Available online at www.sciencedirect.

com

ScienceDirect
Energy Procedia 114 (2017) 5083 – 5099

13th International Conference on Greenhouse Gas Control Technologies, GHGT-13, 14-18

November 2016, Lausanne, Switzerland

Influence of phase behaviour in the well design of CO2 injectors

Luis Acevedoa*, Ajay Choprab
a
Shell UK Ltd, 1 Altens Farm Road, Aberdeen AB12 3PY, United Kingdom
b
ARC Well Enginnering Ltd, Aberdeen, AB32 6AY, United Kingdom

Abstract

The design of CO2 injection wells differs from that of gas and water injectors owing to the effects of the CO2
phase behaviour. Each CO2 injection well requires a design specific to the reservoir conditions, rates and injection
stream characteristics just as there is no one well design that fits all hydrocarbon production requirements but there
are key themes that emerge from our experience in the design of Carbon Capture and Storage (CCS) wells.
In depleted reservoirs, the CO2 injection poses unique challenges when compared to injection for Enhanced Oil
Recovery projects or aquifer injection. The potential to generate low temperatures due to CO2 expansion into vapour
phase during steady state injection can be a concern. Here one technique is to employ a “reverse velocity string” that
is to deploy slim tubing to create back pressure that avoid the production of low temperature during normal injection
conditions. Well operations or transient effects can also generate low temperatures, albeit of short duration.
During the highly unlikely event of loss of containment leading to flow to atmospheric conditions, there could be
a rapid expansion of CO2 in the top part of the well. Such an expansion would be accompanied by a reduction in the
temperature due to the Joule Thomson effect as the CO2 travels down the saturation and possibly the sublimation
line. This cooling potential is one driver for technological advancement of present day well components. Once the
flow is identified, the emergency shutdown system of the well automatically activates (for example closing the tree
valves and/or the Subsurface Safety Valve – SSSV). Modelling has been conducted to understand the limitations of
the current safety critical equipment for different scenarios. The result of modelling indicates that even in low
magnitude releases, the SSSV can be subjected to low temperatures with metal temperatures close to the triple point
even when set deep in the well.
In addition to the normal requirements of the well equipment in terms of pressure, loads and chemical
interactions of the injection fluid; this cooling effect of the CO2 has an impact on the well design. Existing well

* Corresponding author. Tel.: +44 1224 888913; fax:+44 1224 882874

E-mail address: luis.acevedo@shell.com

1876-6102 © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of GHGT-13.
doi:10.1016/j.egypro.2017.03.1663
5084 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099

equipment, including safety critical components (tree valves and SSSV), have limitations in terms of material
embrittlement, qualification envelopes and performance of fluids (annuli and power fluids) at this low temperature.
The placement of some well components, such as SSSV, also may be affected by the cooling effect, to ensure its
ability to close in case of a release.
It is crucial to make a distinction between functionality and integrity requirements, whereby components retain
integrity under this low temperature excursion even though they might lose operability. This differentiation has to be
identified and agreed at an early stage in the well design. Another critical distinction is between the system and
component design rating as some components of the system might be exposed to low temperature subject to
placement in the well.
An area of well design for the lifecycle of the well that requires significant research and development into is
around well intervention, both planned routine intervention with wireline and unplanned remedial intervention via
platform and/or vessel. Again here the CO2 phase behaviour can pose challenges to current equipment and methods.
This paper aims to highlight the effects of CO2 phase behaviour on well design and the considerations for a
robust and optimised well design for CCS.
© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
© 2017 The Authors. Published by Elsevier Ltd.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of GHGT-13.
Peer-review under responsibility of the organizing committee of GHGT-13.
Keywords: CO2; well design; leak; release; phase behaviour; subsurface safety valve; tree valves; CO2 expansion; Joule Thomson; tubing.

1. Introduction

The design of CO2 injection wells differs from that of gas and water injectors owing to the effects of the CO2
phase behaviour [1]. Each CO2 injection well requires a design specific to the reservoir conditions, rates and
injection stream characteristics just as there is no one well design that fits all hydrocarbon production requirements.
This paper concentrates on the impact of the phase behaviour of CO2 on the well design. The paper is based on
the Peterhead Carbon Capture and Storage project in the United Kingdom, although there are key themes that
emerge from our experience in the design of CCS wells. Although the Peterhead CCS project was cancelled at the
end of 2015 at the end of the detailed design phase, it provided learnings that may be useful for future projects.

