Report number Date Security level POL-O-2007-138-A 8 January 2008 Open State-of-the-Art Overview of CO 2 Pipeline Transport with relevance to offshore pipelines Antonie Oosterkamp Joakim Ramsen
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
State-of-the-Art Overview
of
CO2 Pipeline Transport with relevance to offshore
pipelines
Antonie Oosterkamp
Joakim Ramsen
Page 2 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Title:
State-of-the-Art Overview of CO2 Pipeline Transport
with relevance to offshore pipelines
Stoltenberggt. 1 5527 Haugesund
Tlf: 52 70 04 70
Fax: 52 70 04 71
www.polytec.no
Project number:
E-0751
Report number:
POL-O-2007-138-A
Number of pages:
87
Principal investigator:
Antonie Oosterkamp
Security level:
Open
Date:
8th of January 2008
Client:
Research Council of Norway, Gassco
and Shell Technology Norway
Authors:
Antonie Oosterkamp
Joakim Ramsen
Client reference:
182603/I30
Summary:
This report provides the results of a study of the existing experience regarding the design
and operational aspects of CO2 transport by pipeline with relevance to future application
on the Norwegian Continental Shelf. The effect of expected new conditions like higher
pressures, offshore environment and impurities present in the CO2 mixture are taken into
account. The report concludes by summarizing the remaining uncertainties and R&D
needs that were identified in this study. In addition, an overview of competence holders is
given.
Principal Investigator
ANTONIE OOSTERKAMP
Quality Assurance Responsible
GUNN SPIKKELAND HANSEN
Chief Executive Polytec
TORLEIF LOTHE
Page 3 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Table of Contents
1 List of abbreviations ........................................................................................................... 5
Summary .................................................................................................................................... 6
2 Introduction ...................................................................................................................... 12
3 Existing CO2-pipelines ..................................................................................................... 13
4 Properties of Pure CO2 ..................................................................................................... 15
5 Expected Mixtures from Different Sources ..................................................................... 20
6 Effect of Impurities .......................................................................................................... 22
6.1 Density, Viscosity and Vapor Pressure ..................................................................... 22
6.2 Available Measurement Data .................................................................................... 26
6.3 Effects on design and operation ................................................................................. 27
7 Standards/Pipeline Code .................................................................................................. 30
8 Fluid Specifications for Pipeline Transport of CO2 ......................................................... 31
9 Material Aspects ............................................................................................................... 33
9.1 Elastomers ................................................................................................................ 33
9.2 Lubricants and Sealants ............................................................................................ 34
9.3 Coatings (internal) .................................................................................................... 34
9.4 Valve Seats ............................................................................................................... 34
9.5 Gaskets ....................................................................................................................... 35
9.6 Metals ....................................................................................................................... 35
9.7 Engineering Plastics ................................................................................................. 35
10 The Free Water Issue ........................................................................................................ 36
10.1 Corrosion ............................................................................................................... 36
10.2 Hydrates ................................................................................................................ 40
10.4 Water Solubility .................................................................................................... 43
11 Fracture Propagation in CO2 Pipelines ............................................................................. 45
12 Flow Assurance ................................................................................................................ 47
13 Viscosity Relations and Equations of State ...................................................................... 50
14 Metering and measurement .............................................................................................. 54
15 Monitoring and control ..................................................................................................... 56
16 Operational issues ............................................................................................................ 59
16.1 Ready for operation (RFO): .................................................................................. 59
16.2 Packing/depacking the pipeline ............................................................................ 59
16.3 Blowdown/depressurization .................................................................................. 60
16.4 Dynamic effects ..................................................................................................... 62
Page 4 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
17 Maintenance Aspects ........................................................................................................ 63
18 Risk Assessment, Health Environment and Safety .......................................................... 64
19 USA CO2 Pipeline Transport Experience ........................................................................ 66
19.1 Sheep Mountain Facilities ..................................................................................... 66
19.2 Cortez pipeline ...................................................................................................... 67
19.3 Weyburn Pipeline .................................................................................................. 68
19.4 NJED Pipeline ....................................................................................................... 69
20 Conclusion; remaining uncertainties and R&D needs ..................................................... 72
20.1 Material Aspects ................................................................................................... 72
20.2 Available measurement data ................................................................................. 73
20.3 Water Content ....................................................................................................... 74
20.4 Smart Pigging of long offshore CO2 pipeline ....................................................... 74
20.5 Modeling ............................................................................................................... 75
20.6 Fluid Specification ................................................................................................ 76
20.7 Most critical short term needs ............................................................................... 77
21 References ........................................................................................................................ 78
22 Appendix 1 Details about existing CO2 pipelines ........................................................... 82
23 Appendix 2 Overview of Identified Competence Holders .............................................. 86
Page 5 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
1 List of abbreviations
BWRS Benedict-Webb-Rubin-Starling
CCS Carbon Capture and Storage
C-Mn Steel Carbon Manganese Steel
DCG Dakota Gasification Company
DEG Diethylene Glycol
EOR Enhanced Oil Recovery
EoS Equation of State
ERW Electro Resistance Welding
HES Health, Environment and Safety
MAOP Mean Allowable Operating Pressure
MEG Monoethylene Glycol
MMP Minimum Miscibility Pressure
LBC Lohrenz-Bray-Clark
LDS Leak Detection System
LK Lee-Kessler
LNG Liquified Natural Gas
NIST National Institute of Standards
PA Polyamide
PCTFE Polychlorotrifluoroethylene
PMS Pipeline Modelling System
PP Polypropylene
PPS Pressure Protection System
PR Peng-Robinson
PT Patel-Teja
PTFE Polytetrafluoroethylene
PTV Patel-Teja-Valderama
PVDF Polyvinylidene fluoride
PvT Pressure, volume , temperature
R&D Research and Development
RK Redlich-Kwong
RKS Redlich-Kwong- Soave
SCADA Supervision, Control and Data Acquisition
SSC Sulphide Stress Cracking
STEL Short Term Exposure Levels
TEG Triethylene Glycol
VLE Vapour Liquid Equilibrium
Page 6 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Summary
Transport of CO2 by pipeline will be necessary if large volumes of captured CO2 are to be
stored in geological formations at short to medium distance from the capture location. For a
number of countries, including Norway, the preferred storage locations will be offshore,
necessitating offshore pipelines between the capture and storage facilities. This report gives
an overview of the state-of-the-art of pipeline transport of CO2 of relevance for offshore
conditions. It provides an assessment what will be novel in an offshore context. The
implications of new capture sources on CO2 pipeline transmission systems are looked into.
The remaining uncertainties identified in this study concerning offshore transmission of CO2
with impurities present in the CO2 fluid stream are provided. The report concludes with
suggestions for the research and development needs to address these uncertainties.
CO2 transport by pipeline is routinely done in the USA for over 30 years. The existing
pipelines in the USA are land based and divided into relatively short sections; this reduces the
blow down and refilling times and limits the risk to the public in case of leaks. For offshore
pipelines this is expected to be different. The main block valves and metering will be located
at the inlet and outlet only; sectioning of offshore parts of the pipeline may not be a viable
option.
To transport CO2 efficiently by pipeline, the pressure is kept over the critical point and the
fluid is transported in dense phase.
Important properties of CO2 at typical operating conditions (dense phase) are:
Density is relatively high and sensitive to temperature.
Low viscosity.
Non-linearly varying compressibility factor.
Acts as a solvent.
The fluid composition of the CO2 to be transported depends on the source. Typically the CO2
originates from natural deposits and the fluid stream is relatively pure; few other components
are present. A pipeline will be designed for a long life time. Thus it can be expected that the
fluid composition in the pipeline will change when different capture sources are connected to
the pipeline infrastructure. The new capture methodologies can give lead to new compounds
in the captured stream for which there is little or no experience within CO2 transport.
Currently, no CO2 quality requirements have been decided upon that take into account these
new compounds.
Page 7 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Impurities in the CO2 have an effect upon:
Design of equipment like pumps and compressors: specifically setting of suction
pressure and compression strategy to avoid the two phase region.
Toxicity: it can be impurity concentrations that determine the safe exposure limits for
the fluid instead of CO2 concentration.
