SPE 65007

Successful Applications of Anti-agglomerant Hydrate Inhibitors

L.M. Frostman, SPE, and J.L. Przybylinski, Baker Petrolite

Copyright 2001, Society of Petroleum Engineers Inc.

support, however, the whole structure collapses. To
This paper was prepared for presentation at the 2001 SPE International Symposium on Oilfield understand gas hydrates, water must be understood first.
Chemistry held in Houston, Texas, 13–16 February 2001.
The physical properties of water are dominated by the
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
formation of hydrogen bonds between water molecules.
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to Hydrogen bonds are much weaker than chemical bonds, yet
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at much stronger than the dipole-dipole interactions (van der
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
Waals forces) that allow most materials to form condensed
for commercial purposes without the written consent of the Society of Petroleum Engineers is phases. Without the existence of hydrogen bonds, water would
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous be a gas that would be more difficult to liquefy than carbon
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
dioxide, oxygen or nitrogen.
One of the most recognized peculiarities of water is the
fact that it expands on freezing. This clearly shows that solid
Abstract water has a more “open” structure than liquid water. While
While it is important to approach each system individually rare, this phenomenon is not nearly as unusual as the fact that
when dealing with new types of performance chemicals, the the density of water increases as it is heated from 0°C to 4°C.
anti-agglomerants have been found to be suitable for hydrate This behavior shows that this open structure, which is the
control in most of the situations encountered in offshore oil result of hydrogen bonding, persists in liquid water.4
production. Anti-agglomerants have now been used in several When water freezes it releases an extraordinary amount of
field applications in deepwater systems - both in subsea wells heat as additional hydrogen bonds are formed. Clearly the
and dry tree wells - under both flowing and shut-in conditions. formation of solid water is energetically favored, but it is
During these applications, no hydrate problems were opposed by the strict order (entropy) that is created by its
observed. The anti-agglomerants caused no upsets in the formation. Gas hydrates form because the gas molecules
topsides operations; there were no emulsion problems, no dissolved in the water help support an open, ordered system
foaming problems, and no water quality problems. The thereby allowing a solid water phase to form at a higher
products were deployed through the existing methanol temperature. The action of the gas is primarily related to its
injection systems without incident. Application rates were space filling geometry. Molecules of similar size contribute
significantly lower than if methanol were used. Potential similarly to the stability of the gas hydrate. This model is
advantages of the anti-agglomerants over methanol include supported by the fact that the standard entropy of solution for
smaller umbilicals, smaller pumps, smaller storage facilities, the series of hydrocarbons from methane to butane increases
and less frequent supply trips. in absolute magnitude as the size of the molecule increases.5
This implies that larger molecules create more order.
Introduction Of the straight chained hydrocarbons, only methane,
The formation of natural gas hydrates is the most universal of ethane and propane support gas hydrate formation by
the many flow assurance challenges faced during production themselves. The branched isobutane and the cyclic alkanes
of oil and gas from deepwater reservoirs. As noted by Mehta, through cyclohexane also support gas hydrate formation. A
et. al.,1 deepwater production conditions are ideal for the number of inorganic compounds support hydrate formation.
formation of hydrates. Thus there is a need to improve both Among them are oxygen, nitrogen, nitrous oxide, carbon
the system design and operating procedures. dioxide, hydrogen sulfide, sulfur dioxide, chlorine and xenon.
Although the existence of gas hydrates was recognized by One of the few compounds that is small enough, non-volatile
Davy2 nearly two hundred years ago, it took over a hundred enough and sufficiently soluble in water to support hydrate
years before their formation was recognized as a problem in formation at atmospheric pressure is tetrahydrofuran.
gas transmission lines.3 Gas hydrates are essentially Methanol, ethanol and isopropanol are the right size to form
compounds of water in which the gas molecules play, both hydrates, but the fact that these compounds hydrogen bond
figuratively and literally, only a supporting role. Without this
2 L.M. FROSTMAN, J.L. PRZYBYLINSKI SPE 65007

