pc = a c r

(9.9)

Given the boundary conditions show in Fig. 9.3, the differential equation can be solved by simple integration

Fig. 9.3: Schematic diagram of a core in a centrifuge and its boundary conditions.

57

Pc = a c dr
r2

(9.10) (9.11)

Pc (r ) = 2 rdr
r2

The capillary pressure at the outer face of the core is zero, Pc(r2) = 0, so
Pc (r ) =
1 2

2 r22 r 2

) )

(9.12)

and for a continuous phase, the capillary pressure at the inner face of the core is
PcL = Pc (L ) = 1 2 r22 r12 2

(9.13)

Now, the main purpose is to relate the capillary pressure and saturation S for a given core which gives the saturation in the core at equilibrium with the capillary pressure, S = S(Pc). The saturation at a distance h above the outer face of the core can not be measured directly. However, the average saturation, which is the ratio of remaining liquid volume after production to pore volume can be written as S= 1 1 S (r )dr r2 r1 r 2
r

(9.14)

We will have a relationship of saturation as a function of capillary pressure, S = S(Pc), so Eq. (9.14) can be expressed as follows by changing integration variable Pc(r2) = 0 1 S= r2 r1
0

and S (Pc )dPc 2 r PcL

Pc(r1) = PcL

(9.15)

An expression for r is obtained from Eq. (9.12) r = r2 1 and we obtain 1 S= (r2 r1 ) 2 r2

Pc 2 2 2 r2

(9.16)

S (Pc )dPc Pc 1 1 2 r 2 2 2

(9.17 )

58

and with mathematical manipulation it becomes

( 2)
0

PcL

S ( pc )dPc P 1 c sin 2 PcL

(9.18)

where
cos = r1 r2
1 2

cos 2 ( 2 ) =

(1 + cos ) =

r1 + r2 2r2

sin 2 = 1 cos 2 = 1

r12 r22

Eq. (9.18) can not be solved so simply for the unknown function S. For small values of (small core sample), the acceleration gradient along the core can be neglected. Assuming r1/r2 1 then cos 2 ( 2 ) = 1 and sin 2 = 0

Eq. (9.18) is then reduced to S PcL =

S (P )dP
c
0

(9.19)

or in differentiation form
SL = d S PcL dPcL

(9.20)

In this method the cores is saturated with water (or oil) and rotated at increasing speed. The speed is increased in steps, and average fluid saturations at each speed is calculated from observation of liquid produced. The liquid volume is read with a stroboscope while centrifuge is in motion, and the speed of centrifuge is increased stepwise. When the run is over the cores are removed and weighed. The value of PcL for each centrifuge speed are then computed from Eq. (9.13), and the average saturation for each core is obtained from the dry and saturated weights and the corresponding pipette reading. From these data a smooth curve is prepared for each core.

59

Fig. 9.4a shows a typical S PcL as a function of PcL and points indicated on the curve are first, second and third speed.

Fig. 9.4: Graphical differentiation of S PcL - PcL curve (a) to determine S-Pc curve (b).

The value of saturation that goes with each value of PcL, which now represents the capillary pressure, is obtained from this curve by graphical differentiation according to Eq. (9.20). A typical plot of Pc as a function of S is shown in Fig. 9.4b. A complete capillary pressure curve by this method may be obtained in a few hours, and several samples are run simultaneously. The method is claimed to be accurate, to reach equilibrium rapidly, give good reproducibility, and is able to produce high pressure differences between phases.

9.2.3

Mercury Injection (Purcell Method)

The test specimen is evacuated and mercury is injected in increments into the core at increasing pressure levels. When the entry pressure is reached, one can easily determine the bulk volume of the core. A mercury injection apparatus is schematically shown in Fig. 9.5. The equipment consists basically of a mercury injection pump, a sample holder cell with a window for observing constant mercury level, manometers, vacuum pump, and a pressurized gas reservoir.

60

Fig. 9.5: Apparatus for mercury injection method.

Fig. 9.6: Dynamic capillary pressure apparatus.

In this method the mercury injected is calculated as a percentage of pore volume and related to pressure. A practical pressure limit on most equipment is about 15-25 MPa, but equipment for 150 MPa also exists. Two important advantages are gained with the mercury injection method: (1) the time for determining a complete curve is reduce to less than one hour, and (2) the range of pressure is increased compared with the other methods. However, this method is a destructive method and it is difficult to transform the results to reservoir conditions because of the highly unrealistic fluid system and the uncertainty of wetting of mercurysolid. Using mercury-air as the fluid-pair, one will not obtain the irreducible saturation as when displacing water with air.

9.2.4

Dynamic Method

The main feature in this method is the way it controls the capillary pressure at both ends of the sample. This is accomplished by placing the tested core between two membranes, or porous plates, which are permeable only to the wetting fluid (Fig. 9.6). This permits maintenance of a uniform saturation throughout the length of the core even at low flow rates. The membranes permit separate pressure measurements in each of the two phases. The capillary pressure is then equal to the pressure difference between the non-wetting phase and the wetting one at the inflow face. When equilibrium is reached, the sample is removed and its saturation is determined by weighing.

61

9.2.5 Comparison of Methods

In Table 9.1 advantages and disadvantages of the method are listed. The mercury injection method is primarily used for obtaining pore size distribution data, since it is destructive method. Mercury-air capillary pressure curves have been found to be similar to water-air capillary pressure curves when the mercury-air pressure is divided by a constant. The constant can range from 5.8 to 7.5, depending on the nature of the rock. The porous plate method is the simplest method conceptually and it must be regarded as the standard method. Small and large samples can be used, and the choice of fluids is not restricted. A serious drawback is the limitation in pressure, since most equipment is limited to about 5 atm.
Table 9.1: Comparison of the methods.