Transport capacity; impurities reduce the transport capacity of the pipeline.
Vapor pressure: raising of the vapor pressure means that higher minimum entrance
pressure or shorter recompression/booster station intervals are needed to keep the fluid
in dense phase.
Pipeline integrity: The vapor pressure sets the decompression pressure at a pipeline
break. Thus a high decompression pressure can facilitate further propagation of a
fracture. Presence of atomic hydrogen can lead to hydrogen embrittlement of the
pipeline steel or hydrogen induced cracking. Sulphide Stress Cracking (SSC) has to be
taken into account with presence of H2S (requirement for sour service).
Corrosion.
The water solubility and hydrate formation conditions.
For pure CO2 there are developed reference equations of state (EoS) providing highly accurate
calculations. For the relevant CO2 mixtures, there is generally very limited data published
about the applicability of existing EoSs and the applicable mixing rules and parameters.
There is no real consensus which EoS should be used in flow modeling of CO2 pipeline
transport when the CO2 contains impurities. With respect to viscosity calculations, accurate
correlations have been developed for pure CO2. Within the course of this study not many
references were found to viscosity measurement data for the relevant CO2 mixtures. A search
and review of available thermodynamic data for the relevant CO2 mixtures concluded with
that measurements of PvT and VLE data at conditions relevant for CO2 pipeline transport are
few. This was found to be the case for binary mixture data for CO2 with components like H2,
SO2, NO, O2, CO, COS and Ar as well as multi component mixtures.
To transport CO2 in the dense phase has its effect upon the material selection process during
design. Care has to be taken when selecting materials and compounds for gaskets, valve seats,
sealants, coatings and lubricants. CO2 gives rise to higher susceptibility for explosive
decompression of elastomers in seals and gaskets.
When liquid water is present, CO2 will partially dissolve and form carbonic acid. This will
give rise to corrosion problems with the steel alloys commonly used in pipelines. Carbon steel
(C-Mn) can be used in the absence of free water. No corrosion problems have been reported
Page 8 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
where the CO2 is suitably dry or when stainless steel alloys are used. At high partial pressures
the existing models tend to overestimate the corrosion rates. In addition, the concentrations
and types of other impurities present in the CO2 mixture may influence the corrosion rates.
The mechanism of CO2 corrosion in the presence of impurities is not understood entirely.
There are different practices regarding the allowable water contents specifications used for the
existing CO2 pipeline systems. There are remaining uncertainties how to set the maximum
allowable water contents for a CO2 pipeline. The minimum water solubility limit at
operational conditions might be a non-conservative limit. The water contents in the CO2 may
also lead to the formation of hydrates. An offshore pipeline on the Norwegian Continental
Shelf will along most of its length transport CO2 within the hydrate stable region. Hydrates
will form when free water is present but might also form when the water contents is under the
saturation limit. What the effect is of impurities on both water solubility and hydrate
formation is another area of uncertainty.
CO2 pipelines are considered to be more susceptible to fast propagating ductile fractures than
gas pipelines. The first CO2 pipelines in the USA were designed with relatively short distance
between fracture arrestors. Alternatives to address the risk of running ductile fractures are to
increase the wall thickness or through use of material with higher fracture arrest properties.
The existing models for assessing fracture arrest are based upon tests with hydrocarbon
gasses. They are not necessarily directly applicable for use with CO2 pipelines without
additional experimental assessment.
Heat loss and elevation terms must be included in the energy balance calculations underlying
pressure drop estimation.
With respect to flow modeling, the sensitivity of density to temperature will make it more
difficult to predict flow for an offshore CO2 pipeline. Generally, offshore pipelines will have
less measurement points along the way to tune the thermal model used in simulating the flow
conditions. This means that accuracy of capacity calculations and leak detection for long
offshore pipelines will be more reduced compared to a sectored land based system.
Measurement and instrumentation is in principle similar to that used for natural gas pipelines.
For any instrumentation used on CO2 pipelines, the special requirements regarding material
choice of sealants and gaskets for dense phase CO2 have to be taken into account. For a CO2
pipeline made from carbon steel, measurement of actual water contents at the pipeline
entrance is a necessity.
Integrity monitoring of land based CO2 pipelines is typically done by visual inspection and
use of corrosion coupons. Integrity assessment of the pipeline with a smart pig is also viable,
but very few inspections runs with smart pigs are reported. Inspection pigging of CO2
pipelines is not routinely done and regarded as more difficult than natural gas pipeline
pigging. The big concern is the friction wear through the line. Two pipelines with
Page 9 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
computational leak detection systems have been identified within this study that use real time
transient modeling. CO2 is considered as a challenging fluid for computational leak detection.
Most important for leak detection of CO2 is the thermal modelling capability. Uncertainties
remain regarding the interaction of the escaping CO2 from a leak in a subsea pipeline with the
surrounding seawater.
A number of relevant issues regarding the operation of CO2 pipelines were identified:
RFO (Ready For Operation)
Dewatering and drying is even more critical than is the case with a natural gas
pipeline. It is advised to have a dewatering/drying strategy for all components of the
pipeline.
Care has to be taken during initial filling up of the pipeline to avoid rapid cooling
down of the expanding CO2 fluid behind the inlet valve.
Pressurizing with nitrogen after hydrostatic testing will allow for the detection of
remaining leaks that do not show up during hydrostatic testing with water.
Blow down/depressurization
The blow down facility must be specifically designed for CO2.
A blow down should be controlled through slow depressurization and sufficient heat
transfer from the ambience. For offshore pipelines we can expect that this is more
challenging as they are not likely to be sectored as land based pipelines.
The danger exists that low temperatures can cause instability of the pipeline due to the
freezing of the surrounding medium.
Stop and Start procedures
Dynamic effects should be considered.
Care must be taken to avoid large temperature drops over valves.
Health, environmental and safety risks are mostly associated to the release of CO2 to the
ambience. CO2 is an asphyxiant: it has an effect upon the respiratory system already at low
ambient concentrations. It is important to notice that CO2 is heavier than air so it will collect
in low laying terrain. Exposure to a stream of expanding CO2 can cause cold burn of the skin.
The expansion of the gas during the phase transition will also give a thrust, potentially
displacing a pipeline in case of a leak. When leaks occur, toxic impurities can be setting the
safe limits rather than the CO2 itself. This is specifically the case with H2S and SO2.
Page 10 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Conclusions, remaining uncertainties and identified R&D eeds
Within the scope of this study, no real show stoppers for offshore pipeline transport of CO2
were identified. The identified R&D needs are mostly due to uncertainties that arise when the
operating conditions go beyond existing experience or related to the effect of impurities and
that offshore pipelines typically will not be sectored.
New capture technologies are under development that can give CO2 mixtures with several
new compounds. The impact of these impurities on the pipeline transport system should be
evaluated. There are some studies made public that look into the impact of impurities in the
CO2 but these are typically not backed up by experimental data.
The following R&D needs are identified:
Corrosion with use of Carbon (C-Mn) steel.
o Assess the consequences and develop the counter measures (e.g. adding
corrosion inhibitors) on incidences of free water in offshore CO2 pipelines.
o Evaluate the need to perform further studies on corrosion with high partial
pressure CO2 with impurities, and the need to develop suitable corrosion
models.
Non-steel materials (seals and gaskets).
o The need for additional material compatibility testing has to be evaluated in
cases where CO2 transport incorporates higher pressures or pressure variations
than currently employed or in the presence of new impurities.
Available experimental data on thermodynamic and transport properties of CO2 with
impurities.
o There is a need for a more extensive data search and assessment which
additional data should be generated in a follow up measurement campaign.
Review which level of uncertainties in the fluid properties (e.g. density) can be
accepted, followed up by an experimental program addressing the remaining
gap.
Water content.
o Further investigation in to the effect of the impurities on water solubility, the
availability of experimental data and possibly further development of the
thermodynamic models to calculate the solubilities for actual CO2 mixtures.
This with the objective to be able to set safe water specifications.
Page 11 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Inspection Pigging of long offshore CO2 pipelines
o Testing and potentially development of smart pigs suitable for CO2 pipelines
that can operate at high pressure and can travel longer distances. Test on land
based pipelines first if possible.