with water seems to keep them from acting as hydrate formers. The critical size depends directly on the interfacial energy
Charged species also do not support hydrate formation. and inversely on the thermodynamic driving force, which is
Once a hydrate forming gas is present, some other the free energy of formation. The rate of crystal formation is
molecules can participate by occupying larger structures in the low for small driving forces, rapidly becomes larger at
gas hydrate crystal. These molecules include hydrocarbons intermediate driving force and reaches a limiting value at very
that range up to highly branched seven carbon molecules such high driving force.
as 2,2,3 trimethylbutane and the eight carbon cyclooctane. By analogy with the mineral scale formation and other
common crystallization processes, one may consider a gas
Hydrate Control. Since gas hydrates are like water ice in that molecule and the sphere of water molecules it keeps arranged
they consist of stable hydrogen bonded structures, they can be around it as a growth unit. A number of these units may come
prevented in the same way as ice. The simplest means of together by chance. If enough come together the assembly will
doing this is to add salts or other soluble compounds to the become stable and begin to grow. It is not at all clear how to
water. Most common among these are sodium chloride, assign an interfacial energy to this assembly. In order to define
calcium chloride, methanol and ethylene glycol. These such an energy, the surface must be defined. Defining the
compounds lower the formation temperature of gas hydrates surface of such a gas hydrate assembly is more difficult than
by about the same amount as they lower the freezing point of in the case for a mineral scale because the surface of a hydrate
ice. In fact, methanol injection is the preferred hydrate control consists of solvent (water) molecules. Still it is clear that there
method in many systems. However, to get a large decrease in must be some interfacial energy or the growth rate of hydrates
the formation temperature, large amounts must be used. This would be as fast at low driving force as at high driving force.
can create a logistics problem, especially for offshore If a gas hydrate inhibitor can stop the growth of the
locations. assembly when it is small enough, it is more likely to fall apart
For the past ten years or so there has been increasing than to grow. Unlike mineral scale inhibitors, kinetic hydrate
interest in low dosage inhibitors. There were certainly clues inhibitors seem to be unable to check the growth of larger
that such compounds exist even before specific compounds crystals. This may be a fundamental limitation due to the fact
were discovered. Hydrates do not always form plugs when that hydrates may be able to grow on the entire surface rather
conditions are favorable.6 This indicates that there are than just at certain growth sites, or it may be only a limitation
compounds in some natural fluids that inhibit the formation of of the currently available compounds.
hydrate plugs. It is established that certain proteins in some Anti-agglomerant low dosage hydrate inhibitors are
fish and insects prevent the formation of ice crystals.7,8 By believed to specifically interact with the growing hydrate
extension, comparable compounds would be expected to exist crystal surface.9 These products allow hydrates to form but
for inhibiting the growth of hydrates. inhibit agglomeration, deposition, and plugging. It is likely
Low dosage inhibitors have at least two modes of action. that anti-agglomerant LDHI’s are incorporated into the
Those that are referred to as kinetic inhibitors prevent any hydrate crystal lattice and serve to disrupt further growth of
hydrates from forming for a period of time. When they do fail, the crystals. Hydrophobic moieties in the anti-agglomerant
however, they may fail catastrophically. Those that are LDHI’s then serve to keep individual hydrate particles
referred to as anti-agglomerants may not completely prevent separated and dispersed in the liquid hydrocarbon phase.
the formation of hydrates, but they do keep the hydrates from The thermodynamic driving force for hydrate formation is
forming solid plugs. Anti-agglomerants usually also exhibit difficult to define. Christianson et. al., showed one method of
some kinetic inhibition activity. calculating the driving force.10 His data showed a steeply
The mechanism(s) by which low dosage inhibitors work is increasing rate of hydrate growth with increasing driving
not well established. Their means of checking gas hydrate force. This is in agreement with the nucleation model.
formation may be analogous to that of mineral scale inhibitors, For convenience a surrogate quantity, “subcooling,” will
whose mode of action is on much firmer ground. Mineral be used for driving force. This quantity is defined as the
scales form from solution by first forming a chance assembly difference between the equilibrium hydrate temperature and
of growth units. If the radius of an assembly is smaller than a the operating temperature at a fixed pressure and composition.
“critical radius”, the assembly is more likely to fall apart than For a given system, the subcooling gives an intuitive measure
to grow. Eventually some assembly that has a radius greater of the driving force, but it may fail in system to system
than the critical radius will form; it will grow to macroscopic comparisons. Yousif5 found that hydrates formed in a matter
size. of minutes with a subcooling greater than 20°F, but took over
Mineral scale inhibitors operate by blocking growth sites 28 hrs with a subcooling less than 13°F. These data are
on the surface of the assembly. Assemblies that are prevented consistent with the nucleation model and the use of subcooling
from growing as large as the critical size fall apart, and any instead of free energy of formation.
that are greater than the critical size are prevented from Kinetic inhibitors seem to have certain limitations on the
growing any larger. An assembly of critical size may comprise amount of subcooling they can sustain. Kelland et. al. in 1995
from several thousand to hundreds of thousand of growth found them limited to subcoolings of “perhaps 10°C at
units. most”.11 Whereas he found anti-agglomerants “do not appear
to be dependent on the subcooling”. The graphs in Ledehos et.
SPE 65007 SUCCESSFUL APPLICATIONS OF ANTI-AGGLOMERANT HYDRATE INHIBITORS 3

