OTC 17355
Hydrate Prevention using MEG instead of MeOH: Impact of experience from major
Norwegian developments on technology selection for injection and recovery of MEG
S. Brustad, K.-P. Lken, and J.G. Waalmann, Aker Kvrner Engineering and Technology
Copyright 2005, Offshore Technology Conference
This paper was prepared for presentation at the 2005 Offshore Technology Conference held in
Houston, TX, U.S.A., 25 May 2005.
This paper was selected for presentation by an OTC Program Committee following review of
information contained in a proposal submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Offshore Technology Conference and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Offshore Technology Conference, its officers, or members. Papers presented at
OTC are subject to publication review by Sponsor Society Committees of the Offshore
Technology Conference. Electronic reproduction, distribution, or storage of any part of this
paper for commercial purposes without the written consent of the Offshore Technology
Conference is prohibited. Permission to reproduce in print is restricted to a proposal of not
more than 300 words; illustrations may not be copied. The proposal must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, OTC, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
Abstract
The four Norwegian gas/condensate field developments Troll,
Snhvit, Ormen Lange and sgard are used together with
amongst others Britannia Satellites (ConocoPhillips) and Shah
Deniz (BP) to illustrate experience and technology status
related to injection and recovery of MEG (Mono Ethylene
Glycol).
A comparison between MEG and MeOH is presented.
Advantages, limitations and cost elements (CAPEX/OPEX) by
use of glycol (MEG) vs. alcohols (Methanol/Ethanol) are
reviewed and discussed in relation to their applicability.
Based on experience, typical operational problems within
closed loop MEG systems with salt- and water removal are
carbonate scale deposits in pipeline and recovery systems,
accumulation of corrosion products and other small particles
and carry-over/foaming. A holistic approach for proper design
and prevention of these and other incidents are presented.
Introduction
Proper hydrate management is vital for all field developments.
For short and moderate tie-backs, flowline insulation
(maintaining fluid temperature above the hydrate formation
temperature) combined with a depressurisation strategy, is
normally the basic method. For developments with cold well
fluids, systems which are difficult to depressurise or restart
(deep water), and for general improvement of the field
regularity, flowline heating is an additional element. For
deepwater developments, the hydrate management is often
based on displacement with stabilized oil, or built around a
subsea separation concept /4/. For appropriate fluids, Cold
Flow may turn out to be the next quantum leap. However, for
long distance gas-condensate tie-backs with complex subsea
gathering architecture, chemical hydrate inhibitors (glycols
and alcohols) are considered the best option.
Part 1 Selecting hydrate inhibitor: Comparing MEG
and MeOH
Chemical hydrate inhibitors
Chemical hydrate inhibitors can be arranged in two main
categories:
Thermodynamic Hydrate Inhibitors (THI)
Low Dosage Hydrate Inhibitors (LDHI)
The latter category consists of Anti Agglomerants (AA) and
Kinetic Hydrate Inhibitors (KHI). These new and promising
chemicals are becoming popular in the GoM and on the UK
sector. However, for long distance gas-condensate tie-backs,
especially in cold waters, they have serious limitations. AAs
generally require a certain amount of condensate (continuous
oil phase) in order to be effective. KHIs can only give a
limited suppression of the hydrate formation point, and they
are not proven to be fully effective at higher pressures.
Thermodynamic inhibitors cost
This means that THIs are the robust choice for long distance
gas-condensate tie-backs. The thermodynamic inhibitors of
widest use are methanol (MeOH) and monoethylene glycol
(MEG). Other alcohols and glycols may be used, but two main
factors making methanol and MEG the most common
thermodynamic inhibitors are hydrate suppression
performance (see Table 2) and cost.
Rough costs are given in Table 1, where it is seen that the
common thermodynamic inhibitors are relatively cheap per
unit volume. However, since large quantities may be necessary
to suppress the equilibrium temperature below the lowest
operational temperature expected in the system, the
infrastructure cost may add up to significant amounts.
Necessary considerations to make when designing for a
hydrate strategy with thermodynamic inhibitors are storage
volumes and regeneration facilities.
2 OTC 17355

