DYNAMIC SIMULATIONS OF CARBON DIOXIDE PIPELINE TRANSPORTATION FOR

THE PURPOSE OF CARBON CAPTURE AND STORAGE

By

BILAAL YUSUF HUSSAIN

A thesis submitted to the University of Birmingham

For the degree of

ENGINEERING DOCTORATE

Efficient Fossil Energy Technologies Centre

School of Metallurgy and Materials

College of Engineering and Physical Sciences

University of Birmingham

September 2017
 University of Birmingham Research Archive
e-theses repository

This unpublished thesis/dissertation is copyright of the author and/or third

parties. The intellectual property rights of the author or third parties in respect
of this work are as defined by The Copyright Designs and Patents Act 1988 or
as modified by any successor legislation.

Any use made of information contained in this thesis/dissertation must be in

accordance with that legislation and must be properly acknowledged. Further
distribution or reproduction in any format is prohibited without the permission
of the copyright holder.
 Abstract

This Engineering Doctorate project aimed to study the effects of varying flowrates

on the flow dynamics of carbon dioxide within a pipeline for the purposes of carbon

capture and storage. Understanding the flow dynamics of the carbon dioxide within

a pipeline when there are varying inlet flowrates is important in establishing the most

appropriate method to operate such a system.

The researched utilised the software tool gCCS to develop models in which three

different transport systems were simulated. The first scenario looked at the effects

of transporting pure carbon dioxide in both the supercritical phase and the sub-

cooled liquid phase. The outputs from the model showed that when the inlet flowrate

is decreased, the outlet flowrate responds in three distinct phases. The first phase

that occurs has been referred to as the ‘delayed response phase’, the second phase

is the ‘offset phase’ and the final phase is the ‘reduction phase’. The simulations

showed that the flowrate difference between the inlet and the outlet of the pipeline

during the ‘offset phase’ was greater when transporting carbon dioxide in the

supercritical phase when compared to the sub cooled liquid phase.

The second scenario looked at comparing the effects of three different impurities;

hydrogen, nitrogen and oxygen, in the carbon dioxide on the flow response when

the inlet flowrate is decreased. It was found that all three impurities caused an

increase in the offset between the inlet and outlet flowrate during the ‘offset phase’.

The largest difference in the flowrates was observed when hydrogen was present.

The third case that was investigated looked at how multiple sources of carbon

dioxide effected the flowrate within the main trunk pipeline when the flowrate of one

i
of the sources was reduced. It was found that multiple sources of carbon dioxide do

not affect the flow of within the pipeline beyond that of the base case.

The final part of the research compared real pipeline data from the Shell QUEST

pipeline to the model, this enabled validation of the model. It was found that the

model was able to predict the flowrate and pressure of the carbon dioxide with good

accuracy. Temperature predictions were significantly different from the data and it

has been suggested this is due to the restrictions of the thermal conductivity of the

surrounding pipeline material in the model.

ii
 Acknowledgements

This EngD could not have been done without the help and support of numerous

people who provided me with all that I needed to complete the research.

I would firstly like to give greatest thanks to both of my supervisors. My academic

supervisor, Professor Joe Wood who helped drive me to do the best that I could do

and gave me guidance throughout the project. My Industrial supervisor Den

Gammer who is the reason I have had the opportunity to do this project and who I

am indebted to for all the help and support he has provided.

Thank you to the Efficient Fossil Energy Technologies doctoral centre, the EPSRC

and the Energy Technologies Institute who provided the funding to allow this project

to be done. I would also like to thank Andrew Green, Shell Canada and the

Canadian Foreign and Commonwealth Office for allowing me to undertake an 8

month placement which enabled the research to reach new heights that would

otherwise not have been possible. I would specifically like to thank Stephen

Tessarolo at Shell for supporting me while I was undertaking my research in

Canada.

All the people at the Energy Technologies Institute for making the entire experience

enjoyable.

iii
A special thank you to Shamal Crowther, for always being there for me during the

difficult times and uplifting my spirits when I needed it the most and my cousin

Hassan who provided me with the support I needed through the most difficult times.

The most important people in allowing me be able to achieve all that I have, my

parents, who gave me the love and support I needed to be able to get through my

EngD and I will always be grateful to them for all they have given me. My sisters

who have always been there for me when I needed them and my nephews and

niece who always made me smile when I needed it.

iv
 Table of Contents

Chapter 1 - Introduction................................................................................................ 1
1.1 Background ........................................................................................................ 2
1.2 Research Aims ................................................................................................... 2
1.3 Thesis Outline..................................................................................................... 3
Chapter 2 – Literature Review ...................................................................................... 6
2.1 Introduction ......................................................................................................... 7
2.2 Green House Gasses and Global Warming ........................................................ 7
2.3 Carbon Capture and Storage .............................................................................. 8
2.3.1 What is Carbon Capture and Storage? ........................................................ 8
2.3.2 Global CCS Projects .................................................................................. 10
2.4 How CO2 is Transported ................................................................................... 11
2.4.1 Methods of Transporting CO2 .................................................................... 11
2.4.2 Vessel VS Pipeline Transport .................................................................... 11
2.5 CO2 Pipeline Modelling ..................................................................................... 16
2.5.1 Economic Modelling of CO2 Pipelines ........................................................ 16
2.5.2 Technical Modelling of CO2 Pipelines ........................................................ 19
2.6 Modelling Tools Evaluation ............................................................................... 25
2.6.1 Model Requirements.................................................................................. 25
2.6.2 OLGA ........................................................................................................ 28
2.6.3 Aspen HYSYS ........................................................................................... 31
2.6.4 gCCS ......................................................................................................... 33
2.7 Conclusion ........................................................................................................ 35
Chapter 3 – Comparing Variable Flows For Liquid and Supercritical Phase Carbon
Dioxide 39
3.1 Introduction ....................................................................................................... 40
3.2 Hypothesis ........................................................................................................ 40
3.3 gCCS Transport Models ................................................................................... 40
3.3.1 Carbon Dioxide Source Model ................................................................... 41
3.3.2 Pipeline and Well Models........................................................................... 42
3.3.3 Valve Model ............................................................................................... 46
3.3.4 Distribution Header Model ......................................................................... 47

v
 3.3.5 Reservoir ................................................................................................... 48
3.4 Carbon Dioxide Pipeline Model Development ................................................... 50
3.5 Simulation Output Analysis ............................................................................... 54
3.5.1 Liquid Phase Transport .............................................................................. 54
3.5.2 Supercritical Phase Transport.................................................................... 61
3.5.3 Carbon Dioxide Phase Evaluation ............................................................. 67
3.6 Conclusion ........................................................................................................ 72
Chapter 4 – Effects of Impurities and Multiple Sources of CO2 on Pipeline Flow ........ 75
4.1 Introduction ....................................................................................................... 76
4.2 Hypothesis ........................................................................................................ 77
4.3 CO2 with Impurities Case Model Development.................................................. 77
4.3.1 Impurities ................................................................................................... 78
4.4 Results ............................................................................................................. 79
4.4.1 Nitrogen Case............................................................................................ 79
4.4.2 Hydrogen Case.......................................................................................... 83
4.4.3 Oxygen Case ............................................................................................. 85
4.5 Analysis ............................................................................................................ 88
4.6 Multiple Sources of CO2 Model Development ................................................... 89
4.6.1 Pipeline Dimensions .................................................................................. 89
4.6.2 Pipeline Flows ........................................................................................... 90
4.7 Results ............................................................................................................. 91
4.7.1 Flowrate analysis ....................................................................................... 91
4.7.2 Pressure Analysis ...................................................................................... 92
4.7.3 Temperature Analysis ................................................................................ 93
4.8 Conclusion ........................................................................................................ 94
Chapter 5 – Modelling of Shell QUEST CO2 Pipeline ................................................. 96
5.1 Introduction ....................................................................................................... 97
5.1.1 Overview ................................................................................................... 97
5.1.2 The QUEST Carbon Capture Facility ......................................................... 98
5.2 Methodology ..................................................................................................... 99
5.2.1 Pipeline Details ........................................................................................ 100
5.2.2 Pipeline Operation ................................................................................... 106
5.2.3 Simulation details .................................................................................... 107
5.2.4 Data Analysis........................................................................................... 109

vi
 5.2.5 Measurement Devices ............................................................................. 109
5.2.6 Simulation Periods ................................................................................... 112
5.3 Initial Model Development............................................................................... 116
5.3.1 Topology.................................................................................................. 116
5.3.2 Control Schemes ..................................................................................... 117
5.3.3 Schedule ................................................................................................. 117
5.3.4 Initial Model Analysis ............................................................................... 120
5.3.5 Secondary Model Development ............................................................... 120
5.4 Comparison Between Model and QUEST Data .............................................. 123
5.4.1 Flowrate................................................................................................... 123
5.4.2 Pressure .................................................................................................. 128
5.4.3 Temperature ............................................................................................ 134
5.5 Determining ‘Goodness of Fit’ ......................................................................... 137
5.5.1 Coefficient of Efficiency ........................................................................... 137
5.5.2 Index of Agreement ................................................................................. 138
5.6 Conclusion ...................................................................................................... 140
Chapter 6 – Conclusions and Future Work ............................................................... 142
6.1 Conclusions......................................................................................................... 143
6.2 Future Work ........................................................................................................ 149
Appendix A Simulation Code ................................................................................... 152
A.1 Code for Base Case and Impurities Case ........................................................... 152
Appendix B Simulation Code .................................................................................. 153
B.1 Code for Multiple CO2 Sources Case .................................................................. 153
Appendix C Simulation Code .................................................................................. 154
C.1 Code for Shell QUEST Simulation for Time Period 24/10/15 – 28/10/15 ............. 154
C.2 Code for Shell QUEST Simulation for Time Period 31/10/15 – 04/11/15 ............. 155
C.3 Code for Shell QUEST Simulation for Time Period 08/11/15 – 11/11/15 ............. 156
References ................................................................................................................... 157

vii
 List of Figures

Figure 2-1: CO2 Phase Diagram ..................................................................................... 13

Figure 3-1: CO2 Source Variables .................................................................................. 41
Figure 3-2: Pipeline Segment Design Variables .............................................................. 43
Figure 3-3: Pipeline Segment Heat Transfer Variables ................................................... 45
Figure 3-4: Well Heat Transfer Variables ........................................................................ 46
Figure 3-5: ESD Valve Configuration .............................................................................. 47
Figure 3-6: Distribution Header Configuration ................................................................. 48
Figure 3-7: Reservoir Configuration ................................................................................ 49
Figure 3-8: Model Pipeline Topology............................................................................... 51
Figure 3-9: Inlet and Outlet Flowrate Profile for Pipeline 1 and 2 .................................... 56
Figure 3-10: Inlet and Outlet Flowrate Profile for Pipeline 1 and 2................................... 57
Figure 3-11: Flowrate Profile along Pipeline001 .............................................................. 57
Figure 3-12: Inlet and Outlet Pressure for Pipeline 001 ................................................... 59
Figure 3-13: Inlet and Outlet Temperature for Pipeline 001 ............................................. 60
Figure 3-14: Outlet Flowrate and Outlet Temperature for Pipeline 001 ........................... 61
Figure 3-15: Inlet and Outlet Flowrate for Pipeline 001 and 002 ...................................... 63
Figure 3-16: Inlet and Outlet Flowrate for Pipeline 001 and 002 ...................................... 64
Figure 3-17: Flowrate Profile along Pipeline001 .............................................................. 64
Figure 3-18: Inlet and Outlet Pressure for Pipeline 001 ................................................... 65
Figure 3-19: Inlet and Outlet Temperature for Pipeline 001 ............................................. 66
Figure 3-20: Inlet and Outlet Flowrate for Liquid and Supercritical CO2 .......................... 69
Figure 3-21: Flowrate profile along Pipeline001 for Supercritical and Liquid CO2 ........... 69
Figure 3-22: Inlet and Outlet Pressure for Liquid and Supercritical CO2 .......................... 71
Figure 3-23: Inlet and Outlet Temperature for Liquid and Supercritical CO2 .................... 72
Figure 4-1: Pipeline Inlet and Outlet Flowrate (CO2 + N2)................................................ 80
Figure 4-2: Pipeline Inlet and Outlet Pressure (CO2 + N2) ............................................... 81
Figure 4-3: Pipeline Inlet and Outlet Temperature (CO2 + N2) ......................................... 82
Figure 4-4: Pipeline Inlet and Outlet Flowrate (Pure CO2 & CO2 + N2) ............................ 82
Figure 4-5: Pipeline Inlet and Outlet Flowrate (CO2 + H2)................................................ 83
Figure 4-6: Pipeline Inlet and Outlet Pressure (CO2 + H2) ............................................... 84
Figure 4-7: Pipeline Inlet and Outlet Temperature (CO2 + H2) ......................................... 84
Figure 4-8: Pipeline Inlet and Outlet Flowrate (Pure CO2 & CO2 + H2) ............................ 85
Figure 4-9: Pipeline Inlet and Outlet Flowrate (CO2 + O2) ............................................... 86
Figure 4-10: Pipeline Inlet and Outlet Pressure (CO2 + O2) ............................................. 86
Figure 4-11: Pipeline Inlet and Outlet Temperature (CO2 + O2) ....................................... 87
Figure 4-12: Pipeline Inlet and Outlet Flowrate (Pure CO2 & CO2 + O2) .......................... 87
Figure 4-13: Pipeline Topology with Multiple Sources of CO2.......................................... 91
Figure 4-14: Inlet and Outlet Flowrate for Pipeline 001, Pipeline 002 and Pipeline 003 ... 92
Figure 4-15: Inlet and Outlet Pressure for Pipeline 001, Pipeline 002 and Pipeline 003 .. 93
Figure 4-16: Inlet and Outlet Temperature for Pipeline 001, Pipeline 002 and Pipeline 003
....................................................................................................................................... 94
Figure 5-1: QUEST pipeline topography ....................................................................... 101
Figure 5-2: Pressure sensor.......................................................................................... 111

viii
Figure 5-3: Inlet and Outlet mass flowrate of the Quest pipeline between 07/10/15 and
28/11/15........................................................................................................................ 113
Figure 5-4: QUEST pipeline inlet mass flowrate data from 25/10/15 to 28/10/15........... 114
Figure 5-5: QUEST pipeline inlet mass flowrate data from 31/10/15 to 04/11/15........... 115
Figure 5-6: QUEST pipeline inlet mass flowrate data from 08/11/15 to 11/10/15........... 116
Figure 5-7: Initial pipeline model ................................................................................... 119
Figure 5-8: Simplified QUEST pipeline model ............................................................... 122
Figure 5-9: Model and QUEST pipeline inlet and outlet flowrates 24/10/15 – 28/10/15 . 124
Figure 5-10: Model and QUEST pipeline inlet and outlet flowrates 31/10/15 – 04/11/15 124
Figure 5-11: Model and QUEST pipeline inlet and outlet flowrates 08/11/15 – 11/11/15 125
Figure 5-12: Quest pipeline inlet and outlet flowrate 24/10/15 - 28/10/15 ...................... 126
Figure 5-13: Quest pipeline inlet and outlet flowrate 31/10/15 - 04/11/15 ...................... 127
Figure 5-14: Quest pipeline inlet and outlet flowrate 08/11/15 - 11/11/15 ...................... 127
Figure 5-15: Model and QUEST pipeline pressures 24/10/15 – 28/10/15 ...................... 129
Figure 5-16: Model and QUEST pipeline pressures 25/10/15 09:00 - 25/10/15 10:30 ... 130
Figure 5-17: Model and QUEST pipeline pressures 26/10/15 12:00 - 27/10/15 12:00 ... 130
Figure 5-18: Model and QUEST pipeline pressures 31/10/15 – 04/11/15 ...................... 131
Figure 5-19: Model and QUEST pipeline pressures 01/11/15 05:00 – 01/11/15 06:30 .. 131
Figure 5-20: Model and QUEST pipeline pressures 02/10/15 14:00 – 02/11/15 21:00 .. 132
Figure 5-21: Model and QUEST pipeline pressures 08/11/15 – 11/11/15 ...................... 132
Figure 5-22: Model and QUEST pipeline pressures 08/11/15 13:30 – 08/11/15 15:00 .. 133
Figure 5-23: Model and QUEST pipeline pressures 09/11/15 14:00 – 10/11/15 00:00 . 133
Figure 5-24: Model and QUEST pipeline inlet and outlet temperatures 24/10/15 –
28/10/15........................................................................................................................ 136
Figure 5-25: Model and QUEST pipeline inlet and outlet temperatures 31/10/15 –
04/11/15........................................................................................................................ 136
Figure 5-26: Model and QUEST pipeline inlet and outlet temperatures 08/11/15 –
11/11/15........................................................................................................................ 137

ix
 List of Tables

Table 2-1: Cost estimates (€/t CO2) for commercial natural gas-fired power plants with
CCS or coal-based CCS demonstration projects with transported volume of 2.5 Mtpa ... 14
Table 2-2: Cost estimate for large scale networks of 20 Mtpa (€/tonne CO2) .................. 15
Table 2-3: Pre-combustion CO2 impurities from pulverised coal...................................... 26
Table 2-4: Post-combustion CO2 impurities from pulverised coal .................................... 27
Table 2-5: Oxyfuel combustion CO2 impurities ................................................................ 27
Table 3-1: CO2 Source Parameters ................................................................................ 51
Table 3-2: ESD Valve Design Parameters ...................................................................... 52
Table 3-3: Pipeline Section Design ................................................................................. 52
Table 3-4: Choke Valve Design ...................................................................................... 52
Table 3-5: Well Design.................................................................................................... 52
Table 3-6: Reservoir Parameters .................................................................................... 53
Table 4-1: Common Impurities Found in Captured CO2 .................................................. 79
Table 4-2: Pipeline Dimensions ...................................................................................... 90
Table 5-1: QUEST pipeline details ................................................................................ 102
Table 5-2: QUEST pipeline dimensions ........................................................................ 103
Table 5-3: QUEST fluid composition ............................................................................. 104
Table 5-4: Quest pipeline operating conditions ............................................................. 105
Table 5-5: Reservoir Operating Conditions ................................................................... 106
Table 5-6: Soil Material & Thermal Conductivities ......................................................... 135
Table 5-7: Goodness of fit ............................................................................................. 139

x
 List of Symbols

Alphabetic Symbols

A Helmholtz free energy

a1, a2, a3 Adjustable parameters -

ac Acentric factor -

d Index of agreement -

Do Outer pipeline diameter m

E Coefficient of Efficiency -

fD Darcy friction factor -

FL Location factor -

FT Terrain factor -

L Pipeline length m

Loffshore Offshore pipe length m

Lonshore Onshore pipe length m

M Mass flowrate kg s-1

Oi Observed value -

Pc Critical pressure MPa

xi
Pi Predicted value -

q Heat energy J

R Molar gas constant J mol-1 K-1

T Temperature K

Tc Critical temperature K

Tf Fluid temperature K

To Soil temperature K

Uo Overall heat transfer coefficient W m-2 K-1

V Volume m3

Δp Pressure drop mPa

Volumetric flowrate m3 s-1

xii
Greek Symbols

δ Reduced density

δij Binary interaction coefficient

Θ Function

τ Inverse reduced temperature

Φ Dimensionless Helmholtz free energy

Φo Ideal gas behaviour

Φr Reduced fluid behaviour

xiii
Chapter 1 - Introduction

1
1.1 Background
Carbon capture and storage (CCS) will be a critical greenhouse gas reduction

technology for as long as fossil fuels are widely used [1]. Modelling by the Energy

Technologies Institute has found that without CCS the cost of meeting the UK 2050

greenhouse gas targets will double [2]. The development of any CCS capture

technology will require an understanding of how the CO 2 is transported.

Transportation via pipeline is seen as the most viable option to transfer the CO 2

from the source to the site of sequestration. An awareness of the dynamics of the

CO2 fluid is necessary to ensure the pipeline can be operated in the most

economical way.

To develop the knowledge base around CCS, the Efficient Fossil Energy

Technologies (EFET) engineering doctorate centre was established, which is a

collaboration between the University of Nottingham, Loughborough University and

the University of Birmingham along with the Energy Technologies Institute and other

industrial partners, with funding from the Engineering and Physical Sciences

Research Council. The work presented in this thesis reports the application of a

developed modelling tool to simulate a dynamic CO2 pipeline, so as to represent a

real CCS system that is connected to a combined cycle gas turbine power plant.

1.2 Research Aims

The aim of the research is to investigate the effects of variable CO2 flowrates on the

dynamics of the fluid within a CO2 pipeline used for CCS purposes. To carry out this

research, a professionally developed piece of software (gCCS) has been used to

simulate the pipeline dynamics. To validate the simulation outputs, industrial data

obtained from the Shell Quest CO2 pipeline has also been utilised.

2
The research was split into the simulation and then the validation. The simulation

looked at three different scenarios, the first scenario examined a base case where

pure CO2 was transported via a single pipeline and investigated the difference

between transporting in the supercritical phase and the subcooled liquid phase, the

second scenario investigated transporting near pure CO2 and compares the effects

of different impurities on the dynamics of the fluid within the pipeline; the third

scenario simulated multiple sources of CO2 and understanding how changing the

flowrate from one source effects the pipeline.

The validation part of the research involved collaboration with Shell Canada, via an

eight month long placement at the QUEST project in Fort Saskatchewan, Alberta,

with the aim of comparing real CO2 pipeline data to the simulation outputs and

evaluating the accuracy of these ouputs.

For each scenario a hypothesis is presented which is either proven or disproven

through the research.

1.3 Thesis Outline

The thesis contains six chapters. The first chapter introduces the project and the

background of the research along with the aims of the study and an overview of

each of the chapters.

The second chapter reports a critical review of the literature and helps to inform the

direction of the research and where there are knowledge gaps in the area. The

literature covers the broader area of carbon capture and storage and then a more

detailed investigation into the area of CO2 transportation where all types of CO2

transport are examined. The review then covers the current literature on CO 2

3
pipelines, the most economical way of transportation and the modelling and

simulation work that has been done which compares the technical differences

between supercritical and the subcooled liquid phase in steady state operation. The

final part of the literature review addresses which of the current commercially

available modelling tools are the most appropriate for the research to be undertaken

and is based upon the tools ability to predict the physical properties of CO2 and near

pure CO2.

The third chapter provides the details of the process simulation tool that is used and

details specifically the units related to pipeline and CO2 storage and what values

are user defined. The development of the simulations of the scenario of a base case

pipeline model comparing supercritical and subcooled liquid phase CO 2 are

covered. The results from the simulations are then presented and analysed.

The fourth chapter covers two areas of the research, the impacts of impurities on

the dynamics of CO2 flow which looks at the main impurities found within CO2 from

different CCS technologies and the effects of multiple sources of CO2 on pipeline

dynamics. Each scenario is simulated and the results presented and analysed.

The fifth chapter covers the work carried out at the Shell QUEST facility in Canada.