Blow down/Depressurization
o Modelling combined with experimental verification to set safe regimes for
depressurization of long offshore pipelines. If possible, link this to ongoing
work regarding this theme at StatoilHydro and SINTEF.
Fracture Propagation
o An assessment has to be made if fracture propagation is a real threat to CO2
pipelines and if the existing requirements from the design code are enough to
arrest a fracture. This assessment should include the applicability of existing
propagation models for CO2 pipelines and what work should be done to update
the models and the existing requirements.
Equations of State for CO2 with impurities.
o A further assessment should be made which EoS is valid under what
conditions for the relevant mixtures within the applicable temperature and
pressure range for offshore pipelines. The accuracy of the EoS of choice
should be checked against and were necessary EoS and mixing rules be
modified to match the data.
Fluid Specification
o The potential chemical reactions between the impurities under relevant time,
pressures and temperatures and the potential negative effects of the products
need to be mapped. A specification should be made and agreed upon that
specifies for allowable levels of impurities in the CO2 for pipeline transport.
This is currently addressed in the EU Dynamis project. A link between this
project and the governmental projects for CCS from Krst and Mongstad
regarding setting of transport specifications in relation to the end use of the
CO2 would be advisable.
Correspondence to: A.Oosterkamp, Polytec R&D Foundation, Stoltenberggt.1, N-5527, Haugesund, Norway.
E-mail: [email protected]
Reference to part of this report which may lead to misinterpretation is not permissible. The authors and Polytec
disclaim any liability to the client and to third parties in respect of the publication, reference, quoting, or
distribution of this report or any of its contents to and reliance thereon by any third party.
Page 12 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
2 Introduction
A better understanding of the relationship between CO2 emission and climate change has been
gained in the last decades. Therefore several projects around the world investigate the
possibility of capturing the CO2 resulting from burning fossil fuels and injecting it for storage
into geological formations. This can be one of the possible measures to avoid global warming.
The storage location will not necessary be located near the source. Thus an extensive transport
system must be applied. For large volume and short to medium distances, pipeline transport is
usually more cost effective than the other alternative, ship transport. For several countries the
actual storage locations will be located offshore e.g. Norway and the U.K.
Norway has a long experience with offshore pipeline transport of natural gas. In pipeline
transport, CO2 as a fluid shows behavior and properties that differ from natural gas. During
the last three decades, a lot of experience has been gained with land based transport in the
USA. Unfortunately, it is not easy to get a complete overview of design and operational
issues. The available literature describes either only some of the relevant aspects and/or is
relatively old. The only existing offshore pipeline for transporting CO2 is the Snhvit pipeline
which is due for operation. The relatively short timeline in capture projects at Krst and
Mongstad make it necessary to gain more knowledge about CO2 transport before operational
experience from Snhvit becomes available. In this report, the existing knowledge from land
based CO2 transport has therefore been included where relevant.
This report is the result of a 7 month study. The information gathering process comprised a
literature study, personal communication to experts and visits to CO2 pipeline operators in the
USA. The report gives an overview of the state-of-the-art, references to relevant literature,
overview of relevant competence holders and discussion of issues that need to be adressed.
The main focus has been to identify critical issues related to offshore pipeline transport, the
effect of expected impurities in the CO2 and the remaining R&D needs. References to
literature are given in the text. An overview of the relevant competence holders identified in
the course of this study is included in Appendix 1.
Acknowledgements are hereby given to the Norwegian Research Council, Gassco and Shell
Technology Norway for funding this study. We like to thank everybody who has contributed
to this study through interview, personal communication and sharing of their professional
opinion.
Page 13 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
3 Existing CO2-pipelines
An overview of the majority of the existing long CO2 pipelines is given in Table 3-1 below.
Table 3-1: Overview of some existing long CO2 pipelines
Name of pipeline Operator Length
(km)
Diameter
(in)
Capacity
(MT/year)
Country
NEJD Pipeline* Denbury Resources 295 20 USA
Cortez Pipeline* Kinder Morgan 808 30 19.3 USA
Bravo Pipeline BP 350 20 7.3 USA
Transpetco Bravo Pipeline Transpetco 193 12 3.3 USA
Sheep Mountain part 1* BP 296 20 6.3 USA
Sheep Mountain part 2 * BP 360 24 9.2 USA
Central Basin Pipeline* Kinder Morgan - 26 and 16 11.5 USA
Este Pipeline Exxon Mobil 191 12 and 14 4.8 USA
West Texas Pipeline Trinity 204 8 to 12 1.9 USA
SACROC pipeline 354 16 4.2 USA
Weyburn Pipeline* Dakota Gasification
Company 330 12 to 14 4.6
USA
Canyon Reef Carriers Kinder Morgan 225 16 4.6 USA
Bati Raman Turkish Petroleum 90 1.1 Turkey
Snhvit* StatoilHydro 153 8 0.7 Norway
*A more detailed description of these pipelines is provided in Appendix 1
All the pipelines shown in Table 3-1 are land based pipelines except Snhvit, which is the
first offshore CO2 pipeline. The Snhvit pipeline is planned to start-up in the fourth quarter of
2007.
The CO2 transported in the Snhvit pipeline is captured from natural gas from the Snhvit
field. The high CO2-contents from this gas is reduced before it is processed to Liquified
Natural Gas (LNG) at Melkya. One of the first pipelines designed and installed is the
Canyon Reef Carriers which started operation in 1972. The land based pipelines are typically
divided in sections by several valve or compressor stations where instruments are installed to
monitor, pressure and temperature. The CO2 typically originates from natural deposits and is
used for Enhanced Oil Recovery (EOR). The fluid consists of minimum 95% CO2, see
Chapter 8 for quality requirements. The pipelines are designed to operate above the critical
pressure so that two-phase flow is avoided. The pressure is typically below 200 bara. The
short distance (< 30 km) between main block valves, reduces the time for blow
down/depressurization and refilling operations. This also reduces the environmental
consequences of a pipeline rupture.
Page 14 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Figure 3-1: Laying of 8 CO2 pipeline at Snhvit [http://www.caithness.org]
Figure 3-2: Ground entry of a land based CO2 pipeline, Jackson Dome operated
by Denbury Onshore LLC.
Page 15 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
4 Properties of Pure CO2
Properties of pure CO2 are well known and have been extensively studied. In this section, the
relevant properties for pipeline transport will be shown.
The phase diagram for pure CO2 is shown in Figure 4-1 below.
It is important to note that pure CO2 has a triple point at -56.6 C and 5.18 bara. It has its
critical point at 30.9782 C and 73.773 bara. This has its implications for both compression
and transport conditions. Note that above the critical point CO2 will not be able to separate in
two phases (except at very low temperature or high pressure where solid CO2 can form).
Figure 4-1: Phase diagram for pure CO2 (1)
Triple point
Critical point
Page 16 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
The pressure-enthalphy diagram is shown in Figure 5.2 below.
Figure 4-2 Pressure-Enthalpy diagram for pure CO2 (2)
This diagram can be used to estimate how the temperature changes during de-pressurization,
e.g. over a valve. Note the isotherms are almost vertical at low temperatures. This means that
a throttling in this region (liquid) will not alter the temperature significantly as long as the
CO2 is kept in one phase. The Pressure-Enthalpy diagram is also used to visualize the
thermodynamic path for compression and pumping. It can be seen from the diagram that in
the liquid region relatively low energy input is necessary to increase the pressure, compared to
compression of the gas (isentropic lines are steeper).
For all the relevant thermodynamic properties and its viscosity the reference is the NIST
chemistry webbook (3).
Dense phase
Liquid
Gas
Two phase
Page 17 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Figure 4-3 below shows the density as function of temperature and pressure.
Figure 4-3: Density of CO2 as function of temperature and pressure (3)
As can be seen from Figure 4-3 above, the density of CO2 has a stronger dependency on
temperature than pressure at lower temperatures.
It can also be seen that the density is very sensitive to small temperature changes near the
critical point. Density is an important factor in flow calculations. This means that accurate
knowledge of inlet temperature, ambient temperature and heat transfer is necessary to model
the flow correct, especially if conditions are close to the critical point.
