You are on page 1of 9

Rheology of Heavy-Oil Emulsions

Hussein Alboudwarej, Moin Muhammad, and Ardi Shahraki, Schlumberger; Sheila Dubey and Loek Vreenegoor,
Shell Global Solutions (US) Inc.; and Jamal Saleh, Shell International E&P

Summary In addition to formation blockage and general difficulty in the


Water is invariably produced with crude oil. If there is enough separation of oil and water in production facilities, one of the main
shear force when crude oil and produced water flow through the drawbacks of emulsion formation is an increase in the apparent
production path, stable emulsions may be formed. This scenario viscosity of the oil. Viscosity of water-in-oil emulsions increases
may particularly be present during the production of heavy oils, as the water cut increases before the so-called emulsion inversion
where steam is used to reduce the viscosity of heavy oil, or in cases point, beyond which the continuous phase changes to water (i.e.,
in which submersible pumps are used to artificially lift the pro- water-in-oil emulsion switches to oil-in-water emulsion). It has
duced fluids. To efficiently design and operate heavy-oil produc- been shown that the viscosity of the water-in-oil emulsion may
tion systems, knowledge of the realistic viscosities of the emulsi- increase as much as one order of magnitude or even higher over
fied heavy oil, under the actual production conditions, is necessary. the viscosity of the dry oil (Singh et al. 2004). In oil-in-water
This study is an attempt to investigate the effect of water content, emulsions, viscosity decreases with an increase in water content.
pressure, and temperature (i.e., operating conditions on the viscos- Therefore, the maximum apparent viscosity of emulsions occurs at
ity of live heavy-oil emulsions). the emulsion inversion point (Szelag and Pauzder 2003).
Two heavy oil samples from South America were used for this One of the field implications of emulsion inversion and emul-
study. The stock tank oil (STO) samples were recombined with the sion viscosity is the transportation of heavy oil-in-water emulsions.
corresponding flash gases to reconstitute the original reservoir oil It has been shown that under certain conditions, the viscosity of
compositions. Live oil/water emulsions were prepared in a con- the heavy oil-in-water emulsion is considerably lower than the
centric cylinder shear cell using synthetic formation water, under heavy oil (Nunez et al. 2000). Because water production is gen-
predetermined pressure, temperature, and shear conditions. The erally increasing with the life span of the oil production field, it
stability of live emulsions was investigated using a fully visual is also very important to have an accurate measure of the emul-
pressure/volume/temperature (PVT) cell, while viscosities were sion inversion point and viscosity for optimum pipeline and facil-
measured using a precalibrated, high-pressure capillary viscom- ity design.
eter. Viscosities were measured at least in three different flow rates The majority of experimental work on the emulsion stability
at the testing conditions. In addition to live-oil emulsion studies, and viscosity measurements is performed for STO samples. How-
the stability and droplet size distribution of STO emulsions were ever, dissolved gases affect the viscosity of oil and the correspond-
also determined. ing emulsions. For optimum design and operation of the oilfield
Experimental results indicated that the inversion point for the production facilities, knowledge of live-oil emulsion properties is
STO emulsions was approximately 60% water cut (volume), and required. In the present work, stability and viscosity of two South
the average droplet size was increasing with water content. For all American live heavy-oil emulsions were studied. Live-oil emul-
measured cases, viscosities varied with temperature according to sions were prepared in a shear cell and analyzed in a visual PVT
an Arrhenius relation, while viscosities did not indicate any varia- cell for stability and in a capillary viscometer for viscosity mea-
tion with flow rate (shear) within the range of tested flow rates. surement. The information from this study may be used for both
Measured viscosities also increased as pressure decreased below facility-pipeline and artificial lift system design.
the bubblepoint of the sample as lighter hydrocarbon components The study of the performance of ESP systems indicates that
evolved. The measured viscosities increased as much as 500% under these conditions, a given system will not be able to provide
because of the presence of emulsions before a sharp drop in vis- optimum performance if the water cut increases from 0 to 50%.
cosity beyond the inversion point. The variation of viscosity with For a system operating on unchanging parameters, the rate of
water content for live emulsion samples indicated that the inver- production will decrease drastically because of the scale of in-
sion point for live emulsions is similar to that of STO samples. crease in emulsion viscosity with water cut.
The experimental results are also used to analyze and evaluate
Experimental
the performance of an ESP system when water cut increases and
causes emulsion in a well. Two South American live-oil samples were used for this study.
The STO samples were recombined with 5-component synthetic
gas mixtures to reconstitute the reservoir fluid compositions. The
Introduction recombined samples were conditioned at reservoir pressure and
As an oilfield ages, the rate of water production increases. With temperature for a period of 5 days. The recombined samples were
enough shear force (e.g., flow through a downhole pump or a flow then analyzed for composition and physical properties, as shown in
restriction such as a choke valve or orifice), a stable emulsion can Tables 1 and 2 for Oil A and Oil B, respectively.
be formed. Presence of inorganic solids such as sand, clay, and To prepare emulsions, water samples with similar compositions
corrosion products, together with surface-active materials such as to reservoir waters were used. Compositions of formation water
asphaltenes and naphthenic acids, also enhance the stability of samples are provided in Table 3. Note that Water A and B cor-
emulsions (Kokal 2005). Because of the presence of these ele- respond to Oil A and Oil B, respectively. A high-pressure shear
ments, the occurrence of tight emulsions in the production facili- cell was used to prepare live emulsions. The device is a Taylor-
ties is quite common. In some cases, emulsions may also form in Couette flow device in which the annulus between the two con-
the near-wellbore region, leading to emulsion blockage of porous centric cylinders is used to provide the shear environment for the
media (Kokal et al. 2002). fluid. It is rated to pressures up to 7,000 psia, temperatures up to
390°F, and rotational frequencies (through a magnetic coupling)
up to 100 Hz. The operation of the shear cell is controlled auto-
matically to maintain operating conditions. To form emulsions, the
Copyright © 2007 Society of Petroleum Engineers
shear cell is first filled with a water sample, and then pressurized
This paper (SPE 97886) was accepted for presentation at the 2005 SPE/PS-CIM/CHOA and heated up to the desired set points. Water is then displaced
International Thermal Operations and Heavy Oil Symposium, Alberta, 1–3 November. Origi-
nal manuscript received for review 26 August 2005. Revised manuscript received 3 January
under live conditions with the oil sample until desired water/oil
2007. Paper peer approved 8 February 2007. volume ratio is achieved. The rotational speed in the cell then