1.1. CO2 phase behaviour and well operations

In the oil and gas industry, planned operations and uncontrolled events have normally been modelled using the
OLGA software (SPT group - a Schlumberger company), in the module for CO2 single component. The software
considers heat properties in the well, heat exchange with the surroundings and the phase behaviour of CO2 using the
Span and Wagner equation of state.
Injection conditions (section 2), planned operations (section 3) and uncontrolled events (section 4) combined with
the phase behaviour of CO2 will affect the well design (section 5).
The most important aspects of the well design affected by the CO2 phase behaviour are:
x Well Concept
x Material selection
x Equipment suitability

1.2. Peterhead CCS - Project Dimensions and Characteristics

The Peterhead Carbon Capture and Storage (CCS) project planned to separate, capture and permanently store
CO2 [2] arising from combustion of natural gas in a combined cycle gas turbine. It was intended to inject up to one
million tonnes per annum (Mtpa) of 99% purity CO2 over a period of up to 15 years for storage in the U.K.
Continental Shelf within the depleted Goldeneye hydrocarbon field. Following capture, compression and
conditioning at the Peterhead power station the dense phase CO2 would have been transferred into the Goldeneye
 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099 5085

pipeline system at a pressure of approximately 120bar. The Goldeneye platform was to be reused. The installation is
normally unmanned with frequent visits from the operators.
The CO2 would have arrived at the platform at seabed temperature. CO2 arriving at the Goldeneye infrastructure
would have been in liquid state between 2.3 and 10.1 °C depending on the season and at 120bar approximately.
The CO2 flows to an injection manifold where the flow would be directed to one or more wells. It was planned to
re-use the existing hydrocarbon wells with the necessary modifications. The wells each have a non-cemented lower
completion with gravel pack and sand screens, these were to be re-used. The wells are currently completed with
single 7” x 5” tubing used for hydrocarbon production. The dry trees on the Goldeneye platform and the ability of
wells to free flow would have led to the requirement for additional safety equipment such as a subsurface safety
valve in each well.
The CO2 would have been injected into the storage site at a depth >2516m (8255ft) below sea level into the
previously gas bearing portion of the high quality Captain Sandstone Member – in total a 130km long and <10km
wide ribbon of Lower Cretaceous turbiditic sandstone fringing the southern margin of the South Halibut Shelf.
Injection of 15 million tonnes of CO2 would have raised the pressure in the main interval, the Captain reservoir to
around ~259bara (3750psia) at the end of injection, close to normal hydrostatic pressure.

2. Steady State Injection and Well Concept

In depleted reservoirs, CO2 injection poses unique challenges compared to injection for Enhanced Oil Recovery
projects or aquifer injection where the reservoir pressure is close to or higher than hydrostatic conditions.
At the predicted start of the injection the reservoir would have been depleted (~2750psi, ~190bar reservoir
pressure) for a pressure gradient of approximately 0.724 bar/10m (0.32 psi/ft). With injection volumes and time the
pressure in the reservoir would have increased to close to hydrostatic conditions ~ 0.979 bar/10m (0.433psi/ft).

2.1. Free expansion of the CO2

The existing wells in the Goldeneye reservoir are completed for hydrocarbon production with a tubing size of 7”x
5”. The CO2 would have arrived cold at the platform at approximately the seabed temperature ~4°C. Simulations in
OLGA were carried out to evaluate the performance of the CO2 during continuous injection in the existing tubing.
Results indicated low temperature in the top section of the well (with a minimum of -25°C and sub-zero
temperatures above ~2000ft (610m) True Vertical Depth (TVD)) during the continuous injection period. Increases
in the reservoir pressure lead to less severe cooling (Figure 1a.). The reduction in temperature is due to the flashing
of the liquid CO2 to gas/liquid CO2 caused by the low reservoir pressure. Whilst the top part of the well is operated
in two phases (liquid-vapour), the bottom part of the well is operated in single phase (Figure 1b).
5086 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099

ExistingCompletionͲ FreeExpansionofCO2 b ExistingCompletionͲ FreeExpansionofCO2

a FluidTemperature HoldUp
0 0

500 500

1000 1000
TrueVerticalDepth,m

TrueVerticalDepth,m
1500 1500

2000 2000

2500 2500

3000 3000
Ͳ30 Ͳ20 Ͳ10 0 10 20 30 40 0 0.25 0.5 0.75 1
Temperature,°C HoldUp,fraction
2750psi(~190bar)ReservoirPressure 3000psi(~207bar)ReservoirPressure 2750psi(~190bar)ReservoirPressure 3000psi(~207bar)ReservoirPressure
3250psi(~224bar)ReservoirPressure 3500psi(~241bar)ReservoirPressure 3250psi(~224bar)ReservoirPressure 3500psi(~241bar)ReservoirPressure
3830psi(~264bar)ReservoirPressure 3830psi(~264bar)ReservoirPressure

Figure 1. Simulation results of continuous CO2 injection using the existing well completion at different reservoir pressures. This case is
considered an unconstrained expansion of CO2. a) Fluid Temperature at different well depths. b) Hold Up (area proportion of gas in the tubing, 0
means only gas, 1 means only liquid) at different well depths.

2.2. Injection Options and selection of the proposed solution

There are a number of ways that the phase behaviour of CO2 can be managed. The merits of each of these are
specific to each project and must be considered at the conceptual design stage.
The general isenthalpic expansion from liquid CO2 (arrival CO2 conditions to Goldeneye platform) is represented
in Figure 2. If the CO2 is kept in liquid phase, then there will be a small reduction in temperature for large pressure
drops. If the liquid CO2 is allowed to expand down to the saturation line, there will be a significant reduction of
temperature for a small change in pressure. This restriction of the expansion of the CO2 to liquid phase can be
achieved by inducing an extra pressure drop in the well (not only given by the reservoir pressure). Means of
achieving this include the installation of a small diameter tubing creating back pressure by friction loss or a pressure
drop device (downhole choke).
 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099 5087

120

100

Critical Point
80
Pressure, bar
60
Liquid
40

20
Gas

0
-40 -20 0 20 40
Temperature, °C

CO2 Saturation Line (gas/Liquid)

Isenthalpic expansion - Downstream Choke Performance

Figure 2. Isenthalpic expansion of CO2 arriving to the platform (~120bar and 4°C – liquid phase) in the well choke for the low reservoir pressure
case (2750psi - ~190bar reservoir pressure). If the CO2 is kept in single phase (liquid) then the restricted expansion of the CO2 will generate
temperatures compatible with the well design.