The Shell QUEST pipeline is modelled and historical input data from the pipeline is

used to develop the simulation. The simulation and the pipeline outlet data are

compared and a statistical analysis is carried out to measure the goodness of fit of

the model. This chapter allows the model to be validated and identify whether the

model produces any unrealistic results.

4
The sixth and final chapter is the conclusion to the research. This covers all the

work in the other chapters and possible future work that can be carried out on CO 2

transportation. This chapter also investigates how the process simultaion tool could

be updated to give greater accuracy in predicting dynamic CO2 flows.

5
Chapter 2 – Literature Review

6
2.1 Introduction

Carbon capture and storage has been reported to be a technology that will play a

dominant role as a greenhouse gas reduction solution for as long as fossil fuels are

used [1]. A major part of the technology is the transportation of carbon dioxide from

a point source such as a fossil fuelled power station with CO2 capture to the site of

sequestration. This chapter will critically review the current literature on the different

methods for CO2 transportation, the different phases of CO2, how CO2 behaves in

pipelines and the current literature on the technical and economic modelling of CO2

pipelines for the purpose of CCS. This chapter will also inform on the novel

contribution of the work that was carried out and that is presented in this thesis.

2.2 Green House Gasses and Global Warming

Carbon Dioxide is known to be a greenhouse gas as its increased concentration in

the atmosphere has been found to be linked to an increase in radiative forcing [3],

this is the difference between the energy absorbed by the earth and the energy

reflected back out of the atmosphere. The increase of the earth’s temperature by

only a few degrees could have significant impacts on the planet through climate

change. It has been found that since the industrial revolution there has been a

significant increase in the atmospheric CO2 concentration due to humans burning

fossil fuels, this is known as anthropogenic CO2 [4]. The United Nations Climate

Conference in Paris that was held in 2015 concluded with 195 countries agreeing

to a plan to reduce CO2 and other greenhouse gas emissions, in order to limit the

global temperature rise by 2oC [5].

7
Within the UK, energy supply has been the biggest contributor to greenhouse gas

emissions where in 2015 29% of CO2 emissions could be linked to the energy sector

[6]. This therefore makes the energy sector one of the main areas in which

greenhouse gas reduction has been targeted. Research carried out by the Energy

Technologies Institute has found that Carbon Capture and Storage and Bioenergy

are the two main technologies required to reduce the UK’s CO2 emissions from the

energy sector [7]. Without either one of these technologies the cost of reaching the

2050 CO2 reduction targets could increase by 1% GDP [8] which is equivalent to an

extra £1000 on annual average household energy bills for energy and transport [2].

This shows that CCS is a crucial technology to ensure an affordable, reliable and

low carbon energy system.

2.3 Carbon Capture and Storage

2.3.1 What is Carbon Capture and Storage?

Carbon Capture and Storage (CCS) is a technology that can remove and sequester

up to 90% of the carbon dioxide emissions from large point sources, such as fossil

fuelled power plants [9]. The CCS technology consists of three distinct processes.

Firstly, the CO2 is captured from the flue gasses that are produced during the

burning of fossil fuels. The high purity CO2 is then transported either via ship or

pipeline to the storage location. Finally the CO2 is then stored deep underground

where it is sequestered [10]. The idea behind CCS technology is to prevent the CO2

from the burning of fossil fuels from entering the atmosphere so that it does not

contribute to global warming.

8
2.3.1.1 CO2 Capture and Compression

The capture of CO2 from large point sources is the first part of the CCS process.

There are three main processes that are the main methods of CO2 capture are,

post-combustion capture, pre-combustion capture and oxyfuel combustion [11].

Post-combustion capture is a type of technology where the CO2 is separated from

the flue gasses after the fuel has been combusted. The most mature and

economically viable post combustion capture technology is amine-based chemical

absorption [12] which has a technology readiness level (TRL) of 9 [13] which

indicates it is a mature technology. Other capture methods such as adsorption and

membrane technologies have a lower TRL of around 6 [14] which indicates there is

still development required to allow these technologies to be ready for industry.

Amine based chemical absorption is the only one of the three that can be retrofitted

to existing power plants [15]. Pre-combustion capture technology works by the

production of syngas from the fuel and capturing the CO2 from the syngas. After the

CO2 is captured the syngas contains a high concentration of hydrogen which can

be used to produce electricity [16]. The most common pre-combustion capture

technologies are physical and chemical absorption [16]. The third technology is

oxyfuel combustion which is where the fuel is burned in oxygen instead of air, this

produces a flue gas consisting mainly of CO2 and H2O [17]. This allows for a less

energy intensive separation process compared to the other two technologies.

After the capture process there is the purification of the CO2 which is followed by

the compression of the pure CO2. Compression of the carbon dioxide is necessary

to enable it to be easily transported and stored in underground geological storage

9
sites [18]. There are different strategies for CO2 compression with some of the

major ones listed below [19];

• Intercooling compression

• Intercooling compression with subcritical liquefaction and pumping

• Intercooling compression with supercritical liquefaction and pumping

• Shockwave compression

The literature concerning CO2 compression has reiterated how there has been

extensive research and focus on the carbon capture process, as well as injection

and monitoring in geological storage sites [18] [19]. This indicates that there has

been less focus on CO2 compression and CO2 transportation and therefore there

are greater opportunities to further the knowledge in these areas.

2.3.2 Global CCS Projects

There are currently 22 large-scale CCS projects around the world that are either in

operation or under construction, with a combined CO2 capture capacity of 40 million

tonnes per annum [20]. However it should be noted that of the currently operating

projects, only two of them are within the energy sector. These are the Boundary

Dam project in Saskatchewan, Canada and the Petra Nova project in the United

States. The Boundary Dam CCS project began operation in 2014, it is a 115MW

power station with a CO2 capture plant, that is capable of capturing 1.3 million

tonnes of CO2 per annum which is a CO2 reduction of 90% [21]. The Petra Nova

project became operational in January 2017. It is a 240MW power plant that will

capture 1.4 million tonnes of CO2 per annum [22]. There are currently no CCS

10
project in operation in the UK. In 2015 the UK government cancelled a £1 billion

CCS commercialisation project and until 2017 there have been no new proposals

for any large scale CCS projects.

2.4 How CO2 is Transported

2.4.1 Methods of Transporting CO2

The method of transport for the captured CO2 is an important choice when

developing carbon capture and storage. There are different factors that need to be

considered when deciding on the optimal method of transport. There are two

methods in which the pure CO2 can be transferred to the storage site. The first is

through vessels which are filled at the capture facility and then transported through

a mixture of land vehicles, barges and ships. The second method of CO 2 transport

is via pipeline. There have been numerous studies comparing these two methods

from both a technical and economic standpoint. This has given greater knowledge

on when and where each of the two transport options should be used as well as the

technical aspects relation to each transport method.

2.4.2 Vessel VS Pipeline Transport

2.4.2.1 Technical Comparison

From the literature there has been identified two parameters that effect which

transport method should be used. These are; the volume of CO2 that is to be

transported which is based on an annual value and the distance from the capture

facility to the site of sequestration. It should also be noted that the transport method

chosen will influence the phase which the CO2 is transported in. As can been seen

in Figure 2-1[23], CO2 that is transported by vessel should be transported in the

11
liquid phase. This is because the density is comparably higher in the liquid phase

than in the gas phase and therefore a greater mass of CO2 can be transported per

vessel making it more economical than if it is in the gas phase. As a higher density

of CO2 is desirable when transporting via vessel it should be the case that CO 2 be

transported in the solid phase, given the density is approximately 1500kg/m 3 and

therefore smaller vessels would be required or more CO2 could be transported.

However given the complex loading and unloading procedures this method

becomes uneconomical [24]. It has been found that the for economical large scale

transport via shipping, the CO2 should be in a phase close to the triple point of CO2,

approximately 6.5 bar and -52oC [24]. Transporting CO2 via pipeline can be done

in either the liquid phase or the supercritical phase as can be seen in Figure 2-1

[23].

12
 Figure 2-1: CO2 Phase Diagram

In pipeline transport pressure losses need to be taken into account as this will affect

the pumping requirements. Therefore the phase in which CO2 is transported should

be one that reduces pressure losses but also has a relatively high density.

2.4.2.2 Cost Comparison

The deciding factor that dictates what transport option will be used is based around

economics. There have been numerous studies comparing the costs of CO 2

transport via pipeline to that of shipping. A study conducted by the Zero Emissions

Platform (ZEP) has compared the costs of different transportation methods. Table

13
2-1 [25] shows estimated costs of onshore pipelines, off shore pipelines and

shipping for 2.5 Mtpa of CO2 at varying distances.

Table 2-1: Cost estimates (€/t CO2) for commercial natural gas-fired power plants with CCS or coal-based
CCS demonstration projects with transported volume of 2.5 Mtpa

As can be seen in Table 2-1, for distances equal or greater than 500km shipping

becomes more economical when compared to pipeline transport. The further the

distance for transportation of CO2 to the storage site, the more economical shipping

becomes. The study carried out by ZEP also looked at the costs when transporting

20Mtpa of CO2 as a comparison to understand how the mass of CO 2 to be

transported effects the method which is the most economical. This is shown in Table

2-2.

14
 Table 2-2: Cost estimate for large scale networks of 20 Mtpa (€/tonne CO2)

It can be seen that for large masses of CO2 pipeline transport is the most

economical and shipping only becomes the preferable choice when transport

distances are equal or greater than 1500km. What can also be seen in Tables 2-1

and 2-2 is that distance has a greater impact on pipeline transport than for shipping.

Other studies have also made similar conclusions in which CO 2 transport via ship

is preferable when transporting relatively low volumes across long distances and

transport via pipeline is more economical when transporting large volumes over

shorter distances [25]. The reasons behind this difference in costs is because

developing new pipeline infrastructure comes with high capital costs which can be

as much as 90% of the cost of transport whereas capital costs for shipping are

approximately 50% of the cost of transportation [26].

Other studies have suggested a strategy for the development of CCS in which

transportation will initially be carried out by shipping when the volumes of CO2 being

transported are relatively low. When CCS uptake increase then there will be a

switch to pipelines due to the greater volumes of CO2 [25]. This strategy however is

15
not one that has been proposed in many CCS projects. Both CCS projects that were

proposed in the UK for the CCS competition used onshore and offshore pipelines

for CCS transport. The Peterhead project that was proposed by Shell suggested

that current pipeline infrastructure that has until recently been used for natural gas

could be reused for CO2 transport [27]. The theory behind developing single

pipelines for single source to storage usage is that the pipelines would be oversized.

This would then promote the further development of CCS in the area which would

lead to the creation of CCS ‘hubs’ as the oversized pipeline can be used by multiple

users and hence reduce the cost of transportation.

2.5 CO2 Pipeline Modelling

From the analysis of the literature comparing CO2 transport by vessel to that of

pipeline it can be clearly seen that pipeline transport is the preferred choice for CCS

both in the short term, shown by the current projects in operation and the long term,

shown by the economic benefits of pipeline transport when large volumes of CO2

are transported. Given this evaluation it is therefore necessary to have a

comprehensive understanding of CO2 pipeline transport for the purposes of CCS.

This is crucial knowledge in developing safe and economic CO2 pipelines which can

also be operated efficiently. The analysis of the current literature on CO2 pipelines

can be separated into economic analysis and technical analysis.

2.5.1 Economic Modelling of CO2 Pipelines

Economic modelling of CO2 pipelines is necessary to understand costs and how

these costs compare with the rest of the CCS process. Economic models can also

inform investment decisions and the economic viability of a project. The cost of a

16
CO2 pipeline has been found to be highly dependent on the design capacity of the

pipeline and the pipe length. This is shown in one study in which the model

simulation results showed an increase in costs from US $1.16/tonne to

US$2.23/tonne of CO2 when the design capacity was reduced from 5 million tonnes

per year to 2 million tonnes per year [28]. The same study also showed that an

increase in the length of the pipeline from 100km to 200 km increases costs further

to US$ 4.06/tonne of CO2. A more detailed model developed by the IEA GHG found

similar results which show that an increase in the diameter of the pipeline and hence

an increase in the capacity, increases the overall capital cost of the pipeline but will

decrease the cost of the pipeline per tonne CO2 [29]. This study by the IEA GHG

goes into further detail to look at the costs related to pumping stations. It found that

for onshore pipelines it is more economic to use booster stations to maintain the

pressure than to operate large low pressure drop pipelines. However when

considering offshore pipelines it was found to be more economic to pressurise the

CO2 at the shore to the point where the pressure in the pipeline will not fall below

90 barg and then only pressurise again at the sequestration site.

There is limited cost data on CO2 pipelines even though there are over 6500km of

CO2 pipelines in the united states alone [30] many of the pipeline developers keep

such information out of the public domain. The models that predict pipeline costs,

generally use diameter calculations which can be applied to various pipelines

universally however developing a single cost equation to accurately evaluate

pipeline costs for all locations and regulatory structures around the world is not

practical [31].

17
This makes it difficult to validate the cost models that have been developed.

However it has been argued that the available capital cost data for natural gas

pipelines is valid for construction costs which is generally independent of the fluid

being transported [31]. Using this data, simple regression analysis can be carried

out to develop equations that can be used for CO 2 pipelines. Equation 2-1 was

developed by the IEAGHG [32] for the onshore pipeline costs with the addition of

the location factor FL. The location factor is used to account for different economic

locations while the terrain factor FT allows for the equation to take into account cost

inflation caused by complex terrains such as mountainous or populated areas[33].

The value of the location factor is assumed to be 1.00 for the USA, Canada and

Europe while the value of 1.2 has been used for the UK. The terrain factor varies

for different land types for instance a value of 1.00 is used for grassland, 1.05 for

woodland and 1.10 for cultivated land[31]. The constants in equation 2-1 were

developed through the use of a pipeline cost model[34];

𝑃𝑖𝑝𝑒𝑙𝑖𝑛𝑒 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝐶𝑜𝑠𝑡 𝑜𝑛𝑠ℎ𝑜𝑟𝑒 = 𝐹𝐿 × 𝐹𝑇 × 106

× [(0.057 × 𝐿𝑜𝑛𝑠ℎ𝑜𝑟𝑒 + 1.8663) + (0.00129 × 𝐿𝑜𝑛𝑠ℎ𝑜𝑟𝑒 ) × 𝐷𝑜

+ (0.00486 × 𝐿𝑜𝑛𝑠ℎ𝑜𝑟𝑒 − 0.000007) × 𝐷𝑜 2 ]

Equation 2-1: Capital Cost Equation for Onshore Pipelines

The IEAGHG also developed a formula for offshore pipeline costs;

18
 𝑃𝑖𝑝𝑒𝑙𝑖𝑛𝑒 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝐶𝑜𝑠𝑡 𝑜𝑓𝑓𝑠ℎ𝑜𝑟𝑒 = 𝐹𝐿 × 106

× [(0.4048 × 𝐿𝑜𝑓𝑓𝑠ℎ𝑜𝑟𝑒 + 4.6946)

− (0.00153 × 𝐿𝑜𝑓𝑓𝑠ℎ𝑜𝑟𝑒 + 0.0113) × 𝐷𝑜

+ (0.000511 × 𝐿𝑜𝑛𝑠ℎ𝑜𝑟𝑒 + 0.00024) × 𝐷𝑜 2 ]

Equation 2-2: Capital Cost Equation for Offshore Pipelines

These formula for pipeline costs have been said to be only useful for estimates for

pipeline costs and should only be used in feasibility and possibly FEED studies [31].

2.5.2 Technical Modelling of CO2 Pipelines

A greater understanding of the technical parameters of a CO2 pipeline is necessary

to ensure the safe operation of such a system. It has already been mentioned that

there are thousands of kilometres of CO2 pipelines in the USA however given the

confidential information with regards to these pipelines their technical performance

is not publicly available. Therefore there has been a development of CO 2 pipeline

models to understand the performance of potential pipeline systems.

2.5.2.1 Pipeline Depressurisation Modelling

Health and safety is of extreme importance when considering the development of a

CO2 pipeline system. An incident related to a pipeline in which there is loss of

containment is extremely dangerous given that it could impact on the public, as CO2

has a greater density than air, a release of CO2 could cause a blanket effect with

the potential to cause asphyxiation. Therefore, modelling to assess the

consequences of a pipeline rupture will allow for improved safe operation and

design of the pipeline. The modelling of the decompression and the discharge rate

19
of the CO2 helps determine the minimum safe distance to populated areas,

emergency response planning and optimum spacing of emergency shut down

valves [35]. Some of the models developed to investigate the release of CO 2 have

also been validated through experimental results [35] [36]. The validation of such

models allows for greater accuracy of the outputs. One consideration that these

models do not take into account is the presence of impurities in the CO2. It has been

identified that there is insufficient knowledge to enable the correct predictions for

depressurisation of CO2 with impurities [37] however it has been argued that the

inclusion of impurities is very important [38]. It has been found that concentrations

of N2 greater than 1 mole% have significant changes on the depressurisation

thermodynamic trajectory [39].

2.5.2.2 CO2 Phase Modelling

As previously mentioned, the phase of CO2 when it is transported is important.

There is agreement in the literature that CO2 will be transported in either the

subcooled liquid or the supercritical phase [40] given that transporting in the gas

phase is less economic due to the large energy requirements for moving the fluid.

There has been significant research into whether the liquid phase or the

supercritical phase is the optimal for transportation of CO 2. One such study used

ASPEN PLUS 10.1 to simulate and compare the efficiency and costs of transport of

CO2 in the supercritical phase to the liquid phase [41]. It was found that CO2 in the

subcooled liquid phase is ideal for colder climates where energy savings of up to

9% are feasible. For warmer climates refrigeration units may be economical to

ensure the temperature stays below the critical temperature. It is not only economic

reasons that transporting in the subcooled liquid phase have been found to be

20
beneficial over the supercritical phase. One study which compared transporting CO2

in the liquid phase to the supercritical phase used Aspen to model the safe distance

before choked flow occurs, it found that the lower the inlet temperature to the

pipeline the longer the distance before choking occurs and therefore a longer

distance before recompression of the CO2 is required. This indicates that the

subcooled liquid phase provides greater safety in that choked flow occurs at longer

distances than in the supercritical phase [42]. While these modelling and simulation

studies have indicated that transporting CO2 in the subcooled liquid phase has both

economic and safety benefits, much of the research in the modelling of CO 2

pipelines has done so in the supercritical phase. This may be due to CO2 being less

likely to form bubbles in the supercritical phase, which minimises cavitation

problems occurring in components such as booster stations and pumps [43]. There

is also the benefit that if the pressure were to fall to the point where phase change

were to occur in the pipeline, the change from supercritical phase to the gas phase

is less extreme in that there is a smaller density change than from the liquid phase

to the gas phase. This is because the supercritical phase is an intermediate phase

between liquid and a gas.

2.5.2.3 Steady State Transport Modelling

Technical models of CO2 can be separated into steady state and dynamic. Steady

state models assess the technical parameters of a CO2 pipeline with a steady flow

of CO2. This would be the case when a power plant is operating at base load, where

the power output from the plant is constant and therefore the CO 2 output from the

plant would also be constant. Modelling of steady state CO 2 pipelines vary

depending on the modelling tool used and the assumptions made around the design

21
and operation of the pipeline. Research has been carried out that looked at how the

pressure, temperature, density and flow velocity changed along the length of a

150km pipeline for 5 different scenarios [44], these were simulated in the modelling

tool MATLAB/Simulink where algorithms were developed for the simulation of the

CCS system described, the scenarios looked at;

1. Transport and injection; maintaining minimum pressure at the end of the

pipeline above 85 bar;

2. Transport and injection; ensuring the pressure at the bottom of the well is

less than the maximum pressure;

3. Transport, pre-conditioning and injection; avoiding two-phase flow in pipeline

and well;

4. Transport and injection, alternative conditioning measures;

5. Influence of impurities on transport and storage.

The results from scenario 1 show that the inlet pressure to the pipeline must be a

minimum of 98 bar to ensure an exit pressure of 85 bar. The CO 2 temperature

decreased along the length of the pipeline and as a result the density increased

which meant that the flow velocity also decreased. These results were observed in

all scenarios with slight variation of the inlet pressure and the profile of the observed

variables along the length of the pipeline. It was found that for near pure CO 2 with

impurities of O2 and N2 the density of the fluid is lower, where at 40oC the density is

approximately 60% of pure CO2 [45]. This therefore results in lower storage

potential and higher pressure drops.

22
2.5.2.4 Dynamic Transport Modelling

Carbon capture and storage can also be applied to combined cycle gas turbine

plants (CCGT’s). These plants are mainly used as load following power plants.

Therefore the output from these plants varies over time. These types of power

plants are used for this application as they have relatively fast response times and

therefore can provide power when the demand increases. As the load factor

changes the emissions will also change, this means that the flow of CO 2 entering

the pipeline will vary. It is therefore necessary to understand how the variability in

flowrate will affect the operation within the pipeline and whether there are

consequences which may require extra measures of control. Dynamic models are

used to simulate how certain variables change over time. In the case of CO 2

pipelines, dynamic models simulate the change in the inlet flow of CO 2. Simulations

of dynamic CO2 pipelines has been carried out using the software tool Modelica

within the Dymola environment [46]. The researchers modelled a 30km pipeline

transporting supercritical CO2 to a 1.2km depth well. Simulations were developed

to study the effect of load change in which the mass flow was varied to 90%, 15%,

105% and 50% of the nominal reference load of 100kg/s. The study concluded that

there was considerable risk of the occurrence of two phase flow in the well and that

preventative measures will be required to help avoid two phase flow. This study

considered exclusively supercritical CO2 therefore the results from sub-cooled liquid

CO2 could be drastically different, however such a study has not been found within

the literature. This study also contained a flaw in that it did not allow the flowrate at

the end of the pipeline to reach the setpoint before another change to the setpoint

23
was made, this means that a complete understanding of the impacts of making a

change in the flowrate was not established.

2.5.2.5 Multiple Sources of CO2

As previously mentioned, the long term development of CO2 pipeline transport is

expected to form around CCS hubs. These hubs are expected to develop to allow

the sharing of infrastructure between different sources of CO2. By doing this the cost

of CCS can be reduced. The idea is that there will be a main trunk pipeline that is

initially oversized and transports the CO2 to the sequestration site. The producers

of the CO2 will each have smaller pipelines which connect them to the trunk pipeline.

However the dynamics of such a system may be significantly different from a single

source pipeline as the flowrate from one capture site may be constant while the

other may vary. It will be necessary to understand how to operate the trunk pipeline

where both these flows merge. One such study has considered steady state

modelling of multiple sources of CO2 [47]. The study was based on real proposals

of two CCS projects in the Humber region. The first project is known as the Don

Valley Power Project, which is a 650 MW Integrated Gasification Combined Cycle

(IGCC) power plant [48]. The second proposal is known as the White Rose project,

which is a 426 MW oxyfuel power plant [49]. This study used Aspen HYSYS to

model a system with a collecting pipeline for each of the CO 2 sources which join

onto an onshore trunk pipeline, which leads to an offshore pipeline. One of the main

technical findings from the study was that as the higher the velocity of the CO 2 the

higher the pressure drop along the pipeline and hence greater boosting pressure at

the pump station before the off shore pipeline. The fluid within the pipeline is in the

subcooled liquid phase. The study doesn’t consider transportation within the

24
supercritical phase and doesn’t consider dynamic flows of CO 2, both of which are

possible scenarios in the development of CCS hubs.

