SYMPOSIUM SERIES NO.

156 Hazards XXII # 2011 IChemE

CO2 ENHANCED OIL RECOVERY: PROCESS DESIGN SAFETY OF CRITICAL

PARAMETERS

Dr Azzam Younes, ABB Engineering Services, Aberdeen, UK

Carbon abatement technologies are known as Carbon Capture and Storage (CCS), in which the
carbon in fossil fuels is captured (as CO2) either pre-combustion or post-combustion and committed
to long-term storage so that it is not emitted to the atmosphere and does not therefore contribute to
global warming. One option for CO2 storage is through the use of carbon dioxide in enhanced oil
recovery (EOR) techniques. The aim of this paper is to describe the process of CO2 EOR, and the
role safety engineering input during the design, furthermore the paper presents the most important
aspects regarding the safety parameters, which need to be taken into account during the design
stage. It also discusses how Safety engineers have an important role to play throughout the CO2
injection design. A key area that will be discussed is the potential of the combined Flammability,
Detection and Toxicity Analysis through out the CO2 EOR life field on the existing platform moni-
tors and safety of personnel.

KEYWORDS: CO2 injection, Safety, Carbon Capture and Storage, Enhanced Oil Recovery, Process Design
Safety Parameter, Flammability, Detection, Toxicity

INTRODUCTION on hydrogen alongside an existing gas-fired power

In recent years, it has been recognised that the threat of station;
climate change, due to the emission of greenhouse gases, par- . The manufacture of hydrogen – in order to supply the
ticularly carbon dioxide (CO2) is one of the key environ- power station – by reforming North Sea gas and captur-
mental concerns facing modern society. Although aspects ing the resulting carbon dioxide; and
of the science remain the subject of expert debate, there is . The transportation of the captured CO2 via an existing
broad consensus that climate change is occurring and this is offshore pipeline to offshore platform in the North
reflected in both international and national initiatives. Most Sea – and injection into the reservoir to enable the addi-
industrialised countries, including the UK, have signed up tional production of oil that is not currently recoverable.
to international conventions and programmes, in particular
the UN Framework Convention on Climate Change 1992 While each of the component technologies is already
and the Kyoto Protocol 1997. CO2 EOR can increase oil pro- proven, their proposed combination at this scale, in an
duction by approximately 5 – 15% beyond what is typically integrated Decarbonised Fuels system is unique. Figure 1
achievable using conventional recovery methods (Tzimas shows how these different technologies combine to create
et al., 2005), while facilitating the long-term storage of CO2 the project concept.
in the oil reservoir. It does this by dissolving in the oil In this project, natural gas from North Sea gas fields
thereby reducing the oil’s effective viscosity and making it transmitted via the National Transmission System (NTS)
more mobile. The movement of the CO2 front within the and converted to hydrogen and CO2 by a process known as
reservoir can then sweep the oil to the production wells. pre-combustion decarbonisation. The vast majority of the
Extraction of additional hydrocarbons also makes more CO2 will be removed from the gas stream and the remaining
space available for CO2 storage in the long-term. hydrogen rich fuel will then be used to generate electricity in
a specially modified gas turbine. The proposal for this part
of the project is for the construction, of power station, of a
combined cycle gas turbine (CCGT) power plant with an
PROJECT DESCRIPTION
installed capacity of nominally 550 MWe generating nomin-
The use of CO2 as EOR is the leading option for demonstrat-
ally 475 MWe of carbon-free electricity; enough to power as
ing CCS in the UK (DTI, 2004 and House of Commons,
many as a three quarters of a million homes in the UK.
2006), not only on cost grounds and existing experience in
The captured CO2 will be compressed and dehydrated
EOR offshore and CO2 EOR onshore, but also because it
then transported by an existing pipeline to a gas terminal.
is permitted under the current London Convention and the
Where it will undergo further compression to convert
Oslo and Paris Convention (OSPAR). There are also well-
the CO2 from the gas to the liquid phase. From there, the
established frameworks for regulating activities related to
CO2 will then be transported via another existing pipeline
offshore oil and gas production.
to offshore platform in the North Sea. Once at the platform,
The project has three main components:
the CO2 will be injected into the mature oil field rese-
. The generation of ‘carbon free’ electricity through the rvoir under the seabed, thereby facilitating enhanced oil
construction of a gas reformer and new turbines to run recovery (EOR).

450
SYMPOSIUM SERIES NO. 156 Hazards XXII # 2011 IChemE

Figure 1. CO2 enhanced oil recovery concept

CONCEPT SELECTION AND DESIGN those hazards to a level of risk that is as low as reasonably
Initial concept selection for the project focussed on site practicable. Within this scope, hazard management has
selection for a suitable onshore power generation site. The encompassed safety, environment and health and has
onshore site appraisal exercise was undertaken within the resulted in a number of philosophies, which are strategic
context of a number of broad objectives of the project: to the safety of the platform. Underpinning the hazard
management process is the identification of hazards by
. To supply approximately 500 MW of power to the UK
multi-discipline reviews and the development of design
national grid generated from hydrogen rich fuel pro-
philosophies to provide clear guidance to the design
duced from existing North Sea gas supplies;
teams. Engaging all disciplines in this process has devel-
. To build a power plant at a location that is or can be
oped a thriving health, safety and environmental culture
served by existing North Sea gas supplies; and
allowing the free exchange of ideas and concepts. During
. To build a power plant at a location that can export the
the development of the design, opportunities have been
CO2 from the reformers to a depleted oil or gas field in
identified to eliminate hazards at source and deliver an
the North Sea where it could be used for CO2 EOR.
inherently safer design. The Fire & Gas Detection and
ESD systems are fundamental to the safe operation of the
CO2 TRANSMISSION CO2 EOR plant, working closely together with the client
A number of engineering studies were undertaken to inves- as a single team to develop the philosophies required to
tigate available options for compression and transport of support this strategy. Hazard identification studies have
CO2 and a number of options were screened. For the pur- been held at key stages of the design and Multi-discipline
poses of concept selection, the design criteria were required hazard identification studies conducted to identify and quan-
to dehydrate, compress and transport CO2 from 1 barg at the tify the risk to the platform and personnel.
beach to a maximum of 240 barg for reinjection at the plat- During the deign stage, the safety engineer played a
form. The option, which had the least environmental key role in focusing the design team scrutiny on the
impacts as it utilises existing onshore and offshore pipelines, options which were available to them to deliver an inher-
has been taken forward into detailed engineering design. ently safer design. The following are some of the safety
issues, which have been carried out by the safety engineer.