0
100
200
300
400
500
600
700
800
900
1000
1100
0 10 20 30 40 50 60
De
nsi
ty [
kg/
m3
]
Temperature [C]
Density
200 bara
150 bara
100 bara
80 bara
Page 18 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
It is the combination of high molecular weight and low compressibility factor (Z-factor) that
makes CO2 density so temperature dependent. The compressibility factor at different
temperatures and pressures is shown in Figure 4-4 below. Figure 4-4 shows that ideal gas
assumption for CO2 is not applicable. The compressibility factor is used to alter the ideal gas
equation to account for the real gas behaviour. For an ideal gas, the Z-factor will be one,
independent of pressure and temperature. The compressibility factor needs to be taken into
account to give correct density in flow calculations.
Figure 4-4: Compressibility factor (z-factor) at different pressures and temperatures (3)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50 60
Z-f
act
or
[-]
Temperature [C]
Z-factor
80 bara
100 bara
150 bara
200 bara
Page 19 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Figure 4-5: Viscosity of CO2 as function of temperature and pressure (3).
CO2 has a low viscosity compared to some other high density fluids; e.g. olive oil (80 cP),
water (0.89 cP). Viscosity of CO2 versus temperature at different pressures is shown in Figure
4-5. As can be seen from the figure above, also the viscosity of CO2 shows a strong
temperature dependency, especially near the critical point.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0 10 20 30 40 50 60
Vis
cosi
ty
[cP
]
Temperature [C]
Viscosity
200 bara
150 bara
100 bara
80 bara
Page 20 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
5 Expected Mixtures from Different Sources
The fluid composition of the CO2 mixture to be transported will depend on the source. CO2
transported in USA is typically taken from natural sources. The mixtures from these sources
contain, apart from CO2 ,typically also: water, hydrogen sulphide, nitrogen and hydrocarbons,
see Table 5-1.
Table 5-1: CO2 composition transported in existing pipelines (given as vol% if not stated otherwise)
Canyon
Reef
Carriers (4)
Central
Basin
Pipeline (5)
Sheep
Mountain
(6) (7; 8)
Bravo
Dome
Source (9)
Cortez
Pipeline
(10)
Weyburn
(11)
Jackson
Dome,
NEJD
CO2 85-98 98.5 96.8-97.4 99.7 95 96 98.7-
99.4
CH4 2-15
C6H14
0.2 1.7 - 1-5 0.7 Trace
N2
Page 21 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Within these methods different capture technologies can be used which will produce CO2 with
different levels of impurities. For Post-Combustion, the absorption technology using amines
is considered to be a technology that will be used in near future, which provide very pure
CO2. The impurity levels present in the CO2 mixture resulting from Pre-combustion or
Oxyfuel will vary with the capture technologies employed.
Table 5-2 below shows typically compounds that can be present from different technologies.
Their exact concentration will depend on several factors, but without purification and co-
capture of other compounds, the maximum level indicated here can be reached. Especially
SO2 and H2S will normally be present in the final CO2 mixture in much lower concentration
than the maximum levels indicated in this table.
Table 5-2: Compounds from different power production methods with CO2 capture. :ote these are
indicative maximum values and not most likely values.
Post-Combustion1 (12) Pre-Combustion [
(13) (12)
Oxyfuel
(12), (13) (14)
CO2 >99 vol% >95.6 vol% >90 vol%
CH4
Report number
POL-O
6 Effect of Impurities
Impurities in the CO2 affect the design of
Impurities affect the phase behavior
example small amounts of hydrogen in
The relationships between the source
pipeline design and operation are shown schematically in
Figure 6-1: Schematically description of how impuri
6.1 Density, Viscosity and Vapo
The effects on density, viscosity and vapo
Chapter 5 are shown in Figure
using the REFPROP program from NIST. At their website
uses the most accurate equations of state and models currently available
used to produce the diagrams, is
should also be noted that the diagrams given in this section are for illustrational purposes onl
and that they have not been verifi
Report number Date Security level
O-2007-138-A 8 January 2008 Open
mpurities
affect the design of the pipeline and the compression
behavior, the thermodynamic properties and the viscosity. For
ydrogen in the CO2 will increase the vapor pressure significantly.
The relationships between the source, the requirements, the pretreatment and
pipeline design and operation are shown schematically in Figure 6-1.
: Schematically description of how impurities affect the pipeline design and operation
Density, Viscosity and Vapor Pressure
on density, viscosity and vapor pressure of the main impurities
Figure 6-2, Figure 6-3 and Figure 6-4. These figures have been made
am from NIST. At their website (15) NIST states that t
urate equations of state and models currently available
used to produce the diagrams, is 98 mole% CO2 with 2 mole% of the other component.
should also be noted that the diagrams given in this section are for illustrational purposes onl
verified by the authors against actual measurement data.
Page 22 of 87
Security level
Open
compression facilities.
, the thermodynamic properties and the viscosity. For
pressure significantly.
, the requirements, the pretreatment and their effect upon
ffect the pipeline design and operation
impurities identified in
. These figures have been made
NIST states that the program
urate equations of state and models currently available. The composition,
% of the other component. It
should also be noted that the diagrams given in this section are for illustrational purposes only
measurement data.
Page 23 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Figure 6-2: Density at 100 bara with different temperatures for CO2 with 2 mole% of another component.
SO2 is the only component that increases the density compared to pure CO2. The estimated
density for this mixture is very uncertain since no mixture parameters were available. From
this figure it can be seen that H2S has minimal impact on the fluid density while H2 has a
significant impact.
300
400
500
600
700
800
900
1000
0 10 20 30 40 50 60
De
nsi
ty [
kg/
m]
Temperature [C]
Density at 100 bara
CO2 (100%)
CO2-CH4
CO2-H2
CO2-N2
CO2-Ar
CO2-SO2
CO2-H2S
CO2-O2
Page 24 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
In Figure 6-3 below the viscosity are shown for different mixtures of CO2 at 100 bara.
Figure 6-3: Viscosity at 100 bara with different temperatures for CO2 with 2 mole % of another
component.
Note that the SO2 is not included in the diagram since REFPROP did not calculate viscosity
for this mixture.
The figure indicates that impurities typically will reduce the viscosity.
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0.11
0.12
0 10 20 30 40 50 60
Vis
cosi
ty [
cP]
Temperature [C]
Viscosity at 100 bara
CO2 (100%)
CO2-CH4
CO2-H2
CO2-N2
CO2-Ar
CO2-H2S
CO2-O2
Page 25 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Figure 6-4 below shows the vapor pressure for the different mixtures. Again, CO2 is mixed
with 2 mole% of another component.
Figure 6-4: Vapor pressure for different mixtures (98 mole% CO2)
It can be seen that the presence of impurities have a significant effect on the vapour pressure.
Exceptions are H2S and SO2. As with the other properties, the values for CO2-SO2 mixture are
very uncertain since the mixing parameters were estimated and not based on any actual
measurement data (16). The presence of impurities implies that a two phase region will be
present.
.
30
40
50
60
70
80
90
0 5 10 15 20 25 30 35
Pre
ssu
re [
ba
ra]
Temperature [C]
Vapour Pressure
CO2 (100%)
CO2-CH4
CO2-H2
CO2-N2
CO2-Ar
CO2-SO2
CO2-H2S
CO2-O2
Page 26 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
6.2 Available Measurement Data
A literature search combined with communication with relevant experts was part of this
investigation. Only a limited amount of experimental data was identified for high content CO2
mixtures (95 mole% +) and within the typical offshore pipeline operating conditions (100-300
bar and 0-50 C).
Below in Table 6-1 presents the available pressure, volume and temperature (PvT) and vapor-
liquid-equilibrium (VLE) measurement data, as reported by Kunz et al. (17). Only the
measurement data sets where CO2 is present at 95 mole% or higher are included here. For a
total overview we refer to Kunz et al (17).