o
Table 1 - Rough costs for common thermodynamic inhibitors Table 2 - Calculated values of depression of hydrate point ( C) for
based on ICIS-LOR (February 2005) various thermodynamic inhibitors (Kelland 2000 /7/).
Depression of hydrate point [oC]
Chemical Cost USD/Metric Tonne MeOH EtOH MEG DEG TEG NaCl
Concentration
MeOH 300 of inhibitor
MEG 900 [wt%]
Offshore shipping cost is approximately 150-200 USD/tonne 5 2.0 1.4 1.05 0.63 0.46 1.96
of chemical. 10 4.2 3.0 2.25 1.4 1.05 4.3
20 9.3 6.6 5.2 3.3 2.7 10.7
It should be noted that the cost numbers tend to fluctuate over 30 15.3 10.7 9.0 5.9 5.0 15.0
time. Lately the price of MEG has shown a rising tendency 35 18.6 13.0 11.35 7.5 6.5 -
due to an increasing demand from the Chinese textile industry
40 22.2 15.4 14.0 9.3 8.2 -
in particular.
It is also important to realize that the actual OPEX is
significantly governed by the chemical losses and the make-up THI regeneration
required to cover for these losses (see Table 3).
Both MEG and MeOH may be readily regenerated, although
THIs and suppression performance there are limitations regarding salinity of the produced water
for conventional MEG regeneration facilities. However, e.g.
The lower the molecular weight, the better the hydrate Kvaerner Process Systems (KPS) has overcome this with their
suppression performance, hence, MEG yields better continuos reclamation facilities, which is installed on the
performance per mass basis compared to TEG (triethylene sgard B semi. Apart from some possible, field-specific
glycol), and methanol yields better performance than ethanol aromatics and phenols following the water-phase, the water
etc. Figure 1 shows the relative performance of some leaving such a regeneration facility is very clean. Also, as the
thermodynamic inhibitors; NaCl, methanol and MEG. The glycol has rather low solubility in the gas phase, the
more the equilibrium line is shifted to the left, the more contamination of the gas phase using MEG as a hydrate
effective is the inhibitor, and the larger is the safe area inhibitor is insignificant. When it comes to MeOH, the losses
(conditions where hydrate formation will not take place). to the gas phase are significant, and the methanol
concentration in the water leaving the regeneration facilities
must be evaluated against the height of the distillation column.
A distillation process can reclaim both chemicals, but also due
to the low vapour pressure of MEG, a smaller, high-
temperature-recovery column may be used for MEG
regeneration compared to MeOH regeneration.

MeOH regeneration is in operation on a limited number of

Figure 1 - Effect of thermodynamic inhibitors on hydrate stability
for a real fluid. Simulated in PVTsim.
installations in the Gulf of Mexico and in the Far East,
including the Shell Malampaya onshore facility in the
Phillipines and the Williams Canyon Station platform (part of
the Canyon Express system) in the GoM. The industry
knowledge on the cost-effectiveness of these systems is fairly
limited, since recovering a volatile like methanol is not
straight-forward. The presence of aromatics in the condensates
only makes life more difficult due to the formation of
azeotropes.

A rough guide to compare the performance of a number of
thermodynamic inhibitors is given below.
THI losses and product contamination
E. D. Sloan /8/ presents the following rules-of-thumb in order
to account for the inhibitor losses to the gas and condensate
phases:
OTC 17355
Table 3 - Rules-of-thumb for estimating inhibitor losses to the gas
and condensate phases (Sloan 2000 /8/).
Loss to gas phase Loss to condensate
phase
Methanol At 4oC and pressures The methanol
higher than 70 bar, the concentration dissolved in
methanol lost to the condensate is 0.5 weight%.
vapour phase is 16 kg per
106 Sm3 gas for every
weight% methanol in the
water phase.
MEG At 4oC and pressures The mole fraction of MEG
higher than 70 bar, the in a liquid hydrocarbon at
maximum amount of MEG 4oC is 0.03 % of the water-
lost to the gas is 0.3 kg per phase mole fraction of
106 Sm3 gas. MEG.
Based on the above rules-of-thumb, the total methanol loss
from a field, with typical 30 wt% MeOH in the water phase,
will be app. 500 kg per 106 Sm3 gas and 50 kg per 1000 kg
condensate. As Table 3 shows the MEG lost to the gas is only
is 0.3 kg per 106 Sm3 gas, independent of the weight% of
MEG in the water phase. Hence the MEG losses are negligible
compared to the MeOH losses. The MeOH losses also
increases substantially with increased temperaure. At 10oC the
loss to the vapour phase is app. 25 kg per 106 Sm3 gas for
every weight% methanol in the water phase.
This also raises operational and environmental issues with
respect to methanol regeneration. In addition, depending on
the customer, there might be a need for decontaminating the
gas phase because of the high MeOH content (as described
above). This might also introduce a significant investment cost
that must be included in the comparison. Alternatively there
might be a cost penalty due to pollution of the product.
Product contamination and LNG
Because of the large distances between the major gas
producers and the consumers, gas export pipeline cost has
been a showstopper for a number of planned developments.
Over the last few years, the LNG market has opened up,
presenting LNG transport as an enabling technology for
development of remote gas-condensate fields. As a result of
this trend, a large number of the ongoing and planned gas
developments are planned with LNG facilities. Gas
contaminated with MeOH is unacceptable in the cryogenic
Cold Box part of a LNG plant. MeOH has a freezing point
that is well above the temperature in the LNG plant. For this
reason MeOH is generally avoided for LNG based
developments. Should MeOH be chosen, this would require an
additional pre-treatment system removing the MeOH from the
gas (IFPEXOL or similar).
THIs and corrosion management
For long distance tie-backs, the pipelines are major cost
drivers. Material selection and corrosion management are
important elements in the overall Flow Assurance evaluations.
MEG is corrosion protective in itself, and reduces operational
costs with respect to corrosion inhibitors for the pipelines. A
3
number of developments are today basing the corrosion
philosophy on carbon steel pipelines, MEG and pH
stabilization. MeOH gives no benefit on corrosion resistance.
THIs and HSE
MEG is non-flammable, with a flash point at 111 oC. MeOH is
highly flammable with a flash point at 11 oC. This clearly
indicates that MeOH presents a greater safety risk with respect
to handling and storage (especially on offshore installations
with limited area). Furthermore, MeOH burns with an
invisible flame, making fire detection a more difficult
problem.
MEG regeneration plants with de-salting (reclamation) will
produce a waste product that mainly consists of salt. MeOH
regeneration, on the other hand, does not produce any solid
waste. It is however possible to re-dissolve the salts from a
MEG reclaimer before disposal to sea.
THIs and viscosity
MEG is significantly more viscous than MeOH, especially at
low temperatures (which is the case for the most relevant field
developments). This means that a MEG injection system will
require a larger diameter injection line and/or more pumping
horsepower.
THIs combined with KHIs
For the Nuggets development in UK (tie-back to the Alwyn
North platform), Total has investigated a hydrate inhibitor
strategy combining MEG with KHI for the peak water
production period. The reason is a requirement to limit the
MEG injection to a given rate, and add KHI to make up the
difference to the required subcooling. The investigations have
concluded that MEG is better suited than MeOH for this kind
of mixing strategy /1/.
THIs and dehydration
In the GoM, some developments use MEG for hydrate control
in the gathering system, at the same time dehydrating the gas
enough to meet export pipeline specifications (typically 7
lb/MMSCF) without a topside contactor.
THI selection - conclusion
In view of the factors listed above, it is obvious that a general
ranking of the two inhibitors, based on a direct CAPEX
comparison between a standard MEG regeneration unit and a
corresponding standard MeOH regeneration unit, is inadequate
for selection in a specific project. Not only does the loss of
MeOH to the gas phase add substantial OPEX, it may also
impose a price penalty on the product. Furthermore, the
requirements for salt removal (reclamation) in the MEG case
or gas cleaning in the MeOH case will be project specific cost
drivers of significant order.
4 OTC 17355

