OTC-26854-MS

A More Efficient Use of MEG to Fully Inhibit Hydrates with Reduced Cost
Eric M. Grzelak, and Morten Stenhaug, OneSubsea

Copyright 2016, Offshore Technology Conference

This paper was prepared for presentation at the Offshore Technology Conference Asia held in Kuala Lumpur, Malaysia, 22–25 March 2016.

This paper was selected for presentation by an OTC program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Offshore Technology Conference and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Offshore Technology Conference, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the
written consent of the Offshore Technology Conference is 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 acknowledgment of OTC copyright.

Abstract
Monoethylene glycol (MEG) is the primary hydrate inhibitor used in today’s large gas fields. To fully
prevent hydrate formation, high dosage rates of over 1 gallon of MEG per gallon of produced water may
be required. As a result of these high rates, MEG usage can reach into the 100s of tonnes per day for a
field.
The first estimate of MEG requirement stems from subsurface analysis of expected water production
and it becomes the starting point for a design process that may lead to excess. A standard inhibition system
must accomodate large uncertainties due to inaccurate MEG distribution to each well, inaccurate
measurement of water production from each well, and difficulties recovering and reclaiming the MEG
through topsides processing. The compounding of these uncertainties then requires a planned overdose of
MEG to insure total hydrate inhibition. Determining the planned overdose and the necessary equipment
to achieve it often leads to a doubling of system size, which can push the capcity of the MEG system into
the 1000s of tonnes with a concurrent major increase in CAPEX. Likewise, the operation of an oversized
systems leads to a significant increase in OPEX from energy usage for recovery, MEG replacement from
loss, and other factors.
To combat this excess cost and make more fields economical, a hydrate management system has been
identified that ties together industry proven equipment into an intelligently operated package. Uncertain-
ties are reduced through the use of industry leading chemical injection metering valves to pinpoint and
control the distribution of MEG, accurate metering using a variety of technologies to target exactly the
amount of water in need of hydrate inhibition, energy efficient MEG recovery with up to 100% salt and
solids removal, and real-time condition monitoring, performance evaluation, and process modeling to
ensure that the whole system is operating optimally to prevent hydrates under any circumstance.
Tightening the use of MEG for hydrate inhibition with the hydrate management system can create
substantial operational savings over the life of the field and, should it be included early enough in the
design phase, has the ability to reduce overall system size resulting in substantial capital savings as well.

Introduction
Clathrate hydrates are ice-like solid compounds composed of water and natural gas. They form at high
pressures and low temperatures, conditions often found in the subsea production systems around the
2 OTC-26854-MS

globe. They have been a known and planned-for issue in pipelines for many years. However, their
prevention (or mitigation) is still something that continues to be a major cost driver of many projects.
The primary means of hydrate prevention in upstream production systems remains thermodynamic
inhibition chemicals. The associated class of chemicals, thermodynamic hydrate inhibitors (THIs) includ-
ing alcohols and glycols, are used to shift the equilibrium curve of hydrates into even higher pressure and
lower temperature ranges. The ability of THIs to suppress the formation conditions of hydrates is well
known, and physical models can be used to determine the necessary dosage for the desired amount of
protection.
Another class of chemicals used for the prevention of hydrate-related problems is known as Low
Dosage Hydrate Inhibitors (LDHIs). These chemicals are often further classified as either kinetic hydrate
inhibitors (KHIs) or anti-agglomerants (AAs). While the term LDHI is commonly used, the means by
which the chemical inhibits hydrates is dependent on which of the two subclasses it falls under. KHIs are
thought to inhibit either the nucleation or growth of hydrates, yet their actual means of inhibition is still
debated. Common KHIs are variants of polymers like polyvinylcaprolactam or polyvinylpyrrolidone. AAs
are used to prevent already formed hydrate particles from binding together to form larger, jam inducing
particles. Like KHIs, the means by which AAs prevent agglomeration are debated. However, prevailing
theories suggest they work by either altering the wettability of hydrate by water, which prevents capillary
bridging, or assisting in the conversion of free water to hydrates such that there isn’t any water for
capillary bridging. Common AAs are often quaternary salts or other surfactants.
The other means of hydrate prevention is engineering design. When feasible, certain measures can be
designed into a production system to prevent it from entering hydrate-forming conditions. This is often
done through insulation but there are many other techniques. Design aspects can reduce the amount of
inhibitor necessary but, in most cases, some form of chemical inhibition will always be necessary for
transient operations or long term shut-ins.