Table 6-1: Overview of available PvT and VLE data for some binary CO2 mixtures (17)
Y+X Data Number of
data points
Temperature
(C)
Pressure
(Bar)
Mole
Fractiona
CH4 +
CO2
PvT 7 15 55-145 0.96
PvT 91 -48 - 127 21-358 0.98
VLE 6 28 70-77 0.97-0.99
VLE 21 15-20 56-82 0.83-0.99
N2 + CO2 PvT 64 27-57 23-331 0.98
PvT 39 0-200 4-88 0.98
VLE 13 28-30 72-81 0.96-1.00
VLE 22 -40 - 25 37-127 0.63-0.97
VLE 22 15-30 61-103 0.81-0.99
VLE 18 15-20 60-97 0.85-1.00
VLE 15 0 41-118 0.70-0.99
CO2 + H2 PvT 42 5-20 48-193 0.01-0.16
VLE 58 -53 - 17 11-203 0.00-0.14
VLE 42 5-20 48-193 0.01-0.16
CO2 + O2 VLE 72 -50 - 10 10-132 0.01-0.78
VLE 72 -50 - 10 10-132 0.00-0.39
CO2 + Ar PvT 5 15 83-145 0.06
PvT 12 15 57-98 0.06-0.021
VLE 12 15 57-98 0.06-0.17
a Mole fractions of component X in the saturated liquid phase for VLE data.
Page 27 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Below are described data sets for other mixtures obtained and/or found reference to in the
course of the investigation:
CO2-O2
In information received from NIST (16), 33 PvT data points from Muirbrook (18) were found,
but only one data point was below 5 mole% O2; at 0 C and 52 bar.
CO2-H2S
In information received from NIST 169 PvT data points from Stouffer (19) were found for the
lowest concentration of H2S ( 6.07%). The data includes pressure from 1 to 236 bar and
temperatures from 16 to 177 C.
CO2-CO
Kunz et al (17) reports 75 data points but only for 43 mole% CO and low pressure from 1 to
65 bara.
No reference to measurement data for CO2 mixed with SO2, NO or COS were identified.
Some multi-component data for CH4-N2-CO2 and N2-CO2-H2 exist but only at low CO2
content (
Page 28 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Some of the impurities, like NOX, CO and especially H2S and SOx are highly toxic. When a
leak occurs, a cloud of CO2 will disperse. The safe concentration of the dispersed CO2
mixture can be driven by the allowable concentration of these impurities from occupational
and environmental values (12). An example of this is in case of H2S. An assumed STEL
(Short Term Exposure Levels) of 30,000 ppm. for CO2 (20) and 15 ppm. for H2S, means that
for H2S concentrations over 2000 ppm. in the CO2 compositional mixture, H2S will be the
limiting factor.
Impurities have also an effect upon the design and operation of blow down facilities. If a
combustible compound is present which is not allowed to be vented to atmosphere (e.g. H2S)
a possible solution is to connect the blow down facility to a flare (It should also be noted here
that combusting H2S produces SO2 which is also highly toxic). This again implies that a fuel
gas system needs to be incorporated in the design. In such a case, the CO2 needs to be
mingled with enough fuel so combustion can take place.
Impurities have a high impact on the transport capacity. Studies on the qualitative and
quantitative effect of the impurities on the pressure drop and transport capacity are reported in
(8) (21). For example, CO2 plus 5 % methane decreases the flow by 16 % (flow adjusted to
have an 82.7 Pa/m pressure drop at 10 341 kPa and 16 C in 406 mm pipeline) (8). In
addition, impurities take up space in the pipeline that otherwise is utilzed for transporting
CO2. Compared to transporting pure CO2, 5 vol% impurities will reduce the volume of CO2
transported by 5%.
Since CO2 is transported as a dense fluid it will be relatively easy to compensate for losses of
capacity by boosting the pressure using a pump, as long as the pipeline is not already
operating close to Mean Allowable Operating Pressure (MAOP). An investigation into the
effect of impurities on the minimum distance between recompression/boosting stations is
presented in (21).
Especially hydrogen is shown here to have a large effect. A CO2 mixture containing 3 mole%
of hydrogen halves the minimum distance between recompression stations compared to pure
CO2. When recompression is not an option, with a given pressure loss along the pipeline
route, the minimum entrance pressure will have to be raised when the vapor pressure of the
fluid is higher due to the presence of impurities. This in turn can necessitate to design the
pipeline for higher operating pressures leading to for example large pipe wall thickness or
stronger materials.
The impurities can also have an effect upon the pipeline integrity. The vapor pressure sets the
decompression pressure at a pipeline break. Thus a high decompression pressure can facilitate
further propagation of a fracture: This is further described in detail in Chapter 11.
Page 29 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
Presence of atomic hydrogen can lead to hydrogen embrittlement of the pipeline steel or
hydrogen induced cracking. For atomic hydrogen to occur, free water needs to be present. The
underlying mechanism is that atomic hydrogen diffuses into the metal matrix and combines
again to hydrogen molecules. This creates local internal pressure which reduces the ductility
and tensile strength of the steel. The atomic hydrogen may also embrittle the steel through its
interference with the plastic flow during deformation.
Carbon steels used for pipelines can be specified with additional requirements to remediate
this potential problem. Measures can include lower sulphur contents of the steel, limiting the
hardness and alloying of the steel.
Presence of H2S is another issue of concern. Even without the presence of free water H2S
poses a potential problem (with free water also atomic hydrogen is produced). A reaction
between iron and H2S will occur at the pipe inner surface, creating a thin surface of iron
sulphide and atomic hydrogen. This is called Sulphide Stress Cracking (SSC). The sensitivity
for this can be reduced by for example adding nickel to the steel alloy composition. For
pipeline operations, the presence of H2S implicates that the steel has to be specified for so-
called sour service
The presence of Oxygen is considered problematic from a corrosion point of view, especially
when free water is present.
Finally, impurities can affect the water solubility and hydrate formation conditions. This will
be further discussed in Chapter 10.
Page 30 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
7 Standards/Pipeline Code
In the USA, CO2 pipelines fall under the Department of Transport 49-CFR 195 as they are
considered hazardous liquid pipelines and the ANSI/ASME B31.4 pipeline code. For Canada
the Canadian Standard Association Z662 applies. Department of Transport 49-CFR 195 puts
requirements on issues like material compatibility, pipeline integrity, monitoring, reporting of
accidents, etc.
For offshore pipelines the most widely used code is DNV-OS-F101. This code does not
provide special considerations for the transport of CO2. It is described in this code as a non-
flammable, non-toxic gas at ambient temperature and atmospheric conditions. This means it
falls under the codes fluid classification C. With this classification, CO2 pipelines fall under
design criteria for safety class Low or Normal (when in areas with human activity). However
it is questionable if large releases of CO2 (due to pipeline rupture) are as harmless as this
classification would indicate. In addition, there are differences between natural gas pipeline
transport and transport of CO2. On top of this, future parts of the existing hydrocarbon pipeline
infrastructure might be considered to be used for CO2 transport. In order to address this, DNV
is currently conducting a gap analysis and is assessing how to update the DNV-OS-F101 code
for offshore transport of CO2 (22).
Page 31 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
8 Fluid Specifications for Pipeline Transport of CO2
The pipelines described in Table 5-1 are differing with respect to the actual CO2 mixture
transported in it and its specified requirements on purity and water contents.
The pipeline specification Kinder Morgan uses in the USA is shown in Table 8-1. (23)
Table 8-1: Specification for Kinder Morgan operated pipelines
Compound Specification Issue/Remark
CO2 95% Min. MMP concern
Nitrogen 4% Max. MMP concern
Hydrocarbons 5% Max. MMP concern
Water 257 ppm wt. Max Corrosion (Specified as 30lbs/MMscf)
Oxygen 10 ppm wt Max Corrosion
Glycol 4*10-5
l/m3 Max
Operations (Specified as 0.3 gal/MMscf)
Temperature 50 C Max Material limit (Specified as 120 F)
For some of the other pipelines, details about the CO2 mixture are given in Appendix 2.