Surveying the choices made by the operators for recently built

and planned gas-condensate tie-backs, it is evident that MEG
seems to be the preferred inhibitor. The list of MEG-based
developments includes record-breaking developments like
Ormen Lange (Norsk Hydro - Norway), Snhvit (Statoil -
Norway), KG-D6 (Reliance Industries India), Scarab-Safron
(Burullus Egypt), South Pars (Total Iran), Shah Deniz (BP
- Azerbaijan), Britannia Satellites (ConocoPhillips UK),
Gorgon (Chevron Texaco Australia) and finally the ultimate
Subsea to Beach concept: Shtokman (Gazprom - Russian
Barents Sea).

MeOH is on the other hand applicable for intermittent

inhibition during shut-down and start-up. Some companies are
starting to replace MeOH with MEG also for these purposes.
On the Kristin Semi, Statoil has decided to go for MEG as the Figure 2 - Schematic illustration of an on-shore Closed Loop MEG
base case chemical hydrate inhibitor for intermittent use. system

In South America, ethanol is commonly used as a hydrate anti-
freeze since it is cheap and available in large quantities from
sugar fermentation. Generally ethanol has the same
disadvantages as methanol compared to MEG for continuous
inhibition.
Part 2 The holistic approach to the closed loop
MEG system
Although continuous MEG injection is found as the most
reliable and cost-effective method of hydrate inhibition on
numerous gas/condensate field developments, the closed loop
MEG system is complex and represents a number of chemical,
physical and hence operational challenges. Consequently, it is
extremely important to approach a closed loop MEG system
from a holistic point of view. The consequence of neglecting
this will in many cases be severe operational problems leading
to plant downtime and losses of income.
Continuous experience from the conventional MEG
regeneration plant at the Kollsnes onshore facility and the full
reclamation unit on the sgard B semi, combined with
comprehensive investigations performed in the Snhvit and
Ormen Lange projects, have given the Norwegian E&P
industry a frontrunner position with respect to understanding
of the totality in the closed loop MEG system /5/.
A schematic illustration of a typical on-shore Closed Loop
MEG system with a slip-stream desalination unit is shown in
Figure 2.
The Aker Kvaerner approach is built on six main
steps:
1. Design basis
Establishment of the correct design basis figures is extremely
important and must be performed in co-operation with the oil
companies. I.e. reservoir and well flowing conditions,
production profiles over the field lifetime including formation
water production potential, fluid compositions, environmental
conditions etc.
The strategy for hydrate (and ice) prevention, Rich and Lean
MEG content will be based on such input.
2. Subsea architecture and components
Pending on the chosen subsea production system and pipeline
network, utilisation of, or in some instances development of,
proper technology must be done. Separate MEG line(s) or
MEG lines integrated in the umbilical(s) are options for
carrying MEG to the destination for injection. A multipurpose
umbilical, actually an illustration of an Integrated Production
Umbilical (IPU), from Kvaerner Oilfield Products (KOP) is
shown below /2/. For MEG distribution, an ISU (Integrated
Service Umbilical) with a 3-4 centre line may be sufficient.