Typical Hydrate Prevention Systems

Most hydrate prevention systems are composed of a combination of design principles and chemical
inhibition. The optimal system depends on many factors, including type of fluids produced, tieback
distance and depth, reservoir temperature and pressure, and operational philosophies like no-touch time,
etc. The design of a hydrate prevention system incorporates any manner of redundancies and safety factors
to ensure its ability to work at all times and for all potential scenarios. As such, it can be built to sizes with
up to 100% excess capacity [1]. With this excess capacity comes the opportunity to overdose inhibitor.
In some instances, overdosing can be upwards of 80% excess to ensure that hydrates will not cause
problems [2]. However, operational experience over the life of the field can bring that overdosing (and
related cost) down.
It is the compounding of uncertainties in a hydrate prevention system that drives its size and overdosing
to extremes. For example, in a monoethylene glycol (MEG) based system, uncertainties can arise from the
measurement of water being produced, the salinity of produced water, gas composition (and subsequently
hydrate equilibrium curves), ambient temperatures, lean MEG quality from reclamation and recovery, and
MEG injection rates. Operational experience can provide assurances that some aspects of the system are
not uncertain, but so too can existing technologies provide the same assurances. Components of a full
MEG system exist today that can drive down the uncertainties of a system, unleashing the ability to save
capacity, operate tighter, and – if identified early enough – reduce capital costs by reducing capacity.

Water Measurement
The transition to any smarter hydrate management system begins with the accurate measurement of water.
Water is an essential component of hydrate formation and is the substance with which inhibitor dosages
OTC-26854-MS 3

are determined. In previous work, the measurement of water has been attributed as the largest contributor
to uncertainty in hydrate management [2].
A basic level of water measurement begins at the separator. Separator volumes are accurate and
absolute measurements of water production for the total field. Many fields have comingled wells though,
so separator volumes alone are insufficient for determining each well’s water production. Differential well
test data is then added to total volumes, which gives a representative picture of each well’s production.
Due to the nature of differential well tests - the inability to assign a particular salinity to each well and
other complications - uncertainties in water production can range upwards of ⫾30 [2]. In large gas fields
where wells are each producing hundreds of standard cubic feet per day, ⫾30% could be the difference
between hundreds of barrels of MEG.
An alternative to the standard technique would be one in which each well’s tree is fully instrumented.
The suite of commercially available meters, in combination with modelling, can yield a water flow rate
and salinity measurement through the span of 0 to 100% watercut. What follows is a brief summary of
a more detailed paper explaining a means of measuring water production [3].
PVT Models
Most dry gas fields produce initially without the appearance of any liquid water at reservoir conditions.
Assuming there is water contact somewhere in the reservoir, it is a reasonable approximation to assume
that the gas in place is saturated with non-saline water. To capture the water being produced with
otherwise dry gas, one must have an estimate of water saturation level and gas flow rate. The saturation
percent multiplied by gas flow will yield water production. At common reservoir conditions, water
saturations are on the order of 0.5% by mole, which may not seem like a large quantity. However, this
can equate to water production rates of over 75bbl per 100MMscf, which could require over 10 tonnes of
MEG to inhibit. Typical uncertainties in this modeling can be as low as 5% or as high as 100%, so it is
good practice to devote effort to reducing this uncertainty, as it is the primary source of uncertainty in
water measurement until it reaches its detection limit.

Figure 1—Full spectrum of water volume fraction measurement

Detection of Condensed and/or Formation Water

Newly available commercial measurement techniques have reduced the detection limit of liquid water to
as low as 50ppm. This technology uses measurements of conductivity and permittivity to detect the
appearance of condensed water, saline water, and measure any changes in the water’s salinity. Detection
of condensed water permits the confirmation of saturation models with measurements, but also the
transition to measurement from solely using those models. Detection of salinity grants the operator a wider
variety of options upon the production of formation water. One can shut in the well, prepare MEG
reclamation in addition to recovery, alter MEG dosage to incorporate the inhibiting nature of brine, or
other options depending on operational philosophy. Also, the combination of permittivity and conduc-
tivity measurements yields a measurement of salinity, as shown in the figure below. If the conductivity
probe is placed in the proper location it can capture salinity measurements from water continuous wetted
regions of pipe. Knowledge of salinity has many uses in addition to hydrate management. For example,
4 OTC-26854-MS

a strategy in which water flooding is used may result in a scenario where water appears first without
salinity as it condenses out of the gas. The salinity then rises as formation water makes its way into the
well. Then salinity may drop as injection water fingers into the well, indicating a necessary change in
production strategy. Meanwhile, the salinity measurement is revealing just how much inhibitor is needed
for that particular well.