2.6 Modelling Tools Evaluation

2.6.1 Model Requirements

2.6.1.1 Model Inputs and Outputs

Before the analysis of the different software packages, it is necessary to make clear

what is needed from the model. The tool should have the option to vary parameters

with time. For this research the ability to vary the flowrate of CO2 to the inlet of a

pipeline is of crucial importance and is the basis of the work that is presented. For

the design of the pipeline it is necessary to be able to define the pipeline geometry,

heat transfer properties and the elevation. The output data from the simulations

should show the fluid temperature, pressure, density and flowrate at any point within

the pipeline. This allows for the phase of the fluid in the pipeline to be determined

and hence whether any phase boundaries are crossed within the pipeline.

2.6.1.2 Model Calculations

Since the research is investigating CO2 pipelines for CCS purposes it is important

that the fluid entering the pipeline is equivalent to that which is expected from a CCS

capture plant. There has been some experimental work carried out, which has

considered the effects of impurities on the physical properties of CO2. The impurities

that are of most concern for CO2 transport and storage purposes are dependent on

the technology used in the CO2 capture process. For pre-combustion carbon

capture the main impurities are nitrogen and hydrogen, due to the fact that pre

25
combustion is used for coal or biomass gasification which produces a stream

compromising of mainly CO2 and H2. Table 2-3 [50] shows the main components

within a stream of gas captured using different pre-combustion technologies.

Table 2-3: Pre-combustion CO2 impurities from pulverised coal

For post combustion carbon capture technologies the main impurity is nitrogen,

which is due to the burning of either coal or natural gas with excess air. The

impurities in a stream of gas from a post combustion capture technologies can be

seen in Table 2-4. Table 2-4 shows that for both amine PC plant and an MEA PC

plant the amount of Nitrogen within the CO2 stream is between 0.045 and 0.18 %

v/v.

26
 Table 2-4: Post-combustion CO2 impurities from pulverised coal

In oxyfuel combustion technology the main impurities are oxygen, nitrogen and

argon. The presence of oxygen is due to the burning of the fuel in pure oxygen. The

levels of these impurities can be seen in Table 2-5.

Table 2-5: Oxyfuel combustion CO2 impurities

27
For the modelling of CO2 streams it is important for the modelling tools underlying

equations of state to be able to determine the physical properties of the following

mixtures; CO2+N2, CO2+O2 and CO2+H2. This is necessary, to ensure that

modelling of CO2 streams from the different capture technologies is possible with a

high degree of accuracy, as it is known that the presence of these impurities has

significant impact upon the physical properties of a CO2 rich stream. Experimental

studies to determine the physical properties of a CO2 rich stream containing O2, Ar

and N2 have shown that there is an increase in pressure of 3000-5000 kPa for the

single liquid phase region. This is accompanied by an increase in density of as much

as 35% at the same temperature and pressure of pure CO2 [50].

Other impurities which are found in CO2 from different capture technologies and are

in lower quantities include SOx, NOx and H2O which can be seen in tables 2-3, 2-4

and 2-5, these impurities are present in quantities of PPM. Some studies have

suggested that because these impurities are significantly low especially when the

water content is below the solubility limit for pure CO2 the corrosion rates are likely

to be sufficiently low[51, 52].

2.6.2 OLGA

2.6.2.1 Software Background

OLGA is a tool that is traditionally used to simulate the transport of oil, water and

gas as well as mixtures of these components. [53]. It is used for simulating

multiphase systems and has the ability to model a pipeline system from reservoir

pore to process facility [54]. OLGA has been used extensively in the petrochemical

industry by companies such as British Petroleum. One case study using OLGA

resulted in BP saving 12 days of downtime and 6000m3 of diesel, this was done by

28
conducting a scenario in OLGA to identify alternative restarting methods for the

pipeline that uses less diesel. [55]. This shows the impact that simulation tools can

have on the decisions made by large organisations and the economic benefits that

modelling can provide.

2.6.2.2 Equations of State

For CO2 transport OLGA uses the Span and Wagner equations of state to determine

the thermodynamic properties of CO2 at different temperatures and pressures [56].

The Span and Wagner equation of state has been developed for pure CO2 and has

the ability to determine thermodynamic properties up to 30MPa and 523K. This

covers the range in which normal transportation of CO2 occurs so is therefore valid

to use for CO2 pipeline modelling. Within this temperature and pressure region the

uncertainty of the equation ranges from ±0.03% to ±0.05% for the density, ±0.03%

to 1% in the speed of sound and ±0.15% to 1.5% in the isobaric heat capacity.

The Span and Wagner equation of state is expressed as the dimensionless

Helmholtz free energy as shown by equation 2-3. If an expression for the Helmholtz

free energy and its derivatives are known then all other thermodynamic properties

can be derived from the expression [57].

𝜙(𝜏, 𝛿) = 𝜙 0 (𝛿, 𝜏) + 𝜙 𝑟 (𝛿, 𝜏)

Equation 2-3: Dimensionless Helmholtz Energy

The Helmholtz function 𝜙 = 𝐴/(𝑅𝑇) is split into an ideal gas part φ0, and a residual

part φr.

The Span and Wagner equation of state expresses the ideal part of the equation

as;

29
 8
0
𝜙 0 (𝜏,
𝛿) = ln(𝛿) + 𝑎10 + 𝑎20 𝜏 + 𝑎30 ln(𝜏) + ∑ 𝑎𝑖0 ln [1 − 𝑒 (−𝜏𝜃𝑖 ) ]
𝑖=4

Equation 2-4: Helmholtz Energy Ideal Gas Property

The residual part is expressed in equation 2-5 and represents the compressibility of

the fluid[58];

7 34 39
−𝛿 𝐶𝑖 2 −𝛽 (𝜏−𝛾 )2 )
𝜙 𝑟 (𝜏, 𝛿) = ∑ 𝑛𝑖 𝛿 𝑑𝑖 𝜏 𝑡𝑖 + ∑ 𝑛𝑖 𝛿 𝑑𝑖 𝜏 𝑡𝑖 𝑒 + ∑ 𝑛𝑖 𝛿 𝑑𝑖 𝜏 𝑡𝑖 𝑒 (−𝛼𝑖 (𝛿−𝜀𝑖 ) 𝑖 𝑖

𝑖=1 𝑖=8 𝑖=35

39
2 −𝐷 2
+ ∑ 𝑛𝑖 Δ𝑏𝑖 𝛿𝑒 (−𝐶𝑖 (𝛿−1) 𝑖 (𝜏−1) )

𝑖=35

Equation 2-5: Helmholtz Energy Residual

2.6.2.3 Modelling of CO2 Rich Mixtures

The Span and Wagner equation of state was developed specifically for determining

the thermodynamic properties of pure CO2.There have been found to be no studies

on the use of the Span and Wagner equation of state for CO 2 rich mixtures. Its use

in this way would be inappropriate given that the properties of CO2 rich mixtures are

different from just pure CO2. Because of this, the use of OLGA for these types of

systems would give inaccurate results and hence not be sensible to use for the

modelling of CO2 pipelines.

30
2.6.3 Aspen HYSYS

2.6.3.1 Equations of State

For CO2 pipeline modelling Aspen HYSYS has numerous different packages using

various equations of state. Previous studies have decided on using the cubic

equations of state, Peng Robinson as the method of determining the

thermodynamic properties of CO2 within the pipeline [59]. The P-R equation of state

is given by equation 2-6.

𝑅𝑇 𝑎(𝑇)
𝑃= −
𝑉 − 𝑏 𝑉(𝑉 + 𝑏) + 𝑏(𝑉 − 𝑏)

Equation 2-6: Peng-Robinson Equation of State

For pure components, the terms a and b are expressed using critical properties and

acentric factors;

𝑎(𝑇) = 𝑎𝑐 𝛼

𝑅 2 𝑇𝑐2
𝑎𝑐 = 0.45724
𝑃𝑐

1⁄ 1⁄
𝛼 2 = 1 + 𝜅 (1 − 𝑇𝑟 2 )

𝜅 = 0.32464 + 1.54226𝜔 − 0.26992𝜔2

𝑅𝑇𝑐
𝑏 = 0.07780
𝑃𝑐

For mixtures of components the terms a and b are expressed as follows

1⁄ 1⁄
𝑎 = ∑ ∑ 𝑥𝑖 𝑥𝑗 (1 − 𝛿𝑖𝑗 ) 𝑎𝑖 2 𝑎𝑗 2

𝑖 𝑗

31
 𝑏 = ∑ 𝑥𝑖 𝑏𝑖
𝑖

The term δij is the binary interaction coefficient and is determined experimentally.

2.6.3.2 Pure CO2

A study investigated the accuracy of different models at predicting the vapour-liquid

equilibrium of CO2 found that the Peng-Robinson Equation of state was capable of

predicting the VLE data accurately. However this study also shows that there is

deviation of the model when the temperature approaches the critical pressure.

A study simulated a scenario in which pure CO2 within a pipeline enters the gas-

liquid two phase region in which liquid hold-up occurs [60]. It was found that the

Peng-Robinson equation of state diverged from the actual values of pressure,

temperature and hold-up at the outlet of the pipeline, indicating that for pure CO 2

there are situations in which Peng-Robinson cannot be used. It has been explicitly

written that the poor agreement of the Peng Robinson equation of state with the

density measurements of pure CO2 near the critical pressure is unacceptable and

that variants of the Peng-Robinson model also suffer the same limitations. The

explanation behind this is due to the model being developed for separation of

mixtures of natural gas where CO2 is a minor additive [61].

2.6.3.3 CO2-rich Mixtures

Studies have been carried out to determine the accuracy of the Peng-Robinson

equation of state for some of these mixtures. A study that investigated the ability of

the Peng-Robinson equation of state to predict the phase behaviour of both a CO 2-

N2 and a CO2-H2 mixture found that there is good agreement between the model

with the experimental data. However this study also found that the model failed

32
marginally in the critical region [62]. This finding is crucial in determining which

software is most appropriate to use for the research, as the transportation of CO 2

can occur in the supercritical phase so there are possibilities that the CO 2 could

approach the critical region within the pipeline.

2.6.4 gCCS

2.6.4.1 Software Background

gCCS is the newest of the three modelling tools to be evaluated for the use of the

research. It was developed by Process Systems Enterprise for the specific purpose

of modelling a full chain CCS system, from power plant all the way through to

sequestration. It is based upon the gPROMS platform and uses the gSAFT tool to

calculate the properties of CO2. gSAFT uses statistical associating fluid theory

(SAFT) as the method to determine the thermodynamic and the phase equilibrium

properties of fluids.

2.6.4.2 Equations of State

SAFT is based upon Wertheim’s theory of Helmholtz energy expansion. It is known

as thermodynamic perturbation theory. In this method the Helmholtz free energy is

calculated from graphical summation of interactions of different species [63]. The

SAFT equation of state can also be developed for determining the properties for

mixtures of fluids, through a simple extension of the model. Unlike cubic equations

of state e.g. Peng-Robinson, SAFT does not require experimental data to produce

good approximations of fluid properties of mixtures.

2.6.4.3 Pure CO2

Previous research on the ability of SAFT equations of state to determine CO 2

properties has been carried out. One such study looked at the ability of SAFT EoS

33
to determine the vapour liquid equilibrium data and the second derivative

thermodynamic properties of various components related to carbon capture and

storage. These components included CO2, H2S, N2, H2O, O2 and CH4. The

properties that were calculated within this study include the isobaric and isochoric

heat capacity, speed of sound, Joule-Thomson coefficient and isothermal

compressibility. It was found that the model is able to calculate the vapour pressure

and the liquid density of pure CO2 with good accuracy in both the sub-critical and

the super critical regions, however similar to the Peng-Robinson the accuracy of the

model decreases closer to the critical region. This again could pose some problems

when modelling CCS processes as there is possibility that the fluid may approach

the critical region [64].

2.6.4.4 CO2 Mixtures

There is currently limited literature on the use of SAFT equations of state for

predicting the fluid properties of CO2 rich mixtures that include the components that

are expected in CCS processes. Research carried out on using SAFT equations to

predict the phase equilibrium of CO2-H2O mixtures concluded that there was

satisfactory agreement between the calculated and the experimental values [65].

There was found to be no literature on the ability of SAFT equations to accurately

determine the fluid properties of mixtures of CO2-H2, CO2-O2 and CO2-N2. This limits

the ability to critically evaluate the gCCS modelling tool for CO2 pipeline transport,

however in comparison to the Span and Wagner and Peng-Robinson equations of

state SAFT is known to have a greater accuracy in modelling fluid mixtures [66],

especially when there is an absence of experimental data which is required for the

Peng-Robinson equation of state.

34
2.7 Conclusion

A review of the literature on the transportation of CO2 for the purpose of CCS has

given an understanding of the areas that still need to be researched or where more

detailed analysis can take place. From the literature it has become apparent that

CO2 pipelines will be the preferred method of transportation given the greater

economic benefit when transporting large volumes of CO2. This is backed up with

the current CCS projects in Canada also using pipelines compared to vessels. The

financial modelling of pipeline transport is well developed and understood. However

given that pipeline costs are highly dependent on the length, diameter and the

operation of the pipeline the models can only give a rough approximation of the

costs which can be used in the preferred study but would need to be modified for

each CCS project to give greater accuracy. Detailed technical information on CO2

pipelines is not publicly available given that there are operational pipelines in the

USA. This may be due to intellectual property issues with the organisations

operating the pipelines not wanting to share the data publicly. This means that the

current modelling of CO2 pipelines has not been validated alongside actual pipeline

data.

An area within CO2 pipeline modelling which has been the main focus of the

research is in the depressurisation of pipelines in the case of pipeline ruptures. This

is understood to be an extremely important area of research given the health and

safety implications of such a hazard occurring. This area of research also includes

experimental data of pipeline ruptures which allows for more detailed models and

hence is believed to be an area which is well understood. Given that CO 2 can be

35
transported in different phases, modelling has been carried out to understand the

optimum phase of the CO2. Given the research there was no definitive answer on

whether the sub cooled liquid phase or the supercritical phase should be used and

this will be decided for each CCS project separately. Unless there is the option of

reusing pipelines which will provide an economic benefit, gas phase transport was

concluded to be undesirable for pipeline transport given the low density.

In the modelling of CO2 transport there are two main types of models that have been

developed. The first are steady state models, these are simpler and represent the

flow of CO2 from base load power plants where there is a constant supply of CO2 to

the pipeline. The results from these models allow understanding of pressure profiles

along the pipeline and whether phase change occurs for specific inlet conditions.

There has also been the development of steady state models which also take into

consideration near pure CO2 with N2 and O2 as impurities.

Simulations studying dynamic flows of CO2 have been used to determine whether

two phase flow will occur when there is a change in flowrate at the inlet of the

pipeline. This allows the simulation of a pipeline which transports CO 2 from a load

following power plant such as a CCGT. These models however do not show how

the flowrate changes throughout the pipeline, therefore they do not indicate at what

point along the pipeline phase change occurs. The models developed for dynamic

flows also do not compare the difference between transporting in the sub cooled

liquid phase and the supercritical phase. There has also been no research on the

effects of impurities within the CO2 on the dynamics. These areas show that there

is empty space within the research that require to be filled and are areas of novelty.

36
Models simulating multiple sources of CO2 have been developed, these models

have given technical insight into steady state flows of CO2. While these models have

been comprehensive by modelling real proposed projects and taking into account

impurities within the CO2. There is still novel research to be carried out looking at

the effects of dynamic flows from multiple sources of CO2 on the fluid within the

pipeline.

Through the literature review it has been possible to understand what research has

been carried out and what research is still needed, to enable a more comprehensive

understanding of CO2 transport within pipelines. There are key areas of novel

research particularly looking at dynamic flows of both sub cooled and supercritical

CO2. These are the areas in which the research has expanded and developed on.

To carry out the research an analysis of three different modelling tools was carried

out OLGA, Aspen HYSYS and gCCS. The approach taken to evaluate the most

appropriate tool was to assess the equations of state each one uses to determine

the properties of CO2 and of CO2 rich fluids. For pure CO2 the Span and Wagner

equation of state used in OLGA is deemed to be the most accurate, as it was

specifically developed for CO2. Peng-Robinson which is used in Aspen HYSYS was

seen to be the least accurate, especially around the critical region and was

considered by some researchers to be not good enough to model CO 2 for CCS

purposes. The SAFT equation of state was said to have good agreement with

experimental data for CO2 in the sub critical and supercritical region with deviation

occurring closer to the critical point. It was concluded however that it was still

appropriate to use for CO2 modelling.

37
The modelling of CO2 mixtures containing H2, N2 and O2 is important when looking

at CO2 from different capture technologies, as even small amounts of these

impurities can affect the properties of the fluid. The Span and Wagner equation of

state was not developed for CO2 rich fluids and therefore is not capable of modelling

these types of systems. Peng-Robinson uses binary interaction coefficients for fluid

mixtures, which rely on the availability of experimental data. The literature found

that for CO2-H2 and CO2-N2 systems there was good agreement with the

experimental data except around the critical region as with pure CO2. For the SAFT

equation of state there was found to be very limited literature on the modelling of

CO2 and therefore a cull analysis of the model was not possible.

Through the analysis carried out, it was decided that given the specific application

of gCCS it was the most appropriate tool of choice for the modelling of dynamic

flows in pipelines of CO2 and CO2 rich fluids and allowed for another layer of novelty

of the research, in the use SAFT equations of state for CO 2 pipeline modelling. It

was preferred over OLGA because the Span and Wagner equation of state is not

applicable for CO2 rich systems, while it was the greater accuracy compared to the

Peng-Robinson equation of state that was the reason it was chosen over Aspen

HYSYS.

38
Chapter 3 – Comparing Variable Flows For Liquid and
Supercritical Phase Carbon Dioxide

39
3.1 Introduction

Transporting carbon dioxide via pipeline is a part of the process that enables the

sequestration of CO2 from large point sources, including power stations. This can

include both coal fired power stations as well as natural gas combined cycle power

(NGCC) stations. Within the UK NGCCs are operated as load following power

plants. This means that an NGCC with CCS will produce variable flowrates of CO2.

The purpose of this chapter is to report the effects of reducing the inlet flowrate of

CO2 into the pipeline and comparing the outputs when CO 2 is transported in the

liquid phase and the supercritical phase. Understanding the effects of changing

flowrates within a CO2 pipeline could help indicate which phase of CO2 would be

the most beneficial and inform the development of transporting CO 2 in the most

efficient way.

3.2 Hypothesis

For this chapter the following statement presents the hypothesis that will either be

proved or disproved through the modelling and simulations that will be carried out.

“The gCCS model will show that the rate of change in the outlet flowrate of a CO 2

pipeline when the inlet flowrate is reduced, will be greater when the CO 2 is

transported in the subcooled liquid phase compared to the supercritical phase”.

3.3 gCCS Transport Models

A CO2 transport system is constructed within gCCS through the connections of

single models. Each model has specific variables that can either be chosen or

defined; these allow the system to be designed to a certain specification. It is

40
necessary to understand the degrees of freedom within the system to ensure the

simulation will complete. Over specifying or underspecifying a variable will result in

the simulation returning an error message indicating a problem, however it will not

inform the user which variable has been over or under specified.

3.3.1 Carbon Dioxide Source Model

In the case of modelling the transport and storage of CO 2 separately from the rest

of the system, for example without including the capture process itself, a source of

CO2 is specified using the CO2 Source model within gCCS. This model allows for a

user to define the flowrate specifications of CO2 that enters the pipeline. The

manipulated variables for this model can be seen in Figure 3-1.

Figure 3-1: CO2 Source Variables

Within the CO2 source model the thermodynamic properties of the fluid are defined

by specifying the temperature and the pressure if it is not specified anywhere else

41
within the system. For a transport system the CO2 flowrate must be defined here.

The model also allows the composition of the fluid to be specified, giving the choice

of 9 of the most common components that are to be found in captured CO2, for

example impurity gases. There are two ways in which the components can be

specified; firstly through the ‘gCCS standard’ tab which requires a mass fraction to

be entered for each of the 9 components, even if the value is zero. This however

causes the model to run through every calculation for even those which have a

mass fraction of zero and increases the time it takes for the simulation to complete.

The second way in which the components can be specified is through the ‘user

defined’ option which is available in the drop down list of the ‘physical properties’

tab. This method produces a separate dialogue box in which the desired

components of non-zero mass fractions can be specified. Only those components

will require a mass flowrate to be defined. This therefore reduces the time for the

simulation to complete.

3.3.2 Pipeline and Well Models

To build a full CO2 transport system gCCS has within it, a ‘pipeline’ model and a

separate ‘well’ model. The pipeline model represents a pipeline segment whereas

the well model represents the entire well. Figure 3-2 shows the configuration tab for

the pipeline model which is identical to the well model. This allows specification of

the pipeline design. It allows the user to manipulate the pipe length, elevation

changes, internal diameter, pipe thickness and pipe roughness. These variables are

necessary to calculate important parameters for the fluid flowing in the pipeline.

• The pipe roughness is required to obtain the Darcy friction factor which then

allows for the frictional pressure losses to be calculated.

42
 • The internal diameter is required to determine the flow regime i.e. Reynolds

number, which is also used to determine the pressure drop in the pipeline.

• The pipe length and elevation are also used in determining the pressure drop

along the pipe length.

There is also a choice of material for the pipeline, carbon steel and stainless steel

are the options available within the model and are the most common materials for

pipe construction.

Figure 3-2: Pipeline Segment Design Variables

The heat transfer tab for the pipeline model can be seen in Figure 3-3. The factors

that affect heat transfer of the fluid in the pipeline which can be defined include, the

ambient temperature Ta, which effects the soil temperature Ts and hence the

temperature gradient between the fluid and the surrounding material. Further factors

to be taken into consideration are:

43
3.3.2.1 Burial Depth of the Pipeline

The burial depth plays a significant factor in determining the temperature change in

the pipeline. Studies have shown that the effect of the ambient temperature on the

soil temperature changes with burial depth. The deeper the soil the smaller the

effect of the ambient temperature on the soil temperature [67].

3.3.2.2 Material the Pipeline is Buried in

The choices for the surrounding material includes soil, air and water. These choices

are available to simulate buried pipes, above ground pipes and offshore pipelines.

A further choice is available for the type of soil that the pipeline is buried in, this

includes dry sandy soil, soaked sandy soil, dry clay soil or soaked clay soil. The soil

type affects the heat transfer as different soils will have different thermal properties

and therefore affect the soil heat transfer coefficient hs which in turn will affect the

overall heat transfer coefficient U. It has been found that wet soils have a greater

thermal conductivity compared to dry soils and therefore will have a greater heat

transfer coefficient [68]. This will affect the heat losses from the pipeline in two ways,

firstly the effect of the ambient temperature on the ground temperature will be

greater for wet soils and secondly the heat transfer between the pipe wall and the

soil will be greater for wet soils.