Figure 3 - Illustration of an IPU by KOP

OTC 17355
The type of injection valves and proper subsea flow
measurement devices (water cut meter/multiphase flow meter)
must also be chosen.
3. Corrosion
Establishment of a proper corrosion inhibition philosophy
must be based on the input from design basis combined with
the chosen hydrate inhibition strategy. In most cases the
production pipeline(s) will be made of carbon steel due to
cost. Although the MEG itself represents a reduction of the
corrosion rate, additional means are almost always required to
avoid unacceptable pipeline corrosion rates.
Even though a proper field lifetime corrosion protection
scheme is established large amounts of iron will corrode from
a carbon steel pipeline /3/. Alternatives for corrosion control
are pH-stabilisation, addition of a corrosion inhibitor or partly
pH-stabilisation combined with the addition of a corrosion
inhibitor. The alternatives mentioned above can also be
combined with the addition of a scale inhibitor. By injection of
a pH-stabiliser, the formation of a protective iron carbonate
film on the inner pipeline steel surface will be facilitated,
reducing the corrosion rate.
The corrosion inhibition method may well alter over the field
lifetime for many field developments, at least if production of
formation water is initiated. High alkalinity in the lean MEG
may result in severe precipitation of calcium carbonate in the
subsea systems when the formation water rates increase. IFE
(Institute for Energy Technology) has proven to be world
leading and an important partner in evaluating these issues.
It has also been learned that minimising the oxygen level
within the closed loop MEG system is very important to avoid
transformation of iron carbonate to iron oxide(s), avoid an
increasing corrosion rate and avoid possible degradation of the
MEG. Hence, avoiding oxygen ingress to the system is very
important. A qualified oxygen scavenger for injection into the
lean MEG should be considered as a back-up when a vacuum
reclaimer system is selected.
Pictures of pipeline corrosion are shown below.
Figure 4 - Pictures of pipeline corrosion (pitting to the right)
5
4. Precipitation chemistry
Understanding the chemistry and physics of the carbonates
and sulphates in combination with the different divalent
cations (Ca2+, Fe2+, Mg2+, Ba2+, Sr2+, etc.) is important to
design the on-shore/topside MEG plant properly. A picture of
carbonate scale before and after cleaning is shown in the
figure below.
Figure 5 - Picture of carbonate scale before and after cleaning
At arrival on-shore/topside, iron and carbonate will be in the
form of ions, Fe2+ and CO32-, and it will probably be saturated
or maybe supersaturated with respect to FeCO3. Hence, there
is a risk of scaling in different parts of the plant where
operating conditions reduce the solubility of various
substances. Scaling of carbonates (calcium carbonate and iron
carbonate) are expected in those locations where the
temperature is increased and/or there is a reduction in pressure
with subsequent increase in pH due to loss of dissolved carbon
dioxide (CO2).
To minimize problems related to iron, from pipeline corrosion,
and calcium, from the formation water, coming into the on-
shore/topside plant, it is important to control the corrosion rate
to a low level and control the precipitation of carbonates by
designing the MEG process in a way that handles the
formation and deposition of particles. Otherwise precipitation
and deposition at unwanted locations may occur, resulting in
highly increased maintenance/cleaning frequency.
Based upon input from plants processing gas that has been
treated with MEG, it has been experienced that the surface of
heat exchangers will be especially vulnerable to the formation
of carbonate scales.
The solubility/saturation ratio of iron and calcium carbonate
must be established for each section of the MEG plant to
determine where scaling and depositions may occur. As the
solubility is dependent of the pH and thereby the CO2 content
in the MEG solution it is possible to both determine and
control where the carbonates shall precipitate.
5. MEG Regeneration Options
In general there are three overall options for regenerating
MEG in a closed loop system:
a) Conventional Regeneration - by water removal only
All salts and non-volatile chemicals remain and accumulate in
the closed loop MEG system.
6 OTC 17355
b) Full Reclamation - by evaporating the total rich MEG
feed
Boiling at low pressure (0.1-0.15 bara) followed by distillation
to the required Lean MEG specification. All salts and non-
volatile chemicals remain in the boiler. Salts crystallise and
can be removed by e.g. centrifuges/decanters.
c) Slip-stream salt removal
Conventional regeneration combined with a slip-stream salt
removal system (vacuum reclaimer or ion exchange). The slip-
stream rate is generally determined by the allowable amount
of dissolved salts in the lean MEG.
There are several issues that need to be evaluated to determine
the overall regeneration philosophy for a closed loop MEG