No internationally accepted standard for the specification of CO2 mixtures exists for pipeline
transmission systems. The fluid specification will largely depend upon an assessment
performed during the design phase including flow assurance, pipeline integrity and safety, and
the requirements put upon the CO2 purity by the end user/destination.In (24), a very recent
Dutch study of possible barriers of the CCS chain components with respect to coal fired
power plants the impact of impurities upon the transport system was assessed. The study
proposes the following transport conditions (see Table 8-2):
Table 8-2: Proposed transport conditions from Ecofys study (24)
Compound Specification Issue/Remark
CO2 95% Min.
Nitrogen
Page 32 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
The requirements to be put upon the CO2 purity are further addressed in the EU Dynamis
project (25) and the ENCAP project (26). In (12), an initial investigation of the effect of
impurities in the CO2 stream on the transport systems is presented. This work concludes that
strict requirements on CO2 quality should be avoided to reduce the cost of the capture process
Knowledge gaps are identified regarding the effect of the impurities upon the pipeline
transmission system. The initially proposed quality recommendation of the Dynamis project is
shown in Table 8-3.
Table 8-3 Dynamis proposed specifications (25)
Compound Concentration limit Remarks
H2S 200 ppm Health and safety
considerations
CO 2000 ppm Health and safety
considerations
SOx 100 ppm Health and safety
considerations
NOx 100 ppm Health and safety
considerations
H2O 500 ppm Technical limit
O2 Aquifer 100 ppm
Technical limit; storage issue
CH4 Aquifer < 4 vol%, EOR
Page 33 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
9 Material Aspects
Transporting CO2 by pipeline has an effect upon the material choice of the pipeline and
pipeline components during the design process.
Firstly, supercritical CO2 is used as an industrial solvent; for example in the production of
medication and decaffeinated coffee. This solving ability must be taken into account in the
materials selection process.
Secondly, when liquid water is present, CO2 will partially dissolve and form carbonic acid.
This will give rise to corrosion problems with the steel alloys commonly used for long
pipelines. This will be discussed in detail in Chapter 10.
Thirdly, in transient situations involving rapid depressurization of parts of the pipeline the
material can be exposed to temperature drops below the triple point (-56.6 C).
Fourthly, CO2 pipelines are considered to be more susceptible to fast propagating ductile
fractures compared to gas pipelines. This can put additional requirements on the fracture
properties of the material. This will be discussed in more detail in Chapter 11.
When transporting dense CO2, care has to be taken when selecting materials and compounds
for gaskets, valve seats, sealants, coatings and lubricants. In (27) an overview of tests
regarding the effect of supercritical CO2 on materials is provided. Another overview based
upon several sources is provided in (28). Information regarding compatibility of materials
with supercritical CO2 can also be found in (29). In (30) a best practice for injection well
technology for CO2 as used in EOR is given. This includes the material selection process. In
general the following can be said regarding the material selection for use with dense phase
CO2.
9.1 Elastomers
Generally, elastomers do not respond well to exposure to supercritical CO2. Problems have
been reported with the use of standard Nitrile, Polyethylene, some fluorelastomers,
chloroprene and to some extent ethylene-propylene compounds. Swelling of the elastomer is
attributed to the solubility/diffusion of the pressurized CO2 into the bulk material. With dense
phase CO2 explosive decompression of the elastomer can occur. This phenomenon occurs
when system pressure is rapidly decreased and the gases that have permeated or dissolved into
the elastomer expand. In a mild case, the elastomer will only show blistering (due to the
Page 34 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
expansion of the diffused CO2), but potentially rupture may occur. These issues may be more
severe with higher operating pressures and larger pressure differentials.
Ethylene-propylene co-polymers are reported to be a better option. For high pressure gas
filling connectors as used in CO2 tanks and cylinders, EPDM seals are used. These are
reported to show significant swelling during use. Problems can be alleviated by choosing the
right shore hardness for a specific application. Silicon is reported to be suitable but shows a
high rate of permeability for dense phase CO2.
9.2 Lubricants and Sealants
Petroleum based products will dissolve when in contact with supercritical CO2. Some grease
will decompose and create gum like deposits or harden when in contact with CO2; the main
constituents will dissolve leaving behind the hard gum and wax additives of the grease. This
can seize valves and compressor shafts. Special lubricants and greases for use with dense
phase CO2 are available from a number of suppliers.
9.3 Coatings (internal)
Experimental work reported in (27) on epoxy (both force cured and fused), phenolic (both
baked and fused), nylon-epoxy-amide (force cured), glass (fused) and vinyl (cemented)
coatings/linings showed that only force cured epoxy gave rise to de-bonding after long term
exposure testing to supercritical CO2. Fused epoxy was reported to show no adverse affect.
We did not come across sources referring to the use of pipeline coatings in existing CO2
pipeline. At the SACROC unit, powder applied phenolic epoxy and glass fiber reinforced
epoxy has successfully be used to coat carbon steel pipe (30).
9.4 Valve Seats
There is some uncertainty in the literature about the use of nitrile and teflon in valve seats.
Recommended is EPDM, but only in the absence of hydrocarbons. For hard valve seats
chrome plating is recommended. When the valve seats are in contact with CO2 one can use
anodized aluminum. There exist considerable experience regarding the use of valves with
dense phase/supercritical CO2; it is a matter of specifying the use with dense phase CO2 to the
valve supplier. For example Cooper Cameron supplies a whole range of ball valves for use
with CO2 to the pipeline industry. In (31) the observation is made that ball valve seats should
Page 35 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
be made to seal by injecting sealant into the seating area. This is related to the problems
observed with leaking of mechanical connections with gaskets or seals.
9.5 Gaskets
In (31) the use of stainless/asbestos spiral wound gaskets in combination with standard raised
face flanges is advocated.
9.6 Metals
Carbon steel (C-Mn steel) can be used under the provision that there will be no free water
inside the pipeline. When free water is present, stainless steel has to be used. Experience
shows that S316L functions well. There is mixed experience with S304L. In (30) the use of
S410L is reported to have given pitting corrosion problems at the SACROC unit. Regarding
other metals; dry CO2 functions well with aluminum, brass and copper.
9.7 Engineering Plastics
The following engineering plastics are reported to perform satisfactory with CO2 in dense
phase:
PTFE, PCTFE, PVDF, KYNAR , PA, NYLON, PP
Page 36 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
10 The Free Water Issue
The CO2 mixture coming from most sources contains a certain amount of water. The actual
amount is varying with the source. For example, CO2 scrubbed from flue gas by an amine
process has a water content that can easily exceed 5% vol. The water is to a large extent
knocked out during compression and in subsequent dewatering stages (usually TEG).
On the other end of the scale, CO2 originating from coal gasification, separated by the
Rectisol process can have as little as 2 ppm. vol. water.
Water has a limited solubility in CO2, both in the gaseous as well as in the dense liquid phase.
The solubility of water in CO2 will be in effect a function of pressure and temperature, and
will also be influenced by the purity of the CO2. When during transport the solubility limits of
water are exceeded, free water will precipitate inside the pipeline and give rise to problems.
The occurrence of free water has two negative effects upon the design and operation of CO2
pipelines; corrosion and hydrate formation.
10.1 Corrosion
Occurrence of free water will lead to dissolution of CO2. This will form carbonic acid, H2CO3.
The free water will thus in effect be present as a weak acid. This gives rise to corrosion
problems for the carbon steels of choice for pipelines. CO2 corrosion has been studied
extensively and forms a serious problem for pipeline operations where the chosen material is
carbon steel. For longer pipelines, carbon steel is about the only economically feasible
material choice for dense phase transport of CO2, balancing material cost with the mechanical
strength needed to withstand the internal high pressures and the external loads. Dry CO2 (all
water is dissolved in the CO2) has both in laboratory experiments and from pipeline operating
experience shown to give very low corrosion rates for C-Mn steel. For example, in a study
conducted in connection with the design and engineering of the SACROC pipeline, the
experimental corrosion testing program reported corrosion rates of less than 0.0005 mm/yr on
X-60 ERW steel when there is no liquid water present (27). When liquid free water is
present, corrosion of carbon steel will definitely occur. Corrosion reactions are
electrochemical in nature. For the CO2 - Fe corrosion system there are several anodic and
cathodic reactions, because in presence of liquid water as electrolyte, CO2 will partially
dissolve and forms carbonic acid. These components participate in the reaction chemistry as
well. The basics of this are well described in (32). The possible electrochemical reactions
occurring with CO2 corrosion of carbon steel are shown in Figure 10-1 and the resulting
corrosion in Figure 10-2.