August 2007 SPE Production & Operations 285


emulsifies the water under provided shear condition. After shear-
ing the mixture for the desired period of time, the cell content is
displaced with an inert gas isothermally and isobarically to a
sample cylinder equipped with an internal mixing ring. The emul-
sion cylinder is maintained under the same pressure and tempera-
ture and a continuous rocking condition. Emulsion samples are
transferred to a visual PVT cell for stability analysis and to a
capillary viscometer for viscosity measurement as needed. Table 4
summarizes the conditions under which different emulsions were
prepared. Note that the water cut (%vol) is defined as the volume
of water over the total volume expressed as percentage.
Stability analyses for STO samples were performed using
bottle tests; however, for live emulsion samples, a fully visual PVT
cell was used. Live emulsion samples were transferred from a
sample cylinder (with mixing ring) to the PVT cell. The magnetic
coupling in the PVT cell was turned off to eliminate any further
shearing of the sample. For live emulsion samples, only stability of
85% water-cut samples were tested.
A capillary viscometer rated to 10,000 psia and 374°F was used
for viscosity measurements. The capillary viscometer consists of
two, high-pressure cylinders (32 ml each) connected to a 10-ft-
long and 0.03-in. diameter capillary coil. A differential pressure
transducer is used to monitor the pressure drop across the capillary
coil. The fluid sample is pumped from one cylinder to the other
through the capillary coil by an opposed pump. From the measured
fluid-flow rate and pressure drop, the viscosity can be determined
using the Hagen-Poiseuille relationship for laminar flow in tubes,
namely