system. Common early phase field evaluations listed under
Step 1, 3 and 4 above set premises for selection of the MEG
Regeneration system configuration. Other evaluation criteria
are allowable content of dissolved salts in the Lean MEG,
limitations in chemicals consumption and MEG losses and
availability requirements.
The availability requirements must be seen in conjunction with
the MEG chemistry and the operational challenges such as
corrosion, scaling/precipitation and system accumulations.
The selected complexity of the MEG system and thereby the
investment costs must be compared to operational losses in
terms of reduced availability, increased maintenance,
consumption of chemicals and MEG, and also the
infrastructure and environmental requirements for the field. A
general overview of the three overall options for MEG
regeneration and special concerns and selection criteria are
described below.
a) Conventional Regeneration
In conventional regeneration systems water is boiled off at
close to atmospheric pressure. The boiling temperature is
dependent of the required Lean MEG specification and the
back-pressure in the venting system. The main components are
a reboiler, either a kettle-type or a recycle heater, and a
distillation column.
A conventional regeneration system removes water only. This
means that all salts, pipeline corrosion products and
chemicals are accumulated in the closed MEG loop. A system
like this cannot handle a continuous production of formation
water from the wells. As this is very difficult to prevent during
the whole field lifetime, these types of plants would normally
require offshore facilities for water separation prior to MEG
injection. Even these fields would experience accumulation of
particles due to pipeline corrosion and formation water carry-
over, which means that the MEG inventory at certain intervals
must be replaced or cleaned.
Below, some experiences and challenges from start-up and
operation of conventional onshore and offshore regeneration
facilities are listed:
Statoil/Gassco Kollsnes:
The Statoil/Gassco operated Troll Gas (Kollsnes) plant has a
conventional regeneration system, which has been in operation
since 1996. At the offshore Troll A platform, formation water
is separated and MEG injected, before the well stream is sent
through the two 68 km 36 multiphase gas/condensate
pipelines to the onshore facility at Kollsnes. Not only is the
Troll A the largest GBS in the world (the largest man-made
object ever moved), but the Troll development was also the
first development to fully rely on multiphase transport through
large diameter pipelines. In principle the overall arrangement
is quite similar to a Subsea to Beach (S2B) concept including
water separation (re-injection).
The Kollsnes plant experienced a considerable amount of salts
entering the onshore plant in early operation. The plant also
had problems with scaling in the inlet heaters and also the
column/reboiler. High corrosion rate in early operation was
resolved by switching the pipeline corrosion protection
scheme from corrosion inhibitor to pH-stabilisation. A
permanent ion exchange unit is installed for slipstream salt
removal. In addition new particle removal units with high-
speed centrifuges have been introduced to take out corrosion
products and carbonates from the closed loop. The Kollsnes
plant had a common MEG system for the pipelines and for gas
dehydration. Due to severe depositions in the onshore plant
these systems have now been segregated.
Shell Mensa:
For the Mensa field in the GoM, MEG was selected for
combined hydrate control and dehydration (without topside
dehydration provisions). Unexpected condensate production
caused serious topside problems as the condensate formed an
emulsion with the glycol. At the same time, dissolved solids
were detected in the well stream. Even though formation water
production was not expected, completion fluids and formation
water were suggested to be the source of the salts. As a result,
the MEG regeneration facility experienced serious operational
problems (scaling, fouling, plugging). These problems even
propagated to the MEG injection system with plugging of the
subsea injection lines.
Total South Pars:
South Pars is another major development where MEG has
been selected for hydrate control in the two 109 km 32
pipelines. MDEA is mixed in the MEG, thus producing a very
high pH for corrosion protection. Even though formation
water production was not expected, excessive scaling has been
experienced (column, heat exchangers and inlet piping). The
reason is up-concentration of Ca in the closed loop MEG
system, and the main source of Ca is believed to be
completion fluids from the well clean-up.
OTC 17355
b) Full Reclamation
In a full Reclamation system the total rich MEG feed is boiled
off in a Flash Separator followed by distillation to the required
Lean MEG specification. This configuration is applicable
when higher formation water rates is expected from the
producing wells. The MEG plant will then consist of a pre-