Method Porous plate

Fluid type Oil-water Gas-water Gas-oil Hg-air Gas-water Oil-water Gas-oil Gas-water Gas-oil Oil-water

Pc curve type Imbibition Drainage Drainage (Imbibition) Drainage (Imbibition) Drainage

Max. P in lab. 2-5 atm.

Test time Weeks

Mercury injection Centrifuge

100 atm. 10 atm.

Minutes Days

Dynamic

1-10 atm.

Days

The centrifuge can reach capillary pressures higher than the porous plate method, and a large number of samples can be run in a relatively short time.

9.3

Converting Laboratory Data

To use capillary pressure data in reservoir evaluation it is necessary to convert to reservoir conditions. Laboratory data are obtained with a gas-water or an oil-water system, which does not normally have the same physical properties as the reservoir water, oil and gas system. As shown previously, by means of the capillary tube, the capillary pressure is expressed as
Pc = 2 cos r

(9.21)

Considering two specific cases wherein the laboratory and reservoir values are determined with gas-water and water-oil respectively. The capillary pressures become

62

(Pc )L

2 wg cos wg

and

(Pc )R

2 wo cos wo r

(9.22)

where wg = interfacial tension between water and gas used in laboratory test wg = contact angle between water and gas wo = interfacial tension between reservoir water and oil at reservoir conditions wo = contact angle between reservoir, water and oil r = radius of capillary.

Comparing the equations for laboratory and reservoir capillary pressure and assuming the radius of capillary constant with pressure and temperature, it is found that the reservoir capillary pressure is

(Pc )R

wo cos wo (Pc )L = ( cos ) R (Pc )L ( cos ) L wg cos wg

(9.23)

Note that it is difficult to determine the exact value of the contact angle for fluids in a porous matrix, and therefore the contact angles are often neglected. The equation becomes

(Pc )R

R (Pc )L L

(9.24)

9.4
9.4.1

Experiments
Capillary Pressure Measurement using Porous Plate Method (Exp. 10)

Description:

The porous plate method is the most accurate measurement of capillary pressure in homogeneous and heterogeneous cores. Several plugs can be measured at a time. The limitation is that the capillary discontinuity may distort the results. Fig. 9.7 shows the capillary pressure apparatus used in our laboratory.
Procedure:

1. Put the saturated core onto the porous plate in the cell. 2. Adjust the pressure regulator and give the cell the pressure difference according to a measuring plan. 3. After 3 days, stop the pressure supply and take off the cores from the cell. 4. Weigh the core and calculate the water saturation corresponding to the capillary pressure. 5. Repeat 1-4 each 3 days until no appreciable water is produced.

63

Fig.9.7: The capillary pressure apparatus.

Calculations and report:

1. Calculate and fill the data form. 2. Plot capillary pressure curve (Sw-Pc) and pore size histogram (distribution). Core No. Date D: cm, L: cm, Wdry: No. (i) Wwet(i) Pc(i), Hwater (cm) (g) 0 1 2 3 4 5 g, Wsat: g, = r(i) (m) W(i) Wwater

Sw(i) (fraction)

where Pc(i) = capillary pressure of ith measurement, cm of water, reading of U-manometer Wwet = core weight of ith measurement, g Sw(i) = (Wwet(i)-Wdry)/Wwater, ith water saturation of Pc(i) Wwater = Wsat-Wdry, g r(i) = 2g-w/Pc(i), radius corresponding to Pc(i) g-w = 70.0 dynes/cm, interfacial tension of gas-water Pc(i) = 981.Hwater, dynes/cm2 W(i)/Wwater = (Wwet(i-1)-Wwet(i))/Wwater, fraction of the capillaries of r(i) in total pore volume.

64

9.4.2

Capillary Pressure Measurement using Centrifuge Method (Experiment 11)

Description:

The objective of the experiment is to measure capillary pressure in a core sample and plot the capillary pressure versus water saturation curve.
Procedure:

1. Measure the pore volume and bulk volume of a cleaned and dried core sample. Weigh the dry core, Wdry. 2. Saturate the core sample with the brine/water with density w. Weigh the saturated core, Wsat. 3. Place the core sample in a centrifuge core holder and start the centrifuge with rotation rate of 500 RPM. 4. Monitor the displaced fluid volume in the measuring tube while the centrifuge is in motion, by adjusting the strobe light frequency to the rotational speed of the centrifuge. 5. Record the collected fluid volume Vcoll in the graduated tube when the volume of collected fluid shows no further changes. 6. Repeat the procedure described in the steps 4 and 5 at higher rotational speed.
Calculations and report:

1. Calculate pore volume of the core sample: VP =

Wsat Wdry

2. Convert RPM to rad/s: =

3. Calculate capillary pressure at each step: PcL = where r2 = 0.145 m r1 = r2-L L = length of the core sample

w 2 (r22 r12 ) ,

4. Calculate average water saturation S w in the core sample based on the volume V of the water collected at corresponding capillary pressure: S w = 1 coll . VP 5. Plot S PcL versus PcL .

65

6. Draw tangents to the plot of the curve at each point and calculate the slope of each tangent. These slopes are the water saturation values Sw at the corresponding capillary pressure. 7. Plot the capillary pressures as a function of the water saturations. Core No.: RPM 500 750 1000 1500 2000 2500

D:

cm,

(rad/s)

L: PcL (bar)

cm.

Vcoll (cm3)

Sw

S PcL (bar)

Sw from plot

66