␮=
⌬p
Q
冉 冊␲ r4
8L
=
⌬p
Q
k, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)

where ␮ is the fluid apparent viscosity; ⌬p is the pressure drop


across the capillary tube of length, L, and an internal radius r; and
Q is the volumetric flow rate. The tube constant k is determined by
calibrating the viscometer using standards of known viscosity at
test pressures and temperature. Note that the maximum Reynolds
number for this study corresponding to the highest flow rate and
the lowest viscosity was approximately 25, indicating a fully lami-
nar flow. Shear stress at the wall can be calculated as
⌬p D
␶= , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)
4L
in which D is the internal diameter of the capillary tube, L is the
length of capillary tube, and ⌬p is the pressure drop across the
capillary tube. Shear rate at the wall may be calculated as

␥= , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (3)

in which ␶ and ␮ are wall shear stress and fluid apparent viscosity.

Results and Discussion


STO Emulsion Preparation and Stability Analysis. STO emul-
sions were prepared for both Oil A and Oil B samples to determine
the emulsion inversion points. The significance of the inversion
point is that generally, emulsions show the maximum viscosity at
or near the inversion point. Also, the STO inversion points may
give an idea about the corresponding live-oil emulsion inversion

286 August 2007 SPE Production & Operations


points. STO emulsions were prepared by simply adding proper speed is higher. A combination of lower rotational speed (i.e.,
volume ratio of oil and formation water, and hand-shaking the decoupling at higher rotational speed than 70 Hz) and higher vis-
mixture intermittently for at least half an hour. In all cases, both oil cosity leads to a lower shear rate environment for the case of 50%
and water had been equilibrated at 130 °F before mixing. No volume water content emulsion sample. Stability analysis was only
estimate of applied shear rate is available. The samples were kept performed on 85% water cut emulsions for both Oil A and Oil B
at the same temperature for at least 18 to 20 hours before any samples. Water separation could be observed as soon as the emul-
photomicrographs were taken. For the cases of 85% volume water sion sample was transferred from shear cell to the visual PVT cell
cuts, both Oil A and Oil B sample emulsions were unstable, and oil (in less than 15 minutes). Fig. 3 shows the segregated water and oil
and water started segregating instantaneously. in a visual PVT cell under the same conditions. Note that for the
Fig. 1 shows the results of stability bottle tests for STO Oil A case of Oil B, oil is partially adhered to the glass surface and does
emulsion samples with water cuts of 10, 50, 62, and 85% volume, not allow clear observation. The 85% water-cut emulsions were
respectively. The 50% water-cut emulsion was stable at least for unstable even in the presence of shear from PVT cell magnetic
18 to 24 hours, whereas the 62% volume water-cut emulsion stirrer. No image analysis was performed on the 85% water-cut
showed signs of oil/water segregation after a few hours. The in- emulsions, because both STO and live emulsions were unstable
version point for STO Oil A emulsions samples is estimated to be and could not be used for viscosity testing. No stability analysis
60±5% volume water cut. Fig. 2 shows the stability results for was performed on other live emulsions.
STO Oil B emulsion samples with water cuts of 50, 70, and 85% To investigate water droplet size distribution in emulsions, a
volume, respectively. Both 70 and 85% volume water cuts showed subsample of the live emulsions was flashed from the correspond-
instantaneous segregation of oil and water, indicating unstable ing emulsion cylinders to ambient conditions. The samples were
emulsions. The inversion point for STO Oil B emulsion samples is then analyzed under a microscope, and photomicrographs for each
estimated to be 65±5% volume water cut. emulsion were analyzed using image analysis software. Figs. 4
and 6 show the photomicrographs of emulsions for the Oil A and
Live Emulsion Preparation and Stability Analysis. Table 4 Oil B samples, respectively. Figs. 5 and 7 depict the correspond-
shows the conditions at which live emulsions were prepared in the ing droplet size distribution for the same images. Note that the
shear cell. The difference in the applied shear rates for the Oil A resolution of the images is approximately 1 micron, and number
emulsion samples is caused by the viscosity of the samples and the frequency for droplets smaller than 1 micron is merely a calculated
rotational speed. For 10 and 85% volume water contents, viscosity number and may not represent the actual droplets. For Oil A emul-
is lower than the case of 50% volume water content, and rotational sions, the droplet size distribution for both 10 and 50% water cut

Fig. 1—Stability of Oil A (STO) at 130°F. Fig. 2—Stability of Oil B (STO) at 130°F.