al. illustrate that when kinetic inhibitors fail, the rate of 1°C (2°F) of subcooling. Leporcher et. al. reported that a
hydrate formation may exceed the rate seen without kinetic inhibitor replaced methanol with subcooling conditions
inhibitors.12 up to 8°C (14°F).17
Anti-agglomerants do have a fundamental limitation in that Even more recent presentations do not present field trials
they require a liquid hydrocarbon phase to be present in order under more severe conditions. Fu reported field tests of
to transport hydrates that may form. Laboratory studies have poly(VIMA-VCAP), a kinetic inhibitor.18 Without inhibition,
demonstrated the breadth of applicability of anti-agglomerant in the first system, hydrates were detected in eight hours and
hydrate inhibitors.9 These tests indicate that the products plugged the line in two days. Over a total test period of several
perform well in systems containing condensates and black months the subcooling varied from 5 to 16°F. It was found
oils. The oils have included paraffinic, non-paraffinic, that the amount of inhibitor required to control hydrate
asphaltic, and non-asphaltic types. Good performance has formation varied at a greater than linear rate with increasing
been observed in fresh and highly saline produced water with subcooling. In a second application the line was pigged
water cuts up to 70%, subcoolings up to 40°F, and shut-ins up weekly and the maximum subcooling was 10°F. In the
to two weeks. absence of inhibitors, enough hydrates formed in two days to
cause a stuck pig. This trial lasted at least two months and
Commercial Applications of Kinetic and Anti- included a 70 hour shut-in. The kinetic hydrate inhibitor
Agglomerant Hydrate Inhibitors. It is often difficult to controlled hydrates throughout this test.
introduce new types of oilfield production chemicals. With the Mitchell and Talley reported on the successful test of
consequences of failure so severe, operators are reluctant to poly(VIMA-VCAP) in a black oil system at a choke outlet.19
risk field trials of experimental chemicals. This has been Because of the short residence time, they were able to use very
particularly true of low dosage hydrate inhibitors, where the low levels of inhibitor. Maximum subcoolings during flowing
existing technology – methanol injection – is tried and true. conditions were about 7°F. During shut-ins, which did not last
Recently, the shortcomings of methanol injection have been longer than six hours, subcoolings increased up to 14°F.
noted. These include high volumes, high transportation costs, These reports of successful field tests of kinetic inhibitors
high capex costs, and potential effects on downstream leave little doubt that kinetic hydrate inhibitors can be used up
processes. to at least 16°F subcooling if the shut-in time is kept to a
In the past half dozen years there have been a number of minimum.
reports on the successful field application of low dosage Reports of field tests on other types of low dosage
inhibitors. One of the earliest series of field tests was reported inhibitors are much rarer. Several field trials of anti-
by Notz et. al.13 The tests in Wyoming and Texas showed that agglomerant hydrate inhibitors have now been performed.
the first generation kinetic inhibitors could be successfully Reported herein are summaries of three field trials in the
applied “when the fluids are not too far into the hydrate region deepwater Gulf of Mexico using two different anti-
(less than 5 to 10°F subcooling) and/or if the residence time is agglomerant low dosage hydrate inhibitors, LDHI A and
less than a few minutes.” These tests ran up to 36 days. LDHI B. LDHI A is suitable for systems with salinities in
Corrigan et. al. reported on field trials of a “threshold excess of 15,000 ppm TDS. LDHI B performs well even in
hydrate inhibitor”, apparently what we refer to as a kinetic fresh water. It too will work at higher salinities, but it is less
hydrate inhibitor.14 These inhibitors were successfully tested 8 cost effective than LDHI A.
to 9°C (14 to 16°F) inside the equilibrium hydrate curve. The
tests ran up to 17 days. There were indications that hydrates Field Trial #1
formed after 30 hours in the absence of inhibitor and also Our first field trial of LDHI A took place in a deepwater Gulf
formed 50 hours after injection of the hydrate inhibitor was of Mexico subsea flowline. Details of the trial have been
terminated. Corrigan stated that this chemical could be used documented previously.20 The subsea well produced 3600
up to 10°C (18°F) subcooling, but that for more severe BOPD, 0-10 BWPD, and 2.3 MMSCF/D via a 4 mile
subcooling a different type of inhibitor can be used. He uninsulated flowline. Fluids arrived at the platform at 600 psig
referred to these as hydrate dispersants, apparently what we and 42°F, roughly 13°F below the predicted hydrate stability
refer to as anti-agglomerants. temperature. During shut-ins, the flowline pressure can reach
Bloys and Lacey reported a field test of Gaffix VC-713, a 2500 psig, thus posing a severe hydrate formation problem
kinetic inhibitor.15 The total test ran for seventeen days, of (18°F subcooling). Prior to this field test, hydrates were
which three to four days was at a subcooling of 4°C (7°F), controlled by injection of methanol.
four days at 2°C (4°F), and three days at 1°C (2°F). The After an extensive screening process, Baker Petrolite
maximum subcooling at which they were able to control formulated LDHI A for use in this subsea well. Extensive
hydrate formation was 9°C (16°F). At 12°C (22°F) hydrate cell experiments demonstrated definitively that the
subcooling, hydrates formed at steady production rates. product readily controlled hydrates in field oil under
Other field tests reported similar limitations with kinetic conditions similar to those expected in this system.
inhibitors. Argo et. al. reported a test with a total of seventeen After confirming product efficacy, the customer’s next
days of production and thirteen days of shut-in.16 The greatest concern was to ensure the product would have no
maximum subcooling was 8°C (14°F) and the minimum was detrimental effects on the production system. Capillary
4 L.M. FROSTMAN, J.L. PRZYBYLINSKI SPE 65007