Management of the phase behaviour of the CO2 could also be achieved by heating up the CO2 stream at the
platform above a limit at which the resultant expansion of the CO2 (and hence fluid temperature) would not lead to
freezing conditions in the wells. Another concept option was to allow the pipeline to flow in two phases as required
by the injectors well(s) without the requirement to choke the wells.
Considering the specific Peterhead CCS project conditions, it was decided to perform modifications to the wells
by installing small diameter tubing in the well and restricting the expansion of the CO2 to within liquid phase by
creating back pressure during normal injection conditions [3].
Heating up the CO2 was considered very expensive and CO2 intensive as new equipment would need to be
installed on the platform requiring the installation of a new jacket. Two phase flow in the pipeline was complicated
from the control perspective and has operational issues as predicted slugs of CO2 liquid and CO2 gas would have
arrived at the platform and the well(s) would have injected when filled with liquid CO2 and would have attempted to
backflow if the well was filled with gaseous CO2.

2.3. Controlled injection conditions – Proposed solution

As discussed above, the phase behaviour of the CO2 needs to be managed during long term injection to avoid low
temperatures. During normal injection conditions the CO2 should be kept in liquid phase by creating back pressure
in the well by using a small diameter tubing [4]. The well concept is to use a friction dominated scenario by high
velocities. This concept is sometimes used to restrict production from wells.
With appropriate size tubing installed in the upper completion the wellhead pressure will be increased to the
extent that it lies above the saturation line, Figure 3. Varying tubing sizes and lengths can be installed in the well to
meet the injection requirements [3]. A minimum injection rate per well is required to maintain the liquid phase
injection condition because the friction induced backpressure is a function of flow rate. A minimum tubing head
injection pressure was determined by the expected arrival temperature of the CO2 at the platform.
5088 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099

TubingSizeSensitivity
(BasedonGYA01well)
5500
InjectionBottomHolePressure,psi 5000
4500

41/2"tubing

51/2"tubing
4000
3500
3000
2500 27/8"tubing 31/2"tubing

2000 CCPRate

0 10 20 30 40 50 60 70 80
InjectionRate,MMscfd
InflowͲPreservoir2750psi InflowͲPreservoir3750psi
5.5"tubing50bar 5.5"tubing115bar
4.5"tubing50bar 4.5"tubing115bar
3.5"tubing50bar 3.5"tubing115bar
27/8"tubing50bar 27/8"tubing115bar

Figure 3. Performance of different upper completion sizes (based on a single well simulation). Smaller tubing sizes will have a smaller operating
envelope. Different size lengths can be used in combination to the well deviation to select the appropriate tubing size per well.

The concept involves operating the CO2 at average velocities of around 5 to 10 m/s with a high density fluid
(900-970 kg/m3). This was checked against abrasion/erosion of the tubing, vibration and intrinsic instability of the
tubing due to the large fluid momentum. Keeping the tubing in tension and installing a short tail pipe in the
completion were recommended to avoid instability issues.
In order to determine the minimum required number of wells and optimal tubing size for each well, simulations
were carried out for the life cycle of the project considering the increase in reservoir pressure, the expected
injectivity and the CO2 delivery rates from the capture plant. Further sensitivities were run regarding the arrival CO2
temperature to the platform, frictional parameters in the steel (roughness) and injectivity. Overlapping of the well
envelopes was planned in case of failure of a particular well to crate redundancy.

3. Planned Well Operations

Well operations or transient effects can also generate low temperatures, albeit of short duration [5].

3.1. Closing in and re-starting injection operations

During transient operations (closing-in and re-starting injection operations), a temperature drop is observed at the
top of the well for a short period of time. This is due to the reduction of friction caused by a lower injection rate.
The duration of these operations dictates the extent to which the various well elements are affected. The sequence of
steady state injection, closing-in operation (30 minutes), closed-in time (10minutes), and starting-up operation
(30minutes) was simulated for the low reservoir pressure case using OLGA, Figure 4.
 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099 5089

Temperatureatthetopofthewell
SequenceofclosingandopenningaCO2injectorwell

Injection ClosingͲin ClosedͲ ReͲStartingInjection Normal

Operation In Operation Injection
10
CO2ͲFluid

0 Tubing

Temperature AͲannulus
atwelltop
°C Ͳ10
Production
casing

BͲannulus
Ͳ20

Surface
Casing
Ͳ30
0 20 40 60 80 100 120
Time,min

Figure 4. Temperature at different well elements in the top of the well for the sequence of closing-in a well after a long term injection period, well
closed-in for a short period of time and re-starting injecting for a low reservoir pressure case (2750psi - ~ 190bar). CO2 injection rates during the
closing-in and the re-starting operations are proportional with time.