August 2007 SPE Production & Operations 287


Fig. 3—Stability of live 85% water-cut live emulsions of Oil A
and Oil B.

are very similar, and only the number frequency in the 50% water
cut is slightly higher. For Oil B, there is a slight shift in the droplet
size distribution of 50 and 60% emulsions compared to the 15%
emulsion. Comparing the 50 and 60% emulsions of Oil B, it seems
that 50% emulsion has a higher-number frequency of smaller drop-
lets, while the 60% emulsion has a higher number of larger drop-
lets. Note also that for 60% water cut, droplet size distribution
suggests the existence of another size mode, thus a near bimodal
distribution [e.g., partially unstable emulsions (Salager et al.
2000)]. The viscosity measurements for the same oil at 60% water
cut also shows a partially unstable system. Fig. 4—Photomicrographs of Oil A live emulsions (sample
flashed to ambient conditions).
Viscosity Measurements. The viscosities of the neat Oil A and
Oil B samples, together with their corresponding emulsions, were
determined using a capillary viscometer. Measured viscosities for
both dry Oil A and Oil B samples, at various pressures, are shown
in Fig. 8. Data follow an Arrhenius-type temperature trend, at least
for the measured range of temperature. Also, the neat oil viscosi-
ties decrease as the amount of solution gas increases with increas-
ing pressure. The viscosity of each sample was measured at least
at three different volumetric flow rates (shear rates), and the re-
ported viscosity values are the arithmetic average of measured
viscosities for each sample. There was no significant variation of
viscosity with shear rate. Tables 5 and 6 show the range of applied
shear rates (s−1) for each measured viscosity point. The standard
deviation of viscosity with shear rate varied between 0.1 and 1.4%
for Oil A and between 0.02 and 1.4% for Oil B. The overall
average standard deviations of measured viscosities over the range
of shear rate were 0.5 and 0.4% for Oil A and Oil B, respectively.
In all cases, reducing pressure below the bubblepoint (and sepa-
ration of lighter hydrocarbon components) increased the measured
viscosities for both neat oils and emulsions, as expected. At the
same time, an increase in water content of emulsions (up to and
near inversion point) increased the viscosities (Fig. 9). For the case

Fig. 5—Droplet size distribution of Oil A live emulsions (sample Fig. 6—Photomicrographs of Oil B live emulsions (sample
flashed to ambient conditions). flashed to ambient conditions).