stability tests were thus performed at both high (190°F) and sufficiently. Given the small diameter of the capillary strings
low (40°F) temperatures. Any increase in viscosity, solids (0.277 inch ID), the maximum methanol injection rate was
deposition, or tendency to plug would manifest itself by a insufficient to completely prevent hydrate formation during
rising differential pressure across the capillary column. LDHI shut-in at the existing water production rates for Wells A and
A showed no such rise in differential pressure at either B. While neither well has ever plugged with hydrates, the
temperature, thus verifying the product’s suitability for operator deemed the risk of this procedure too high.
injection through a long umbilical and downhole capillary. To eliminate this risk, the fluids in the well after shut-in
Extensive corrosion and materials compatibility tests verified were displaced below the SCSSV with diesel. This required 98
that the product was compatible with all materials present in bbl diesel/well and resulted in the wells being shut-in for
this system. Chemical compatibility tests demonstrated that longer periods of time. Because of the complex operating
LDHI A was compatible with both methanol and the paraffin procedures and extra downtime associated with this technique,
inhibitor in use in the subsea well. With a viscosity of 23 cP at the operator sought an LDHI trial.
40°F, LDHI A could readily be applied through the umbilical Prior to the field test, numerous laboratory experiments
with a small chemical injection pump. On-site bottle test in the were run to ensure that LDHI A was suitable for application in
produced fluids suggested no emulsion problems would be Wells A and B. The field test was successful: the wells were
created by employing LDHI A. shut-in for six days and no problems were encountered on
The 1.5 month field test successfully demonstrated the startup. A total of 160 gal of LDHI A was used between the
applicability of LDHI A in a subsea tieback. After the two wells. Following the field test, modeling of the hydrate
methanol injection line was flushed with LDHI A, stability curve aided interpretation of the results. Described
approximately 100 gal of LDHI A were injected into the below are the laboratory tests, the field test, and the hydrate
flowline. Even at this high concentration (two orders of modeling used to interpret the test.
magnitude higher than the normal treating rate), LDHI A did
not upset the production system. The injection rate was Laboratory Testing.
quickly reduced to 1 gal LDHI A per barrel of water produced Hydrate Inhibition. The gas hydrate test apparatus
(the lowest injection rate achievable with the existing pump). consists of a high-pressure sight glass cell containing a
No hydrate problems were detected during LDHI A injection, stainless steel ball. In a typical experiment, the cell is charged
even during two brief shut-ins. LDHI A did not adversely with brine and condensate, then pressurized to the desired
affect overboard water quality: both oil & grease counts and level with a gas mixture. The cell is then rocked (to simulate
aquatic toxicity tests remained within the normal operating flow conditions) or held static (to simulate shut-in) in a
range of this platform. LDHI A did not cause any emulsion temperature-controlled bath. The cell is checked periodically
problems and BS&W counts remained low. for the formation of hydrates; this includes monitoring the
pressure and visual observation through the sight glass.
Field Trial #2 The absence of hydrates at the end of a test is considered a
Background. The second anti-agglomerant LDHI trial took “pass.” Additionally, the test is considered a “pass” if hydrates
place in two dry tree wells located on a deepwater Gulf of form but do not agglomerate, do not stick to the glass or ball,
Mexico platform in 2900 ft of water. At the time of the field and do not cause a plug anywhere in the vessel. Any evidence
test, 15 dry tree wells and 1 subsea well were operating, with of hydrates sticking to the glass or ball, forming a plug, or
total production rates of 156,000 BOPD, 178 MMSCFD, and freezing the ball in place is considered a “fail.”
1300 BWPD. Wells A and B were the only ones producing Prior to the field test, only oil from Well A was available
significant amounts of water, and so were the subjects of this for hydrate testing. This oil was believed to contain about 11%
test. Well A produced roughly 4200 BOPD while Well B water emulsified in it. Because this test procedure relies on
produced over 11,000 BOPD. Both wells had GOR’s of ~1100 visual observations, it is usually necessary to have a water cut
scf/stb and each made approximately 500 BWPD. Well A had of approximately 50% when testing black oils. At lower water
a water cut near 11% with a salinity of 18 wt% salt, while cuts, it is difficult to see the hydrate phase. Thus, supplemental
Well B’s water cut was 4% and the salinity 15 wt%. Because water was added to the oil to bring the water cut up to 50%.
the total amount of water produced was less than 1% of the Since these test cells are limited to 1400 psig at room
total oil volume produced, the water was not separated from temperature, the degree of subcooling achievable is also
the oil. Conditions during flowing and shut-in conditions are limited. Rather than trying to mimic the high salinity produced
shown in Table 1. water, a lower salinity brine (3.5% salt) was used for the
During shut-in and restart, Wells A and B may be in supplemental water. This helped to keep a higher level of
danger of forming a hydrate blockage. Both wells have subcooling in the test. It should be noted that LDHI A
completion fluid in the annulus, so there is little insulation typically performs better at higher salinities and lower water
from the ambient water temperature. The normal procedure cuts, so substituting a low salinity / high water cut system
was to dose the wells with methanol prior to shut-in, and gives a conservative model of the field. A high propane gas
continue pumping methanol for at least an hour after shut-in. was used to raise the subcooling in the test. The test conditions
Methanol injection was resumed at least an hour before the are summarized in Table 2. The cells were held static for 23
wells were started up and continued until the well warmed
SPE 65007 SUCCESSFUL APPLICATIONS OF ANTI-AGGLOMERANT HYDRATE INHIBITORS 5

hours. After shut-in observations were recorded, the cells were treating rate for LDHI A is 0.72 gal/bbl of water. Since both
agitated for 6 hours. wells were making about 500 BWPD, the target treating rate
A total of three tests were performed: one blank and two was 360 gpd LDHI A under normal flowing conditions. At
containing LDHI A. The LDHI A sample used in these tests this rate, the estimated time to fill each capillary was 60
was taken from the actual batch sent to the field trial. The minutes. Prior to shut-in it was necessary to treat the tubing
dosage rate used for LDHI A was equivalent to 0.72 gal/bbl volume above the SCSSV so that any fluids remaining above
water. The contents of all three test cells remained fluid after the SCSSV during shut-in would not form a hydrate plug. The
the 23 hour shut-in and for the first 30 minutes of rocking. estimated times to do so were 30 minutes for Well A and 12
After 6 hours of rocking, the blank test cell was plugged solid minutes for Well B.
– no free fluid remained. The contents of both cells treated Both the Well A and Well B capillaries were filled with
with LDHI A were still fluid. As noted previously, these tests LDHI A. Data collected during the field test are shown in
were at more severe conditions than those predicted for the Figures 2-3 for Well A and Figures 4-5 for Well B. Note that
field test. The good performance seen in the laboratory sug- during shut-in, the wellhead pressure gauges are downstream
gested that LDHI A would control hydrates in Wells A and B. of the closed SSV valves, so the measured pressures (Figures
Viscosity Measurements. As noted above, the viscosity of 2 and 4) are not representative of the shut-in tubing pressure.
LDHI A is relatively low, thus allowing sufficient product to Wells A and B were dosed with LDHI A at 415 gpd and
be delivered through the capillary strings. 370 gpd respectively. The methanol pump was unstable at
Materials Compatibility. Compatibility of LDHI A lower flowrates. In general, the higher the injection pressure,
(sampled from the batch sent to the field trial) with 22 Cr was the lower the flowrate which the pump could achieve. Note
tested to ensure the product could be applied through the that the injection pressure for Well B was higher than for Well
capillary string. No weight loss (0.0 mpy) was seen at 140°F A, thus the rate for Well B was closer to the target injection
and 3 weeks. When examined under the microscope, the rate. Produced fluid samples from each well were collected
coupon showed no localized attack. In addition, the fluid and the presence of LDHI A was confirmed in each. Once it
remained bright and clear, free of solids. Thus LDHI A was was clear that each well had been properly dosed, the SSV
deemed appropriate for use with 22 Cr under the expected was closed, followed by the SCSSV. LDHI A injection was
conditions. then stopped. Between the two wells, a total of 63 gal of LDHI
Effect on Defoaming. Prior to the field test, it was A was injected prior to shut-in.
necessary to demonstrate that LDHI A would have no adverse The shut-in lasted a total of six days. On the fifth day of
effects on defoaming at the platform. Laboratory sparge tests the shut-in, the SCSSV’s of all platform wells were opened in
were thus run on Well A oil, with and without LDHI A, and anticipation of the restart. This was accomplished by
with and without the platform’s preferred defoamer. Sparge pressuring up on top of each SCSSV using diesel. A day later,
tests were performed using 20 psi nitrogen at ambient the platform was restarted well by well.
temperature. Well A oil was added to a fine frit column. LDHI A injection was started as the Well B choke was
Nitrogen sparging was started, with a regulator upstream of being opened at the beginning of the 4 hour ramp up. Initially,
the column set to give an appropriate gas flow rate. Foam when the well was flowing very slowly, the LDHI A injection
growth was recorded at 30 second intervals. After the foam was tuned as low as possible (125 gpd). At this point, the well
height leveled off, the nitrogen was turned off and the foam was producing only gas. As the well production increased, the
decay was recorded at 15 second intervals. LDHI A injection rate was increased. The minimum wellhead
LDHI A has appreciable solubility in the oil phase. The temperature recorded was 60.3°F at 75 minutes after initiating
LDHI A dosage rate for these tests was determined by the restart. At this point, the well had begun producing oil, but
assuming 100% of the LDHI A partitioned to the oil phase. the measured BS&W was still zero. About halfway through
Figure 1 shows the results of these tests. Note that the the ramp up, the well had begun to produce a small amount of
expansion factor is the total height of the foam divided by the water (0.5% BS&W) and the wellhead temperature had
initial liquid height. As shown in Figure 1, LDHI A does not warmed to 69.7°F. LDHI A injection was stopped soon there-
increase the foaming tendency of the Well A oil, nor does it after. A total of 14 gal LDHI A was used to restart Well B.
adversely affect the performance of the defoamer. The procedure followed for restart of Well A was similar
Effect on Emulsion Tendencies. Since produced water at to that of Well B, except that the scheduled ramp up time for
the platform was not separated from the oil, emulsion Well A was only 2 hours. LDHI A injection was started
tendency was not a primary concern prior to the field test. coincident with the opening of the choke. The minimum
Emulsion tendency tests were thus not performed. It was noted injection rate achieved was 380 gpd LDHI A; below this, the
however during the course of the laboratory tests that LDHI A pump was unstable. A sample taken at 45 minutes later
appears to help demulsify both the Well A and Well B oils. showed oil, but no water. The minimum wellhead temperature
This is consistent with previous observations.20 of 61.4°F was recorded 55 minutes after restart. Within
another 45 minutes, the wellhead temperature had warmed to
Field Test Summary and Analysis. Injection of LDHI A was 71.4°F. LDHI A injection was stopped soon thereafter.
accomplished via the methanol injection capillary which Samples taken at this time had an average water content of 30-
enters the tubing just above the SCSSV. The recommended 38%. A total of 51 gal LDHI A was used to restart Well A.
6 L.M. FROSTMAN, J.L. PRZYBYLINSKI SPE 65007