44
 Figure 3-3: Pipeline Segment Heat Transfer Variables

The heat transfer tab for the well model is different from that of the pipeline model.

In exchange for a choice of the surrounding pipeline material, the well model instead

allows for a value for the overall heat transfer coefficient to be specified. The heat

transfer tab for the well model is shown in Figure 3-4.

45
 Figure 3-4: Well Heat Transfer Variables

3.3.3 Valve Model

As part of a CO2 pipeline system it is necessary to have Line Block Valves (LBV)

along the pipeline as a safety feature. If a leak occurs anywhere along the pipeline

the valves can be quickly closed to limit the release of CO2 to the atmosphere.

The LBV valve acts as a safety valve and is not optimised for tight control of the

flow. The addition of the valve allows a system to be modelled that is closer to what

is seen in real CO2 pipelines. The valve model takes into account the pressure and

temperature changes that occur through the valves. While the temperature changes

across the valve may be relatively small, depending on the liquid flow coefficient the

pressure drop across the valve can be significant and is an important parameter to

take into account.

Figure 3-5 shows the valve model interface within gCCS. The parameters which

can be changed within the model include the flow coefficient which determines the

46
pressure loss through the valve, the stem position which relates to how far open or

closed the valve is and the leakage fraction.

Figure 3-5: ESD Valve Configuration

3.3.4 Distribution Header Model

The distribution header model is used to simulate the distribution of captured CO 2

among several wells. It allows for a single inlet of CO2 from the pipeline and has

connections to allow for several wells to be attached. The model has inputs for the

header length, the upstream pipe diameter and the rate of heat input. Figure 3-6

shows the configuration tab for the distribution header.

47
 Figure 3-6: Distribution Header Configuration

3.3.5 Reservoir

The reservoir model simulates the storage of CO2 in underground reservoirs and is

modelled as a pressure vessel. There are two methods in which the reservoir

pressure can be specified, either by specifying the pressure directly in which case

the model will maintain a constant reservoir pressure, or the pressure can be

determined via an external file which contains data on how the reservoir pressure

will change depending on the mass of CO2 injected into the reservoir.

48
 Figure 3-7: Reservoir Configuration

A study has been conducted to compare the different effects of changing the inlet

flowrate of a CO2 pipeline, simulating CO2 being in both the liquid and supercritical

phase. Along with comparing the different phases of CO 2, the addition of different

impurities in the CO2 stream has been studied. The need for this research comes

from the lack of work carried out in this area, most importantly how varying flows

into the pipeline system effect the flow within the pipe. This will be of upmost

importance for the cases of load following power plants with CCS. The aim is to gain

greater knowledge in this area to help in the operation of CO2 pipelines and storage

sites.

The study was carried out using a simulation tool known as gCCS which has been

developed by Process Systems Enterprise as a tool to model full chain CCS

systems. For the purpose of this study only the transport, injection and storage

models have been used. The software is based on gSAFT to predict the

thermodynamic properties of the fluids in question. As previously mentioned gSAFT

is a predictive tool and is seen as the most appropriate method in simulating CO 2

49
pipelines due to the lack of experimental data available which is a requirement for

other cubic equations of state such as PengRobinson.

3.4 Carbon Dioxide Pipeline Model Development

The setup of the model can be seen in Figure 3-8. The source of the CO2 is taken

from the compression system which is located at the site of the power plant and the

capture plant. There is an Emergency Shut Down (ESD) valve directly before a

52000 m pipeline and another ESD valve located directly after this pipeline. There

is then another 52000m pipeline and ESD valve which is followed by a distribution

header which is connected to a 1200m well. The vertical well leads to a reservoir

where the CO2 is stored.

The technical specifications for each part of the system can be seen in Tables 3-1

to 3-6. The pipeline setup for the simulations done in both the liquid and supercritical

phase were identical to allow for accurate comparison between the results. The

initial simulation was carried out using pure carbon dioxide to present a base case

in which all others could be compared. It was necessary to repeat the simulations,

each with a longer simulation time than the previous until the system reached a

steady state. The initial time period tested was 100,000 seconds (27.78 hours) with

an increase of 100,000 seconds (27.78 hours) for each subsequent simulation.

Steady state in the pipeline was assumed when the outlet flowrate of the pipeline

matched that of the inlet flowrate.

50
 Figure 3-8: Model Pipeline Topology

The model was set to simulate a case in which there is a linear reduction in flowrate

at the inlet of the pipeline, with a starting flowrate of 100 kg s-1 which was then

programmed to decrease to 50 kg s-1 over a period of 12.5 minutes.

Table 3-1: CO2 Source Parameters

CO2 Source
Property Value Unit
CO2 Mass Fraction 1 -
Temperature 293 K
Ramp Rate 4 Kg s-1min-1

The CO2 source indicates a single point source of CO2. In this case the fluid is pure

carbon dioxide.

51
 Table 3-2: ESD Valve Design Parameters

ESD Valves
Property Value Unit
Leakage Fraction 0.001 -
Liquid Flow Coefficient 100 gpm/psi0.5
Stem Position Setting 1 -

Table 3-3: Pipeline Section Design

Pipe Sections
Property Value Unit
Material Carbon Steel -
Pipe Section Length 52000 m
Elevation Change 0 m
Pipe Internal Diameter 0.6096 m
Pipe Thickness 0.01 m
Pipeline Depth 1 m

Table 3-4: Choke Valve Design

Choke Valve
Property Value Unit
Leakage Fraction 0.001 -
Liquid Flow Coefficient 100 gpm/psi^0.5
Stem Position 1 -
Table 3-5: Well Design

Well
Property Value Unit
Material Carbon Steel
Pipe Section Length 1200 m
Elevation Change 1200 m
Pipe Internal Diameter 0.3 m
Pipe Thickness 0.01 m
Overall Heat Transfer Coefficient 11 W m-2K-1

52
 Table 3-6: Reservoir Parameters

Reservoir
Property Value Unit
Injectivity Index 3.3333x10-5 Kg s-1Pa-1
Specified Reservoir 150 bar
Pressure

The code for the process to produce the change in flowrate can be seen in Appendix

A.1. The code describes a process in which the initial setup is allowed to continue

for 30 seconds. At this point the inlet flowrate from the CO 2 source is reduced. The

change is a ramp down in the mass flowrate at a rate of 4kg s-1 min-1. This means

it will take the system 750 seconds to reach the set point of 50kg s-1. The system is

then allowed to continue for 500,000 seconds (139.89 hours). This value has been

used as through repeating this test it was found that it takes approximately 500,000

seconds (138,89 hours) for the entire system to completely settle and reach an

equilibrium. This simulation time is taken as the base case for this project. A single

ramp down was used to show how the system responds when the simplest changes

are made to the system. It is also necessary to mention that a ramp down in the

flowrate of CO2 is a consequence from a changing output from a power plant,

therefore this case simulates expected conditions for a load following power plant

with CCS.

53
3.5 Simulation Output Analysis

3.5.1 Liquid Phase Transport

The initial simulation was set up for transporting CO2 in the liquid phase, with an

inlet temperature of 20oC. The areas which are of significance in the transport of

CO2 are the flowrate, pressure, temperature and the density profiles. These

variables show how the CO2 is flowing and whether there is any change of phase

along the pipeline.

3.5.1.1 Flowrate

Figure 3-9 shows how the flowrate changes at the inlet and the outlet of each

pipeline over the simulation time period. The inlet flowrate to Pipeline001 is

equivalent to the specified flowrate from the CO2 source and the outlet flowrate of

Pipeline001 is equal to the inlet flowrate to Pipeline002. There are three distinct

phases that occur that can be seen from Figure 3-9, the first phase which is

highlighted in the blue dashed lines, shows that as the inlet flowrate drops to the set

point the outlet flowrate from Pipeline001 falls at a slower rate. Figure 3-10 shows

a close up of the initial drop in CO2 flow and shows more clearly the difference in

time taken for the inlet and outlet flowrates of each pipeline to reach a steady state.

This delayed response can be explained through the physical properties of CO2 as

there is a temperature drop as the flowrate decreases at the outlet of the pipeline

yet the temperature at the inlet remains constant. The large difference in

temperature means that the density at the inlet is lower than the density at the outlet

and hence the flowrates at these two points will be different. The current literature

on CO2 pipelines has not identified this significant time delay when a flowrate

54
change occurs and is hence a novel finding in how a CO2 pipeline reacts to changes

in the inlet flowrate. This initial response to the change in the flowrate will be referred

to as the ‘delayed response phase’.

The second phase shown by the red dashed box in Figure 3-9 shows that the outlet

flowrate reaches an initial steady state that is approximately 4kg/s higher than the

inlet flowrate. This offset lasts for a period of 23,000 seconds (6.39 hours). This will

be referred to as the ‘offset phase’.

After the offset phase the outlet flowrate declines further from 54 kg s-1 to 50 kg s-1.

This same response is shown in Pipeline002 however it takes a further 230,000

seconds (63.89 hours) for the flowrate to fall from 54 kg s-1 to 50 kg s-1. This will be

referred to as the ‘reduction phase’.

To understand what is occurring within the pipeline, Figure 3-11 shows the flowrate

profile along the axial length of Pipeline001. Figure 3-11 shows that a wave like

flowrate profile develops within the pipeline, this wave then travels along the pipe

length. When the wave reaches the end of the pipeline the outlet flowrate then falls

to the set point. The wave then carries on to the second pipe length until it again

reaches the outlet of the pipe. The reason behind this phenomena is unknown and

there are two possibilities behind such outputs; the wave like profile could be a

consequence of the compressibility of the CO2 in which there is a large density

change of the fluid over a small distance of pipe length. However there is also a

possibility that this phenomena is a result of the model itself and is not a real

consequence of changing the CO2 flowrate.

55
 Variable
1 00 Inlet Flowrate
Pipeline001 Outlet Flowrate
Pipeline002 Outlet Flowrate

90
Flowrate (kg/s)

80

70

60

50

0 20 40 60 80 1 00 1 20 1 40
Time (h)

Figure 3-9: Inlet and Outlet Flowrate Profile for Pipeline 1 and 2

56
 Variable
100 Inlet Flowrate
Pipeline001 Outlet Flowrate
Pipeline002 Outlet Flowrate

90
Flowrate (kg/s)

80

70

60

50

0 2 4 6 8 10 12 14
Time (h)

Figure 3-10: Inlet and Outlet Flowrate Profile for Pipeline 1 and 2

Variable
54 Time-50000s
Time-100000s
Time-150000s
Time-200000s
Time-250000s
53
Flowrate (kg/s)

52

51

50

0 10000 20000 30000 40000 50000

Axial Position (m)

Figure 3-11: Flowrate Profile along Pipeline001

57
3.5.1.2 Pressure

Within the model the inlet pressure of the fluid is determined by the reservoir

pressure. At the initial flowrate of 100kg/s the pressure at the inlet to the pipeline is

14187 kPa; this is greater than the critical pressure of CO2 of 7.39MPa. The

pressure at the outlet of pipeline001 is 14121 kPa giving a pressure drop along the

length of the pipeline of 66 kPa. Since there is no elevation of the pipeline the static

pressure losses amount to zero, therefore the pressure drop can be attributed to

the frictional pressure loss alone. Figure 3-12 shows that as the flowrate at the inlet

drops both the inlet and the outlet pressures also fall. There is a difference between

the response of the flowrate and the response of the pressure when a drop in

flowrate occurs. Unlike the outlet flowrate, the outlet pressure drops at a similar rate

as the inlet pressure and no offset is observed like that of the flowrate. An important

observation from Figure 3-10 is that the pressure of the fluid does not fall below the

critical pressure of CO2. This indicates that two phase flow doesn’t occur within the

pipeline.

58
 1 5000 Variable
Inlet Pressure
Outlet Pressure
1 4000

1 3000
Pressure (kPa)

1 2000

1 1 000

1 0000

9000

8000

0 20 40 60 80 1 00 1 20 1 40
Time (h)

Figure 3-12: Inlet and Outlet Pressure for Pipeline 001

3.5.1.3 Temperature

The temperature profiles at the inlet and outlet of Pipeline001 can be seen in Figure

3-13. The pipeline inlet temperature is designed to stay constant and ensures that

CO2 doesn’t move between the liquid and the supercritical phase. The temperature

along the pipeline can vary through heat losses or gains to and from the surrounding

pipeline material. Figure 3-14 shows that the outlet temperature of Pipeline001 falls

as the flowrate drops, with a drop in fluid temperature of 6K. When the outlet flowrate

initially reaches an equilibrium the rate of change of the temperature decreases,

until the point at which the outlet flowrate begins to fall. As the flowrate drops from

50kg/s to 54kg/s the temperature increases, until the point at which the flowrate

reaches the set point at which point the temperature settles at approximately

294.8K. The temperature at the outlet of the pipeline does not change intuitively as

59
there is a decrease and then an increase in the temperature. Heat loss from the

fluid over a section ∆x is calculated using the following equation [69].

4𝑈𝑜 𝑑𝑜
𝑞= (𝑇𝑜 − 𝑇𝑓 )
𝑑2

This equation does not explain the pipeline outlet temperature change. It is possible

however that there is an unknown interaction occurring within the model that cannot

be explained without the detailed code that underpins the software.

298 Variable
Inlet Temperature
Outlet Temperature

296
Temperature (K)

294

292

290

0 20 40 60 80 1 00 1 20 1 40
Time (h)

Figure 3-13: Inlet and Outlet Temperature for Pipeline 001

60
 296 Variable
1 00
Outlet Flowrate
Outlet Temperature
295

90
294
Temperature (K)

293

Flowrate (kg/s)
80

292

70
291

290
60

289

50
288
0 20 40 60 80 1 00 1 20 1 40
Time (h)

Figure 3-14: Outlet Flowrate and Outlet Temperature for Pipeline 001

3.5.2 Supercritical Phase Transport

In comparison to transporting the CO2 in the liquid phase, the same simulations

were repeated however this time the CO2 at the inlet was modelled to be in the

supercritical phase. The one difference here is that the temperature of the CO2 at

the inlet of the pipeline is above the critical temperature of 304.25K (31.1 oC). To

ensure the CO2 is in the supercritical phase an inlet temperature of 313K was used.

3.5.2.1 Flowrate

Figure 3-15 shows the flowrate profiles for Pipeline001 inlet, Pipeline001 outlet and

Pipeline002 Outlet. Similar to the liquid phase flowrate profile there is delay between

the inlet flowrate and the outlet flowrate showing that the model predicts that the

delayed response phase also occurs when CO2 is transported in the supercritical

61
phase. The extent of the delay between the inlet and the outlet flowrate can be seen

in Figure 3-16 again.

As with the simulation with liquid CO2 the same three distinct phases, ‘delayed

response phase’, ‘offset phase’ and ‘reduction phase’ occur and are shown by the

blue, red and green areas on Figure 3-16. The delayed response phase can be

justified through the same phenomena that was discussed for the liquid phase

transport and can also be used to explain this occurrence in the supercritical phase,

in that due to the temperature drop at the outlet of the pipeline there is a density

increase which through the continuity equation explains why there is a higher

flowrate at the outlet of the pipeline. For supercritical CO2 there is a higher

temperature drop compared to that observed for liquid CO2 and hence there is a

larger time delay between the change at the inlet and the outlet reaching a steady

state.

The model again predicts that there is an offset phase which occurs between the

inlet and the outlet flowrate which lasts for approximately 53 hours. The time in

which the offset phase occurs is shorter than the offset phase for the liquid phase

model. This is due to the higher velocity observed when in the supercritical phase

which in turn is due to the decrease in the density from the liquid to the supercritical

phase.

Figure 3-16 shows the first 14 hours of the simulation, this illustrates the extent of

the delay as the flowrate within the pipeline drops between the inlet of pipeline001

and the outlet of pipeline001 and pipeline002.

62
 Variable
1 00 Inlet Flowrate
Pipeline001 Outlet Flowrate
Pipeline002 Outlet Flowrate

90
Flowrate (kg/s)

80

70

60

50

0 20 40 60 80 1 00 1 20 1 40
Time (h)

Figure 3-15: Inlet and Outlet Flowrate for Pipeline 001 and 002

63
 Variable
1 00 Inlet Flowrate.
Pipeline001 Outlet Flowrate
Pipeline002 Outlet Flowrate

90
Flowrate (kg/s)

80

70

60

50

0 2 4 6 8 10 12 14
Time (h)

Figure 3-16: Inlet and Outlet Flowrate for Pipeline 001 and 002

Variable
60 Time-50000s
Time-100000s
Time-150000s
Time-200000s
58 Time-250000s
Flowrate (kg/s)

56

54

52

50

0 10000 20000 30000 40000 50000

Axial Position (m)

Figure 3-17: Flowrate Profile along Pipeline001

64
3.5.2.2 Pressure

Figure 3-18 shows how the pressure changes as the flowrate at the inlet of the

pipeline drops. The required initial inlet flowrate is calculated at 15600 kPa. As the

flowrate decreases, the pressure at both the inlet and the outlet of the pipeline also

decreases until it reaches approximately 10000 kPa. The pressure initially falls at a

rate of 990 kPa/h-1 for approximately 7 hours, at which point the rate of change of

the pressure, at both the inlet and the outlet of the pipeline slows and the pressure

remains between 9400 and 9800 kPa. A significant observation from Figure 3-18 is

that the pressure does not fall below the critical pressure, indicating that two phase

flow doesn’t occur within the pipeline as a result of the inlet flowrate falling by 50%.

1 6000 Variable
Inlet Pressure
Outlet Pressure
1 5000

1 4000
Pressure (kPa)

1 3000

1 2000

1 1 000

1 0000

9000
0 20 40 60 80 1 00 1 20 1 40
Time (h)

Figure 3-18: Inlet and Outlet Pressure for Pipeline 001

65
3.5.2.3 Temperature

Figure 3-19 shows the inlet and outlet temperature profiles for Pipeline001. To

ensure that the CO2 remains in the supercritical phase the temperature was

specified in the CO2 source at 313K and was maintained at this temperature

throughout the simulation period.

31 4 Variable
Inlet Temperature
Outlet Temperature
31 2

31 0
Temperature (K)

308

306

304

302

300

0 20 40 60 80 1 00 1 20 1 40
Time (h)

Figure 3-19: Inlet and Outlet Temperature for Pipeline 001

Figure 3-19 shows that while the temperature at the inlet remains above the critical

temperature the temperature at the outlet of the pipeline drops below 304.25K as

the flowrate falls. The model is therefore predicting that the CO 2 will transition from

the supercritical phase, into the liquid phase. The model however didn’t indicate that

there was two phase flow within the system at any point during the simulation. To

explain this, it can be reasoned that the supercritical phase is an intermediary phase

between the gas phase and the liquid phase as it has a density similar to a liquid

66
but a viscosity more similar to a gas and therefore the transition between the two

phases is more subtle and the changes in the physical properties are less extreme

than between the gas phase and the liquid phase.

The temperature losses between the inlet and the outlet of the pipeline can be

explained by the heat losses between the pipe wall and the surrounding material

which had a specified temperature of 278 K.

As shown with the liquid phase CO2 simulation, there is a correlation between the

temperature and the flowrate which can again be explained by the effect of

temperature on the fluid density, however as with the liquid CO 2 simulation the

change in temperature at the outlet of the pipeline cannot be explained with the heat

loss equations.

3.5.3 Carbon Dioxide Phase Evaluation

The aim of this study was to compare and contrast the response of liquid and

supercritical phase CO2 to reducing the flowrate to the inlet of a pipeline. The areas

in which the two phases are to be compared are flowrate, pressure and

temperature.

3.5.3.1 Flowrate

Figure 3-20 shows the inlet and outlet pipeline flowrate for both the supercritical and

liquid phase simulations. There are two observable differences between the

response in the outlet flowrate from a drop at the inlet flowrate. Both phases show

an offset between the inlet flowrate and the outlet flowrate however the difference

between the two is that in the size of the offset. The liquid phase simulation has

shown an offset of approximately 4kg/s whereas the supercritical phase simulation

67
shows a much greater offset of approximately 10kg/s. It has already been reasoned

that the offset is caused by the temperature drop causing a change in the density

and hence affecting the flowrate. Figure 3-23 shows a greater temperature drop

when the CO2 was transported in the supercritical phase when compared to the

temperature drop in the liquid phase. The reason behind this is due to the greater

temperature difference between the fluid in the supercritical phase and the

surrounding material, therefore there is a greater driving force for heat loss which

results in a larger temperature drop along the pipeline.

The second difference the model shows between transporting CO 2 in the

supercritical phase and the liquid phase is the time it takes for the offset at the outlet

of the pipeline to decrease. Figure 3-21 shows that the flowrate ‘wave’ in the

supercritical phase travels at a greater velocity along the pipe than the ‘wave’ shown

in the liquid phase simulation, this results in the outlet flowrate reaching the set point

faster when transporting supercritical CO2 compared to liquid CO2.

68
 Variable
1 00 Pipeline001 Inlet Flowrate
Pipeline001 Outlet Flowrate (Liquid)
Pipeline001 Outlet Flowrate (Supercritical)

90
Flowrate (kg/s)

80

70

60

50

0 20 40 60 80 1 00 1 20 1 40
Time (h)

Figure 3-20: Inlet and Outlet Flowrate for Liquid and Supercritical CO2

Variable
60 (Supercritical CO2) Time-50000s
(Supercritical CO2) Time-100000s
(Supercritical CO2) Time-150000s
(Supercritical CO2) Time-200000s
(Supercritical CO2) Time-250000s
58
(Liquid CO2) Time-50000s
(Liquid CO2) Time-100000s
Flowrate (kg/s)

(Liquid CO2) Time-150000s

(Liquid CO2) Time-200000s
56
(Liquid CO2) Time-250000s

54

52

50

0 10000 20000 30000 40000 50000

Axial Position (m)

Figure 3-21: Flowrate profile along Pipeline001 for Supercritical and Liquid CO2

69
3.5.3.2 Pressure

The pipeline inlet and outlet pressures can be seen in Figure 3-22 for both

supercritical phase and liquid phase CO2 models. The model shows a higher inlet

pressure requirement for CO2 in the supercritical phase that the in the liquid phase.

The reason behind this is due to the greater pressure drop along the pipeline when

transporting in the supercritical phase compared to the liquid phase, which is due

to supercritical CO2 having physical properties between that of a liquid and a gas.

This can be further explained with the Darcy-Weisbach equation which shows that

the pressure drop along a pipeline is a function of the fluid density and the flow

velocity squared. As the density of the liquid CO2 is higher than that of supercritical

CO2 this will increase the pressure drop, however the flow velocity of the

supercritical phase CO2 will be higher than that of the liquid phase when the mass

flowrates of both are the same, which is the case in this scenario. As the flow

velocity in the Darcy-Weisbach equation is squared, this will have a greater impact

on the pressure drop than the fluid density and is therefore the reason that a higher

pressure is necessary for supercritical phase CO2 compared to when it is in the

liquid phase.