288 August 2007 SPE Production & Operations


Fig. 7—Droplet size distribution of Oil B live emulsions (sample
flashed to ambient conditions).
Fig. 8—Viscosities of Oil A and Oil B (no water). Triangles for
100 psia, circles for 1,000 psia, diamonds for 2,000 psia, and
squares for 3,000 psia.
of the Oil A sample, viscosities increased up to 30% for 50%
water-cut emulsions at 40°F, and viscosity variation was more
sensitive to pressure than to water content. In comparison, the Case Study
viscosity of Oil B emulsions showed higher variation with water
content than with pressure. For the case of Oil B emulsions, vis- An ESP system performance analysis under these experimental
cosities increased more than 5 times at 50% water cut and 40°F. results was conducted. For this analysis, the measured data for Oil
For Oil B at 60% water cut, the viscosity dropped drastically B at 2,000 psia and 70°F is used. The measured data under the
compared to the 50% water cut. This behavior might be an indi- conditions was used to calibrate the calculated viscosity. The in-
cation that the emulsions were beyond the inversion point. formation and data for the reservoir and the well are presented in
Table 7 and Fig. 12.
Comparison of Emulsion Viscosity Correlations. Four single- Fig. 13 displays the well performance curves for oil (with no
parameter correlations were chosen for the comparison. Taylor’s emulsion). For this case, the flowing bottomhole pressure (FBHP)
(1932) and Vand’s (1948) correlations were used as the pioneer was 511 psia. For the flow rate, a pressure of 1,076 psia is required
and most commonly referred correlations. The correlations of Ron- to lift the fluid to the surface. This difference in pressure clearly
ningsen (1995) and Yaron and Gal-Or (1972) were chosen based indicates that the well requires artificial lift to produce 950 B/D of
on two separate performance studies (Johnsen and Ronningsen liquid, and an ESP system is the most suitable candidate for this
2003; Pal 2001). Although multiple-parameter correlations are situation. To achieve this flow rate, an additional 565 psi of pres-
available for more accurate prediction of emulsion viscosities, the sure is required. The pressure profile (pressure gradient) plot, seen
single-parameter equations were chosen for the sake of simplicity. in Fig. 14, summarizes the available and required lift pressures.
Multiple-parameter viscosity correlations need experimental data The change in water cut that changes the viscosity affects the
for tuning purposes. The emulsion viscosity data may not be avail- whole system twofold.
able a priori. Details of the correlations are provided in the Ap- First, increase in viscosity drastically changes the tubing per-
pendix. formance. Achieving the same rate requires much higher pressure
Figs. 10 and 11 depict the results of such a comparison for Oil to lift the fluid to the surface when water cut is increased. The
A and Oil B, respectively. The emulsion viscosities for Oil A were change in profile is shown in Fig. 14, and the friction component
best predicted with Taylor’s correlation with an absolute average that bears the effect of the viscosity is shown in Fig. 15.
error of 13%; Oil B emulsion viscosities were best predicted by the Second, the viscosity affects the performance of the pump. As
correlation of Yaron and Gal-Or with an average absolute error of the viscosity increases, the efficiency of the pump decreases and
21%. Note that only emulsion viscosity data were considered in the requires more stages to produce the same flow rate. Table 8 shows
error analysis. For Oil A, single-parameter correlations tend to that the number of stages, and required power increases drastically
overpredict the viscosity of emulsions, while for Oil B, no par- as water cut increases. The table also shows that the head factor for
ticular trend could be distinguished. This comparison between the each stage decreases and the power factor increases. The lower the
selected correlations, although limited to single-parameter corre- head correction factor, the less head or lift will be provided. On the
lations, shows that the predicted viscosity of emulsions with no a other hand, the higher the power factor, the more power will be
priori measurement may be largely inaccurate. required by the pump to operate and lift the same rate (these
factors are 1.0 for water). Therefore, the pump provides less head
or lift and requires more power (e.g., a larger motor) as emulsion
viscosity increases. This increase indicates that if the increase in
water cut is expected, the system should be designed for approxi-
mately the inversion point that causes the highest viscosity. As
emulsion viscosity increases with the water cut, more lift is re-
quired to move the fluid to the surface, and also the increase in
viscosity reduces the efficiency of the pump. In Table 9, the best
efficiency rate of the pump decreases as the viscosity increases.
In this analysis, the number of stages required to continue
producing the same rate was determined. In actual practice, if the
number of stages is set, then the production rate will decrease as
water cut increases, as expected.