Figure 2—Conductivity probe domain of measurement

Multiphase Flow Measurements

In order to make use of water concentration and salinity measurements, or PVT models of saturation,
additional rate measurements are necessary. These measurements of subsea wet gas flow rates have
become more standard as fields become bigger, deeper, or farther out. The primary means of measuring
gas rates is using pressure drop, e.g. with a Venturi. The addition of a gamma densitometer with high
frequency acquisition completes the measurement, yielding gas flow rates with accuracy down to 2.5%
and water volume fractions with accuracies down to 0.2%. The sensitivity of the measurement to water
volume fraction can be 0.05%, which is then the threshold where measurement priority switches from
models/conductivity to the meter. Below this threshold there is not a detection technique to take a direct
measurement, which is why the models and conductivity measures are used. Before this switch occurs, the
gas flow measurements and saturation models can establish a reference water volume fraction with which
uncertainties can be further lowered.
Technology exists and is in use that can provide the detection, salinity, and flow rate of water per well
in large bore gas fields spanning the spectrum of no-free-water up to 100% water volume fractions. This
can be accomplished by combining saturation models, conductivity and permittivity measurements, and
more traditional wet gas flow metering via Venturi and densitometry.
OTC-26854-MS 5

Figure 3—Measurements and their accuracies encompasing a hydrate management system. A full analysis is available in Ref [3].

MEG Regeneration and Distribution

Another contribution to the large uncertainties in a MEG-based hydrate prevention system comes from the
MEG regeneration, reclamation, and distribution portion of the process. This accumulated uncertainty can
range upwards of 30%, with the greater portion specifically attributed to MEG distribution [2].
Chemical Injection Metering Valves
MEG distribution to each well can be both controlled and measured through the use of subsea chemical
injection metering valves (CIMVs). A third generation technology exists by which the flow of inhibitor
is throttled through a low pressure loss device and measured through acoustic techniques. The ultrasonic
flow measurement is performed bi-directionally, with and against flow, which then provides a flow rate
measurement with lower dependence on fluid properties. This Doppler Effect produces a measurement
with an uncertainty of only 3% measured value over the entire range of flow rates (fractions of a liter per
hour to over 25,000 liters per hour depending on CIMV size).

Figure 4 —Ultrasonic flow measurement schematic

In terms of MEG usage, the use of CIMVs provide both the ability to dose MEG with precision to wells
of different pressures, but also a means of corroborating other measurements. For example, a known
quantity of MEG can be injected into the production stream to push the volume of the aqueous phase into
6 OTC-26854-MS

its detectable range. Then the known amount of MEG can be reduced from the measurement, yielding a
free water measurement. This information can be fed back into the determination of MEG needs and
CIMV set points.
In the absence of a CIMV, MEG may only be controllable by regulating pressure drops from the
distribution system to the production system. This stipulates that lower pressure wells receive more MEG
than higher pressure wells, irrespective of which wells are producing more water. MEG injection can only
be measured by these pressure drops as well, leading to an uncertain measurement that further compounds
uncertainties.
MEG Recovery Units
One of the primary reasons MEG is chosen over other thermodynamic inhibitors is because it is more
easily recovered at the platform for reuse. MEG and water are miscible and pulled from the production
stream by common processes. MEG and water separation is also well understood. It is typically
accomplished by some form of vacuum distillation where the vacuum is pulled to keep operating
temperatures below those at which MEG degrades. The real difficulty in recovering and reclaiming MEG
occurs when impurities like mono- and divalent salts appear from formation water.
Monovalent salts, those composed of alkali metals, are precipitated through concentration to saturation.
Divalent salts, those composed of alkaline earth metals, are not so easily removed. New technology on the
market contains a process by which these hard-to-remove divalent salts are reacted and precipitated out
prior to entering the MEG recovery unit. This reduces the MEG lost to the process which is equivalent
to reducing the amount of MEG needed to replenish the system, lowering overall costs. Likewise, the
removal of divalent salts combined with removing monovalent salts from brine, as opposed to MEG, can
further reduce MEG losses and also lower energy needs for MEG recovery. Using the best technology to
regenerate and recover MEG means the purity can be controlled. Knowing the purity of the lean MEG
means a determination of precise dosage can be calculated so the right amount of inhibitor can be used
with certainty.