treatment system where the rich MEG is heated and
depressurised to remove hydrocarbons in a 3-phase separator,
followed by Storage facilities, before the MEG is sent to the
Flash Separator. The Flash Separator operates under vacuum
(0.10-0.15 bara) in order to vaporise the MEG at a temperature
low enough to avoid decomposition of the MEG. Essentially
all the salts and particles are removed, since the whole rich
MEG feed exits the Flash Separator as vapour. Non-volatile
chemicals will also remain in the Flash Separator. Salt
accumulates in the concentrated liquid phase, crystallises and
can be separated in a Decanter Centrifuge. Heat for
evaporation is added through a heater in a recycle stream. The
vapour from the Flash separator is distilled into MEG and
water in the downstream Distillation Column.
A full Reclamation system has shown to be a proven design
for fields with larger quantities of formation water entering the
on-shore/topside plant. Kvaerner Process Systems (KPS),
which is part of the Aker Kvaerner group, has two reference
plants in operation; STOS (Shell Todd, New Zealand) and
sgard B (Statoil, Norway). Another two full Reclamation
systems, designed by KPS, are under construction; Shah Deniz
(BP, Azerbaijan) and Britannia Satellites (ConocoPhillips,
UK).
Below, some experiences and challenges from engineering,
start-up and operation of onshore and offshore full
Reclamation facilities are listed:
Statoil sgard B:
KPS performed a full scale testing of the sgard B MEG
Reclamation unit at Kvaerner Egersund yard, spring 1999
prior to installation offshore. The tests were performed with
anticipated salt compositions and demonstrated the specified
salt removal capacity. However, after start-up offshore there
was insignificant production of formation water. The result of
this has been that only minor particles from the pH-stabiliser
and pipeline corrosion has precipitated in the Flash Separator.
Sodium carbonate and iron products have shown to be too
small to remove efficiently in the Decanter Centrifuges and
have therefor accumulated in the Flash Separator. With
additional accumulation of corrosion inhibitors, hydrocarbon
carry-under from upstream three-phase separation and flashing
in the liquid phase of the Flash Separator these are believed to
be the reasons for periodically problems with foaming and
carry-over form the Flash Separator. The lessons learned from
this experience are:
- New gas tight Centrifuges adapted for smaller particles
- Flashing in the liquid phase of the Flash Separator should
be avoided
- Hydrocarbon removal should be improved upstream the
reclaimer
- Corrosion inhibitor and other inhibitors should be
qualified/selected to ensure compatibility with the boiler
7
in terms of minimizing foaming. An anti-foaming agent
should be considered/qualified as back-up.
- Accumulation of inhibitors and acetates should be
controlled below critical levels
- If it is expected low formation water rates during start-up
or field lifetime, i.e. the concentration of small particles
may be high compared to the large NaCl particles, means
to remove the small particles should be considered.
sgard B has also experienced some problems with scaling of
heaters, especially the Flash Separator Recycle Heaters
(cleaning frequency reported to vary between monthly and 6-
monthly). It is anticipated that boiling of water, from the
Centrifuge flushing system, in the heaters have increased the
scaling. Some level and flow instruments have been replaced
due to high content of iron. Fine filters have proven not to be
applicable for removal of small particles in the rich MEG feed.
Figure 6 Picture of sgard B MEG Reclaimer prior to installation
(KPS)
BP Shah Deniz / ConocoPhillips BritSats:
Two full Reclamation systems, designed by KPS, are now
under construction, namely Shah Deniz (BP, Azerbaijan) and
Britannia Satellites (ConocoPhillips, UK). The experiences
from sgard B have been carefully evaluated and several
changes to design of these new plants are incorporated. The
Flash Separators have been designed to minimise the potential
for foaming, mainly by changing the inlet arrangement. The
facilities for removal of hydrocarbons prior to boiling have
been greatly improved. Institute for Energy Technology (IFE)
has performed qualification programs for inhibitors to ensure
compatibility with the reclaimer. Foam inhibitor has been
included as a back-up. IFE has also performed corrosion tests
for material selection in both plants.
Both projects have considered the possibility of low
formation water production. If the concentration of small
particles is high compared to the large NaCl particles, it is
expected that the Centrifuge will have difficulties in removing
all, and some accumulation of small particles could be
expected. These particles will mainly be different carbonates
and iron oxide, e.g. Na2CO3, CaCO3, FeCO3, and Fe3O4.
Adequate separation equipment/processes for these particles
(e.g. filters) are large and expensive, operator intensive, and
8 OTC 17355