70
 16000 Variable
Inlet Pressure (Liquid)
Outlet Pressure (Liquid)
15000
Inlet Pressure (Supercritical)
Outlet Pressure (Supercritical)
14000

13000
Pressure (kPa)

12000

11000

10000

9000

8000

0 20 40 60 80 100 120 140

Time (h)

Figure 3-22: Inlet and Outlet Pressure for Liquid and Supercritical CO2

3.5.3.3 Temperature

Figure 3-23 shows the temperature profiles at the inlet and outlet of pipeline001 for

both the liquid and supercritical phase CO2. The temperature at the inlet of the

supercritical phase is 13K higher than that of the liquid phase CO2. This is because

both these values were predefined in the model to ensure that the fluid entered the

pipeline in the desired phase.

From observing the temperature difference between the inlet and the outlet of the

pipeline it can be seen that there is a greater temperature drop along the length of

the pipeline when transporting in the supercritical phase, where at time 0 there is a

temperature drop of 4K whereas in the liquid phase, there is a smaller temperature

drop of 1.96K. The reason behind this observation is that there is a greater

temperature difference between the supercritical phase CO2 and the surrounding

71
material of the pipeline; this therefore provides a greater driving force for heat loss

from the fluid which consequently causes a larger temperature drop.

Comparing the temperature change over time at the outlet of the pipeline, for the

liquid and the supercritical phase CO2, it is observed that there is a greater

temperature drop in the supercritical phase CO2 when the flowrate is reduced but

the temperature rises again after a shorter period of time when compared to the

liquid phase CO2. This difference is comparable to the differences observed in the

flowrate where there is a larger offset but lasts for a shorter duration.

31 5 Variable
Inlet Temperature (Liquid)
Outlet Temperature (Liquid)
Inlet Temperature (Supercritical)
31 0 Outlet Temperature (Supercritical)
Temperature (K)

305

300

295

290

0 20 40 60 80 1 00 1 20 1 40
Time (h)

Figure 3-23: Inlet and Outlet Temperature for Liquid and Supercritical CO2

3.6 Conclusion

The simulation of a pipeline transporting CO2 in the liquid and supercritical phases has

given results which show how the flowrate, pressure and temperature of the CO 2 change

when there is a drop at the inlet of the pipeline. The outputs from the model show that the

72
flowrate at the outlet of the pipeline in both the liquid and supercritical phase falls at a rate

slower than the inlet and an offset in the flowrate occurs between the inlet and the outlet.

The duration of the offset is determined by the velocity of the fluid, this explains why the

offset occurs for a shorter duration when transporting in the supercritical phase when

compared to the liquid phase as the same inlet mass flowrate was prescribed for

both phases which results in the velocity of the fluid being greater for the

supercritical phase CO2. The results from the modelling also mean that the

hypothesis can be accepted as the model indicates that the gCCS model will show

that the rate of change in the outlet flowrate of a CO2 pipeline when the inlet flowrate

is reduced, will be greater when the CO2 is transported in the subcooled liquid phase

compared to the supercritical phase.

The simulation outputs show that the pressure of the CO2 transported in the

supercritical phase will be required to be 12 bar higher at the pipeline inlet than

when transported in the liquid phase, this has been explained with the use of the

Darcy-Weisbach equation. This result can be justified by the requirement of the

reservoir to maintain a constant pressure of 150bar. This higher pressure at the inlet

makes up for the lower pressure increase as the fluid flows through the well.

The results from the simulations show that there is a temperature drop along the

pipeline for both liquid and supercritical phase CO2. The reason behind this

temperature drop is due to the temperature of the surrounding soil being lower than

the fluid temperature. There is a greater temperature drop of the fluid along the

pipeline when the CO2 is transported in the supercritical phase. This is due to the

higher inlet temperature of the supercritical CO2 which means there is a greater

73
temperature difference between the fluid and the surrounding soil and hence a

greater driving force for heat loss and a greater temperature change.

There are some outputs from the simulations that cannot be explained such as the

pipeline outlet temperature change, however it can be theorised that this is

occurring due to unknown interactions within the code that underpins the software.

However given that there is limited access to this code, a precise understanding of

why the model behaves in this way cannot be developed.

74
Chapter 4 – Effects of Impurities and Multiple Sources of CO2 on
Pipeline Flow

75
4.1 Introduction

Chapter 3 covered a CO2 transport scenario in which pure carbon dioxide was

transported via a single pipeline from a single CO2 source. This allowed for a base

case scenario to be developed in which other scenarios can be compared. In this

chapter, two scenarios will be investigated that will build on the work presented in

Chapter 3.

In a real CCS process there is likely to be impurities in the CO2 which are known to

effect the physical properties of the fluid. The impurities within the CO 2 vary

depending on the capture process used. It is therefore important to understand how

these impurities effect the flow of CO2 within the pipeline. It has been argued in the

literature review that the gCCS software is the most appropriate software as it has

the capability to model impurities within CO2 with greater accuracy than other

commercial software packages as it uses statistical associating fluid theory to

determine the equations of state, which are believed to be more accurate than cubic

equations of state in the absence of experimental data. It is for this reason that

gCCS was the software of choice to model CO2 pipelines.

In the development of carbon capture and storage, it has been found that the

building of clusters of CO2 sources which share transportation infrastructure is the

most economic way to perpetuate the expansion of the technology. The concept

behind this is that multiple sources of CO2 whether from IGCC plants or CCGT’s will

connect via a branch pipeline to a trunk pipeline that will transport the CO 2 to the

site for sequestration. As there are different sources of CO2 it is expected that each

will operate differently from each other. For the research presented here, the flow

76
of CO2 from an IGCC and a CCGT are modelled. The reason behind modelling

these two different plants is that IGCC plants are expected to operate as base load,

given the inflexibility of this technology and hence the CO2 flowrate will remain

relatively consistent. CCGT’s are operated as peaking plants, so the flow of CO2 to

the pipeline is expected to be variable. This chapter will investigate the effects of a

consistent flow of CO2 into a pipeline and a variable flow.

4.2 Hypothesis

For this chapter two hypothesis will be investigated. The first will cover the effects

of impurities on the flow of CO2 in the pipeline and will try to prove or disprove the

following;

‘The gCCS modelling tool will demonstrate that impurities within the carbon dioxide

transported for CCS will cause a different response in the fluid dynamics to

changing the inlet flowrate compared to when transporting pure CO2’

The second part of this chapter will aim to prove or disprove the following

hypothesis;

‘The gCCS modelling tool will show that varying the flowrate of one of two sources

of CO2 will have a different effect on the fluid dynamics of the CO2 within the trunk

pipeline, compared to when there is only a single source of CO 2’

4.3 CO2 with Impurities Case Model Development

The model that has been developed to test the effects of impurities on the flow of

CO2 within a pipeline is the same as the model in Chapter 3 where the base case

scenario of pure CO2 in a single pipeline was analysed. The only difference between

the scenarios is the composition of the CO2 entering the pipeline. Keeping all other

77
parameters the same allows for a direct comparison between transporting pure CO2

and CO2 with impurities. The topology for this scenario is the same as that in Figure

3-8.

4.3.1 Impurities

The impurities to be investigated are those which are likely to be found within the

captured CO2. The three CCS capture technologies contain different impurities.

Post combustion capture contains nitrogen as the main impurity, this is due to the

combustion of the natural gas in air in CCGT power plants. Pre-combustion carbon

capture contains hydrogen as the main impurity which is due to the gasification

process which produces hydrogen as the product. After the capture process not all

the hydrogen is separated from the CO2 and remains in small quantities. Oxyfuel

combustion contains oxygen as the impurity due to the combustion of the fuel with

pure oxygen instead of air. These three components are therefore the ones of

interest to understand the effects of impurities on the flow of CO2. While it has been

mentioned in Chapter 2 that there are other impurities within the CO 2 these are in

significantly low quantities and have therefore not been included in the analysis in

this research.

To develop the scenario it is also important to know what proportion of the fluid

entering the pipeline is the impurity. This has been covered in the literature and is

summarised in Table 4-1. However to allow for a direct comparison between each

impurity the same percentage was used for each in the simulation. A value of 0.2

mol/mol was used as the input into the model for each case as this was the lowest

concentration of any of the impurities and therefore indicates whether the smallest

concertation of any of the impurities would have any effect on the system.

78
 Table 4-1: Common Impurities Found in Captured CO2

Capture Technology Main Impurity % vol/vol

Post-combustion capture Nitrogen 0.2

Pre-combustion capture Hydrogen 1.5

Oxyfuel combustion Oxygen 3.2

The simulation carried out for this scenario is the same as that of the base case

developed in chapter 3, in which the flowrate into the pipeline begins at a steady

state of 100 kg s-1. The flowrate then falls to 50 kg s-1 at a rate of 4 kg s-1 min-1. The

code used to produce this simulation is shown in Appendix A.1.

4.4 Results

4.4.1 Nitrogen Case

The results for the nitrogen impurities case are shown in Figure 4-1 to 4-4. Figure

4-1 shows the flowrate change at the inlet and outlet of pipeline 001. As with the

base case there are three distinct phases in the outlet flowrate of the pipeline which

result from the change in the inlet flowrate. The three phases are the ‘delayed

response phase’, the ‘offset phase’ and the ‘reduction phase’. This indicates that

the presence of nitrogen as an impurity does not affect the shape of the response

at the outlet of the pipeline.

Figure 4-4 shows the results from the nitrogen impurity scenario along with the base

case scenario. Observing Figure 4-4 shows that the presence of nitrogen in the CO2

79
does impact on the fluid dynamics. The presence of nitrogen in the CO 2 is shown

to cause an increase in the offset from the base case. The increase in the offset

between the base case and the nitrogen impurity case is approximately 1kg/s.

Figure 4-1: Pipeline Inlet and Outlet Flowrate (CO2 + N2)

80
Figure 4-2: Pipeline Inlet and Outlet Pressure (CO2 + N2)

81
 Figure 4-3: Pipeline Inlet and Outlet Temperature (CO2 + N2)

Figure 4-4: Pipeline Inlet and Outlet Flowrate (Pure CO2 & CO2 + N2)

82
4.4.2 Hydrogen Case

Figures 4-5 to 4-8 show the flowrate results for the H2 and CO2 scenario. Figure 4-

5 shows that the three phases as seen in the base case are also observed when

there is hydrogen as an impurity. Figure 4-8 Shows that the addition of hydrogen in

the CO2 does impact the response at the outlet of the pipeline when compared to

the base case of pure CO2. In the same manner that the addition of nitrogen caused

an increase in the offset during the ‘offset phase’ the addition of hydrogen also

caused this same increase in the offset. However the addition of hydrogen caused

a greater impact than the nitrogen with an increase in the offset of approximately

2kg/s.

Figure 4-5: Pipeline Inlet and Outlet Flowrate (CO2 + H2)

83
 Figure 4-6: Pipeline Inlet and Outlet Pressure (CO2 + H2)

Figure 4-7: Pipeline Inlet and Outlet Temperature (CO2 + H2)

84
 Figure 4-8: Pipeline Inlet and Outlet Flowrate (Pure CO2 & CO2 + H2)

4.4.3 Oxygen Case

Figures 4-9 to 4-12 show the results from the carbon dioxide and oxygen simulation.

As with the nitrogen and the hydrogen simulations the response in the flowrate,

temperature and pressure at the outlet of the pipeline follows the same profile as

that of the base case simulation. The difference in the offset phase between the

pure CO2 and the oxygen impurity is again observed. In the case for the presence

of oxygen as in impurity, there is an increase in the offset at the outlet of the flowrate

of approximately 1kg/s which is similar to that of the nitrogen impurity case.

85
Figure 4-9: Pipeline Inlet and Outlet Flowrate (CO2 + O2)

Figure 4-10: Pipeline Inlet and Outlet Pressure (CO2 + O2)

86
 Figure 4-11: Pipeline Inlet and Outlet Temperature (CO2 + O2)

Figure 4-12: Pipeline Inlet and Outlet Flowrate (Pure CO2 & CO2 + O2)

87
4.5 Analysis

The three scenarios that have been developed and modelled show the impact of

impurities in the CO2 on the flow dynamics when the inlet flowrate to the pipeline is

reduced. In all three cases there is the same three phases as observed in the base

case scenario. The difference between the base case and the impurities case is the

difference in the ‘offset phase’. The modelling of the three scenarios shows that the

addition of nitrogen, hydrogen or oxygen cause an increase in the offset between

the inlet flowrate and the outlet flowrate. This effect can be explained through the

impact of impurities on the physical properties of the fluid entering the pipeline. As

nitrogen, hydrogen and oxygen all have different molecular weights to carbon

dioxide they will affect the total density of the fluid and hence affect the flowrate.

There is also a difference between each of the impurities, with oxygen having the

smallest impact on the offset and hydrogen having the greatest impact on the offset.

This again can be explained through the difference between each of the impurities

compared to carbon dioxide, as oxygen has the closest molecular mass to carbon

dioxide and hydrogen has the biggest difference in molecular mass to carbon

dioxide. This theory is supported by how both nitrogen and oxygen have the same

approximate impact on the offset and both have the closest molecular mass out of

the three impurities, with nitrogen having a molecular mass of 28 g mol-1 and oxygen

having a molecular mass of 32 g mol-1. This therefore explains why the results from

the hydrogen impurity scenario would deviate the most from the base case with

pure carbon dioxide. This therefore indicates that the three carbon capture

technologies; pre-combustion, post-combustion and oxyfuel combustion will all

88
have different flow dynamics during transportation, due to the presence of the

specific impurities present in the CO2. It is expected that the greater the amount of

an impurity in the CO2, the greater the effects of the presence of the impurity will

be.

The results from the simulations show that the original hypothesis was correct, as

the model shows that the presence of impurities that are likely to be found within

captured CO2 do impact on the flowrate at the outlet of the pipeline.

4.6 Multiple Sources of CO2 Model Development

4.6.1 Pipeline Dimensions

To develop the model it was first necessary to understand how a pipeline receiving

CO2 from multiple sources would be designed. To avoid over complicating the

model, the simplest scenario of two CO2 sources was chosen as the modelling case.

A previous study looking at CO2 flow from multiple sources modelled the system

with two parallel pipelines from the two sources which then mix the flows

downstream and enter a larger diameter single pipeline which then transports the

CO2 to a reservoir to be sequestered. A similar approach was taken here to develop

the pipeline model. As the two parallel pipelines from the CO 2 sources are

transporting a smaller amount of fluid than the main trunk line further downstream,

the pipeline diameter of these two lines is smaller which gives greater economic

benefit than having oversized pipelines for the branch lines. Table 4-2 shows the

length and internal diameter of the pipelines shown on Figure 4-13. The well

dimensions used in this model are the same as those used for the two previous

scenarios, with the reservoir also having the same inputs.

89
 Table 4-2: Pipeline Dimensions

Pipeline segment Internal Diameter (m) Length (m)

Pipeline 001 0.3048 2000

Pipeline 002 0.3048 2000

Pipeline 003 0.6096 52000

Pipeline 004 0.6096 52000

4.6.2 Pipeline Flows

The temperature and pressure of the pipelines was set up the same as the base

case pipeline model in chapter 3. The flowrate of the fluid however was set so that

the initial total flowrate in the trunk pipeline was the same as the flowrate within the

pipeline in in chapter 3. The flowrate of the branch pipelines was split equally so

that each CO2 source has an outlet flowrate of 50 kg s-1. The simulation was set up

so that the flowrate from Source 001 was reduced from 50 kg s-1 to 25 kg s-1 while

the flowrate from Source 002 was kept constant throughout the simulation, this

therefore gives a total flowrate in Pipeline 003 of 75 kg s-1. This approach was taken

to simulate a base load power plant with CCS and a load following power plant with

CCS which both feed CO2 into the same trunk pipeline.

90
 Figure 4-13: Pipeline Topology with Multiple Sources of CO2

4.7 Results

The results from the simulation are focussed on the input and output variables of

Pipeline 001, Pipeline 002 and Pipeline 003. The flowrate, temperature and

pressure are the areas of interest and where the analysis of the simulation is

focussed.

4.7.1 Flowrate analysis

Figure 4-14 shows the flowrate profiles at the inlets and outlets of the two branch

pipelines, Pipeline 001 and 002 and then the main trunk Pipeline 003. The results

show that the flowrate at the outlet of Pipeline 001 falls at the same rate as the inlet

flowrate is reduced and reaches the output within minutes of the step change

occurring and does not show an ‘offset phase’ as seen when the flowrate is reduced

in the previous scenarios. The reason behind this is likely due to the pipe length of

Pipeline 001, which is 50 km less than the pipeline in the single CO 2 source

scenarios. Therefore the flowrate at the outlet of the pipeline can reach the set point

91
in a period of time before the ‘offset phase’ develops. The flowrate at the outlet of

Pipeline 002 stays constant throughout the simulation. This demonstrates that the

change in flowrate in Pipeline 001 does not cause any back flow within Pipeline 002.

The flowrate profile at the outlet of Pipeline 003 follows that of the base case

simulation, where there are three distinct phases; the ‘delayed response phase’, the

‘offset phase’ and the ‘reduction phase’.

Figure 4-14: Inlet and Outlet Flowrate for Pipeline 001, Pipeline 002 and Pipeline 003

4.7.2 Pressure Analysis

Figure 4-15 shows the pressure profile at the inlets of Pipeline 001, Pipeline 002

and Pipeline 003 and the outlet pressure of Pipeline 003. This results of the

simulation show that a change in the flowrate in Pipeline 001 means that there is a

requirement for the pressure at the inlet of Pipeline 002 to decrease at the same

rate as the pressure in Pipeline 001. The reason that this response is observed is

due to the way the model has been developed with the inlet pressures being

92
determined by the well pressure. This also ensures that there is no back pressure

in Pipeline 001.

Figure 4-15: Inlet and Outlet Pressure for Pipeline 001, Pipeline 002 and Pipeline 003

4.7.3 Temperature Analysis

Figure 4-16 shows the temperature profile at the inlet and outlet of Pipeline 001,

Pipeline 002 and Pipeline 003. The results from the simulation show that the CO2

at the outlet of Pipeline 001 increases by 1.5K when the flowrate is reduced.

However this is on contrast to the temperature of the CO2 at the outlet of Pipeline

003 which follows the same profile as the base case developed in chapter 3. The

difference in the response in temperature between the outlet of Pipeline 001 and

Pipeline 003 can be linked to the differences in the flowrate profile at the outlets of

each pipeline. While Pipeline 003 shows an offset phase between the inlet and the

outlet, Pipeline 001 does not show such a response.

93
 Figure 4-16: Inlet and Outlet Temperature for Pipeline 001, Pipeline 002 and Pipeline 003

4.8 Conclusion

The results from the impurities study have proven the original hypothesis correct as

the model has shown that the presence of any of the impurities; nitrogen, hydrogen

and oxygen, in the carbon dioxide case a different response in the outlet flowrate

than when only pure carbon dioxide is transported. The modelling of the impurities

scenarios has shown that the presence of any of the three impurities causes and

increase in the difference between the inlet and the outlet of the flowrate during the

‘offset phase’, it is understood that the reason for this effect is due to the impact of

the impurities on the physical properties of CO2, as the impurities tested within this

study have been found to decrease the density of the CO 2 [70] and from the

comparison in chapter 3 between liquid and supercritical phase CO2 a lower density

fluid will have a greater offset. The modelling has also shown that hydrogen as an

94
impurity in carbon dioxide causes a larger difference between the inlet and the outlet

of the flowrate than either the oxygen or nitrogen. It is believed that this is due to

the difference in the molecular weights, with H2 having a smaller molecular mass

than N2 and O2 and therefore having a greater impact on the density of CO2. This

is also supported by the fact that oxygen which has the closer molecular mass to

carbon dioxide than nitrogen, has a smaller impact on the flowrate.

The results from the multiple sources of CO2 modelling have proven the original

hypothesis incorrect, as two sources of CO2 entering the pipeline did not cause a

different response in the flowrate from the base case, of the main trunk line when

there is a reduction in the flowrate of one of the sources. This is because the inlet

flowrate to Pipeline 003 has the same profile as the set flowrate change to the inlet

of Pipeline 001. The results also show that the offset did not occur in Pipeline001,

this will be due to one or a combination of three factors; the pipeline diameter, the

pipeline length or the pipeline initial flowrate as these are the three factors in which

there is a difference between Pipeline 001 and Pipeline 003. The affects of varying

these factors has not been investigated and is therefore an area in which further

research can be carried out. The modelling also showed that the temperature

effects observed in Pipeline 003 were not observed at the outlet of Pipeline 001. It

is understood that the reason for the difference in temperature profiles is due to the

difference in the flowrate profile at the outlets of Pipeline 001 and Pipeline 003. The

results from the simulation also reveal that under these conditions there is no

reverse flow in any of the pipelines and that varying the flowrate from one source

does not effect the flowrate in another branch pipeline.

95
Chapter 5 – Modelling of Shell QUEST CO2 Pipeline

96
5.1 Introduction

5.1.1 Overview

The research that has been presented in previous chapters has modelled and

simulated pipeline systems which are theoretical, the values that have been used

in these models have been based on assumptions. The outputs from the models in

the previous chapters have not been able to be validated as there is extremely

limited public data on the operation of CO2 pipelines. The majority of CO2 pipelines

are owned by large organizations who have not published their data. Models which

have not previously been validated against real pipeline data are able to predict

trends as a response to changes in input variables, but are not able to accurately

simulate CO2 flows. Therefore, in order to be of use for predictive capability, the

models require validation using pipeline CO2 flow data.

During the last year of the PhD an 8 month placement with Shell Canada was

organised to facilitate the collection of data from real operating CO 2 pipelines at

their recently commenced CCS project in Alberta. Given that it was not feasible to

set up a pipeline experiment to produce experimental data for the model, it was

necessary to model an existing operational CO2 pipeline which could be validated

against real operational data.

The Shell Quest project is an operating carbon capture and storage facility which

captures CO2 from a hydrogen production facility. Once captured the CO 2 is

compressed and then transported via a 65km pipeline to 3 different wellheads

where the CO2 is then sent approximately 2km underground where it is stored in

the reservoir.

97
The aim of the work described within this chapter is to model the Quest pipeline

using the gCCS software that has been applied for the simulations reported in

earlier chapters. The parameters within the model that were tuned to match that of

the Quest pipeline. The model was then set to simulate several different operating

situations including start-up, shut-down and flowrate ramps. The output from the

model was then compared to real data from the Quest pipeline. This is the first time

that the outputs from a CO2 pipeline model have been compared to real data. This

gives a greater indication of the accuracy of the outputs, from the work carried out

in previous chapters. This also indicates the accuracy of the SAFT equations of

state as well as the gCCS software in general.