Page 37 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
10-1: Corrosion reactions of water saturated CO2 with carbon steel (32)
10-2: Example of CO2 corrosion attack on a carbon steel pipeline segment (32)
Page 38 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
When free, liquid, water is present, corrosion will occur and at very high rates. Rates up to
and beyond 10 mm/year are reported (33). Furthermore, the corrosion mechanism is an on/off
process somewhat complicating prediction of corrosion rates. Corrosion attacks will be
typically localized at initial initiation sites, this is attributed to the galvanic effect, leading to
high local corrosion rates that may lead to weaknesses or leaks in the pipe wall within short
periods of time.
CO2 corrosion has been researched extensively and is relatively well understood. Major
studies have been conducted regarding CO2 corrosion in oil and gas pipelines for
hydrocarbons containing several mole% CO2. However, very little experimental work has
been carried out regarding CO2 corrosion in pipelines at the high partial pressures encountered
when transporting high purity CO2. A multitude of corrosion models have been developed for
hydrocarbons containing CO2, but it has been registered by (32) (34) that the results can vary
by a factor 100. This is attributed to corrosion, and the corrosion effect of CO2, to be linked to
a multiple of mechanisms. Several CO2 dependent chemical, electrochemical and mass
transport processes occur simultaneously. These are depending on a variety of parameters,
including CO2 partial pressure and temperature. All this will have to be accounted for in the
models. At high partial pressures the existing models tend to overestimate the corrosion rates.
All this makes it a challenging task to specify a corrosion allowance based upon incidental
occurrence of free water in a CO2 pipeline. In addition, the concentrations and types of other
impurities present in the CO2 mixture will influence the corrosion rates. The presence of O2,
H2S, SO2 and NOx all have an influence towards higher corrosion rates. In (34) it is
underlined that the mechanism of CO2 corrosion in the presence of impurities is not entirely
understood. Setting a corrosion allowance is therefore probably not a suitable way to deal
with the danger of CO2 corrosion in a carbon steel CO2 pipeline.
There are however several other ways to mitigate this for natural gas transmission pipelines:
Reduce the contents of acidic gasses (which is not an option here).
Avoid liquid water to wet the pipeline surface, for example with a coating or build up
of an oil or wax layer.
Glycol addition, this reduces the solubility of acid gasses and reduces water
concentration; this has proven to reduce corrosion rates on bare steel surfaces
Inhibition
pH stabilization: can be done by adding a base to glycol, raising the pH and thus
reducing the solubility of FeCO3.
Avoid presence of free liquid water.
Material choice.
Page 39 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
The reported experiences with CO2 corrosion with likely pipeline alloy candidates are shown
inFigure 10-3.
10-3: overview of corrosion rates and experiences reported in the literature for candidate steels for CO2
pipelines (34)
The state of the art approach to this problem is to set and maintain a specification on
maximum water contents of the CO2 in the pipeline. The objective is to avoid forming of free
water within the operational window (pressure and temperature) of the pipeline.
This will still leave open the question what measures should be taken in case there is
incidental free water ingress into the pipeline. This can for example happen when the capture
plant delivers moisture saturated CO2 to the pipeline (for example due to dehydration
equipment failure). In (35) MEG or use of a commercial corrosion inhibitor is suggested. In
(34) the recommendation is given that for transportation of CO2 in C-Mn pipelines the risks of
corrosion at actual flow conditions should be investigated. The consequence of impurities and
an evaluation of inhibitors should be included. Impurities in the CO2 will both affect the water
solubility and the corrosion mechanisms.
Another source of free water can be the reactions between impurities; during this study we did
not encounter any work that describes potential chemical reactions between impurities at
pipeline operating conditions. For example; impurities containing molecular hydrogen (like
for example H2S) might under certain conditions react with O2 creating additional water.
Page 40 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
10.2 Hydrates
Forming of hydrates (clathrate hydrates) during CO2 transport is similar to that occurring in
natural gas transport. Clathrate hydrates are solids in which gas molecules occupy a vacancy
in a cage made up of hydrogen bonded water molecules. CO2 is just one of the light molecular
gas weight molecules that can form hydrates together with water. Some of the impurities that
can be present in the CO2 can also form hydrates, like CH4, H2S, N2, Ar as well as some
higher hydrocarbons (like C2H6 and C3H8). These hydrates are actually not chemical
compounds; their formation and decomposition are actual phase transitions. In appearance
hydrates remind of ice. Contrary to ice, it should be noted that hydrates can form at
temperatures above 0 C. This phase transition point is also highly pressure dependent.
Hydrates become more stable with increasing pressure. Hydrates can form plugs in pipelines,
either blocking valves, fouling up instrumentation or in the extreme case block the entire bore
of the pipeline at a certain location. During depressurization the acceleration of a hydrate plug
can cause structural damage to the pipeline wall in small radius bends. Hydrates have been
observed to have an affinity for building up at the pipeline walls. When they occur they can
be decomposed by either lowering the pressure, increasing the temperature or reduce the
water content.
Measures that can be taken to avoid hydrates:
Use of thermodynamic inhibitors; MEG and DEG to lower the hydrate forming
temperature (effectively an antifreeze). This works for natural gas hydrates, but no
references were found how well this will function with CO2.
Use of kinetic inhibitors; they slow down the kinetics of formation.
Use of anti-conglomerants; they allow hydrates to form but not to stick together
Raising the operational temperature window of the pipeline (maintaining the pressures and
temperatures outside the hydrate formation conditions)
Lowering the water dew-point in the hydrate forming regime (dehydration)
Research addresses the development of so-called low dosage gas hydrate inhibitors that
either act to delay nucleation or prevent growth while being present at low concentrations
(typically less than 1% of the water contents) (36).
Page 41 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
The Phase equilibrium diagram for the system CO2/H2O is shown in Figure 10-4.
10-4: Phase equilibrium diagram CO2/H20 (37)
The CO2-H2O hydrate system has been studied by for example (37), (38).
As Figure 10-4 illustrates, within the operating window for offshore CO2 pipelines, hydrates
will be stable when the temperature is below 283 K (9.85C). It should be noted that the
experimental work behind this figure is based upon water saturated mixtures of CO2. The
presence of free water will definitely enhance hydrate formation. There are indications that
hydrates also can form under conditions above the water dew point of the actual CO2-H2O
system. These are reported to be difficult to create under laboratory conditions. The following
conditions are necessary to get hydrates (39).
The right combination of temperature and pressure (low temperature, high pressure)
Presence of hydrate forming molecules
A suitable amount of water.
Furthermore, hydrate forming is enhanced by turbulence (especially in connection with choke
valves, presence of nucleation sites like welding spots and pipe fittings and the presence of
free water.
Page 42 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
The equilibrium in the hydrate region depends on the amount of water present:
For a large amount of water, the equilibrium is between water and the hydrate
For a small amount of water, the equilibrium is between a gas and the hydrate
If the mixture is very lean on water, no hydrate will form.
As the figure shows, water saturated CO2 will form hydrates below 283 K (9.85 C).
The water saturation condition can occur when there is a temperature drop of the CO2 in
relation to pressure reduction. This is encountered for example during the blow down of a
pipeline or downstream a valve during part filling of the pipeline. Dehydrating the mixture is
likely to reduce the hydrate forming temperature. We did not find reports of occurrence of
hydrates for the existing land based CO2 pipelines in the USA. This can be due to the fact
some of them operate most of the time above the hydrate temperature (although winter
temperatures in the USA can be low enough to get into the hydrate stable region) or due to the
low water contents (much lower than theoretically needed to avoid free water.
The Weyburn pipeline is expected to encounter low enough temperatures during winter, but
due to the separation process, the CO2 here has a very low water contents. When sufficient
water is present as is the case in injection of CO2 for EOR in oil wells, hydrate formation is
taken into account and dealt with. An offshore pipeline on the Norwegian Continental Shelf
will along most of its length transport CO2 within the hydrate stable region.