The effect of the CO2 temperature dropping for a short period of time to ~ -20°C at the tubing head is observed in
the tubing and to a lesser extent in the A-annulus. The production casing and other well elements are not affected by
these transient activities due to their short duration.
This temperature drop occurs in the top part of the well. The temperature in the bottom part of the well is
unaffected by these short term events as the reservoir pressure prevents the CO2 from expanding. Figure 5 illustrates
the fluid conditions during the well operations described here at time=55minutes of Figure 4.

5090 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099

ProposedCompletionͲ ControlledCO2Expansion
FluidTemperature
0

500

1000
TrueVerticalDepth,m

1500

2000

2500

3000
Ͳ30 Ͳ20 Ͳ10 0 10 20 30 40
Temperature,°C
ContinuousInjection
Surfacecoldestduringtransients(t=55min)

Figure 5. Fluid temperature in the well for a continuous injection and the coldest temperature at the top of the well in a closing-in operation for a
low reservoir pressure (2750psi, ~ 190bar). Temperature drops occurs in the top of the well whilst the bottom part of the well is all but
unaffected.

The reservoir pressure affects the temperature calculation during the transient calculations. The lower the
reservoir pressure, the lower the surface temperature expected during transient operations and hence the higher the
stresses/impact on well design.

3.2. Testing SSSV

Testing the SSSV is a controlled operation [5]. Typically, in hydrocarbon wells, the pressure above the valve is
bled off to 10% of the Closed-In Tubing Head Pressure (CITHP); then the surface pressure is monitored for leaks at
the valve.
For CO2, this implies a change in the phase of the CO2 in the top of the well. To conduct an unambiguous SSSV
test, the surface pressure needs to be monitored in a single phase which is normally gaseous CO2. This is a delicate
and lengthy operation because the CO2 temperature may be very low as the CO2 interface (gas/liquid) propagates
down the well (similar to the uncontrolled scenario) during the bleed off operation for which natural boiling-off of
the CO2 is required. By maintaining the pressure (above a certain value), the temperature of the interface is also
maintained above a certain limit.
Surface temperature (easily monitored) will not indicate the temperature down at the interface level. Distributed
Temperature Sensing (DTS) can help identify the interface as it travels down to minimise the duration of the test.
Re-commencing injection after the SSSV test requires that the pressure in the tubing to be relatively high to avoid
low temperatures. To overcome low temperatures during the test and considering the current operability limitation
of the valve (section 5.2.4) a minimum bleed off pressure is to be stipulated.
 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099 5091

4. Highly unlikely events (loss of control)

A loss of control to atmospheric conditions (leak/blowout or release) is highly unlikely. The strategy is to prevent
these incidents by installing the suitable well equipment for pressure containment. Although the potential for such a
well incident is extremely low, this eventuality must be considered in the well design [6].
If an uncontrolled flow were identified, the emergency shutdown system of the platform/well would be activated
(for example closing the tree valves and/or the SSSV).
One important factor is the ability to limit the amount of CO2 released to the atmosphere. Normally the well
would be closed at the tree valves initially to stop the flow. Under certain circumstances, for example damage to the
tree, the tree valves may not be able to stop the flow. Limiting the release volume under these scenarios could be
achieved by the installation of a Sub Surface Safety Valve (SSSV). This valve would be usually set deep in the well.
Questions to answer when contingency planning for uncontrolled leaks to atmospheric conditions are around the
ability to shut valves in the well and the potential to exceed the temperature rating of the equipment which could
possibly lead to brittle fracture during the process. The influence of the low temperature on the different well
elements will vary depending on the leak rate (determined by the leak geometry) involved, the volume, the duration
and the heat transfer from the surroundings of the well and internal elements of the well.
Modelling was conducted to understand the limitations of the current safety critical equipment for different
scenarios. Favourably low CO2 temperatures are always associated with CO2 expansion to low pressures.

4.1. Can the SSSV close?

During a highly unlikely event of loss of containment leading to flow to atmospheric conditions, there would be a
rapid expansion of CO2 in the top part of the well [7] [8]. This expansion would be accompanied by a reduction in
the temperature due to the CO2 expansion as the CO2 travels down the saturation and possibly the sublimation lines.
This pressure and temperature reduction propagates down the well until equilibrium is reached limited by either
the inflow from the reservoir or the leak geometry. The top part of the well would be at very low temperature while
the bottom part would be not affected (close to the reservoir temperature or bottom hole injection temperature
depending on the well status before the incident).
One of the criteria for setting the SSSV in a CO2 well is the ability to identify the leak and close the valve before
the cold front reaches it. This drives the valve to be set deep in the well.