No defoamer was injected into either Well A or Well B. According to operations personnel, the temperature-
No foaming problems were noted during restart of the pressure profiles observed during this test were typical of
platform. Overall, the application of LDHI A did not cause other untreated wells at the platform. When the wells were
any difficulties in restarting the wells. No signs of hydrates first opened, the wellhead pressure dropped rapidly, only gas
were observed. As noted previously, water was not separated was produced, and the wellhead temperature remained fairly
from the oil during production at this platform. Thus, constant. When the pressure dropped sufficiently low, oil pro-
overboard water quality was not an immediate concern during duction began and pressure increased as the well started to un-
this field trial. However, ecotoxicity testing of water samples load. Once oil production began, the fluids were cooled by the
taken during the injection of LDHI A indicated that the water ambient water, resulting in a decreasing wellhead temperature.
would have been dischargeable under the current guidelines in As the choke was opened further, the flowrate increased,
the Gulf of Mexico. lowering the wellhead pressure while warming up the tubing.
Untreated Well A and B oil samples were well emulsified The onset of water production gave rise to further increases in
and showed little or no water drop in the absence of a wellhead temperature and decreases in wellhead pressure.
demulsifier. In contrast, all samples treated with LDHI A In the past, the Well A tended to load up with water during
showed ready water separation. shut-in. The same was observed here, with water cuts reaching
LDHI A injection pump pressures were lower than at least 48%, well above the 11% typically seen prior to shut-
expected during the trial. This was likely due to using too in. The variation in water cut during restart seen in both Wells
conservative an estimate (40°F) for the average temperature of A and B raises some issues for optimizing hydrate inhibitor
the capillary. Based on the field data, it is estimated that a injection during restart. Both wells were likely overtreated
maximum of 2200 gpd LDHI A could be injected into either during this field test, as neither well showed significant water
of these wells prior to shut-in. On restart, the LDHI A will production until the wellhead temperature had begun to rise,
certainly be colder and thus of higher viscosity. It is estimated yet the target injection rate was based on the normally
that the maximum injection rate would be 1300-2000 gpd produced water cut. Some degree of overtreating does seem
LDHI A initially on Well A and 1100-1700 gpd LDHI A prudent given the lack of real-time water production measure-
initially on Well B. As noted during the field trial, the water ment and the tendency of in situ water cuts to be higher than
production rate is initially slower than prior to shut-in. As the the measured water cut at the wellhead.. The extent of
wells warm up, and coincidentally the water cut rises, the overtreating should be based the operator’s comfort level.
maximum injection rate should rise back to 2200 gpd for each Of bigger concern would be the case of a subsea
well. application. In both wells, it took just a few hours for the
Thus the use of the anti-agglomerant significantly extends wellhead temperature to reach 70°F. A subsea flowline could
the ability to treat for hydrates via downhole chemical take significantly longer to warm up. If the restart hydrate
injection in these dry tree wells. Whereas methanol injection inhibitor injection rate is selected based on the typical
was insufficient to handle 500 BWPD (11% and 4% water cut produced water cut, and the well loads up with water during
on Wells A and B, respectively), use of LDHI A would be shut-in, there is a risk that at some point during restart the
feasible for up to ~3000 BWPD (42% and 21% water cut on fluids will be undertreated. Note that this concern applies to
Wells A and B, respectively, at the current oil production methanol as well as LDHI’s.
rates). At that point, an optimized reformulation of LDHI A,
LDHI D, could be employed. LDHI D provides an even better Interpretation of Field Test Results: Hydrate Stability
deliverability due to a lower viscosity and roughly doubles the Predictions. The Multiflash gas hydrate modeling software
volume of water which can be treated (6000 BWPD) from Infochem allows the thermodynamic gas hydrate stability
compared to LDHI A. This would extend the treating curve to be generated for a given hydrocarbon composition.
capability of this procedure up to 58% and 35% water cut on These results predict the conditions (pressure and temperature)
Wells A and B, respectively, at the current oil production at which gas hydrates will be stable. Thus for a particular field
rates. situation, the degree of subcooling experienced can be deter-
LDHI A residuals in the produced water were measured mined, where subcooling is the difference between the hydrate
both on site (pre-shut-in) and in the laboratory (pre-shut-in and stability temperature and the actual operating temperature.
restart). Measured water phase LDHI A residuals just prior to The reservoir fluid composition for Well A was obtained
shut-in and during restart were 440-500 ppm in Well A and from the operator. Since both Wells A and B flow from the
350-380 ppm in Well B. These measured values are consistent same formation, the same hydrocarbon composition was
with past laboratory work and the previous field trial.20 Note assumed for both. The effect of salinity was also accounted
that the actual application rates were 19000 ppm and 17000 for, using 18 wt% salt for the Well A water and 15 wt% salt
ppm based on the water for Well A and Well B, respectively, for the Well B water.
so over 97% of the LDHI A partitioned to the oil phase under The hydrate stability curves for these reservoir fluids are
these conditions. Note that the partitioning of LDHI A shown in Figure 6 as a function of salinity. As shown in Table
between the oil and water phases does not appear to affect 1 and Figure 6, the FWHT and FWHP coordinates for both
performance. Rather, it is the ratio of total LDHI A applied to wells lie to the right of the corresponding hydrate stability
water present that determines performance. curves. Thus hydrates are unstable at these conditions. During
SPE 65007 SUCCESSFUL APPLICATIONS OF ANTI-AGGLOMERANT HYDRATE INHIBITORS 7