Within the scope of this study we found very little information about hydrate forming
conditions in both pure CO2 and CO2 with impurities when there is an absence of free water.
The question remains if the water solubility limit is a non-conservative limit for the allowable
water contents of the CO2. When taking into account that CO2 hydrates are likely to form
below 283 K rather than free water (although we can expect equilibrium between these two as
well), the question remains what the maximum allowable water content is to avoid formation
of stable hydrates.
We can at this stage not rule out that these water levels are below the limit needed to avoid
liquid water drop out. As the fluid is exposed to turbulent flow, possibly more factors are
governing hydrate formation than the thermodynamic equilibrium alone. This is an area that
should be further investigated, including how the impurities affect the formation of hydrates
in unsaturated (dry) CO2.
Page 43 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
10.4 Water Solubility
The water solubility in dense phase CO2 increases with pressure and temperature. This is
different from supercritical natural gas which has a decreasing water content with increasing
pressure. It is therefore not safe to set a dew point requirement at the highest operating
pressure.
Available measurement data for the CO2-H2O system are mostly at higher temperatures; only
a small amount of experimental data can be found within the pressure-temperature range of
interest for offshore pipeline transport of CO2. An overview of available experimental data on
water solubility of H2O in pure CO2 is provided in (40).
Figure 10-5 below shows the relation between pressure and maximum water contents based
upon data from (38) (41) within the temperature and pressure regions applicable for an
offshore pipeline. The water content is in mole%. The graph shows that the water solubility
decreases with pressure and temperature.
10-5: maximum water solubility in CO2 (38).
It must also be noted that the water solubility of CO2 shows a minimum value within the gas
phase at pressures directly under the vapor pressure (35). The solubility of water in gaseous
CO2 at a given temperature will decrease to a minimum with increasing pressure. Further
increasing the pressure will lead to the phase transition to dense phase/liquid CO2 and the
water solubility increases again. This can be used to remove water during compression, but
can also lead to free water when pressure releasing CO2 in the dense liquid phase. The
presence of impurities in the CO2 will also influence the water solubility. Some of the
impurities will lower the water solubility. This is for example the case with H2S and CH4. In
(35) experiments and calculations for the solubility of water in pure CO2 are compared with
0
0,05
0,1
0,15
0,2
0,25
0,3
0,35
0,4
-20 -15 -10 -5 0 5 10 15 20
Mole %
water
Degrees Celcius
62.1 bar
82.8 bar
103.4 bar
137.9bar
Page 44 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
that of a mixture containing methane. The water solubility in the dense phase is shown to be
significantly lower. Measurements presented in (42) show that in the dense phase the
solubility can be as much as 30% lower for a mixture containing 5.3% CH4 compared to pure
CO2. In (43) the results of a literature investigation indicate that for low concentrations (up to
200 ppm) the effect of H2S on water solubility is not significant, while CH4 decreases water
solubility significantly. The same publication also reports that it did not find evidence of
cross-effects of O2 and N2 on water solubility. In (44) a water saturation prediction model is
described for CO2 mixtures with up to 5 mole% in total of CH4 and N2 in the pressure range
0.1-27.7 MPa. The model predicts water saturation values in the non-hydrate region only.
This work concludes with that at constant temperature and pressure, dilution of the CO2 by
CH4 and/or N2 will reduce the water saturation value.
To predict the solubility of H2O in CO2 it is important to note that the original form of the
Redlich-Kwong-Soave (RKS) nor the Peng-Robinson (PR) equation of state accurately
reproduces the vapor pressure of water (45). Therefore modifications of these equations must
be used, which are available (45) (35).
In (45), the results from a commercial tool using a two fluid approach are compared to
existing measurement data for pure CO2 and show reasonable agreement. A recent study (40)
from Sintef and StatoilHydro compared the calculated results for water solubility using RKS
EoS with both the Van der Waals and Huron Vidal mixing rules and the CPA (Cubic Plus
Association) with literature collected experimental data for CO2 mixtures containing CH4.
The conclusions from this study were that SRK model with Huron Vidal mixing rule is able to
calculate the solubility of H2O in these mixtures most accurate (from 3 to 9.3% average
deviation). The RKS model with Van der Waals mixing rule was found not to be able to
calculate the mutual solubilities correctly. The CPA model was found to be able to calculate
the solubilities but with less accuracy (from 9 to 35% deviation).
For the different carbon dioxide pipelines in operation, different practices for maximum
allowable water contents are used (Table 5-1). It is therefore not clear what the optimum
water contents specification is, especially with the presence of impurities. Further
investigation in to the effect of the impurities on water solubility, the availability of
experimental data and possibly further development of the thermodynamic models to
calculate the solubilities for actual CO2 mixtures will be needed to set safe water
specifications.
Page 45 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
11 Fracture Propagation in CO2 Pipelines
CO2 pipelines are considered to be more susceptible to fast propagating ductile fractures
compared to gas pipelines. Fast propagating ductile fractures are fractures that, once initiated,
travel for a long distance along the pipeline. In pipeline design, obviously the probability of
the occurrence of fracture has to be addressed and minimized through material selection and
dimensioning. However, once a fracture occurs, it is necessary that arrest of the fracture also
has been built into the design. This means that fast propagating ductile fractures must be
avoided through material selection and dimensioning during the design phase. In (22) a good
description of the basic mechanisms in play during fast propagating ductile fractures is
provided. The mechanisms are described as follows:
It starts with the initiation of a fracture (for example through external force)
When the driving force (internal pipeline pressure) is above a certain level, the crack
will propagate in one or both directions along the pipeline.
As long as the driving force in the region of the crack tip is above this threshold level,
the crack tip will continue to propagate. The propagation velocity is close to the speed
of sound in the pipeline steel.
As soon as the crack opens, and the fluid medium starts to leak, a pressure relief front
starts propagating in both directions.
If the pressure relief front moves faster than the crack propagates, at some point in
time the driving force at the crack tip disappears and the crack arrests
It is thus a race between these two velocities; the speed of crack propagation versus
the speed of the pressure relief front. When the pressure relief front catches up with
the crack tip, the fracture might arrest. Otherwise the fracture will keep on running
until some other barrier arrest it.
The fracture arrest properties at a given temperature and pressure depend on the wall
thickness and the material properties, particularly fracture arrest toughness. In some of the
existing CO2 pipelines the risk of fast running ductile fractures is addressed through the use of
fracture arresters. These are normally rings of metal, tightly bonded to the outer surface of the
pipeline. They function as a local increase of the wall thickness. In addition, on segmented
lines the housings of the inline block valves can double as fracture arrestors.
As mentioned above, a ductile fracture will not propagate if there is insufficient driving force
in the system to overcome the resistance to propagation of the fracture. The Batelle Two
Curve Methodology is a model of the fracture process, expressing this balance in terms of the
fracture velocity and decompression velocity curve. Supercritical CO2 decompresses as an
Page 46 of 87
Report number Date Security level
POL-O-2007-138-A 8 January 2008 Open
elastic liquid first and then as a two phase fluid (this differs from natural gas dense phase
which will decompress as a gas only, while rich gas will decompress as a gas first and then as
a two phase fluid). The implication of this is that CO2 decompresses first rapidly down to the
saturation pressure (as a liquid) and then maintains the saturation pressure during
decompression as a two phase fluid (for quite some time; there is no rapid depressurization as
with a natural gas pipeline). The risk is that the high, sustained, vapor pressure maintains the
driving force for fracture propagation. The implication of this decompression characteristic is
that the necessary toughness to arrest a fracture can be estimated by using saturation and
arrest pressure (the arrest pressure is the pressure in the pipe below which a running ductile
fracture cannot occur: it is a function of the strength and toughness of the pipeline steel, and
the diameter and wall thickness of the pipe). This approach is in reality a simplification,
ignoring the decompression of the two-phase fluid and leads to a conservative estimate. This
is further described in (46) (47).
In order to stop a fast running ductile fracture, the saturation pressure must be less than the
arrest pressure. The initial temperature and pressure at fracture initiation as well as the
presence of impurities will al