4.2. What happens after the SSSV is closed?

In the case where the barrier between the store and the atmospheric conditions is provided by the SSSV, the
volume of CO2 between the tree and the SSSV would be released, i.e. emptying of the tubing above the SSSV.
The leak size will determine how fast the tubing above the SSSV is emptied. The CO2 liquid/gas interface would
travel down in the well until it reaches the SSSV and stay there until all liquid CO2 boils off into gas at low pressure,
Figure 6.
5092 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099

Tubing full of CO2 Leak created CO2 leak G/ L interface Tubing empty
Dense phase CO2 leak G/ L interface at SSSV level
(worst case DS leak to 1bar travelling down
More CO2 mass)

SSSV Closed

Figure 6. Process of emptying the tubing in a release to atmospheric conditions scenario. Communication with the reservoir is prevented by
closing the SSSV. Once the valve is closed the process of emptying the tubing starts with a gas-liquid interface developing and traveling to the
SSSV.

The temperature at the tree upstream of the release would depend mainly on the leak diameter, Figure 7.
Modelling shows small holes will not pose temperature issues to the tree/wellhead equipment. For significant leaks
(>1 in diameter) the CO2 will have temperatures from -30 to -70 °C for approximately 3 hours. This cooling drives
the requirement for the tree to be rated to API 6A temperature class ‘K’.

5 mm
1 mm
28 mm
10 mm

50 mm

100 mm

Figure 7. Fluid temperature upstream of the leak point at different hole sizes (diameter) for a release scenario. The released CO2 volume is the
volume of CO2 contained between the SSSV and the tree. (Released is started at T=2h)
 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099 5093

However, due to the limited release volume and the mass of the wellhead, the metal temperature of the wellhead
would remain significantly higher than the CO2 temperature. This was confirmed with detailed thermo flow
modelling of these well elements under a worst case scenario.
The transient flow modelling indicates that even in low magnitude releases, the SSSV can be subjected to low
temperatures (Figure Figure 8) with metal temperatures close to the CO2 triple point temperature. When the
interface reaches the SSSV, it can create localised low temperatures at the valve for a couple of hours. Modelling
indicates that even for small leaks (>5mm diameter) the CO2 temperature at the valve can reach -78.5°C for 1-1.5h
with the metal temperature close to the triple point temperature (-56.6°C). For very small holes (< 1mm diameter,
associated with weeps and seeps), heat from the surroundings will warm the CO2 and limit this effect.

5 mm

10 mm
50 mm 28 mm

Figure 8. Fluid temperature directly above the SSSV at different hole sizes for a release scenario. CO2 release starts at T=2h

Depending on the well conditions (e.g. fluid temperature below the critical point), this effect could also be
observed in deep set valves where the phase behaviour of the CO2 is dominant over heat transfer from the formation.

4.3. What happens in the well after the release?

SSSVs normally have an acceptable leak rate according to the API-RP-14B. In a CO2 well, this might create
localised cooling of the primary well elements.
After the SSSV is shut and the CO2 inventory is emptied, the pressure above the valve would be close to
atmospheric. Below the valve, the CO2 would be at high pressure given by the reservoir. Any CO2 that leaks across
the valve would suffer a JT expansion to atmospheric conditions. This expansion can reduce the CO2 temperature
down to -78.5°C. This effect is localised across the valve.
Even under the small acceptable leak rates, with time, the well elements localised around the SSSV would be
cooled significantly because the CO2 expands inside the tubing (and not to the atmosphere as in the case of a surface
release) and the leak is continuous. This localised reduction of temperature would not occur in the case of a perfectly
sealing valve.
5094 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099

5. Well Design

5.1. Well Design Philosophy

The design of a CO2 injection well may initially present itself to be similar to that of a standard injector well and
rightly so as the principles involved in the design of a hydrocarbon production or other injection well apply here too
[9]. The normal requirements of understanding the well loads, equipment pressure ratings and chemical interactions
are all relevant [10]. Where the difference lies is in the design of the well to manage the CO2 phase behaviour and
the potential cooling effect of the CO2.
Equipment in the oil industry is normally designed and qualified for higher temperatures, in recent years much of
the technological advancement has been in the HPHT (high pressure and high temperature) field. Equipment
currently available is not designed or rated for use at the low temperatures discussed here. It is not possible or an
effective use of resources to qualify all the well components to -78.5°C (CO2 isenthalpic expansion to atmospheric
conditions).

5.1.1. Well design and CO2 phase behaviour

As part of the Peterhead CCS project it was planned to workover (re-complete) the Goldeneye wells as some
elements needed to be modified for CO2 injection service (section 2.1). The concept adopted for the Peterhead CCS
project was to maintain the CO2 in liquid phase by utilising small diameter tubing (section 2.3).
Another factor driving the well design considerations is the ability of the well to safely withstand the cooling
effects during transient conditions such as well start-ups and shut downs (section 3.1). During transient conditions
some cooling is expected (in the top part of the well) due to the CO2 expansion, to some extent this can be managed
by employing specifically engineered well operating procedures. It is possible to procure equipment capable to
withstand these short duration temperature fluctuations.
An extreme factor that also needs to be considered is the drastic cooling that may take place in an unplanned
condition such as a loss of well control (section 4). At these low temperatures, material embrittlement would be a
major concern, however for well design it is important to remember that the low temperature is always associated
with CO2 expansion to low pressures.
Rather than designing and qualifying all well components to the predicted isenthalpic CO2 expansion to
atmospheric conditions (-78.5°C), modelling techniques were employed to indicate minimum potential temperatures
at each of the well components and planning was undertaken accordingly.
Safety critical components were to be designed and rated for the low temperature conditions. Individual
components were selected based on the temperature modelling predicted they would be subject to [11]. Well
integrity was prioritised over system operability. A well schematic for the proposed well re-completion phase is
included here (Figure 9).
 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099 5095