Figure 5—A MEG recovery process with fullstream reclamation

OTC-26854-MS 7

Decision Making Components

The discussion thus far has focused on the hardware necessary for a MEG-based hydrate management
system. It entails the physical components necessary for the full MEG loop, including distribution,
measurement, and recovery. What have been missing are the tools necessary for decision making and
dosage determination.
To determine how much inhibitor is necessary, no matter the inhibitor being used, one must have
knowledge of both the operating conditions of the hardware and thermos-physical conditions of the entire
production system. The most hydrate favorable conditions visited within a system are most often not
found where the inhibitor is injected. They are usually found downstream where no instrumentation is
installed.
A means of building together both decision-making tools and the collection and verification of data is
available through the combination of multiphase simulation and condition/performance monitoring
services. The multiphase simulation, which is the tool most often used in determining hydrate manage-
ment philosophies, resides on the same computational platform with which equipment condition and
performance monitoring is conducted. As such, the principles of hardware monitoring can be applied to
the simulated system that serves to both maintain accuracy and confidence in its utility as a tool.

Condition and Performance Monitoring

Condition and performance monitoring of equipment is vital to maximizing their availability, and thus
their utility. With experienced monitoring, one can be assured that the measurements being used by a
management system are accurate. Drifting values can be identified long before they cause a fault.
Diminished performance can be identified long before a failure. Quality condition monitoring can
significantly reduce measurement uncertainty, to which upwards of 30% uncertainty has been attributed
[2]; however, a full quantification of this uncertainty reduction has not yet been ascertained.
Monitoring experience can also be applied to virtual systems that are used to make decisions, either
manually or automatically, that impact how MEG is used for hydrate prevention. The principles of
monitoring can be applied to systems whether or not they physically exist. For example, multiphase
simulation enables an estimate of properties between points of measurement – like having a temperature
and pressure measurement where sensors do not exist. This temperature and pressure estimate could be
monitored as if the data were coming from sensors, inducing the same warnings and alarms that would
be invoked should the estimates deviate from expectation. As such, monitoring is extended to the entire
production system and MEG management system, feeding information from the physical to the virtual and
using results from the virtual to operate the physical.
8 OTC-26854-MS

Figure 6 —Condition and performance monitoring dashboard

Multiphase Flow Simulation

The other logical component of a hydrate management system, no matter the inhibitor chosen, is the
process simulator. In the case of most offshore, upstream, hydrate management systems, this simulator
would be one that can simulate transient and multiphase fluid flow. There are a few such simulators
available commercially, with one in particular being the defacto industry standard. It can be run offline,
as it was originally intended to manually investigate multiphase production systems or, in the case of an
optimized hydrate management system, it can be run online in real-time.
A real-time production system simulation can provide insight into what is happening between the
points of measurement. Changes in production are captured and forecasted before they may occur, which
can lead to preventative actions. One need not update their hydrate management strategy annually, as is
often done. Instead it can be updated on-the-fly or even automatically. If more water breaks through in
a particular well, this is captured by conductivity and multiphase meter measurements, which then update
the online simulator and indicate more MEG is necessary for full inhibition. A fully updated simulation
tool also enables a constant assessment of worst-case-scenario hydrate forming conditions, typically
ambient seafloor temperature and shut-in wellhead pressure. These are the conditions to which many large
offshore gas fields are inhibited against, but they are constantly changing. Having an updated model of
constantly changing conditions enables the constant update of the hydrate management strategy.

Figure 7—Subsea architecture that is modeled for hydrate management

OTC-26854-MS 9

To keep the simulation up-to-date requires periodic maintenance. This maintenance need not be
routine, rather it can be initiated through customized warnings and alarms from the monitoring of its
condition. Whereas a typical update to process models would occur annually, and with a large effort, the
real-time simulation on a condition monitoring platform would indicate when it required maintenance and,
if conducted promptly, will have issues resolved before they compound into a larger need. A standard
multiple-weeks long project of updating process models per year can then become a task requiring only
one day per month. A constantly updated model enables one to have confidence in its results and make
decisions based upon them.
Conclusions
The pieces necessary to construct an intelligent hydrate management system are readily available in the
market today. For example, metering technology exists that can precisely measure water production as it
emerges from each well. We can then target that water with a precise quantity of MEG, tuned to the needs
of the well as determined by live process simulation, because we know the necessary dosage in real-time
and the condition of the equipment determining that dosage. It is no longer required to dose as much
inhibitor as possible to all upstream points in a production system. We can instead tighten the system and
save on capacity and dosage, and therefore cost, to bring otherwise uneconomical fields to profitable
production.

References
1. Wilson, A.W., Overaa, S.J., and Holm, H. Offshore Technology Conference, OTC 16555, Houston, TX, May 3-6
(2004).
2. Gupta, G., and Singh, S.K. Society of Petroleum Engineers Oil and Gas India Conference and Exhibition, SPE
153504, Mumbai, India, March 28-30 (2012).
3. Leong, Y.S., Lupeau, A., Smith, M.T., Grzelak, E. South East Asia Flow Measurement Conference 2016, Kuala
Lumpur, Malaysia, March 16-17 (2016).