normal flowing operations, neither Well A nor Well B fluids Results and Conclusions. To summarize the findings of this
enter the hydrate region. field trial:
The shut-in temperature and pressure profiles in the wells 1) LDHI A passed all performance tests and satisfied all
are unknown. However, the range of possible conditions can criteria for use in Wells A and B, including:
be determined. The coldest temperature during shut-in would a) LDHI A showed good performance as an anti-
be near the mudline at 40°F. The shut-in tubing pressures at agglomerant in laboratory tests of the Well A.
the wellhead can also be estimated from the wellhead b) The viscosity of LDHI A was low enough to pump
pressures after the SCSSV was reopened: 1555 psig for Well through the capillary string at sufficient flow rate.
A and 1845 psig for Well B. Presumably, the worst-case c) LDHI A was compatible with the 22 Cr capillary
hydrate formation conditions will occur near the mudline string.
where the temperature is coldest. The pressure at this point can d) LDHI A did not adversely affect defoaming in the
be estimated knowing the wellhead pressure and the weight of Well A oil.
the fluid column above the mudline. The minimum pressure at e) LDHI A was soluble in methanol.
the mudline, assuming the SCSSV is closed, would occur if f) LDHI A tended to demulsify the oils from Wells A
the tubing above the mudline contained 100% gas. Similarly, and B.
the maximum pressure at the mudline would occur if the 2) A total of 160 gal LDHI A was used to protect the two
tubing above the mudline contained 100% oil. (Note that wells during shut-in and restart. Less chemical could have
different conclusions would be drawn if the SCSSV were left been used with a smaller pump.
open during shut-in). The results of these calculations for both 3) Both wells were restarted without plugging. No problems
wells are shown in Table 1. The actual pressures for the two were encountered during the operation.
wells would then lie in between these minimum and maximum 4) Water production during restart was uneven. It would
pressures. therefore be difficult to optimize restart application rates.
Using these pressure ranges and the mudline temperature Note that this problem is the same for methanol as for
of 40°F, the hydrate temperature at the mudline can be LDHI A.
estimated from Figure 6. As shown in Table 1, this gives 5) The use of an anti-agglomerant for control of hydrates in
hydrate subcooling ranges of 12 – 16°F for Well A and 17 – these wells should continue to be an option, even if the
20°F for Well B near the mudline. This is certainly severe water production increases twelve-fold.
enough to form hydrates near the mudline. The severity of The operator indicated that they considered this a
hydrate formation conditions would decrease as you move successful test. The operator further noted that using the LDHI
away from the mudline. At both the wellhead and the SCSSV, A allowed them to produce Wells A and B for an extra twelve
no hydrates are expected. hours compared to the previous technique of bullheading
The question remains whether sufficient water was present diesel into the wells. This resulted in an extra 7500 bbls of
near the mudline to cause a plug in the absence of hydrate accelerated oil production. The operator has since instituted
inhibitor. This is a difficult question to answer. A combination use of LDHI A as the standard operating procedure for
of transient and steady state modeling could give an estimate planned shut-ins and restarts on this platform.
of the water distribution during shut-in and restart. Still this
information would only suggest whether a plug was possible. Field Trial #3
At present, it is exceedingly difficult to predict plugging Our third field trial of an anti-agglomerant, and first of the low
tendency. An alternative would be to do a “blank test,” where salinity product LDHI B, also took place in a dry tree well on
the system is shut-in and restarted using normal protocols, but a deepwater Gulf of Mexico platform. This was a brand new
no hydrate inhibitor is injected. If the system plugs, then well in 4800 feet of water that initially produced 9000 BOPD
hydrate inhibition should certainly be used for all subsequent and 13 MMSCFD at less that 0.5% water cut. As with the
shut-ins. Even if the system does not plug, the test is only previously described dry tree wells, this well does not need
valid for the current water cut. As the water cut rises, plugging hydrate control during normal flowing conditions. Again, due
tendency will tend to increase, assuming all other conditions to the limited methanol injection capability, the operator
stay the same. Thus the system would need to be reevaluated decided to pursue the option on using an anti-agglomerant
over time. Clearly, such “blank tests” are risky and can be inhibitor for planned shut-ins and startups. The FWHP is
costly in terms of deferred production and remediation costs. ~2600 psi and the SITP, while not directly measured, is
Most operators choose a more conservative route and merely expected to exceed 3000 psi. Coupled with the expected
assume that hydrate protection is needed. seafloor temperature of less than 40 F, it is clear that the riser
Even less is known about the hydrate formation conditions is well within the hydrate region during shut-in.
during restart, as the only temperature and pressure sensors are Initially, it was unclear how high the salinity would be, so
on the wellhead. Again, transient modeling could at least pre- LDHI B was chosen. Oil samples from this well were not
dict the temperature and pressure throughout the tubing as a available prior to the field trial, so an analogue oil was used to
function of time, but plugging tendency would still be verify the performance of LDHI B under conditions expected
unknown. in this well. Capillary stability tests further demonstrated that
LDHI B was suitable for injection downhole via a capillary.
8 L.M. FROSTMAN, J.L. PRZYBYLINSKI SPE 65007