Proposed Goldeneye CO2 Injection Well

Tubing Hanger

4 1/2" 13.5# Vam Top HT Tubing S13Cr

Control Line protectors

1/4" SSSV Control Line (low pour point fluid) 11x11mm
4 1/2" SSSV (non self equalising)

Casing XO 10 3/4" 55.5# x 9 5/8" 53.5#

4-1/2" 13.5# Vam top HT Tubing 13Cr
XO 4 1/2" 13.5# x 3 1/2" 10.2#
3 1/2" 10.2# Vam Top Tubing 13Cr
3 1/2" PDGM for PDG (tubing gauge)

Proposed 1/4" Hybrid gauge line (electric + FO) 11x11mm

3 1/2" PDGM for PDG (tubing gauge)

3 1/2" 10.2# Vam Top Tubing 13Cr
3 1/2" PDGM for PDG (tubing & annulus gauge)
3 1/2" 10.2# Vam Top Tubing 13Cr
9 5/8" x 3 1/2" Packer (cut to release)
3 1/2" 10.2# Vam Top Tubing 13Cr
3-1/2" Sliding Sleeve (down to open)
XO 3 1/2" 10.2# Vam Top Tubing X dummy seal assembly
Sand Control Packer 13Cr (existing equipment)

G22 dummy seal assembly

Wireline re-entry guide
Formation Isolation Valve (existing lower completion)

4.00" Screens (existing lower completion)

Figure 9. Proposed Goldeneye CO2 Injection Well Schematic, illustrating the new proposed upper completion and existing lower completion

5.1.2. Functional and integrity requirement

It is crucial to establish a distinction between functionality and integrity requirements. At low temperatures, it
may be necessary that a component maintains its integrity however it may not be essential for that component to be
functional at the low temperature.
Safety critical components are to be qualified in line with modelling indications. Non safety critical components
are not required to function/operate at low temperatures, however, they are required to maintain their integrity if
they form part of a well barrier.
Example 1: SSSV. Section 4 explains the events in a well during a release of CO2, here it is critical that the SSSV
closes and limits the release of CO2. The valve does however close before the cold front arrives at the SSSV setting
depth and once the valve is cooled it is not required to be able to open it while cold. It may therefore be derived that
though the valve needs to be able to maintain integrity through a cold excursion it does not essentially require to be
functional at the low temperature.
Example 2: Well monitoring system. During a release scenario the functionality of well monitoring such as the
pressure and temperature gauges, though desirable, is not an absolute necessity. Under such a condition, if these
devices cease to provide reliable data, this would not be a concern. It would however be essential that the system
maintains its integrity. Therefore the connectors and the control line which form a part of the well barrier system
would have to continue and retain integrity. This is different for the transient conditions where the data from the
5096 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099

well monitoring system may be highly valuable. It may therefore be implied that the well monitoring system is
required to operate down to the temperatures anticipated during the transient conditions however, past these
conditions, it only requires to retain its integrity and not its operability.

5.1.3. System and component rating

Another important distinction to make is between system and component rating, Taking the example of a well
monitoring system: for the Peterhead CCS wells this was to include a wellhead outlet, hanger terminations, gauge
control line (electric and fibre optic), control line protectors, pressure and temperature gauges and a gauge mandrel.
Each component in the system would have been exposed to different temperatures. The components deeper in the
well such as the gauges and gauge mandrel which are typically set close to the production packer would have been
in a warm environment and even in a release of CO2 scenario would not have seen extensive cooling. These
therefore would not have required to be qualified or rated for low temperature conditions. The gauge control line on
the other hand travels the entire length of the tubing to surface this would see low temperatures as the gas-liquid
interface travels down the well. Similarly, control line protectors/clamps used to secure the gauge cable to the tubing
would be subject to low temperatures and these would have to have been specified to suit.

5.2. Well Components

It is clear that the extent to which the well components are affected varies and is dependent on their placement in
the well and whether they are in contact with the cold CO2, this is evident from Figure 4 and Figure 5.

5.2.1. Casing and Cement

During steady state injection and transient conditions the cement and the casing within their normal operating
envelope. It is only in the highly unlikely release of CO2 scenarios where the casing strings may be subject to
cooling which may cause concern. Even then modelling followed by robust material selection will ensure material
embrittlement concerns can be ruled out.
The Peterhead CCS project had intended to re-use the carbon steel casing strings as certificates supporting the
quality of the installed casing were available and indicated Charpy values demonstrating adequate toughness down
to -40°C. In order to ascertain the thermal effect of CO2 injection in the cement bond, Diana software - a specialist
mechanical cement model - was used.
In the case described in section 4.3 after a prolonged acceptable CO2 leak across the SSSV the casing and cement
adjacent to the SSSV setting depth is subject to significant cooling. An engineering criticality assessment was
conducted for the casing body and thread connections taking into consideration the stress concentration at the thread
roots and API acceptable flaws. Even with a conservative approach, the casing string proved adequate. That said a
precautionary principle was adopted and it was recommended that the SSSV was to be set mid joint of a casing joint
i.e. away from the legacy casing connections.
Additional study into the effects on the cement under this scenario is required.

