Загрузил Кирилл Дубовик

6. Rheological Characterization of Hydrate Suspensions in Oil Dominated Systems

Rheological Characterization of Hydrate
Suspensions in Oil Dominated Systems
Français du Pétrole, Division Chimie et Physicochimie Appliquées,
1 et 4 Avenue Bois-Préau, 92852 Rueil Malmaison Cedex, France
TOTAL, Cedex 47, 92069 Paris–La Défense, France
ABSTRACT: In the work described in this paper, the rheological behavior of
hydrate suspensions in an asphaltic crude oil was studied. A P-T cell adapted
to operate as a double cylinder configuration rheometer was used. The results
obtained show that the hydrate suspensions studied exhibit shear thinning
behavior and thixotropy. This behavior is more visible in suspensions with
higher particle content and seems to be related to colloidal interactions
between the crystals, which causes aggregation between crystals at low shear
The development of offshore oil exploitation at increasing water depths has highlighted the need for better knowledge about hydrate formation and transportation in
oil dominated systems. Indeed, in the deep subsea flowlines that link oil production
wells to their platforms the conditions needed for hydrate formation are found, such
as high pressures, low temperatures, and the presence of natural gas and water.
In oil dominated systems the injection of a thermodynamic inhibitor to avoid
hydrate formation is not feasible, due to the substantial water volume produced in
advanced phases of field production. Solutions may then involve the injection of low
dosage hydrate inhibitors (LDHI) such as kinetic inhibitors and dispersant additives,
which are presently in research and development phases. Dispersant additives do not
avoid hydrate formation, but prevent crystal agglomeration and, consequently, allow
crystals to be transported as a hydrate suspension.
It is well known that crude oils have natural surfactants, such as resins and
asphaltenes that stabilize water-in-oil emulsions.1,2 If these heavy polar components
also exert a dispersing effect, hydrate crystals formed in such a water-in-oil emulsion
would be transportable as a suspension, without the need for further chemical injection. This is the context in which studying the rheology of hydrate suspensions is
seen as one of the possible methods that may allow the prediction of hydrate transportation capabilities of crude oils under realistic production conditions.
Once the transportation of a given hydrate suspension is considered feasible, a
better knowledge of its rheological behavior is essential to improve the design of
aTelecommunication. Voice: 00 33 1 47 52 71 55; fax: 00 33 1 47 52 70 58.
multiphase pipelines, since viscosity is a very important variable in pressure drop
The main purpose of the work is to analyze the rheological behavior of water-inoil emulsions and of the hydrate suspensions formed from them, under the same
pressure and temperature conditions. A crude oil, rich in heavy polar components,
has been used.
A pioneer device and an experimental procedure have been specially developed
for this rheological investigation. The rheological cell is a pressure and temperature
cell that has been adapted to allow the evaluation of rheological characteristics of
hydrate suspensions.
The rheological cell allows one to work at a pressure up to 9 MPa and with temperatures ranging from 258.15 to 338.15 K. It has a double cylinder configuration
and works with imposed shear rate and measured couple. The rotation speed of the
mobile cylinder ranges from 0 to 600 rpm. The couple captors are able to perform
measurements between 0 and 100 mN ⋅ m, but magnetic coupling is only capable of
up to 35 mN ⋅ m.
F IGURE 1 shows a sketch of the rheological cell. The electric motor, at a constant
angular velocity, drives the mobile external cylinder by means of a magnetic coupler,
the couple needed to maintain this velocity being measured by two parallel couple
captors. The 100 cm3 transparent cell is located inside a thermoregulated compartment
FIGURE 1. Rheological cell sketch.
that provides temperature control. The emulsion bottle is used to inject the previously
prepared water-in-oil emulsion into the cell. The pressure in the cell is kept constant
by injecting gas from the gas bottle to compensate for the pressure drop due to temperature reduction and hydrate formation.
One important characteristic of this rheometer is that there are in reality two gaps
that contribute to the measured couple. There is one gap between the internal fixed
cylinder and the external mobile cylinder (internal gap), and another gap between
this external cylinder and the transparent cell wall (external gap). Thus, the viscosity
η of a Newtonian fluid in mPa ⋅ s, is given by
η = ------------------------------------- ,
4πω 0 ( G i – G e )
where C is the couple in mN ⋅ m and ω0 is the angular velocity in rad/s. Gi and Ge are
geometrical factors related to the two gaps that are given by
G i = ------------------------- and G e = ------------------------(2)
 ---- ----- – ------
- – ------
 R 2 R 2
 R 2 R 2
where the subscript i stands for the internal gap and e for the external gap. The height
of liquid is represented by h, R1 is the internal cylinder radius, R2 and R3 are, respectively, the internal and the external radius of the mobile cylinder, and R4 is the transparent cell radius. All these values are given in meters.
Since the rheological cell is an adaptation of a P-T cell and not a true rheometer,
it was necessary to perform calibration experiments in order to estimate its accuracy
FIGURE 2. Results from rheological cell calibration with Newtonian fluids.
and to be acquainted with its limitations. Two sets of experiments were performed.