The viscosity of LDHI B was low enough (14 cP at 40 F) to

allow sufficient delivery through the small (0.277 inch ID) Conference on Petroleum and Gas Phase Behaviour and
capillary string. The recommended initial injection rate was Fouling, Copenhagen, Denmark (August 2000).
4.9 gal LDHI B per bbl water produced. 10. Christianson, R.L., Bansal, V., Sloan, E.D. Jr.: “Avoiding
Application of LDHI B was implemented as the standard Hydrates in the Petroleum Industry: Kinetics of Formation”,
SPE 27994, presented at the University of Tulsa Petroleum
operating procedure for planned shut-ins and restarts of this Engineering Symposium, Tulsa, OK, (August 29-31, 1994).
well. To date, two shut-ins have been conducted using LDHI 11. Kelland, M.A., Svartaas, T. M., Dybvik, L.: “A New Generation
B. Each lasted approximately 9 hours. No problems have been of Gas Hydrate Inhibitors”, SPE 30695, presented at the SPE
encountered during the use of LDHI B. As it turns out, the Technical Conference, Dallas, Texas, (October 22-25, 1995).
produced water has a sufficiently high enough salinity to allow 12. Lederhos, J.P., Long, J.P., Sum, A., Christiansen, R.L., Sloan,
LDHI A to be used in subsequent shut-ins and restarts. In E.D. Jr., “Effective Kinetic Inhibitors for Natural Gas
addition, the use of an anti-agglomerant has also been Hydrates”, Chem. Eng. Sci. (1996), 51(8), 1221.
specified for the next well to be brought on at this platform. 13. Notz., P. K., Bumgardner, B.D., Schaneman, B.D. and Todd,
J.L., “Application of Kinetic Inhibitors to Gas Hydrate
Problems”, SPE Production & Facilities, (November 1996),
Conclusions 256.
Three field trials have now been conducted using two different 14. Corrigan, A., Duncum, S.N., Edwards, A.R., Osborne, C.G.:
anti-agglomerant hydrate inhibitors in deepwater Gulf of “Trials of Threshold Hydrate Inhibitors in the Ravenspurn to
Mexico systems. All three trials have been successful in that Cleeton Line”, SPE Production and Facilities, (November
not only were no hydrate problems seen, but the use of the 1996), 250.
anti-agglomerants did not have any adverse effects on any 15. Bloys, B. and Lacey, C.: “Laboratory Testing and Field Trial of
other aspect of operating the platforms. The low dosage rates a New Kinetic Hydrate Inhibitor”, OTC 7772, presented at the
of the anti-agglomerants and their ability to handle high 27th Annual OTC in Houston, Texas, (May 1-4, 1995).
subcoolings significantly extend the ability of existing 16. Argo, C.B., Blain, R.A., Osborne, C.G. and Priestly, I.D.,
“Commercial Deployment of Low Dosage Hydrate Inhibitors in
production systems to control hydrates via chemical injection. a Southern North Sea 69 Kilometer Wet-Gas Subsea Pipeline”,
SPE 37255, presented at the SPE International Symposium on
Acknowledgements Oilfield Chemistry, Houston, Texas, (February 18-21, 1997).
The author wishes to thank the management Baker Petrolite 17. Leporcher, E.M., Fourest, J.M., Labes-Carrier, C. and Lompre,
for permission to present this paper, the Account M.: “Multiphase transportation: A Kinetic Inhibitor Replaces
Representative for obtaining well data and collection of Methanol to Prevent Hydrates in a 12-inc. Pipeline”, SPE
samples, and members of the Baker Petrolite Technology 50683, presented at the SPE European Petroleum Conference,
group for their help in generating supportive data. The author The Hague, (October 20-22, 1998).
also wishes to acknowledge Dr. Veet Kruka for useful 18. Fu, B.: “A Novel Kinetic Hydrate Inhibitor for Hydrate
Control”, presented by R. W. Pongar at the NACE NAWC,
discussions and suggestions instrumental in developing this Saskatoon, (February 29 – March 2, 2000).
anti-agglomerant LDHI. 19. Mitchell, G.F. and Talley, L.D.: “Application of Kinetic
Hydrate Inhibitor in Black-Oil Flowlines”, SPE 56770,
References presented at the SPE Annual Technical Conference and
1. Mehta, A.; Walsh, J.; Lorimer, S.: “Hydrate Challenges in Deep Exhibition, Houston, Texas, (October 3-6, 1999).
Water Production and Operation,” Ann. New York Acad. Sci. 20. Frostman, L.F.: “Anti-Agglomerant Hydrate Inhibitors for
(2000) 912, 366. Prevention of Hydrate Plugs in Deepwater Systems,” SPE
2. Davy, H., Philos. Trans. R. Soc. London, (1811), 101, 1. 63122 presented at the SPE Annual Technical Conference and
3. Hammerschmidt, E.G.: “Formation of Gas Hydrates in Natural Exhibition, Dallas, Texas (October 1-4, 2000).
Gas Transmission Lines”, Ind. Eng. Chem. (1934), 26, 851.
4. Ben-Na'im, A., Water and Aqueous Solutions, Plenum Press,
New York, (1974), 223.
5. Yousif, M.H.: “The Kinetics of Hydrate Formation”, SPE
28479, presented at the SPE Technical Conference and
Exhibition, New Orleans, LA, (September 25-28, 1994).
6. Fadnes, F.H.,: “Natural Hydrate Inhibiting Components in
Crude Oils”, Fluid Phase Equilibria, (1996), 117, 186.
7. Sicheri, F., Yang, S. C.: “Ice-binding Structure and Mechanism
of an Antifreeze Protein from Winter Flounder”, Nature, (June
1, 1995), 375, 427.
8. Allen, W.: “Beetle Juice: UMSL Chemist Heads Team Seeking
Nature’s De-icers”, St. Louis Post-Dispatch, (September 6,
1994).
9. Frostman, L.M. and Downs, H.: “Control of Natural Gas
Hydrates in Production Systems Using Anti-Agglomerant
Hydrate Inhibitors,” presented at The 2nd International
 SPE 65007 SUCCESSFUL APPLICATIONS OF ANTI-AGGLOMERANT HYDRATE INHIBITORS 9