5.2.2. Tree and Wellhead

As a result of the potential low temperatures during transient conditions and even lower temperatures
encountered under the highly unlikely release of CO2 scenarios, where possible, it is recommended to install a API
6A temperature class ‘K’ (-75°F to 180°F) rated system.
For the Peterhead CCS project it was intended to re-use the Goldeneye wells for CO2 injection hence
computational flow modelling was conducted to analyze and support the utilization of the existing wellhead system.
It was planned to replace the trees with API 6A temperature class ‘K’ trees. In addition, a detailed assessment of
each of the tree and valve seals was conducted to analyze the cause and effect of a leak. Based on this assessment,
recommended design modifications and additional qualifications were specified.
 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099 5097

To allow the valves to be safely operated in a plume of CO2 if necessary, it was decided to automate the upper
master valve and flow wing valve. It is also crucial to select hydraulic power fluid and grease for performance at
low temperatures.

5.2.3. Tubing

As detailed previously it was planned to replace the tubing in the Goldeneye wells with small diameter (4-1/2” x
3-1/2”) tubing for effective management of the CO2 phase behaviour.
The CO2 phase behaviour and associated cooling under the highly unlikely release of CO2 scenario led to the
decision to install S13Cr tubing above the SSSV. It was initially planned to install alloy 825 tubing directly above
the SSSV due to the localized cooling effects noted in a prolonged release of CO2 event (section 4.3). However,
positive preliminary results from an engineering criticality assessment into the use of S13Cr allowed for this
requirement to be relaxed.
For the 4-1/2” tubing VAM Top HT connections were selected as the higher make-up torque compensates for the
relative shrinkage effect between the box and pin. The thread connection test envelopes do not extend to the low
temperatures that they may be subject to in a CO2 well. Additional validation may be required. The effects of the
low temperatures on thread dopes has not been studied it has however been reviewed by the suppliers and is
understood to be acceptable.

5.2.4. Subsurface Safety valve (SSSV)

A SSSV may not be required in all CO2 injection wells; however, the specific conditions of the Peterhead CCS
project dictated the need to install this additional safety device. SSSVs currently available on the market have a
lower end temperature rating of -7°C.
The existing temperature rating of SSSVs places a limitation on the planned testing of SSSVs installed in a CO2
injection well (section 3.2). During a highly unlikely release of CO2 event, the temperatures at the SSSV would be
significantly reduced (section 4.2). Due to the acceptable leak rate across the SSSV flapper (section 4.3), the
temperature at the SSSV could reduce further. Modelling predicted temperatures as low as -78.5°C at the SSSV.
However, to date, a SSSV has not been designed and qualified to these low temperatures.
It was decided to perform minimal design modifications to existing SSSVs and qualify them to withstand the
potential low temperature excursion. The well design philosophy was applied to create a distinction between the
SSSVs operating temperature envelope and its integrity temperature envelope.
The end result of the SSSV development project would have been a valve capable of closing before the cold front
arrived at the SSSV setting depth, and once closed, the valve would have been capable of maintaining a seal through
the low temperature excursion. Only once the well warmed back up would the valve need to be functionally open.
SSSV vendors demonstrated confidence in this approach although the development of such a valve is required.
It was planned to install a ¼” alloy 825 control line with 11mm x 11mm Teflon or Tefzel encapsulation. A
hydraulic fluid with low pour point is required.

5.2.5. Production Packer and Annulus fluid

During the transient conditions, a drop in temperature towards the top section of the well is expected. Generally,
the production packer is set at a depth where it would not be affected by the CO2 expansion. This is certainly the
case for the Goldeneye wells as indicated in Figure 5.
The drop in temperature would be even more pronounced during a highly unlikely release of CO2 as illustrated in
Figure 8 (section 4.2). This leads to concern around potential freezing of the annulus fluid and subsequent damage to
the casing and tubing strings.
For the Goldeneye wells, it was planned to install an oil based fluid in the A annulus (for corrosion avoidance in
case of a tubing to annulus leak) capable of undergoing a low temperature excursion without freezing. The high
thermal expansion and contraction of the fluid was to be mitigated by the inclusion of a Nitrogen (N2) cushion in the
top section of the annulus. The N2 also helps thermally insulate the upper portion of the casing string.
5098 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099

5.2.6. Well monitoring

The well monitoring system can provide valuable data to better understand the behaviour of CO2 in the well and
assist in reducing the duration of well operations such as testing of the SSSV (section 3.2).
For the Goldeneye wells, it was planned to install fiber optic monitoring for distributed temperature sensing in
addition to strategically placed pressure and temperature gauges in the well. Within Shell, such monitoring systems
have been used in applications with temperatures similar to that modelled for the Peterhead CCS project.
During the selection of the well monitoring system the design philosophy was applied to differentiate between
the functional and integrity requirement of the system and the temperature requirements of the individual
components versus the overall system (sections 5.1.2 and 5.1.3)