August 2007 SPE Production & Operations 289


Conclusions Nomenclature
Viscosities of live recombined Oil A and Oil B samples and their D ⳱ internal diameter of capillary tube
different water-cut emulsions were measured using a capillary vis- k1–k4 ⳱ constants in Eq. A-4
cometer. Experimental results indicated that for all measured K ⳱ dispersed-phase viscosity/continuous-phase viscosity
cases, viscosities varied with temperature according to an Arrhe- L ⳱ length of capillary tube
nius relation, while viscosities did not indicate any variation within Q ⳱ volumetric flow rate
the range of tested flow rates, suggesting that fluids are Newto-
r ⳱ internal radius of capillary tube
nian. Measured viscosities also increased as pressure decreased
below the bubblepoint of the sample, inline with the amount of T ⳱ temperature
dissolved gases. Viscosity measurements for dry oil and emulsions ␥ ⳱ shear rate
of Oil A samples indicated that emulsion viscosities have in- ⌬p ⳱ differential pressure
creased by 30%, while for the case of Oil B samples, the increase ␪ ⳱ volume% water cut in Eq. A-4
in the emulsion viscosities was as high as 500%. ␮ ⳱ apparent emulsion viscosity
Regarding emulsion stabilities, the bench top bottle tests indi- ␮c ⳱ continuous phase viscosity
cated that the inversion point for STO Oil A emulsions was ap- ␮r ⳱ relative viscosity defined in Eq. A-1
proximately 60±5% volume water cut, and, for STO Oil B emul- ␶ ⳱ shear stress
sions, was approximately 65±5% volume water cut. For live-oil ␾ ⳱ volume fraction of dispersed phase
samples, the stability of only 85% volume water cut for both Oil
A and Oil B samples was tested, and both emulsion samples were
unstable. A systematic study to determine the inversion point of Acknowledgment
live-oil emulsions was not performed. However, the viscosity mea- The authors wish to thank Schlumberger and Shell Global Solu-
surements for the Oil A sample suggest that the inversion point tions (U.S.) Inc. for permission to publish this work. The authors
may be approximately 55±5% volume water cut at the tested con- would also like to thank Craig Borman, Trevor Lockyer, Rob
ditions (10% volume water cut lower than that of the STO sample). Fisher, and Abdulai Dawodu for performing experimental work.
A more definitive experimental scope of work is required to de-
termine the inversion point for the live emulsion samples. References
Comparison between four emulsion viscosity single-parameter Johnsen, E.E. and Ronningsen, H.P. 2003. Viscosity of ‘Live’ Water-in-Oil
correlations and measured viscosities indicated that the absolute Emulsions: Experimental Work and Validation of Correlations. Jour-
average error in the predicted emulsions viscosities could be as nal of Petroleum Science and Engineering 38: 23–36.
high as 200% for the tested fluids, as shown in Table 9. Single- Kokal, S. 2005. Crude-Oil Emulsions: A State-of-the-Art Review. SPEPF
parameter emulsion viscosity correlations should only be used 20 (1): 5–13. SPE-77497-PA. DOI: 10.2118/77497-PA.
with an expectation of large errors.
If higher water cut is expected during the lifespan of the ESP
system operation, the performance of the ESP should be consid-
ered very carefully. As the results show, either the system should
be designed for the worst-case scenario, which is the inversion
point, or change in the ESP system should be considered at dif-
ferent stages of the well’s life as water cut increases. But these
evaluations require reliable data for the emulsion viscosity of the
fluid.

Fig. 9—Viscosities of Oil A and Oil B and correspond-


ing emulsions.

290 August 2007 SPE Production & Operations


Fig. 10—Comparison between measured and calculated emul- Fig. 11—Comparison between measured and calculated emul-
sion viscosities for Oil A. sion viscosities for Oil B.

Kokal, S. et al. 2002. Productivity Decline in Oil Wells Related to As- Appendix
phaltene Precipitation and Emulsion Blocks. Paper SPE 77767 pre- In the following equations, the relative viscosity is defined as:
sented at the SPE Annual Technical Conference and Exhibition, San
Antonio, Texas, 29 September–2 October. DOI: 10.2118/77767-MS. ␮
␮r = , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-1)
Nunez, G.A., Sanchez, G., Gutierrez, X., Silva, F., Dalas, C., and Rivas, H. ␮c
2000. Rheological behavior of concentrated bitumen in water emul-
sions. Langmuir 16 (16): 6497–6502. in which ␮r is the relative viscosity, ␮ is the emulsion apparent
viscosity, and ␮c is the continuous-phase viscosity.
Pal, R. 2001. Single-Parameter and Two-Parameter Rheological Equations
The celebrated Taylor equation for the relative viscosity of very
of State for Nondilute Emulsions. Industrial and Engineering Chem-
dilute emulsions of nearly spherical noncolloidal droplets is:
istry Research 40: 5666–5674.
Ronningsen, H.P. 1995. Correlations for Predicting Viscosity of W/O
Emulsions Based on North Sea Crude Oils. Paper SPE 28968 presented
at the SPE International Symposium on Oilfield Chemistry, San An-
␮r = 1 + 冉 5K + 2
2K + 2 冊
␾, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-2)