SI Metric Conversion Factors

cP x 1.0* E – 03 = Pa• s 7
Blank
mile x 1.609 E + 00 = km 6 with LDHI A
feet x 3.048 006 E – 01 = m with Defoamer

in x 2.54*
with LDHI A + Defoamer

Expansion Factor
E + 00 = cm 5

mil x 3.93 E + 01 = mm 4
MSCF x 2.831 685 E + 04 = m3
gal x 3.785 434 E + 00 = L 3
bbl x 1.589 882 E + 02 = L
(°F – 32) x 5.55*
2
E – 01 =°C
psi x 6.894 757 E + 00 = kPa 1

* Conversion factor is exact. 0

0 200 400 600 800 1000 1200
Time (seconds)

TABLE 1 – SYSTEM CONDITIONS Fig. 1 – Sparge tests of Well A oil.

Well A Well B
Water Depth 2900 ft 2900 ft 1700 110

Temperature at Mudline 40°F 40°F 1500 100

Wellhead Temperature (°F)

Wellhead Pressure (psig)
Brine Salinity 18 wt% 15 wt% 1300 90
FWHP 870 psig 760 psig Shut-in
1100 80
FWHT 79°F 96°F Temperature
900 70
Shut-in WHP 1555 psig 1845 psig
Minimum Pressure at 1794 psig 2083 psig 700 Pressure (Note: pressure
gauge is downstream of
60

Mudline 500 closed safety valve) 50

Maximum Pressure at 2709 psig 3000 psig
300 40
Mudline
10:00 12:00 14:00 16:00 18:00 20:00 22:00 0:00
Hydrate Temperature at 52°F 57°F Time (hours)
Minimum Pressure
Fig. 2 – Well A wellhead conditions when initially shut-in.
Hydrate Temperature at 56°F 60°F
Maximum Pressure
1700 110
Range of Subcooling 12 – 16°F 17 – 20°F
during Shut-in subcooling at subcooling at 1500 100

Wellhead Temperature (°F)

Wellhead Pressure (psig)

mudline mudline
1300 90
Temperature

1100 80
TABLE 2 – HYDRATE TEST CONDITIONS Sample Results:
900 70
1: 0% BS&W
Brine entrained produced water plus supplemental 2: 30-38% BS&W
water (3.5% salinity) 700 3: 48% BS&W 60
1
Water cut 50% 500 2 3 50
Pressure
Oil Well A field sample 300 40
Gas 15 mol% propane in methane 16:00 18:00 20:00 22:00 0:00 2:00 4:00
Time (hours)
Filled to 1400 psig
Fig. 3 – Well A wellhead conditions during restart.
Shut-in 1250 psig and 40°F
Subcooling ~25°F
10 L.M. FROSTMAN, J.L. PRZYBYLINSKI SPE 65007

2000 110 5000

Shut-in No Salt
1800 100 15 wt% Salt

Wellhead Temperature (°F)

Wellhead Pressure (psig)

4000 18 wt% Salt

1600 Temperature 90

Pressure (psia)
1400 80 3000

1200 70
2000

1000 60

Pressure (Note: pressure 1000

800 gauge is downstream of 50
closed safety valve)
600 40 0
8:00 9:00 10:00 11:00 12:00 13:00 14:00 30 40 50 60 70 80
Time (hours) Temperature (°F)
Fig. 6 – Hydrate stability curves.
Fig. 4 – Well B wellhead conditions when initially shut-in.

2000 110

1800 100
Wellhead Temperature (°F)
Wellhead Pressure (psig)

Temperature
1600 90

1400 Sample results: 80

1: Gas only
2: 0% BS&W
1200 70
3: 0.5% BS&W
1 3 4: 2% BS&W
1000 4 60
2 Note: at 1:00 am, 5% BS&W
800 50
Pressure
600 40
12:00 14:00 16:00 18:00 20:00 22:00 0:00
Time (hours)
Fig. 5 – Well B wellhead conditions during restart.