prone to failure/clogging. Analyses performed by IFE have

shown that if the concentration of small particles can be kept
relatively low compared to the NaCl concentration, most of
the small particles will tend to stick to the large NaCl crystals,
and thereby be taken out in the Centrifuge. This was also
indicated in the tests of the sgard B plant at Egersund prior
to installation. The Shah Deniz plant is therefor provided with
back-up facilities for injection of NaCl into the Flash
Separator. It is also possible to operate the reclaimers as
conventional reboilers by boiling off water only, thus not
remove any salts or chemicals from the MEG. At BritSats
there will be back-up facilities for injection of a flocculating
agent into the Centrifuge slipstream to induce particle
flocculation, thereby creating larger particles and enabling a
simpler particle removal. The final qualification programs for Figure 7 - 3-D model overview of the full stream MEG reclaimers at
these flocculents are ongoing at IFE. Shah Deniz (KPS)
For both projects water analyses have shown an excess of
divalent cations (Ca, Mg, Ba, Sr, Fe) in the formation water.
This means that the concentration of 2+ ions is higher than the
concentration of alkalinity (SO4 and HCO3), hence only a part
of the divalent cations will precipitate as carbonates and
sulphates in the Flash Separator. When the concentration
increases, the density, viscosity and boiling point of the MEG
will increase. This will reduce the capacity of the Reclaimer,
and could also result in increased foaming and carry-over as
described above. A shutdown / cool-down could also cause
some of the accumulated divalent cations to precipitate as
chlorides. If the concentration is high this could cause major
start-up problems, as it will be impossible to recycle and heat
up the content again to dissolve the chlorides. For the above
reasons it is important to keep control of the alkalinity in the
Flash Separator. For both plants alkalinity in the form of
bicarbonate/carbonate can therefor be added in order to
precipitate divalent cations.
For both the Shah Deniz and the BritSats projects the
solubility / saturation ratio of iron- and calcium carbonate has ConocoPhillips: BritSats MEG Regeneration and Desalting
been determined for all sections of the plants. The design has
been carefully selected to avoid scaling and to control Figure 8 - 3-D model overview of the full stream MEG reclaimer at
precipitation of carbonates in the bulk where scale deposits Britannia Satellites (KPS)
and cleaning/maintenance can be minimised. Also based on
solubilities / saturation ratios some heaters and control valves
have been duplicated to avoid stop of production if scaling c) Slip-stream salt removal
should occur. This configuration consists of a conventional regeneration, as
described earlier, combined with a slip-stream salt removal
3-D model overviews of the KPS full stream MEG reclaimers system (reclaimer or ion exchange). The slip-stream rate is
at Shah Deniz and Britannia Satellites are shown in the figures generally determined by the allowable amount of highly
below. soluble salts in the lean MEG. A slip-stream arrangement is
applicable for low/intermediate formation water rates. The
ongoing projects Ormen Lange (Norsk Hydro - Norway) and
Snhvit/Hammerfest (Statoil - Norway) are based on this
configuration. The main advantage of this design is that
inhibitors and pH-stabilisers are re-used, as opposed to a full
stream reclaimer where all salts and non-volatile chemicals are
removed. It should be noted that the full stream reclaimers at
Shah Deniz are designed for also handling a no salt mode
where all chemicals are re-used. Another advantage with a
slip-stream arrangement is that the heat- and cooling
requirements are reduced, as only a smaller portion of the
MEG is vapourised. In addition, dependent of the slip-stream
rate, the investment cost could be lower than for a full stream
OTC 17355 9

Reclaimer. The main disadvantage is that impurities will large amounts of NaAc could be dissolved in the Flash
accumulate in the closed MEG loop. The impurities are mainly Separator (more than 300 g/l). It is important to control the
solid particles and salts. The salts will exist either as ions or in concentrations to avoid increasing the viscosity and the boiling
precipitated form. The solubility of the different salts in the temperature and also to avoid massive precipitation during
MEG/water solution will vary, and therefore require different shut-down and cool-down.
treatment for removal. In addition different salts will
A 3-D model overview of the KPS slip-stream reclaimer at
precipitate in different parts of the system. The MEG/water
Ormen Lange is shown in the illustration below.
phase particles will mainly consist of corrosion particles,
precipitated salts from formation water and completion fluids,
salts from acetic acids and fines. Although it is possible to run
the system with some particles in the regenerated lean MEG,
most of the particles must be removed to avoid problematic
accumulation in process equipment and flowlines.
Scale, in this context, consists of precipitated carbonates and
sulphate salts. These salts are partly soluble and the solubility
will vary with fluid pH and temperature. When the
MEG/water feed is depressurised or heated CO2 will boil off
and thus increasing the pH. To control the precipitation and
minimise scaling on locations giving operational problems or
being labour intensive with respect to cleaning is very
essential for a slip-stream salt removal plant. In the MEG
regeneration process scaling on hot surfaces in heaters and
precipitation on the column internals are the main concerns.
Since the solubility of iron- and calcium carbonate (FeCO3 and
CaCO3) is low, most of the incoming iron and calcium will
precipitate at the operating conditions of a regeneration unit /
reboilers. The Ormen Lange plant will therefore be designed
to provoke precipitation of FeCO3 and CaCO3 upstream the
reboilers. The Flash Drum and the Rich MEG Tank will be Figure 9 - 3-D model overview of the slip-stream reclaimer at
operating at high temperature (app. 80 oC). Combined with Ormen Lange (KPS)
long retention time, due to uncertainties regarding
supersaturation/kinetics, it is anticipated that all the iron and 6. System design
most of the calcium will precipitate in the two tanks. The high System design includes calculation of MEG demand, from day
operating temperature in the Rich MEG Tank will also one and in the future, determination of corrosion and scale
increase the settling velocity of particles due to a lower liquid inhibition philosophy over the field lifetime, and design of
viscosity. In other words; the Ormen Lange Rich MEG Tanks MEG injection points and injection devices. Design of the
will be designed for the purpose of being a buffer tank but also MEG distribution network, pumps, particle removal
a tank for settling of small particles. The particles, present as a alternatives and MEG storage tanks, in many cases being part
Rich MEG slurry in the bottom of the tank, can be removed of the plants regularity philosophy, are also contributors to an
during regular maintenance and sent to special treatment. The overall system. Finally, the selection of a proper MEG
Rich MEG slurry may also be sent to high speed Centrifuges. regeneration process must be done. System design for the
These full flow Centrifuges is placed in the rich MEG stream whole loop also includes sparing philosophy for critical
downstream of the Rich MEG Tank to normally remove more systems and sufficient capacity margins.
of the particles that are not settling in the tank. At Snhvit the For sizing of the MEG tanks and the recovery capacity, input
particles will be removed by micro filtration. from pipeline sizing governing liquid accumulation is
The slip-stream Reclaimer at Ormen Lange is sized to important. In this context, the sizing of slug catchers, and the
maintain a maximum salt content in the lean MEG of 20 g/l. liquid handling capacity from slug catchers to the Rich MEG
The reclaimer shall also ensure a maximum acetate tank are important factors in the total loop logistics /4/.
concentration of below 2.36 g/l in lean MEG to control top of The required redundancy and sparing of components, rich and
line corrosion after a change in corrosion strategy from pH lean MEG storage capacity and the MEG plant capacity
stabiliser to corrosion inhibitor. The design of a slip-stream margins must be based on availability analysis. OREDA data
reclaimer is more or less based on the same evaluations as a has shown to be insufficient for a MEG system, hence the
full stream reclaimer. The main difference between the availability analysis must be based on actual experience from
reclaimers described earlier and the Ormen Lange slip-stream the plants in operation together with input from the oil
reclaimer is that the distillation column is not required as companies.
water is removed in the conventional reboilers. An additional
consideration made at Ormen Lange is the high acetate
concentration in the feed. Organic acids will accumulate in the
Flash Separator. Experiments performed by IFE shows that
10 OTC 17355