tonio, Texas, 14–17 February. DOI: 10.2118/28968-MS. in which K is the viscosity ratio of dispersed phase to continuous
Salager, J-L., Marquez, L., Pena, A.A., Rondon, M., Silva, F., and Tyrode, phase, and ␾ is the dispersed-phase volume fraction.
E. 2000. Current Phenomenological Know-How and Modeling of Vand’s correlation is a theoretically based exponential function
Emulsion Inversion. Industrial and Engineering Chemistry Research in the form of

冉 冊
39: 2665–2676.
2.5␾
Singh, P., Thomason, W.H., Gharfeh, S., Nathanson, L.D., and Blumer, ␮r = exp , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-3)
D.J. 2004. Flow Properties of Alaskan Heavy-Oil Emulsions. Paper 1 − 0.609␾
SPE 90627 presented at the SPE Annual Technical Conference and
in which ␾ is the volumetric concentration of dispersed phase.
Exhibition, Houston, 26–29 September. DOI: 10.2118/90627-MS.
Variations of Vand’s equation have been used. The most common
Szelag, H. and Pauzder, B. 2003. Rheological properties of emulsions modification is to replace 0.609 with a constant that can be deter-
stabilized by acylglycerol emulsifiers modified with sodium carboxy- mined by fitting experimental data.
lates. Colloids And Surfaces A—Physicochemical And Engineering As- Ronningsen correlation is represented as
pects 219 (1–3): 87–95.
Taylor, G.I. 1932. The Viscosity of Fluid Containing Small Drops of ln ␮r = k1 + k2T + k3␾ + k4T␾, . . . . . . . . . . . . . . . . . . . . . . . (A-4)
Another Fluid. Proceedings of Royal Society A 138: 41–48.
Vand, V. 1948. Journal of Physics and Colloidal Chemistry 52: 217.
Yaron, I. and Gal-Or, B. 1972. On Viscous Flow and Effective Viscosity
of Concentrated Suspensions and Emulsions. Rheologica Acta 11:
241–252.

Fig. 12—Well trajectory.

August 2007 SPE Production & Operations 291


Fig. 14—Pressure profile for different water cuts, at 950 B/D.
Fig. 13—Inflow and outflow curves for 0% water cut.

National Iranian Oil Company (NIOC) at both onshore and


in which k1 to k4 are constants that have been determined at dif- offshore production fields as a process/production engineer.
ferent shear rates (30, 100, and 500 s−1) and based on a set of North He holds PhD and an MSc degrees from the University of Cal-
Sea oil database, T is temperature (°C), and ␪ is the volume per- gary, and a BS degree from the Abadan Institute of Technol-
cent water cut. The constants for 500 s−1 are: k1⳱−0.06671, ogy, all in chemical engineering. Moin Muhammad is Manager
k2⳱−0.000775, k3⳱0.03484, k4⳱0.00005. At constant tempera- of the DBR Technology Center with Schlumberger in Edmon-
ton, Canada. He has more than 12 years of operational, re-
ture, Eq. A-3 reduces to a single-parameter equation.
search, and management experience in reservoir fluid prop-
Yaron and Gal-Or’s analysis gives the following correlation for erties, flow assurance, and business development-related po-
the relative viscosity of concentrated emulsions: sitions (Canada and U.S.). Moin holds an MS degree from the
University of Alberta, Canada and a BS degree from Punjab
␮r = 1 University, Pakistan, both in chemical engineering. Ardeshir