5.1.2 The QUEST Carbon Capture Facility

The Shell Quest project is a CCS facility which takes approximately one third of the

CO2 emissions from a hydrogen production facility based in Fort Saskatchewan,

Canada. This equates to 1.2 million tons per annum. The hydrogen is produced via

steam methane reforming and produces a mixture containing hydrogen, carbon

dioxide, carbon monoxide and water. The hydrogen is separated using pressure

swing adsorption and produces a stream of approximately 99% purity which is used

in upgrading of the oil sands. The rest of the gasses are then sent to the CCS facility

for separation and compression. The capture process uses an advanced amine

solvent to separate the CO2 from the flue gasses and uses traditional absorption

and stripping units to produce a high purity CO2 stream. The CO2 is then

compressed and dehydrated in an eight-stage compressor until the CO2 is in the

supercritical phase and at the conditions appropriate for transport along the pipeline.

Once compressed to the desired pressure and temperature the high purity CO2

98
stream is transported via pipeline to three different wells where it is then pumped

2km underground into a saline aquifer. Even though the three wells are located at

different locations, they each pump the CO2 into the same aquifer. This CCS project

is set to sequester approximately 1.2MT of CO2 per year. The project began

operating in September 2015.

The Quest CCS project is one of the first of its kind and is one of the first to use

saline aquifers for the storage of the CO2 instead of the more popular method of

using oil wells with enhanced oil recovery. The cost of Quest including pre-FID,

capital and 10 years OPEX was approximately CND $1.4 billion. The capital ratio

was; 80% capture, 10% transport and 10% wells. Approximately CND$120 million

was provided by the Canadian federal government and CND$745 million was given

by the Albertan provincial government [71]. Since most of the costs have been

provided by government some of the information regarding Quest has become

publicly available. It should be noted however that the publicly available information

would not allow for complete modelling of the QUEST pipeline and hence closer

relationships with Shell were necessary.

5.2 Methodology

To carry out the research, the first step was to obtain all the relevant information

required to produce the model. Some of the design data such as pipeline size, well

depths and topography are publicly available and can be found in documents on

the Canadian government website. However to model the pipeline accurately it was

necessary to obtain data directly from the operators working on the Quest site, as

the way in which the system is operated effects how the model needs to be

developed. Once the model was formulated, it then underwent testing. The testing

99
compares the outputs from the model to real data available from the pipeline; this

data was provided through a system only accessible through Shell.

5.2.1 Pipeline Details

5.2.1.1 Process Unit

The first step in producing a model for the Quest pipeline was to connect models of

the process units together in gCCS. The input of the model is the source of the CO 2,

as systems upstream of the pipeline were not considered as part of the model in

this case. The Quest pipeline includes a 65km pipeline, 6 line-break valves, 3 well

heads, 3 × 2km wells each with a choke valve and a single reservoir. Within gCCS

there is no specific model for a line-break valve therefore the emergency shutdown

(ESD) valve model was used as a substitute. As the three wells are located several

kilometres from each other, lateral pipelines that offshoot from the main pipeline

were also incorporated into the model. Even though the wells are at different

locations the reservoir is considered to homogeneous, it is for this reason only a

single reservoir unit was used in the model.

5.2.1.2 Topography

Knowledge of the pipeline topography is necessary for an accurate representation

of the Quest pipeline. The topography shows how the pipeline elevation changes

along its length. Previous models have simulated straight pipelines with no changes

in elevation. The changes in elevation are incorporated to simulate the actual

features of the pipeline more realistically than the simple straight pipe models are

able to. Figure 5-1 [72] shows the topography of the main 65km pipeline along with

the location of the 6 line block valves. As can be seen in Figure 5-1 there are

elevation changes between each of the valves. Table 5-1 gives more detailed

100
information regarding the elevation changes between each valve as well as the

length and volume of each pipeline segment. The elevation changes vary greatly,

with the maximum change of 16.96m in elevation being between LBV#4 and LBV#5.

The changes in elevation can be modelled within gCCS to produce an accurate

representation of the pipeline.

Figure 5-1: QUEST pipeline topography

101
 Table 5-1: QUEST pipeline details

Length [m] Elev Change [m] ID [mm] Volume [m3]

Scotford-LBV#1 14539 -2.88 299.7 1,026

LBV#1-LBV#2 10190 -0.65 299.7 719

LBV#2-LBV#3 7672 3.28 299.7 541

LBV#3-LBV#4 4404 -4.73 299.7 311

LBV#4-LBV#5 13817 16.96 299.7 975

LBV#5-WS#2 9058 6.65 299.7 639

WS#2-LBV#6 4790 2.11 299.7 338

Lateral 1 1590 0.37 146.3 27

Lateral 2 1962 -0.49 146.3 33

Lateral 3 4727 -3.25 146.3 79

Total 72749 4,687

102
 5.2.1.3 Main and Branch Pipeline Dimensions

Table 5-2 shows the dimensions for both the main pipeline and the lateral pipelines

that branch off towards the wells.

Table 5-2: QUEST pipeline dimensions

Main Trunk Line Well Branches

Diameter-OD [mm] 323.9 168.3

Diameter-IN [mm] 299.7 146.3

Wall Thickness [mm] 12.1 11

Minimum Burial depth [m] 1.5 1.5

Average above ground length at LBV or well pads 20 25

103
5.2.1.4 Fluid composition

Table 5-3: QUEST fluid composition

Component Normal Operation Mole% Upset Condition Mole%

CO2 99.2 95

CO 0.02 0.15

N2 0 0.01

H2 0.68 4.27

Methane 0.09 0.57

Water <52ppm 52ppm

5.2.1.5 Pipeline Operating Conditions

Table 5-4 shows the operating conditions of the pipeline for both winter and summer

conditions. It is necessary to model the conditions for the seasons separately. Due

to the geographic location of the pipeline; temperatures can vary from an average

low of -19.5oC in January to an average high of 23.4oC in July [73]. The difference

in atmospheric temperatures means that the CO2 temperature can change between

the seasons and hence cause different dynamics within the system. Table 5-4 also

gives information regarding full operation and turn down which will allow for accurate

outputs when simulating the transition from one operating state to another.

104
 Table 5-4: Quest pipeline operating conditions

Winter Summer

Conditions Conditions

Pipeline Inlet Temperatures [C] 43 49

Operating Pressure [barg]

Normal Min 80 80

Normal Max 110 110

Maximum Design 140 140

Flowrate Rated Capacity Turndown

Flow into Pipeline [kg/hr] 152207 45662

3 Wells Operating [kg/hr/well] 50736 N/A

2 Wells Operating [kg/hr/well] 76104 N/A

1 Wells Operating [kg/hr/well] 120497 45662

Ambient Temperature -40 35

Ground Temperature at pipeline burial depth [oC] 0 11

Heat Transfer Coefficient [BTU h-1 ft2 oF-1]

Minimum 0.35 0.35

Maximum 1 1

105
5.2.1.6 Reservoir Operating Conditions

As previously stated, the reservoir, as well as the pipeline, was modelled as part of

the simulations. Within gCCS little information is required for the reservoir model

due to its relative simplicity. The data in Table 5-5 shows the information that has

been used for the modelling work carried out by Shell prior to start-up of the Quest

project. As this is not actual data from the reservoir the injectivity values are given

as a range. Within this study low, high and middle values were used in three different

simulations. The middle value was taken as the average of the high and low values.

Table 5-5: Reservoir Operating Conditions

Reservoir Characteristics

Reservoir Temperature [deg C] 60

Reservoir Pressure [bar] 200

Max allowable bottomhole 280

pressure

Reservoir Injectivity [m3/d/Mpa]

Low 300

High 3,000

5.2.2 Pipeline Operation

The way the pipeline conditions are controlled has a significant effect on how the

model was developed. There are three main parameters within the Quest pipeline

that are controlled; the inlet pipeline temperature, the pressure before wellhead 1

and the flowrate before wellhead 2. The temperature at the inlet of the pipeline is

106
controlled to stay at approximately 316K, at this temperature the CO 2 is in the

supercritical phase. The pressure before wellhead 1 is maintained at a pressure

above the critical pressure to avoid the CO2 entering the gas phase within the

pipeline; the pressure is set to remain at approximately 8700kPa. The flowrate at

the second well has been set at 70,000 kg/h, any variability in mass flowrate at the

inlet is absorbed into the flowrate going to the first well. The variables are maintained

using PID control schemes. The model was produced with control schemes similar

to that observed on the Quest pipeline, the PI control schemes are shown in Figure

5-7.

5.2.3 Simulation details

With the model set up as stated, simulations were developed for different process

operations: ramp-up, ramp-down and changes in inlet temperatures. These

simulations have been set with input conditions to match those recorded in the

pipeline to allow comparison of the outputs with the real data. The limitation of the

modes of the simulations is due to the lack of data available from the Quest pipeline

for certain operations. While start-up procedure can be simulated and verified using

historical plant data there hasn’t to date been any need for shut down of the pipeline.

While the shut-down of the pipeline can still be simulated, there is no data available

for verification.

5.2.3.1 Pipeline Data

The data from the Quest pipeline have been collected through a program known as

PI ProcessBook. This software allows access to all networked electronic

measurement devices on the plant including the pipeline and contains all historical

107
data. PI Process Book allows all data to be transferred into Excel, where it can be

manipulated and plotted. There are some limitations to the data in that the

measurement devices are located at specific places along the pipeline. It is

therefore not possible to obtain data along the full length of the pipeline. Flow and

temperature measurements are recorded from the pipeline inlet and at points before

each reservoir. Pressure measurements are recorded at the inlet and on either side

of each valve.

After collecting the data set, the next step was to determine how the simulation

would be run. To do this all historical flowrate data of the pipeline was plotted. The

simulations have been designed to simulate the effects of flowrate ramp-up, ramp-

down, steady state and start-up. By looking at the historical flowrate data, certain

time periods were chosen in which the simulated inlet flowrate can match that of the

pipeline. Figure 5-3 shows the inlet flowrate data for the dates between the 8th

October 2015 and the 30th November 2015. As can be seen, the flowrate is variable

and has several options to simulate the desired operation. To simulate steady state

mode, only a single point is needed to compare the outputs of the model to that of

the Quest pipeline. When simulating the dynamics of the flowrate, dynamic data

over a period of time is required from the Quest pipeline. To simulate a ramp down

and a ramp up in the CO2 flow there are three possible times which could be

simulated as shown in Figure 5-3. Figures 5-4 to 5-6 show the time-periods and the

changes in inlet flowrate; which have been simulated to compare the model’s ability

to produce accurate results.

The time periods that are shown in Figures 5-4 to 5-6 show both an initial ramp

down followed by a period of steady state operation and then a ramp-up. The three

108
time-periods give similar changes in flowrate with approximately 20% reduction in

flow.

5.2.4 Data Analysis

Before running the model it was necessary to analyse the data that is obtained from

the pipeline to gain an understanding of how the flowrate varies. There is a limit to

how much information can be obtained about the pipeline from the measurement

devices, as they are confined to certain locations. The flowrate measurement

devices are limited to the inlet of the pipeline and the inlet to the well, the same is

for the temperature sensors. Pressure sensors are located at the inlet of the

pipeline, on either side of each of the line break valves and at the entry to the well.

However the temperature readings were recorded from the line block valves using

an infra-red gun as the valves are above ground so approximate temperature

readings can be made.

5.2.5 Measurement Devices

5.2.5.1 Flowrate measurement device

The measurement devices that are used to measure the mass flowrate are Coriolis

meters. A Coriolis meter measures the mass flow directly; it works on the principle

of changes in the vibration of the flow meter as the mass flow increases and

decrease. The Coriolis meters also measure density, along with the mass flowrate

a volumetric flowrate is calculated through the following relationship;

𝑚
𝑣=
𝜌

Equation 5-1: Density Equation

109
Two types of Coriolis meters are used. The one measuring the inlet mass flow has

an accuracy of +/-0.05% of the reading. The ones measuring the mass flow at the

entry to the well have an accuracy of +/- 0.10% of the operating range. As the mass

flow is measured at both wells there are two measurement devices, therefore

doubling the possible error. The operating range for these devices is 0 – 120,000

kg/h. Coriolis meters are known for having high degrees of accuracy and it can be

seen in Figure 5-4 to 5-6 that the errors are generally insignificant when it comes to

analysing the mass flow data.

5.2.5.2 Temperature measurement device

The temperature sensors used at the inlet and outlet of the pipeline are resistance

temperature detectors (RTD) also known as resistance thermometers. RTD’s are

made from metallic conducting materials, platinum, copper or nickel. Platinum 100Ω

RTD’s are used as the temperature sensors at both the inlet and outlets of the

pipeline. These have a normal operating range between 70 and 870K with an

accuracy of +/-0.4K. The temperature of the CO2 at the pipeline inlet is kept constant

at approximately 316K. The temperature of the CO2 decreases along the length of

the pipeline however at no point during operation has it decreased below the lower

operating range of the sensor. The RTD works by utilizing the increasing electrical

resistance with temperature in metals.

5.2.5.3 Pressure measurement device

The pressures at the inlet and along the length of the pipeline are measured using

sensors that work using the piezoresistive effect. The piezoresisitive effect is the

change in electrical resistivity of a metal when mechanical strain is applied. Figure

5-2 [74] shows how the specific pressure sensor used on the Quest pipeline works.

110
 Figure 5-2: Pressure sensor

1. Silicon measuring element, substrate

2. Wheatstone bridge

3. Channel with fill fluid

4. Metal process isolating diaphragm

This sensor works by the operating pressure deflecting the process isolating

diaphragm and the fill fluid transfers the pressure to a resistance bridge

(semiconductor technology). The pressure dependent change in the bridge output

voltage is measured and evaluated [74]. The pressure sensor has an accuracy of

+/- 0.075%.

111
5.2.6 Simulation Periods

Figure 5-3 shows the inlet mass flowrate and the combined mass flowrate entering

each of the two wells for the 8th October to the 30th November 2015. The errors

associated with each of the devices shown by the error bars in black. These indicate

that the errors associated with the measurement devices are significantly small.

One of the main observations from Figure 5-3 is that the inlet flow doesn’t always

directly match the total outlet flow at a given time. It can also be seen that there is

a point where the flowrate becomes negative, this was due to an unplanned

shutdown of the pipeline which impacted on the measurement device.

112
 139000

119000

99000
Mass Flowrate (kg/h)

79000

59000

39000

19000

-1000
08-Oct-15 00:00:00 18-Oct-15 00:00:00 28-Oct-15 00:00:00 07-Nov-15 00:00:00 17-Nov-15 00:00:00 27-Nov-15 00:00:00
Time (h)

Figure 5-3: Inlet and Outlet mass flowrate of the Quest pipeline between 07/10/15 and 28/11/15

113
Figures 5-4 to 5-6 show the data from the quest pipeline that is to be simulated within

the model. As can be seen the mass flowrate contains a significant amount of small

changes, for reasons of practicality these small changes were not simulated. The purpose

of the simulations is to understand the effects of significant changes in the system. As

shown in Figure 5-4 the initial flowrate lies between 120000kg/h and 125000kg/h. Within

the model an initial flowrate of 34.6kg/s (124,560 kg/h) was used.

130000

125000

120000
Inlet Flowrate (kg/h)

115000

110000

105000

100000

95000

25-Oct-15 26-Oct-15 27-Oct-15 28-Oct-15

Time (h)

Figure 5-4: QUEST pipeline inlet mass flowrate data from 25/10/15 to 28/10/15

114
 140000

130000
Flowrate (kg/h)

120000

110000

100000
31-Oct-15 01-Nov-15 02-Nov-15 03-Nov-15 04-Nov-15
Time (h)

Figure 5-5: QUEST pipeline inlet mass flowrate data from 31/10/15 to 04/11/15

115
 145000

140000

135000
Flowrate (kg/h)

130000

125000

120000

115000

110000
08-Nov-15 09-Nov-15 10-Nov-15 11-Nov-15
Time (h)

Figure 5-6: QUEST pipeline inlet mass flowrate data from 08/11/15 to 11/10/15

5.3 Initial Model Development

5.3.1 Topology

Using the Quest design and operating information a pipeline model was produced which

is schematically shown in Figure 5-7. The pipe dimensions are the same as those

reported in Table 5-1, all parameters shown within Table 5-2 and Table 5-3 are also

included within the model. The ground temperature at burial depth is taken as the winter

conditions of 273K. This was chosen as the time-periods in which the simulations

replicated are closer to the winter season and therefore the ground temperature is

expected to be colder. However it should be noted that this may not be the actual ground

temperature at pipeline burial depth but is an assumption that is made based on the flow

assurance report developed by Shell [72]. It is reasonable to assume that the ground
116
temperatures used in the flow assurance report have the greatest reliability amongst the

literature on ground temperatures, as it is specific to the area where the QUEST pipeline

is buried. Actual ground temperatures could not be obtained as there is no available

equipment to measure the ground temperature at burial depth. Other parameters such as

pipe wall roughness, valve leakage fraction and liquid flow coefficient were not obtainable,

therefore the default values within gCCS were utilized.

5.3.2 Control Schemes

5.3.2.1 Pressure Control

The pressure to the well is controlled to maintain a constant pressure of 87,000kPa. This

is to ensure that the pressure within the pipeline does not fall below the critical pressure

of CO2 and therefore does not enter the gaseous phase. To control the pressure a PI feed

forward controller was modelled, which is the same type of control used at the Quest site.

5.3.2.2 Flow Control

The flowrate to the second well is controlled to maintain a constant flowrate of 70,000kg/h.

The reason for controlling the flowrate to the second well was to allow the testing of the

pipe flow and the effects on the reservoir. To control the flowrate a feed forward PI

controller is used.

5.3.3 Schedule

The model was simulated to observe the effects of changing the inlet flowrate on

parameters within the pipeline, including flowrate, temperature, pressure and density.

The schedules (Appendix C.1 – Appendix C.2) were written to imitate the changes in

flowrate shown in Figures 5-4 to 5-6. To allow for the entire system to settle, the end of

117
the schedule was written to give 200,000 seconds (55.55 hours) for equilibrium to be

reached.

118
Figure 5-7: Initial pipeline model

119
5.3.4 Initial Model Analysis

The model was set up to directly represent the Quest CO 2 pipeline, as closely as the

software would allow. This included control schemes that maintained the flowrate and the

pressure within the pipeline at desired set points. There were two problems that occurred

with this set up of the model; the first was related to the pressure controller in that there

is an over specification with the reservoir pressure and the pressure controller. As the

pressure is being specified in two parts of this system the simulation would not allow this

to be completed and produced an error when simulating. The control scheme for the

pressure control was removed and the simulation initialized again. The second problem

that occurred was related to the flow controller. The simulation was able to complete,

however when observing the response from the control valve used to maintain the mass

flowrate, the stem position remained constant and the flowrate through the valve varied

as the pipeline inlet flowrate changed. It has been understood that the reason for this

behaviour of the flow controller is due to flow constraints caused by the split of the CO 2

flow between lateral 001 and lateral 002.

5.3.5 Secondary Model Development

To model the pipeline in a way that is representative of how the pipeline is operating, a

new set up was developed in which the well and reservoir were not incorporated into

model. The outlets of the two lateral pipelines were replaced with CO 2 sink models. The

sink connected to lateral 1 specifies a pressure of 87MPa while the sink connected to

lateral 2 specifies a flowrate of 69,840 kg/h. Developing the system in this way means

that the set points for the pipeline outlets can be maintained even when there are changes

120
in the system, therefore operating similar to the action of a controller. The topology for the

model can be seen in Figure 5-8.

Using the same schedule as used previously, the simulations were repeated. With the

new model the system was able to operate in accordance with the Quest pipeline,

whereby the outlet flowrate and pressure specifications are the controlled variables

121
Figure 5-8: Simplified QUEST pipeline model

122
5.4 Comparison Between Model and QUEST Data

5.4.1 Flowrate

The three different scenarios have been modelled and the three primary parameters have

been compared with that of the Quest pipeline. Figures 5-9 to 5-11 show the comparison

between the model data and the Quest data for all three scenarios. The flowrate is taken

from three points on the pipeline, at the inlet and the two outlets where the flow would enter

the well-head. From Figures 5-9 to 5-11 it can be seen that the inlet flowrate profile from the

model almost perfectly matches that of the Quest pipeline in all three cases. The inlet flowrate

was one of the two controlled variables, the second being the outlet to Lateral002 which also

show a close match between the model and Quest. The dependent variable when observing

the flowrate change is at the outlet to Lateral001. Figure 5-9 shows that the model gives a

flowrate profile for the outlet of Lateral001 close to that of Quest however it can be observed

that the reaction time of the model is quicker than that of the actual pipeline whereby the

flowrate in the model drops at a faster rate. Figure 5-10 shows that the initial drop in flowrate

to Lateral001 is modelled tightly with the Quest data however as the model settles at an

approximate flowrate of 37000kg/h the Quest data shows a further decrease and settles at

approximately 28000kg/h. The flowrate then increases and settles at the model value of

37000kg/h.

123
 140000
Variable
Pipeline Inlet flowrate
Flowrate to Well 1
120000 Flowrate to Well 2
Model Pipeline Inlet flowrate
Model Flowrate to Well 1
Model Flowrate to Well 2
100000
Flowrate (kg/h)

80000

60000

40000

20000
26-Oct-15 27-Oct-15 28-Oct-15
Time (h)

Figure 5-9: Model and QUEST pipeline inlet and outlet flowrates 24/10/15 – 28/10/15

140000 Variable
Pipeline Inlet flowrate
Flowrate to Well 1
Flowrate to Well 2
120000
Model Pipeline Inlet flowrate 1
Model Flowrate to Well 1
Model Flowrate to Well 2
100000
Flowrate (kg/h)

80000

60000

40000

20000
02-Nov-15 03-Nov-15 04-Nov-15
Time (h)

Figure 5-10: Model and QUEST pipeline inlet and outlet flowrates 31/10/15 – 04/11/15

124
 150000
Variable
Pipeline Inlet flowrate
Flowrate to Well 1
Flowrate to Well 2
Model Pipeline Inlet flowrate 1
125000
Model Flowrate to Well 1
Model Flowrate to Well 2
Flowrate (kg/h)

100000

75000

50000

09-Nov-15 10-Nov-15 11-Nov-15

Time (h)

Figure 5-11: Model and QUEST pipeline inlet and outlet flowrates 08/11/15 – 11/11/15

To compare the inlet flowrate change with the outlet flowrate dynamics, the outlet flowrates

to both wells were summed for both the model and the Quest data. Figures 5-12 to 5-14 give

a greater indication on how the outlet flowrate responds to a change at the inlet. An analysis

of this shows that there is a time delay between in the decrease in flowrate from the inlet to

the outlet. There is also a difference in the gradient whereby the inlet flowrate drops at a

faster rate than the outlet flowrate. Taking a point at which the inlet flowrate has reached

98186 kg/hr at time (25-Oct-2015 11:30:00) the outlet flowrate reaches a flowrate of

98299kg/hr at time (25-Oct-2015 14:30:00). This is a difference of 3 hours for the flowrate to

propagate through the pipeline.