Conclusion 5. Jan Gunnar Waalmann et al.: An Efficient

When reviewing the advantages, limitations and cost elements
(CAPEX/OPEX) by use of glycol (MEG) vs. alcohols
(methanol/ethanol), it seems evident that MEG is the preferred
inhibitor. This is supported by the choices made by the
operators for recently built and planned long distance gas-
condensate tie-backs. The list of MEG-based developments
includes record-breaking developments like Ormen Lange
(Norsk Hydro - Norway), Snhvit (Statoil - Norway), KG-D6
(Reliance Industries - India), Scarab-Safron (Burullus
Egypt), South Pars (Total Iran), Shah Deniz (BP -
Azerbaijan), Britannia Satellites (ConocoPhillips UK),
Gorgon (Chevron Texaco Australia) and Shtokman
(Gazprom - Russian Barents Sea).
Wellstream Booster Solution for Deep and Ultra
Deep Water Oil fields, OTC paper no 16447,
Offshore Technology Conference, Houston 2004.
6. Stig Brustad, Karl-Petter Lken and Jan Gunnar
Waalmann: MEG Enhances Flow Assurance in
Subsea to Beach Developments, World Oil
Supplement January 2004, Norwegian innovation
7. M. A. Kelland: Chemicals for Gas Hydrate
Control, Conference: Hydrater - oljeindustriens
usynlige problembarn? NTNU, Trondheim 2000
8. E. D. Sloan Jr: "Hydrate Engineering", Society of
Petroleum Engineers Inc, 2000.

A closed loop MEG system is complex and represents a

number of chemical, physical and hence operational
challenges. Consequently, it is extremely important to
approach a closed loop MEG system from a holistic point of
view. The consequence of neglecting this will in many cases
be severe operational problems leading to plant downtime and
losses of income.

Continuous experience from the conventional MEG

regeneration plant at the Kollsnes onshore facility and the full
reclamation unit on the sgard B semi, combined with
comprehensive investigations performed in the Snhvit,
Ormen Lange, Shah Deniz and Britannia Satellites projects,
have greatly improved the understanding of the chemistry and
the totality in a closed loop MEG system.

Acknowledgements

The authors wish to thank Institute for Energy Technology

(IFE), Norsk Hydro, Statoil, BP and ConocoPhillips for their
support and cooperation during the ongoing Aker Kvrner /
Kvrner Process System MEG projects.

References

1. F. Fournie, A. Cripps and D. Atkin: The Nuggets

Development Successful Operation and Plans for
the Future, GPA Europe, Paris Conference 2003
2. O. Heggdal et al.: Integrated Production Umbilical
for the Sognefjord (20 km tieback). Qualification and
Testing., Deep Offshore Technology Conference,
New Orleans 2004.
3. A. Dugstad and Marion Seiersten: pH-stabilisation, a
Reliable Method for Corrosion Control of Wet Gas
Pipelines, 1st International Symposium on Oilfield
Corrosion, Aberdeen May 2004, SPE 87560
4. Karl-Petter Lken, Jan Gunnar Waalmann et al.:
Deepwater Subsea to Beach Gas Developments
Flow Assurance Challenges and Solutions, Deep
Offshore Technology Conference, New Orleans
2004.