+ 再 5.5关4␾7 Ⲑ 3 + 10 − 共84 Ⲑ 11兲␾2 Ⲑ 3 + 共4 Ⲑ K兲共1 − ␾7 Ⲑ 3兲兴


10共1 − ␾10 Ⲑ 3兲 − 25␾共1 − ␾4 Ⲑ 3兲 + 共10 Ⲑ K兲共1 − ␾兲共1 − ␾7 Ⲑ 3兲
冎 ␾,
(Ardi) Shahraki is a senior production engineer with Schlum-
berger in the Artificial Lift segment at Schlumberger Reservoir
Completions campus in Rosharon, Texas. He has been with
Schlumberger since 2002. He previously worked for IHS Energy
. . . . . . . . . . . . . . . . . . . . . . . . (A-5) as a principal engineer/project manager. His interests include
multiphase fluid flow, well-system analysis, and modeling and
in which K is the viscosity ratio of dispersed phase to continuous predicting the performance of artificial lift systems for the de-
phase, and ␾ is the dispersed-phase volume fraction. sign and diagnostics of systems. He holds BS and MS degrees in
petroleum engineering from the University of Louisiana at
Lafayette, and a PhD degree in petroleum engineering from
Hussein Alboudwarej is a senior research project engineer with New Mexico Tech. Sheila Dubey is a senior staff research en-
Schlumberger at DBR Technology Center, Edmonton. The fo- gineer at Shell Global Solutions (U.S.) Inc. at the Westhollow
cus of his research is on wax and asphaltene flow assurance Technology Center in Houston and has been with Shell Oil
and the rheology of waxy and heavy crude oils. His research Company for 27 years. She is currently the U.S. team lead for
interests also include development of experimental tech- Scale Emulsions Rheology Foam (SERF) in Flow Assurance. She
niques and tools for complex phase behavior and flow assur- brings significant experience to these areas of flow assurance
ance studies. Before his graduate studies, he worked for the because of her background in physical organic chemistry,
analytical chemistry, chemical engineering, and strong prac-
tical laboratory skills in the areas of E&P, chemicals, and refin-
ing. Her interests also include EOR and asphaltenes. She holds
an MS degree from MUN in Newfoundland and a PhD degree
from the University of Calgary, both in chemistry, as well as an
MS degree in chemical engineering from the University of Tulsa.
Loek Vreenegoor joined Shell in 1990, working at the Shell Re-
search and Technology Center Amsterdam (SRTCA) in The
Netherlands with a focus on R&D and Technical Services. He
started in the area of polymer processing, then becoming in-
volved in modeling and computational fluid dynamics (CFD)
of viscoelastic flows. After a 1-year assignment at Westhollow
Technology Center (WTC) in Houston, where he worked on 3D
CFD of degrading polymers and multiphase flow, he returned
to SRTCA as the team leader in multiphase flow and became

Fig. 15—Friction component of pressure profile for 15 and 50%


water cut and 950 B/D.

292 August 2007 SPE Production & Operations


involved in three-phase flow modeling and slug mitigation. He Delft University of Technology, and he completed his PhD thesis
then joined the Flow Assurance group, working out of Amster- on the mathematical modeling of two-phase bubbly flows on
dam and Houston. Managerial responsibilities were combined the same faculty. Jamal Saleh has been a senior process/flow
with providing technical support, both in the design and op- assurance engineer with Shell International Exploration & Pro-
erational phase of upstream projects, varying from the Ormen duction since 2002. Before joining Shell, Jamal worked as a
Lange deepwater gas/condensate development in Norway senior flow assurance engineer at Intec Engineering for 2 years
to plugged flow lines in the Gulf of Mexico to heavy oil devel- and a process simulation engineer at Epcon International for 4
opments off the coast of Brazil. Vreenegoor holds a cum laude years. He holds a PhD degree in chemical engineering from
degree obtained at the Applied Mathematics Faculty of the Lamar University.

August 2007 SPE Production & Operations 293

You might also like