The model shows a shorter time for the flowrate to initially propagate through the system

compared to the outlet flowrate of the Quest pipeline however an offset between the model

125
and the real value is observed between times (25-Oct-15 14:50:00) and (26-Oct-15 20:00:00)

at which point the Quest outlet flowrate has reached and settled at the inlet flowrate.

130000
Variable
Pipeline Inlet Flowrate
Total Outlet Flowrate
125000

120000
Flowrate (kg/h)

115000

110000

105000

100000

95000

25-Oct-15 26-Oct-15 27-Oct-15 28-Oct-15

Time (h)

Figure 5-12: Quest pipeline inlet and outlet flowrate 24/10/15 - 28/10/15

126
 1 40000 Variable
Pipeline Inlet Flowrate
Total Outlet Flowrate

1 30000
Flowrate (kg/h)

1 20000

1 1 0000

1 00000

90000
31 -Oct-1 5 01 -Nov-1 5 02-Nov-1 5 03-Nov-1 5 04-Nov-1 5
Time (h)

Figure 5-13: Quest pipeline inlet and outlet flowrate 31/10/15 - 04/11/15

145000
Variable
Pipeline Inlet Flowrate
Total Outlet Flowrate
140000

135000
Flowrate (kg/h)

130000

125000

120000

115000

110000
08-Nov-15 09-Nov-15 10-Nov-15 11-Nov-15
Time (h)

Figure 5-14: Quest pipeline inlet and outlet flowrate 08/11/15 - 11/11/15

127
5.4.2 Pressure

Along the Quest pipeline there are pressure sensors located at the inlet to the pipeline, on

either side of each LBV valve and at the well sites. The only pressure specification that was

implemented was the pressure to well head 1 which was set at 87,000kPa. All other values

of pressure were calculated by gCCS. Figure 5-15 to 5-17 shows the Quest pipeline

pressures and the model determined pressures. For ease of observation only, the pressures

downstream of each of the valves are shown. An initial observation shows that the model

predicted pressures upstream have greater agreement with that of the Quest pipeline when

compared with that of the pressures downstream. Figure 5-16, 5-19 and 5-22 show the first

10 data points of figures 5-15, 5-18 and 5-21 respectively. These charts show the pressures

at the pipeline inlet, upstream of LBV2 and upstream of LBV5 and provide clearer images

for more detailed comparison of how the pressure changes along the ppeline. From observing

how the pressures change over time, the initial pressure drop gives a tight relationship

between the model and the observed data. When the pressure has reached the minimum the

model and the observed data start to deviate with the model giving a constant slight decline

while the pressure at Quest shows a staggered increase. The largest difference between the

model and the data is at the point when the pressure starts to increase, these can be seen

in more detail in Figures 5-17, 5-20 and 5-23. The third significant observation from Figure 5-

15, where the observed pressure has increased to a maximum and slowly starts to decrease

while the predicted pressure shows a steady increase.

The lower accuracy of the model the further downstream of the pipeline is caused by the

model calculating larger pressure drops compared to the observed data. This can be

explained through the differences of the pressure drop through the valves, as any differences

will accumulate and cause greater deviations the further along the pipeline. The discrepancy

128
between the model and the Quest data over the time period may be caused by actions taken

by the pipeline operator trying to increase the pressure of the fluid.

9500
Variable
Inlet Pressure
9400 LBV1 Upstream Pressure
LBV2 Upstream Pressure
LBV4 Upstream Pressure
9300
LBV5 Upstream Pressure
Model Inlet Pressure
9200 Model LBV1 Upstream Pressure
Pressure (kPa)

Model LBV2 Upstream Pressure

9100 Model LBV3 Upstream Pressure
Model LBV4 Upstream Pressure
Model LBV5 Upstream Pressure
9000

8900

8800

8700

8600
26-Oct-15 27-Oct-15 28-Oct-15
Time (h)

Figure 5-15: Model and QUEST pipeline pressures 24/10/15 – 28/10/15

129
 9500 Variable
Inlet Pressure
9400 LBV2 Upstream Pressure
LBV5 Upstream Pressure
9300 Model Inlet Pressure
Model LBV2 Upstream Pressure
Pressure (kPa)

9200 Model LBV5 Upstream Pressure

91 00

9000

8900

8800

8700
0 0 0 0 0 0 0
:0 :0 :0 :0 :0 :0 :0
:00 :1
5
:30 :4
5
:00 15 30
0 9 09 0 9 09 1 0 1 0: 1 0 :

- 15 -1
5
- 15 -1
5
- 15 -1
5
- 15
ct ct ct ct ct ct ct
-O - O - O - O - O - O - O
25 25 25 25 25 25 25
Time (h)

Figure 5-16: Model and QUEST pipeline pressures 25/10/15 09:00 - 25/10/15 10:30

9500 Variable
Inlet Pressure
9400 LBV2 Upstream Pressure
LBV5 Upstream Pressure
9300 Model Inlet Pressure
Model LBV2 Upstream Pressure
Model LBV5 Upstream Pressure
Pressure (kPa)

9200

91 00

9000

8900

8800

8700
0 0 0 0 0 0 0
:0 :0 :0 :0 :0 :0 :0
12 16 20 00 04 08 12
5 5 5 5 5 5 5
-1 -1 -1 -1 -1 -1 -1
ct ct ct ct ct ct ct
6-O - O -O -O - O -O - O
2 26 26 27 27 2 7 27
Time (h)

Figure 5-17: Model and QUEST pipeline pressures 26/10/15 12:00 - 27/10/15 12:00

130
 Variable
10000 Inlet Pressure
LBV1 Upstream Pressure
LBV2 Upstream Pressure
9750 LBV4 Upstream Pressure
LBV5 Upstream Pressure
Model Inlet Pressure
Model LBV1 Upstream Pressure
9500
Pressure (kPa)

Model LBV2 Upstream Pressure

Model LBV3 Upstream Pressure
Model LBV4 Upstream Pressure
9250 Model LBV5 Upstream Pressure

9000

8750

8500
02-Nov-15 03-Nov-15 04-Nov-15
Time (h)

Figure 5-18: Model and QUEST pipeline pressures 31/10/15 – 04/11/15

9500 Variable
Inlet Pressure
9400 LBV2 Upstream Pressure
LBV5 Upstream Pressure
9300 Model Inlet Pressure
Model LBV2 Upstream Pressure
Pressure (kPa)

9200 Model LBV5 Upstream Pressure

91 00

9000

8900

8800

8700

8600
0 0 0 0 0 0 0
:0 :0 :0 :0 :0 :0 :0
:00 :15 :30 :45 : 00 :15 :30
0 5 05 05 05 06 06 06
1 5 1 5 1 5 1 5 1 5 1 5 1 5
- - - - - - -
ov ov ov ov ov ov ov
- N - N - N - N - N - N - N
01 01 01 01 01 01 01
Time (h)

Figure 5-19: Model and QUEST pipeline pressures 01/11/15 05:00 – 01/11/15 06:30

131
 Variable
9800 Inlet Pressure
LBV2 Upstream Pressure
LBV5 Upstream Pressure
9600 Model Inlet Pressure
Model LBV2 Upstream Pressure
Model LBV5 Upstream Pressure
Pressure (kPa)

9400

9200

9000

8800

8600
0 0 0 0 0 0 0 0
:0 :0 :0 :0 :0 :0 :0 :0
14 15 16 17 18 19 20 21
5 5 5 5 5 5 5 5
-1 -1 -1 -1 -1 -1 -1 -1
ov ov ov ov ov ov ov ov
-N -N -N - N - N - N - N - N
02 02 02 02 02 02 02 02
Time (h)

Figure 5-20: Model and QUEST pipeline pressures 02/10/15 14:00 – 02/11/15 21:00

Variable
9800 Inlet Pressure
LBV1 Upstream Pressure
LBV2 Upstream Pressure
LBV4 Upstream Pressure
9600
LBV5 Upstream Pressure
Model Inlet Pressure
Model LBV1 Upstream Pressure
Pressure (kPa)

9400 Model LBV2 Upstream Pressure

Model LBV3 Upstream Pressure
Model LBV4 Upstream Pressure
Model LBV5 Upstream Pressure
9200

9000

8800

09-Nov-15 10-Nov-15 11-Nov-15

Time (h)

Figure 5-21: Model and QUEST pipeline pressures 08/11/15 – 11/11/15

132
 9700 Variable
Inlet Pressure
9600 LBV2 Upstream Pressure
LBV5 Upstream Pressure
9500
Model Inlet Pressure
Model LBV2 Upstream Pressure
9400
Pressure (kPa)

Model LBV5 Upstream Pressure

9300

9200

91 00

9000

8900

8800
0 0 0 0 0 0 0
:0 :0 :0 :0 :0:0 :0
:30 :45 00 30 :1
5 45 00
1 3 1 1 3 4: 14
:
14 14
:
15
:

- 15 - 15 - 15 - 15 - 15 - 15 - 15
ov ov ov ov ov ov ov
- N - N - N - N - N - N - N
08 08 08 08 08 08 08
Time (h)

Figure 5-22: Model and QUEST pipeline pressures 08/11/15 13:30 – 08/11/15 15:00

Variable
9800 Inlet Pressure
LBV2 Upstream Pressure
LBV5 Upstream Pressure
9600 Model Inlet Pressure
Model LBV2 Upstream Pressure
Model LBV5 Upstream Pressure
Pressure (kPa)

9400

9200

9000

8800

0 0 0 0 0 0
:0 :0 :0 :0 :0 :0
14 16 18 20 22 00
5 5 5 5 5 5
-1 -1 -1 -1 -1 -1
ov ov ov ov ov ov
- N -N - N -N -N -N
09 09 09 09 09 10
Time (h)

Figure 5-23: Model and QUEST pipeline pressures 09/11/15 14:00 – 10/11/15 00:00

133
5.4.3 Temperature

Along the pipeline there is either a rise or a fall in temperature of the fluid depending on the

temperature difference between the fluid and the surrounding material. Since the temperature

at the pipeline inlet is greater than the ambient temperature of the surrounding soil, the

temperature of the fluid would decrease over the pipeline length. Figures 5-18 to 5-20 shows

the model and the Quest temperatures at the inlet to the pipeline and before each well. The

inlet temperature remains constant throughout the time period and is one of the variables that

is decided by the user, hence a high degree of accuracy between the model and the observed

data. When comparing the temperatures from before the wells it is clear that the model and

the Quest data have a high degree of disparity. Figures 5-18 to 5-20 show that the model

predicts a smaller temperature drop along the pipeline to the recorded value, with the model

predicting values of the temperature at the outlet of the pipeline greater than 25oC and the

data from the QUEST pipeline showing temperatures lower than 20 oC. Since the model

calculates the surrounding soil temperature based on the ambient temperature, pipeline

depth and soil type it is a reasonable assumption that the reason behind the discrepancy is

the difference in the temperature gradient along the soil depth. This determines the

temperature of the soil surrounding the pipeline and hence the driving force for heat loss

along the pipeline. This is likely to be one of the most difficult parameters to determine as the

soil thermal properties can vary significantly. Previous studies have shown that there are

several factors that affect the thermal properties of the soil. These include the composition,

the volume, soil density, porosity and water migration. The main components of soil can be

seen in Table 5-6[75] along with their thermal conductivities.

134
 Table 5-6: Soil Material & Thermal Conductivities

Material Thermal Conductivity (W m-1 K-1)

Quartz 8.4

Soil Minerals* 2.9

Soil Organic Matter* 0.25

Water 0.6

Air 0.026

*Approximate Average Values

From Table 5-6 it becomes apparent that the composition of the soil can have significant

impacts on the total thermal conductivity of the soil, with the thermal conductivity of the

individual components ranging from 0.026 W m-1K-1 to 8.4 W m-1K-1. The model however has

pre-determined values for the thermal conductivity related to each soil type which cannot be

changed. This is one of the limitations of the modelling software and is an improvement that

would allow the model to have greater accuracy in determining the temperature losses along

the pipeline.

135
 45
Variable
Quest Inlet Temp
Quest Branch 1 Outlet Temp
40 Quest Branch 2 Outlet Temp
Model Inlet Temp
Model Branch 1 Outlet Temp
35 Model Branch 2 Outlet Temp
Temperature (oC)

30

25

20

15

10
26-Oct-15 27-Oct-15 28-Oct-15
Time (h)

Figure 5-24: Model and QUEST pipeline inlet and outlet temperatures 24/10/15 – 28/10/15

45
Variable
Quest Inlet Temp
Quest Branch 1 Outlet Temp
40 Quest Branch 2 Outlet Temp
Model Inlet Temp
Model Branch 1 Outlet Temp
35 Model Branch 2 Outlet Temp
Temperature (oC)

30

25

20

15

10
02-Nov-15 03-Nov-15 04-Nov-15
Time (h)

Figure 5-25: Model and QUEST pipeline inlet and outlet temperatures 31/10/15 – 04/11/15

136
 45
Variable
Quest Inlet Temp
Quest Branch 1 Outlet Temp
40 Quest Branch 2 Outlet Temp
Model Inlet Temp
Model Branch 1 Outlet Temp
35 Model Branch 2 Outlet Temp
Temperature (oC)

30

25

20

15

10
09-Nov-15 10-Nov-15 11-Nov-15
Time (h)

Figure 5-26: Model and QUEST pipeline inlet and outlet temperatures 08/11/15 – 11/11/15

5.5 Determining ‘Goodness of Fit’

Goodness of fit is a way in which the accuracy of models can be measured quantitatively. It

compares how well the model is able to predict output values with changing input conditions.

There are different methods to determine goodness of fit, each with their own benefits and

limitations. An evaluation on different methods of ‘goodness of fit’ for time series data, shows

that the coefficient of efficiency and the Index of agreement are both methods that are

suitable to use to determine goodness of fit of the model. The parameters that will be tested

for goodness of fit are those that are determined by the model and not those that have been

user defined.

5.5.1 Coefficient of Efficiency

The coefficient of efficiency has a range between minus infinity to 1.0, with higher values

indicating better agreement. The CoE is given by the following formulae:

137
 ∑𝑁
𝑖=1(𝑂𝑖 − 𝑃𝑖 )
2
𝐸 = 1.0 − 𝑁
∑𝑖=1(𝑂𝑖 − 𝑂̅)2

Equation 5-2: Coefficient of Efficiency

The CoE is the ratio of the mean squared error to the ratio of the variance in the observed

data minus unity. If the mean squared error is the same as the variability in the observed data

then we get a value of E=0. A value of E < 0 indicates that the mean is a better predictor

than the model. A limitation of this method is that it can be sensitive to outliers due to the

square of the error.

5.5.2 Index of Agreement

The Index of Agreement (IoA) is given by the formula:

∑𝑁𝑖=1(𝑂𝑖 − 𝑃𝑖 )
2
𝑑 = 1.0 − 𝑁
∑𝑖=1(|𝑃𝑖 − 𝑂̅||𝑂𝑖 − 𝑂̅|)

Equation 5-3: Index of Agreement

The IoA varies from 0 to 1 with higher values indicating better agreement between the model

and the observations.

138
 Table 5-7: Goodness of fit

24th Oct – 28th Oct 31st Oct – 3rd Nov 7th Nov – 11th Nov Mean

Outlet Flowrate 001 0.8761 0.9612 0.9631 0.9335

Inlet Pressure 0.9130 0.4559 0.7050 0.6913

LBV1 UP 0.8680 0.3858 0.7152 0.6563

LBV2 UP 0.8100 0.2717 0.5009 0.5275

LBV3 UP 0.7732 0.2314 0.4453 0.4833

LBV4 UP 0.6017 0.1818 0.3309 0.3715

LBV5 UP 0.5126 0.1215 0.2265 0.2869

Outlet Temperature 0.003788 0.007534 0.03211 0.01448

001

Table 5-7 shows the index of agreement values for the determined parameters in the model.

The values for the outlet flowrate indicate that the model is able to determine the change in

the flowrate with high accuracy given that the average index of agreement for the three time

periods is 0.9335.

The goodness of fit for the inlet pressure shows that there is good agreement between the

model and the data with an average index of agreement of 0.6913. The index of agreement

for the pressures upstream of each valve show a decrease along the pipeline. This clearly

indicates that the model is failing to predict the pressure drop along the pipeline. This

discrepancy can be explained with the Darcy-Weisbach equation

139
 ∆𝑝 𝜌 𝑣2
= 𝑓𝐷 × ×
𝐿 2 𝐷

Equation 5-4: Darcy-Weisbach Equation

The Darcy-Weisbach equation shows that the pressure drop along a pipeline is dependent

on the density of the fluid. The density of the CO 2 is affected by the temperature of the fluid.

The model determined outlet temperature is shown to have a low index of agreement which

indicates that the density of the fluid in the model at this point would also show low agreement

with the data and therefore explain why the pressure drop along the pipeline deviates from

what is occurring.

5.6 Conclusion

The development of the Shell Quest pipeline within gCCS has allowed for the specific gCCS

CO2 pipeline model to be compared to real industrial data. This has not been found in any of

the current literature and is novel to this research. The pipeline was created in gCCS using

the specific pipeline models, the use of historical pipeline data allowed for three scenarios to

be simulated that covered three time periods in which the inlet flowrate of the CO 2 to the

pipeline varied. The outputs from the model were then compared to the data, looking at

temperature, pressure and flowrate. To analyse the differences between the model and the

data a goodness of fit was determined through calculating the Index of Agreement. The Index

of Agreement showed that the model was good at determining the outlet flowrate of the

pipeline with a consistently high value across all three time periods that were modelled. The

inlet pressure to the pipeline was also shown to have good agreement between the model

and the data. The analysis did show that along the length of the pipeline the model deviates

from the data with the Index of Agreement getting lower. There are two reasons for the model

predictions of the pressure getting worse along the length of the pipeline, it could be either

due to the temperature losses in the model along the pipeline being much lower than the real

140
data. This difference in fluid temperature effects the density of the CO 2 and therefore impacts

on the pressure drop along the pipeline. The second possible reason for the pressure

discrepancies is the difference between the friction factor in the model and the actual QUEST

pipeline. Through this validation process it can be concluded that the gCCS model struggles

to predict the temperature losses along the pipeline which has repercussions for determining

the pressure drop. It has been argued that the reason behind this significant temperature

difference between the model and the data is due to the thermal conductivity of the soil, as

the model has predetermined values for each of the soil options. The accuracy of the model

outputs could be improved by allowing the user to define the thermal conductivities of the soil

which could be analytically determined for each case that is modelled.

141
Chapter 6 – Conclusions and Future Work

142
6.1 Conclusions

The aims of the research presented in the thesis were to investigate the effects of changing

the inlet flowrate of CO2 into a pipeline for the purposes of carbon capture and storage, which

would allow a greater understanding how a CCS transport system operates when attached

to a load following power plant with CCS. The method to understand how the transport

system responds when a step change occurs was carried out using the modelling tool gCCS,

which was evaluated and concluded to be the most appropriate tool for the requirements of

the research. The first stage of the modelling was to develop a base case scenario in which

the outputs from more detailed scenarios could be compared. The second stage of the

modelling was to look at the impacts of two important developments in the scenario; the first

investigated the effects of impurities in the CO2 on the dynamics of the model, the second

scenario that was modelled examined the effects of a system with multiple sources of CO2.

The final stage of the research was comparing the model outputs to the data from a real CCS

pipeline in Canada, this allowed for a greater examination of the accuracy of the modelling

tool.

6.1.1 Base case scenario

The base case scenario was set up in gCCS to investigate the effects of reducing the inlet

flowrate from a single source of pure CO2 on the pipeline dynamics. The following hypothesis

was tested;

‘The gCCS model will show that the rate of change in the outlet flowrate of a CO2 pipeline

when the inlet flowrate is reduced, will be greater when the CO2 is transported in the

subcooled liquid phase compared to the supercritical phase’

143
Through analysis of the flowrate, temperature and pressure within the pipeline certain

conclusions could be made about the modelling tools understanding of CO 2 pipeline

transport:

• The effect of reducing the inlet flowrate to the pipeline was shown to have an impact

on the pipeline outlet flowrate, with three distinct phases in the response being

observed. The three phases have been to referred to as the ‘delayed response

phase’, the ‘offset phase’ and the ‘final reduction phase’.

• The hypothesis was demonstrated to be correct and the phase of the CO2 was shown

to impact the response of the fluid within the pipeline when the inlet flowrate is

reduced. CO2 in the supercritical phase was shown to have a greater impact on the

dynamics of the fluid in the pipeline when compared to CO2 in the liquid phase.

• CO2 in the supercritical phase was shown to cause a larger offset between the inlet

and the outlet flowrate during the offset phase of the response to a change in the inlet

flowrate.

• The modelling shows that there is a drop in the pressure of the CO 2 at the inlet and

outlet of the pipeline when there is a reduction in the flowrate. The pressure change

did not follow the same profile as the flowrate as there was no distinct phases during

the fall in the flowrate with the outlet pressure drop following the same profile as the

inlet pressure drop. When comparing the pressure profiles of supercritical and liquid

phase CO2 it was found that the inlet pipeline pressure for supercritical CO 2 was

higher than when in the liquid phase. This was explained through the Darcy-Weisbach

equation which shows that the flow velocity has a greater impact on the pressure drop

than the density.

144
 • The temperature at the outlet of the pipeline dropped sharply by 6 oC when the flowrate

was reduced but then increased during the ‘final reduction phase’ with a sharp

gradient to the temperature before the flowrate change. The results from the

simulation indicate that the temperature and flowrate are closely related and have

been explained by the continuity equation however there is also the possibility that

the results are a peculiarity of the modelling tool and the interactions between the

underlying equations.

Overall the results from the simulations have shown that there should be no real technical

difficulties between transporting CO2 in the liquid phase or transporting it in the supercritical

phase and that the decision on which phase would be preferred is down to economical

considerations. However in cases where CO2 is being bought or sold such as for enhanced

oil recovery there may need to be some consideration given to the difference in flowrates

between the inlet and the outlet of the pipeline when there is a step change in the flowrate as

there may be discrepancies between the quantity of CO 2 that is thought to have been

provided and what was actually provided .

6.1.1 CO2 with impurities scenario

Developing the scenario from the base case, the impact of different impurities in the CO 2 on

the flow dynamics was investigated. Three impurities were studied, that are known to be

present in the three main types of CO2 capture technology. These impurities are; nitrogen,

hydrogen and oxygen. To allow for a comparison with the base case results the composition

of the fluid entering the pipeline was the only parameter that was changed, all other

parameters in the model where kept constant. The hypothesis that was tested in this part of

the research was as follows:

145
‘The gCCS modelling tool will demonstrate that impurities within the carbon dioxide

transported for CCS will cause a different response in the flowrate to changing the inlet

flowrate compared to when transporting pure CO2’

Through the analysis of the results from the model the following conclusions were made

regarding this scenario;

• The hypothesis was proven to be correct, in that the addition of any three of the

impurities caused an observable difference in the response of the outlet flowrate

when the inlet flowrate was reduced.