The first was performed with calibrated Newtonian oils, the second with shearthinning non-Newtonian liquids.
The results of the experiments with the Newtonian fluids are shown in FIGURE 2
where the experimental values were calculated using Equation (1). In this figure,
each point is the result of a complete rheological test and represents the arithmetical
mean of the values obtained for the viscosity at each angular velocity. These calibration experiments were carried out before starting the tests with hydrate suspensions
and also between the hydrate tests in order to check that the device was still calibrated. Even though the rheological cell accuracy is far from what is expected for a true
rheometer, it is more than satisfactory for the purposes of this research.
Three aqueous solutions of a high molecular weight polyacrylamide were used
for calibration in a set of experiments with non-Newtonian liquids. In order to calculate the apparent viscosity µ as a function of the shear rate γ̇ the power law model
expressed by
µ = kγ̇ n – 1
was assumed. In (3), k and n are constants that characterize the rheological behavior
of a given fluid and were calculated from measured data. F IGURE 3 shows a comparison between the results obtained with the rheological cell and those obtained by a
Carri-Med rheometer, confirming that the rheological cell is also able to determine
the shear-thinning behavior of a given fluid.
FIGURE 3. Comparison between the results obtained by the Carri-Med rheometer and
by the rheological cell for non-Newtonian fluids (aqueous solutions of polyacrylamide with
concentrations of 1.0, 1.5, and 2.5%).
TABLE 1. Crude oil main properties
Density at 288.15 K
Dead oil viscosity at 293.15 K
mPa ⋅ s
Saturates + aromatics content
% weight
Resins content
% weight
Asphaltenes content (in heptane)
% weight
A complete experimental run includes emulsion preparation, its saturation in gas
at 298.15 K, cooling the fluids to the desired temperature, one rheological test for the
emulsion at that temperature (before hydrates formation begins), and the complete
process of hydrate formation, during which several rheological tests are performed.
A rheological test consists in recording the pair of values (C, ω0) while the angular
velocity slowly increases from 0 to 70 rad/s and, afterwards, decreases back to zero.
Between the rheological tests, the mobile cylinder is maintained at a constant angular velocity, usually 35 rad/s. The values of couple, pressure, temperature, angular
velocity, and gas consumption are recorded throughout the experiment as a function
of time. The pressure and temperature conditions used in this work are 8 MPa and
280.65 K.
The systems studied have three phases: liquid hydrocarbon, water, and gas. The
liquid hydrocarbon is an asphaltenic heavy crude oil from a Brazilian deep water
field. Its main properties are shown in TABLE 1. The water phase is salted water
FIGURE 4. Photomicrograph of a water-in-oil emulsion. Water content, 30% weight.
Amplification: 200 ×.
(33 g/l NaCl). A mixture of 90% molar of methane and 10% ethane was chosen as
the gas phase.
The method used to prepare the emulsion was to agitate the crude oil with a high
shear mixer Ultra-Turrax Model T25 running at 8,000 rpm for 180 seconds, while
slowly pouring the water on it. This procedure results in a very tight emulsion, with
a characteristic water-drop diameter of 2.5 µm (see FIGURE 4). Emulsions with 15,
30, and 50% of water content (by weight) were used.
It is very important to have a homogeneous fluid between the gaps. Thus, the
emulsions prepared must be very stable and also present a very low droplet settling
velocity. In order to check whether those requirements were satisfied, separation and
sedimentation tests were performed with the prepared emulsions, letting them rest in
graded test tubes maintained at 293.15 K. After a period of one month, neither separated water nor emulsion segregation were observed.
While working with crude oils, hydrate crystallization is not a visible phenomenon, thus we must be sure that it really did occur. F IGURE 5 shows how we realize
that hydrates were formed. After a period of time at the desired pressure and temperature conditions, we observe a sharp increase in the measured couple (and therefore
in the viscosity) caused by crystal formation, concomitant with a pronounced
decrease in cell pressure, due to the related gas consumption. The cell pressure
decreases because its connection with the gas bottle, which permits automatic pressure control, is kept closed during hydrate formation.
FIGURE 5. Hydrate formation evidence. Water content, 50%; T, 280.65 K; angular
velocity, 35 rad/s.
FIGURE 6. Rheological test for the pure oil. T, 280.65 K; P, 8 MPa.
As can be seen in FIGURE 6, the pure oil presents a Newtonian behavior, with a
viscosity of roughly 35 mPa ⋅ s.
The results for the rheological tests with 15% of water content (see F IGURE 7)
suggest a Newtonian behavior for both the emulsion and the hydrate suspension.
Nevertheless, it is known that emulsions are in general non-Newtonian fluids.1
Hence, the conclusion is that a possible non-Newtonian behavior for both emulsion
and hydrate suspension is not strong enough to be detected by the rheological cell.