5.3. Well Interventions

An understanding of the lifecycle of the well will allow for the well design to adequately cater for planned
activities such as shut-ins, interventions, stimulations and end of life well abandonment.
Standard well intervention equipment is not designed or rated for use at the low temperatures that it could be
subject to during a release of CO2. Additional qualification of intervention equipment and procedural change is
required in order to safely conduct routine and planned well intervention activities.
Manufacturers of surface pressure control equipment (PCE) such as blow out preventers (BOPs) have indicated
the potential to modify existing designs and carry out low temperature qualification. It is also recommended to use
flanged metal to metal connections and to minimise connections between the tree and the pressure control
equipment.
For the Goldeneye wells, it was planned to include a shear valve which was to be qualified for use at extreme low
temperatures. Simple procedural changes such as placement of the emergency shut-down device during intervention
activities would help to ensure these devices are accessible and out of the potential CO2 plume in an emergency
situation.
Depending on the intervention activity, it may also be possible to fill the surface system and tubing volume with
a fluid other than CO2 prior to commencing the activity, although this may not always be an option.

6. Conclusions

There are a number of ways that the phase behaviour of CO2 can be managed to avoid potential integrity and
operational issues in the wells/facilities. The merits of each of these are each project specific and must be considered
at the conceptual design stage.
For the depleted Goldeneye field in the Peterhead CCS project, it would have been managed by creating back
pressure with friction by using small diameter tubing. The CO2 is managed in the top part of the well by restricting
its expansion within the liquid phase under continuous injection conditions.
During transient conditions, some cooling is expected due to the CO2 expanding to two phases (liquid/gas), and
to some extent this can be managed by employing specifically engineered well operating procedures; however
diligence is required in the well design especially in the top section of the well where the cooling is expected to take
place.
In an extreme scenario such as an uncontrolled release of CO2 from a well, the temperatures of the components
can drastically drop depending on the placement in the well relative to the liquid and gaseous CO2 interface. Low
temperatures are always associated with CO2 expansion to atmospheric conditions and hence low pressures prevail.
It is important to understand what temperature each component may be subject to during the operating conditions
and potential uncontrolled events as the tree valves, tubing material, tubing connections etc. would all see a drop in
temperature as a result of the CO2 cooling.
Standard equipment available from suppliers should be able to perform during continuous and transient
conditions though some additional qualification may be required or a divergence from standard stock may be
required to equipment rated for use in uncontrolled scenarios.
 Luis Acevedo and Ajay Chopra / Energy Procedia 114 (2017) 5083 – 5099 5099

Differentiation has to be noted between the operability and integrity requirements as it is not an effective use of
resources to qualify all components to the temperatures discussed in this paper.
Cooling caused by the phase behaviour of CO2 means that it is necessary to evaluate the suitability of well safety
critical equipment. A SSSV suitable for CO2 releases does not currently exist in the market. A SSSV capable of
maintaining its integrity under cold conditions needs to be developed. An API 6A temperature Class ‘K’ tree should
normally be selected for CO2 injection wells.

References
[1] L. Paterson, J. Ennis-King and S. Sharma, “Observations of Thermal and Pressure Transients in Carbon
Dioxide Wells,” SPE 134881, 2010.
[2] Peterhead CCS Project, “11.128 Storage Development Plan,” 2015.
[3] Peterhead CCS Project, “11.097 Well Completion Concept Select Report,” 2014.
[4] M. Haigh, “Well Design Differentiators for CO2 Sequestration in Depleted Reservoirs,” SPE 124274, 2009.
[5] Peterhead CCS Project, “11.104 Well Operation Guidelines,” 2015.
[6] A. Baklid, R. Korbol and G. Owren, “Slepiner Vest CO2 Disposal, CO2 Injection into a Shallow
Underground,” SPE 36600, 1996.
[7] L. Skinner, “CO2 blowouts: An emerging problem,” World Oil, 2003.
[8] L. Paterson, “Numerical Modeling of Pressure and Temperature Profiles including phase transitions in Carbon
Dioxide wells,” SPE 115946, 2008.
[9] J. Bellarby, Well Completion Design, Elseiver, 2009.
[10] T. Syed and T. Cutler, “Well Integrity Technical and Regulatory Considerations for CO2 Injection Wells,” SPE
125839, 2010.
[11] Peterhead CCS Project, “11.099 Well Technical Specification,” 2015.

Nomenclature

13Cr 13% Chromium metallurgy

API American Petroleum Institute
BOP Blowout Preventer
CCS Carbon Capture and Storage
CFD Computational Fluid Dynamic
CITHP Closed-in Tubing Head Pressure
CO2 Carbon Dioxide
DTS Distributed Temperature Sensing
HPHT High Pressure - High Temperature
Mtpa Million tonnes per annum
N2 Nitrogen
PCE Pressure Control Equipment
PDG Permanent Downhole Gauge
S13Cr Super 13 Cr metallurgy
SSSV Sub Surface Safety Valve
TVD True Vertical Depth
VAM Top HT Type of tubing with HT connections