• The difference between the base case scenario and the scenario when impurities

were introduced to the system was the size of the offset during the ‘offset phase’. The

addition of any of the impurities was shown to cause an increase in the offset between

the inlet flowrate and the outlet flowrate. This phenomena could be explained with the

effect of these impurities on the density of the CO2 and therefore impacting on the

flowrate at which the flow wave propagates through the pipeline. This is supported by

the fact that there was a larger offset when CO2 was transported in the supercritical

phase, which also has a lower density than liquid CO2.

• When comparing the three impurities against each other it was found that hydrogen

had the greatest impact on the offset, while oxygen had the least impact on the size

of the offset when compared to the base case. It is believed that the reason for this

difference is due to the difference in the molecular mass when compared to carbon

dioxide. Hydrogen is the most different from carbon dioxide while oxygen is the most

similar.

146
6.1.1 Multiple sources of CO2 scenario

The final scenario that was modelled investigated the effects of two sources of CO2 into a

single pipeline. This scenario was investigated as it is expected that hubs of CO 2 sources will

develop in the future which will eventually share transport infrastructure. If there is variability

in the CO2 output from any of these sources it is necessary to understand how this might

impact on the entire transport infrastructure. For this scenario the following hypothesis was

tested:

‘The gCCS modelling tool will show that that varying the flowrate of one of two sources of

CO2 will have a different effect on the flowrate of the CO2 within the trunk pipeline, compared

to when there is only a single source of CO2’

The analysis of the outputs from the modelling in gCCS resulted in the following conclusions

regarding this scenario;

• The results from the modelling show that the hypothesis was incorrect and that there

is no difference in the response of the flowrate when there are multiple sources of CO 2

then when there is a single source of CO2 entering a pipeline.

• The flowrate profile at the inlet and the outlet of the trunk pipeline is the same as the

base case scenario as the flowrate at the outlet of the branch pipeline has the same

profile as the setpoint change. The difference in the flowrate profile of the branch

pipeline and the trunk pipeline can be explained by the difference in either the pipeline

diameter, pipeline length or the fluid flowrate.

• The change in the flowrate of one of the sources didn’t cause any back flow in the

branch pipeline of the other CO2 source. However due to the method in which the

software calculates the pressure, the pressure in the flowrate source was forced to

reduce to ensure the reservoir pressure was stable. This is a limitation of the software

147
 and is therefore knowledge gained about the use of gCCS rather than how the pipeline

will actually respond when there are multiple sources of CO2.

6.1.1 Modelling of the Shell QUEST CO2 pipeline

Due to an arising opportunity during the research, real CO 2 pipeline data was able to be

obtained from the Shell QUEST CCS pipeline. A flowsheet was developed in gCCS that

represented the QUEST pipeline with the same pipeline dimensions, topology and fluid

parameters. Historical data was then used to develop a simulation that could then be directly

compared to the pipeline data. To evaluate how well the gCCS model predicts the variables

of interest a statistical value known as the Index of Agreement was used. The index of

agreement was determined for each variable of interest to understand the goodness of fit

between the model and the industrial data. The following conclusions were developed;

• The model was able to predict with a high degree of accuracy the outlet flowrates of

the pipeline with an average Index of agreement value of 0.9335. However,

observation of the QUEST pipeline data indicated that the three separate phases that

were predicted by the model do not occur.

• The model was able to predict the pressure at the inlet of the pipeline with a reasonable

degree of accuracy with a mean index of agreement of 0.6913. However the further

along the pipeline, the accuracy of the model got increasingly worse with a declining

index of agreement along the length of the pipeline. This indicates that the model has

difficulty in determining the pressure drop along the pipeline when there is a change

in flowrate at the inlet of the pipeline. The reason behind this difference can be

attributed the friction factor, from the Darcy-Weisbach equation the pressure drop

148
 along the pipeline is a function of the friction factor and is the only parameter in the

equation that there could be a discrepancy between the model and the actual pipeline.

• There was a significant difference between the pipeline data and the model predictions

when determining the temperature loss along the pipeline. The mean value for the

index of agreement was 0.01448, meaning that the model was not able to predict the

temperature at the outlet of the pipeline with any degree of accuracy. The reasoning

behind this discrepancy between the QUEST data and the model is that the thermal

conductivity of the soil surrounding the pipeline is understood to be variable depending

on the composition of the soil. Since the model does not allow the user to define the

thermal conductivity of the surrounding soil the model is limited and therefore cannot

take into account varying soil types, which can lead to differences in the heat loss

between the model and the actual pipeline being modelled.

6.2 Future Work

The research has shown that the phase of the CO 2 effects the way in which the CO2 in the

pipeline responds, that the presence of impurities effects the offset between the inlet and the

outlet of the pipeline and that multiple sources of CO2 does not impact on the way in which

the outlet flowrate reacts to a change in the inlet flowrate. There are two areas in which further

research needs to be conducted to improve the developments in CO 2 pipeline transportation.

The first being the scenarios which are modelled and the second the improvements to the

gCCS modelling tool.

6.2.1 Further scenarios analysis

The research carried out investigated three scenarios of CO2 pipeline transport. The first

scenario was developed a base case scenario which other scenarios could be compared to.

The second scenario investigated the effects of impurities in the CO2 on the flowrate

149
dynamics and the third scenario examined how multiple sources of CO 2 effect the flowrate.

Future work will take the scenarios further and investigate the impacts of other parameters

on the flowrate of CO2. The areas which should be investigated are those that further

represent real CO2 pipeline infrastructures, such as;

• Multiple sequestration sites. Within this study, the impact of multiple sources of CO2

on the flowrate within a pipeline were investigated however it is also anticipated that

there will be multiple wells which the CO2 is delivered to and is therefore an area of

research that would further inform the operation of CCS transport infrastructures.

• Varying pipeline diameters. Within the modelling, certain parameters were kept

constant throughout each scenario, this included the pipeline diameter. It is likely that

for different CCS projects different pipeline diameters will be required based upon

technical and economic factors. It is therefore necessary to develop a greater

understanding of how the pipeline diameter effects the flow of CO2 when there are

variable inlet flows.

• Offshore pipelines. For all scenarios investigated the modelling looked specifically at

onshore pipelines. Given that offshore storage of CO2 is most likely to be used within

the U.K. it will be essential to understand how CO2 flows within offshore pipelines given

that the surrounding material of the pipeline will be significantly different to onshore

pipelines which may affect the heat transfer of the CO2.

6.2.1 gCCS model development

150
From the validation work carried out with the support of the Shell QUEST project, suggestions

for improvements to the modelling tool gCCS are able to be made. Throughout the research

the temperature profile in the pipeline reacted the least intuitive and had the greatest

disagreement with the data from the QUEST pipeline compared to all other variables. It has

been suggested that the reason behind this is due to the thermal conductivity of the

surrounding material of the pipeline, which is pre-defined within the model. Allowing an extra

degree of freedom so that the user can define the thermal conductivity of the surrounding

material would allow for greater accuracy when the model attempts to determine the heat

loss or gain of the CO2. This could significantly improve the model in determining the

temperature of the CO2 for real CCS pipeline projects.

151
Appendix A Simulation Code
A.1 Code for Base Case and Impurities Case
SEQUENCE
CONTINUE FOR 30
REASSIGN
Flowsheet.Source_CO2001.F:=OLD(Flowsheet.Source_CO2001.F)-50*(Time-30)/750;
END
CONTINUE FOR 750
REASSIGN
Flowsheet.Source_CO2001.F:=OLD(Flowsheet.Source_CO2001.F);
END
CONTINUE FOR 500000
REASSIGN
Flowsheet.Source_CO2001.F:=OLD(Flowsheet.Source_CO2001.F);
END

152
Appendix B Simulation Code
B.1 Code for Multiple CO2 Sources Case
SEQUENCE
CONTINUE FOR 30
REASSIGN
Flowsheet.Source_CO2001.F:=OLD(Flowsheet.Source_CO2001.F)-25*(Time-30)/375;
END
CONTINUE FOR 375
REASSIGN
Flowsheet.Source_CO2001.F:=OLD(Flowsheet.Source_CO2001.F);
END
CONTINUE FOR 500000
REASSIGN
Flowsheet.Source_CO2001.F:=OLD(Flowsheet.Source_CO2001.F);
END

153
Appendix C Simulation Code
C.1 Code for Shell QUEST Simulation for Time Period 24/10/15 – 28/10/15
SCHEDULE
SEQUENCE
CONTINUE FOR 30
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) - 7.6*(TIME -
30)/9960;
END
CONTINUE FOR 9960
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) ;
END
CONTINUE FOR 122400
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) + 8.4*(TIME -
132390)/23880;
END
CONTINUE FOR 23880
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) ;
END
CONTINUE FOR 200000
END

154
 C.2 Code for Shell QUEST Simulation for Time Period 31/10/15 – 04/11/15
SCHEDULE
SEQUENCE
CONTINUE FOR 30
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) - 8*(TIME -
30)/7200;
END
CONTINUE FOR 7200
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) ;
END
CONTINUE FOR 109728
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) + 8*(TIME -
116958)/21600;
END
CONTINUE FOR 21600
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) ;
END
CONTINUE FOR 200000
END

155
 C.3 Code for Shell QUEST Simulation for Time Period 08/11/15 – 11/11/15
SEQUENCE
CONTINUE FOR 30
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) - 8.25*(TIME -
30)/18900;
END
CONTINUE FOR 18900
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) ;
END
CONTINUE FOR 68640
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) + 7.0024*(TIME -
87570)/25740;
END
CONTINUE FOR 24740
REASSIGN
Flowsheet.Source_CO2001.F := OLD(Flowsheet.Source_CO2001.F) ;
END
CONTINUE FOR 200000
END

156
References
1. Agency, I.E., Technology Roadmap: Carbon capture and storage. 2013.
2. Day, G., Building the UK carbon capture and storage sector by 2030 - Scenarios and actions. 2015,
Energy Technologies Institute.
3. J.T. Houton, G.J.J., J. J. Ephraums, Climate Change: The IPCC Scientific Assessment. 1990.
4. IPCC, Climate Change 2007: Synthesis Report. 2007, IPCC: Geneva, Swizerland.
5. Anderson, T.R., E. Hawkins, and P.D. Jones, CO2, the greenhouse effect and global warming: from the
pioneering work of Arrhenius and Callendar to today's Earth System Models. Endeavour, 2016. 40(3):
p. 178-187.
6. 2015 UK greenhouse gas emissions: final figures - statistical summary, E.I.S. Department for
Business, Editor. 2017.
7. Milne, S., Options Choices Actions - UK scenarios for a low carbon energy system. 2015, Energy
Technologies Institute.
8. Newton-Cross, G. and D. Gammer, The evidence for deploying bioenergy with CCS (BECCS) in the UK.
2016, Energy Technologies Institute.
9. Leung, D.Y.C., G. Caramanna, and M.M. Maroto-Valer, An overview of current status of carbon
dioxide capture and storage technologies. Renewable and Sustainable Energy Reviews, 2014. 39: p.
426-443.
10. Association, C.C.a.S. What is CCS? [cited 2016 5th November]; Available from:
http://www.ccsassociation.org/what-is-ccs/.
11. Intergovernmental Panel on Climate Change, Special Report on Carbon Dioxide Capture and Storage.
2005: Cambridge, UK.
12. Liang, Z., et al., Review on current advances, future challenges and consideration issues for post-
combustion CO2 capture using amine-based absorbents. Chinese Journal of Chemical Engineering,
2016. 24(2): p. 278-288.
13. IEAGHG, Assessment of Emerging CO2 Capture Technologies and their Potential to Reduce Costs.
2014.
14. Global CCS Institute, The Global Status of CCS: 2011. 2011: Canberra, Australia.
15. Wang, M., et al., Post-combustion CO2 capture with chemical absorption: A state-of-the-art review.
Chemical Engineering Research and Design, 2011. 89(9): p. 1609-1624.
16. Jansen, D., et al., Pre-combustion CO2 capture. International Journal of Greenhouse Gas Control,
2015. 40: p. 167-187.
17. Uchida, T., et al., Oxyfuel Combustion as CO2 Capture Technology Advancing for Practical use -
callide Oxyfuel Project. Energy Procedia, 2013. 37: p. 1471-1479.
18. Modekurti, S., et al., Design, dynamic modeling, and control of a multistage CO2 compression
system. International Journal of Greenhouse Gas Control, 2017. 62: p. 31-45.
19. Pei, P., et al., Waste heat recovery in CO2 compression. International Journal of Greenhouse Gas
Control, 2014. 30: p. 86-96.
20. Institute, G.C. Large-scale CCS facilities. [cited 2016 10th July]; Available from:
https://www.globalccsinstitute.com/projects/large-scale-ccs-projects.
21. SaskPower. Boundary Dam Carbon Capture Project. [cited 2016 23rd July]; Available from:
http://www.saskpower.com/our-power-future/carbon-capture-and-storage/boundary-dam-carbon-
capture-project/.
22. Energy, O.o.F. Petra Nova - W.A. Parish Project. [cited 2016 23rd July]; Available from:
https://energy.gov/fe/petra-nova-wa-parish-project.
23. Paul, S., R. Shepherd, and P. Woolin, Selection of materials for high pressure CO2 transport, in Third
International Forum on the Transportation of CO2 by Pipeline. 2012: Newcastle, UK.
24. Aspelund, A., M.J. Mølnvik, and G. De Koeijer, Ship Transport of CO2. Chemical Engineering Research
and Design, 2006. 84(9): p. 847-855.
25. platform, Z.e., The Costs of CO2 Capture, Transport and Storage.

157
26. Svensson, R., et al., Transportation systems for CO2––application to carbon capture and storage.
Energy Conversion and Management, 2004. 45(15): p. 2343-2353.
27. Technology, M.I.o. Peterhead Project Fact Sheet: Carbon Capture and Storage Project. [cited 2016
28th July]; Available from: https://sequestration.mit.edu/tools/projects/peterhead.html.
28. McCoy, S.T. and E.S. Rubin, An engineering-economic model of pipeline transport of CO2 with
application to carbon capture and storage. International Journal of Greenhouse Gas Control, 2008.
2(2): p. 219-229.
29. GHG), I.G.G.R.D.P.I., Upgraded calculator for CO2 pipeline systems. 2009.
30. Noothout, P., et al., CO2 Pipeline Infrastructure – Lessons Learnt. Energy Procedia, 2014. 63: p. 2481-
2492.
31. Ghazi, N. and J.M. Race, Techno-economic modelling and analysis of CO2 pipelines. The Journal of
Pipeline Engineering, 2013. 12(2): p. 83-92.
32. Institute, G.C., IEAGHG, Building the cost curves for CO2 storage; European Sector. 2005.
33. Ecofys, Global carbon dioxide storage potential and costs. 2004.
34. IEAGHG, Transmission of CO2 and Energy. 2002.
35. Brown, S., et al., Modelling the non-equilibrium two-phase flow during depressurisation of CO2
pipelines. International Journal of Greenhouse Gas Control, 2014. 30: p. 9-18.
36. Martynov, S., et al., Modelling three-phase releases of carbon dioxide from high-pressure pipelines.
Process Safety and Environmental Protection, 2014. 92(1): p. 36-46.
37. M. Bilio, S.B., M. Fairweather, H Mahgerefteh, CO2 pipelines material and safety considerations, in
IChemE Symposium. 2009, Institute of Chemical Engineers.
38. Ibrahim, M. and I.S. Ertesvåg, PVTx Modeling of CO2 Pipeline at Depressurization Conditions Using
SPUNG Equation of State (EoS) with a Comparison to SRK. Energy Procedia, 2014. 63: p. 2467-2474.
39. Clausen, S., A. Oosterkamp, and K.L. Strøm, Depressurization of a 50 km Long 24 inches CO2 Pipeline.
Energy Procedia, 2012. 23: p. 256-265.
40. Kaufmann, D.K.-D., Carbon Dioxide Transport in Pipelines - Under Special Consideration of Safety-
Related Aspects, in 3rd Pipeline Technology Conference 2008. 2008: Hannover, Germany.
41. Zhang, Z.X., et al., Optimization of pipeline transport for CO2 sequestration. Energy Conversion and
Management, 2006. 47(6): p. 702-715.
42. Witkowski. A, R.A., Majkut. M, Stolecka. K, The Analysis of Pipeline Transportation Process for CO2
Captured From Reference Coal-Fired 900 MW Power Plant to Sequestration Region. Chemical and
Process Engineering, 2014. 35: p. 497-514.
43. Witkowski, A. and M. Majkut, Analysis of Transportation Systems for CO2 Sequestration, in Advances
in Carbon Dioxide Compression and Pipeline Transportation Processes. 2015, Springer International
Publishing: Cham. p. 73-93.
44. Nimtz, M., et al., Modelling of the CO2 process- and transport chain in CCS systems—Examination of
transport and storage processes. Chemie der Erde - Geochemistry, 2010. 70: p. 185-192.
45. Kunz, O. and G. Groupe Européen de Recherches, The GERG-2004 wide-range equation of state for
natural gases and other mixtures. 2007, Düsseldorf: VDI Verlag.
46. Liljemark, S., et al., Dynamic simulation of a carbon dioxide transfer pipeline for analysis of normal
operation and failure modes. Energy Procedia, 2011. 4: p. 3040-3047.
47. Luo, X., et al., Simulation-based techno-economic evaluation for optimal design of CO2 transport
pipeline network. Applied Energy, 2014. 132: p. 610-620.
48. Technology, M.I.o. Don Valley Power Project Fact Sheet: Carbon Dioxide Capture and Storage Project.
[cited 2016 9th August]; Available from:
https://sequestration.mit.edu/tools/projects/don_valley.html.
49. Technology, M.I.o. White Rose* Project Fact Sheet: Carbon Dioxide Capture and Storage Project
(*Formerly UK Oxy CCS Project). Available from:
https://sequestration.mit.edu/tools/projects/white_rose.html.
50. Chapoy, A., et al., Effect of impurities on thermophysical properties and phase behaviour of a CO2-
rich system in CCS. International Journal of Greenhouse Gas Control, 2013. 19: p. 92-100.

158
51. Cole, I.S., et al., Corrosion of pipelines used for CO2 transport in CCS: Is it a real problem?
International Journal of Greenhouse Gas Control, 2011. 5(4): p. 749-756.
52. Dugstad, A., M. Halseid, and B. Morland, Experimental Techniques used for Corrosion Testing in
Dense Phase CO2 with Flue Gas Impurities, in CORROSION 2014. 2014, NACE International: San
Antonio, Texas, USA.
53. Technology, I.f.E. The history of OLGA. [cited 2016 9th August]; Available from:
https://www.ife.no/en/ife/departments/process_and_fluid_flow_tech/historienomolga.
54. Schlumberger. OLGA Dynamic Multiphase Flow Simulator. [cited 2016 11th August]; Available from:
https://www.software.slb.com/products/olga.
55. Shlumberger. Case Study: BP Saves 12 Days of Downtime and 6,000-m3 Diesel Managing Deepwater
Pipeline with the OLGA Simulator. [cited 2016 16th August]; Available from:
http://www.slb.com/resources/case_studies/software/cs_bp_olga_deepwater_pipeline.aspx.
56. Span, R. and W. Wagner, A New Equation of State for Carbon Dioxide Covering the Fluid Region from
the Triple-Point Temperature to 1100 K at Pressures up to 800 MPa. Journal of Physical and Chemical
Reference Data, 1996: p. Medium: X; Size: pp. 1509-1596.
57. Giljarhus, K.E.T., S.T. Munkejord, and G. Skaugen, Solution of the Span–Wagner Equation of State
Using a Density–Energy State Function for Fluid-Dynamic Simulation of Carbon Dioxide. Industrial &
Engineering Chemistry Research, 2012. 51(2): p. 1006-1014.
58. Leachman, J., et al., Thermodynamic Properties of Cryogenic Fluids. 2 ed. International Cryogenics
Monograph Series. 2017: Springer International Publishing. 213.
59. Witkowski. A, M.M., Rulik. S, Analysis of pipeline transportation systems for carbon dioxide
sequestration. Archives of Thermodynamics 2014. 35(1): p. 117-140.
60. L, R., CO2 Transportation with Pipelines - Model Analysis for Steady, Dynamic and Relief Simulation.
The Italian Association of Chemical Engineering, 2014. 36: p. 619-624.
61. Demetriades, T.A. and R.S. Graham, A new equation of state for CCS pipeline transport: Calibration of
mixing rules for binary mixtures of CO2 with N2, O2 and H2. The Journal of Chemical
Thermodynamics, 2016. 93: p. 294-304.
62. Fandiño, O., J.P.M. Trusler, and D. Vega-Maza, Phase behavior of (CO2+H2) and (CO2+N2) at
temperatures between (218.15 and 303.15)K at pressures up to 15MPa. International Journal of
Greenhouse Gas Control, 2015. 36: p. 78-92.
63. Economou, I.G., Statistical Associating Fluid Theory: A Successful Model for the Calculation of
Thermodynamic and Phase Equilibrium Properties of Complex Fluid Mixtures. Industrial &
Engineering Chemistry Research, 2002. 41(5): p. 953-962.
64. Diamantonis. N, E.I., Evaluation of SAFT and PC-SAFT EoS for the calculation of thermodynamic
derivative properties of fluids related to carbon capture and sequestration, CO2PipeHaz, Editor.
2011: Energy & Fuels.
65. Islam, A.W. and E.S. Carlson, Application of SAFT equation for CO2+H2O phase equilibrium
calculations over a wide temperature and pressure range. Fluid Phase Equilibria, 2012. 321: p. 17-24.
66. Diamantonis, N.I., et al., Evaluation of Cubic, SAFT, and PC-SAFT Equations of State for the Vapor–
Liquid Equilibrium Modeling of CO2 Mixtures with Other Gases. Industrial & Engineering Chemistry
Research 2013. 52(10): p. 3933-3942.
67. Florides, G. and S. Kalogirou, Measurements of Ground Temperatures at Various Depths. 2004.
68. Johansen, O., Thermal Conductivity of Soils. 1977, U.S. Army Cold Regions Research and Engineering
Laboratory.
69. Bratland, O., Single-phase Flow Assurance. 2009. 341.
70. Institute, E., Good Plant Design And Operation For Onshore Carbon Capture Installations And
Onshore Pipelines. 2010, Energy Institute London.
71. Peplinski, S. QUEST CCS PROJECT. in PCCC2 Conference. 2013. Bergen, Norway.
72. Peters, D., Pipelines Flow and Flow Assurance Report - Final. 2014.
73. Climate: FORT SASKATCHEWAN. 2015 [cited 2015 November 19]; Available from: https://en.climate-
data.org/location/949/.

159
74. Endress+Hauser. Absolute and gauge pressure Cerabar PMP71. [cited 2016 March 23]; Available
from: https://www.uk.endress.com/en/Field-instruments-overview/pressure/Absolute-Gauge-
Cerabar-PMP71.
75. Farouki, O.T., Thermal properties of soil. 1981, U.S. Army Cold Regions Research and Engineering
Laboratory.

160