Test A was performed at the firsts stages of hydrate formation. Tests B and C were
performed after the end of the crystallization process (end of gas consumption). An
increase in hydrate suspension viscosity was observed as more hydrates were
FIGURE 7. Rheological tests for the water-in-oil emulsion prepared with 15% water
content. T, 280.65 K; P, 8 MPa.
FIGURE 8. Rheological tests for the water-in-oil emulsion prepared with 30% water
content; T, 280.65 K; P, 8 MPa.
formed. The emulsion viscosity is 85 mPa ⋅ s, whereas the hydrate suspension viscosity is 260 mPa ⋅ s.
The rheological behavior of the 30% water content emulsion also seems to be
Newtonian (see FIGURE 8). However, a shear-thinning behavior for the hydrate suspension can be observed. This shear-thinning behavior, as well as the apparent viscosity itself, both increase with the quantity of hydrates formed (from tests A to C).
An increasing thixotropic behavior—expressed by higher viscosity while increasing
the angular velocity than while decreasing the angular velocity—was also observed.
FIGURE 9. Rheological tests for the water-in-oil emulsion prepared with 50% water
content. T, 280.65 K; P, 8 MPa.
This suggests an aggregation tendency between hydrate crystals at low angular
velocities, forming a structure that is broken at higher shear stresses. Accordingly, at
high angular velocities the apparent viscosity does not decrease any further and a
Newtonian region is reached. The emulsion viscosity is 200 mPa ⋅ s. After the end of
the hydrate formation process, the hydrate suspension had an apparent viscosity in
the Newtonian region of 400 mPa ⋅ s (test C).
Concerning the emulsion prepared with 50% water content, there has not been
any problem in performing rheological characterization of the emulsion. We can
assume that this fluid has a Newtonian behavior and its viscosity is 600 mPa ⋅ s (see
FIGURE 9). Nevertheless, at this water content, some instability problems were
observed after hydrate formation (see F IGURE 10). It seems that the high solid concentration suspension formed from this emulsion is a very heterogeneous system,
where direct contacts between particles play an important role in governing flow
characteristics. It can also be observed that, after a period of time, the viscosity
drops, reaching an even lower value than the emulsion viscosity. The most likely reason for this behavior is the sedimentation of the solid particles, since the contact
between them enlarges their typical diameters and, consequently, increases their settling velocities. Of course, a rheological test during the instability or sedimentation
phases would be meaningless. However, in FIGURE 9, test A was performed between
the first and the second hydrate formation process, hence, before the problems
referred to above began. The general tendencies of increasing the viscosity and more
pronounced shear-thinning behavior and thixotropy at higher water contents have
been confirmed in this rheological test.
In FIGURE 11, the results of apparent viscosity for each water content have been
gathered. For the cases in which a shear-thinning behavior was found, the values
at the Newtonian region have been used (Tests C). This value was obtained by
extrapolation in the case 50% of water content and represents only a rough estimate.
FIGURE 10. Apparent viscosity evolution during one experimental run with the
water-in-oil emulsion prepared with 50% water content; T, 280.65 K.
FIGURE 11. Comparison between emulsion and hydrate suspension apparent viscosity at different water contents; T, 280.65 K; P, 8 MPa.
The exponential growth in the hydrate suspension apparent viscosity with the water
content suggests that its transportability is limited by high viscosity before being
limited by the capacity of maintaining the crystal dispersion.
Despite the difficulties usually found in rheological characterization of emulsions
and suspensions at high pressures, this work allows some important preliminary conclusions. Among them, the main conclusions are:
• Besides being a very good emulsifier, natural surfactants contained in the oil
studied act also like a highly efficient dispersant additive for the hydrate suspension. Consequently, crystal transportation is expected under realistic production conditions, at least up to a given water content level.
• The water-in-oil emulsion viscosity increases as water is converted into
• As expected, the larger the water content, the higher the viscosity of both the
water-in-oil emulsion and the hydrate suspension. This suggests a limitation
on transportability of hydrate suspensions in pipes due to high viscosity.
• Hydrate suspensions show a shear-thinning behavior that is more pronounced
in more concentrated suspensions.
• The shear-thinning behavior seems to be related to colloidal interactions
between the solid particles, which causes aggregation between them at low
shear rates. This possibility is reinforced by a thixotropic behavior found in
the hydrate suspensions analyzed.
• Shear-thinning behavior in the hydrate suspension means difficulties for
restart operation after production shut down (very high viscosity at low shear
This work was performed within a collaboration program involving IFP, PETROBRAS, and TOTAL. The authors would like to thank these companies for their financial support and for permission to publish these results.
1. M ANNING, F.S. & R.E. T HOMPSON. 1983. Oilfield Processing Volume Two: Crude Oil.
PennWell Books, Tulsa.
2. M OURAILLE, O. et al. 1998. Stability of water-in-crude oil emulsions: role played by
the state of solvatation of asphaltenes and by waxes. J. Dispersion Sci. Technol. 19
(2/3): 339–367.