HAZARD MANAGEMENT AND SAFE

PLATFORM DESIGN FLAMMABILITY ANALYSIS
The Hazard Management design process requires all disci- Flammability refers to the ability of a mixture of fuel and
plines to identify hazards and propose methods to reduce air to sustain combustion when ignited. Hydrocarbon

451
SYMPOSIUM SERIES NO. 156 Hazards XXII # 2011 IChemE

molecules will react with oxygen (burn) if heated suffi- The flammability analysis generally showed a
ciently, for instance by a spark or similar ignition source. decrease in the flammable range of the gas compositions
Vapour/air mixtures are flammable only over a limited over field life due to increasing CO2. It can be seen from
range of vapour concentrations. This range is defined by Figure 2 that the CO2 has the greatest effect on the higher
the lower and upper flammability limits. Mixtures outside pressure systems (Production Manifolds, Export Gas Com-
this range are described as, respectively, too ‘lean’ or too pressor Outlet stream (Produced Gas stream), and the HP
‘rich’ for ignition. The change in flammable range refers Separator Gas stream) as they contain a proportionately
to the change in the difference between these two flammable higher proportion of methane. A small change is observed
limits. These limits are defined as follows; in the flammable range of the MP and LP Separator gas
streams over the life of the field.
LFL The lower flammable limit (LFL) is the minimum The Fuel Gas stream maintains a presence in all levels
concentration of vapour or gas in air below which of the platform and is flammable throughout the field life.
propagation of flame does not occur on contact This confirms the need for flammable gas detectors. Flam-
with a source of ignition. Below the LFL there is mable gas detectors will also be needed for the MP and
too little combustible fuel to sustain a flammable LP separator gas streams, as they remain flammable
mixture. throughout the field life.
UFL The upper flammable limit (UFL) is the maximum The HP Separator Gas stream, Export Gas Compres-
concentration of vapour or gas in air above which sor stream and the manifolds show a gradual increase in
propagation of flame does not occur on contact CO2 content over the field life. This observation confirms
with a source of ignition. Above the UFL there is the need for CO2 detectors.
too little oxygen to sustain a flammable mixture.

The inert properties of water, CO2 and nitrogen present in GAS DETECTION ANALYSIS
the streams were considered by the flammability analysis. The main objectives of the gas detection analysis were to
Inert gases play no part in combustion reactions but determine the
absorb heat when present in a hydrocarbon/air mixture.
. order of detection of the hazardous components (HC,
For that reason, adding inerts to a mixture tends to reduce
CO2, H2S) in the process streams.
the spread between LFL and UFL until finally the mixture
. requirements for gas detection based on the detection
is no longer flammable.
limits of the gas components in air as defined by the
The main objectives of the flammability analysis on
client and HSE guidelines.
the stream compositions were to gain a clear understanding
of how the flammable range of mixtures containing multiple The order of detection was required to determine
hydrocarbons and multiple inerts, such as carbon dioxide the gas concentrations at first detected gas and also for
(CO2), nitrogen (N2) and water vapour vary with increasing evaluating the toxic properties of the gas at the point of
CO2. A flammability analysis would also enable conclusions detection.
to be drawn regarding the types of detectors required over The detection analysis also sought to clarify issues
the life of the field life and provide an indication of when regarding the possibility of detecting H2S by inference in
these detectors will be required. low H2S content streams

100

90

80
% Flammable range

70

60
Manifolds LP Sep Gas
50 HP Sep Gas Export Comp Outlet
40 MP Sep Gas Fuel Gas
30

20

10

0
2005 2010 2015 2020 2025 2030 2035
Year of operation

Figure 2. Variation of flammable range over field life

452
SYMPOSIUM SERIES NO. 156 Hazards XXII # 2011 IChemE

BASIS OF ANALYSIS molecules of the hydrocarbon gas components equals the

The gas detection analysis was carried out under the follow- light absorbed by propane gas at the calibration concen-
ing assumptions tration. The mechanism of detection is thus additive.
. Presence of all three gas detectors (HC, CO2 and H2S) at
the source of release. POINT DETECTORS
. Perfect mixing around the three detectors. Point detectors operate by measuring the concentration of
The Fire and Gas Detection Safety Philosophy for the the gas at the sampling point of the instrument. The unit
platform is recommends the use of two types of infrared of measurement can be:
flammable gas detectors. . % volume ratio;
. Open Path Detectors . % lower flammable limit (LFL) for a flammable gas;
. Point Detectors . ppm or mg/m3 for low level concentrations (primarily
used for toxic gases).
Open Path Detectors are most suitable for monitoring
lighter hydrocarbon gas components and are recommended The analysis was based on the Simtronics GD10P
for general process area coverage in modules. Point Detec- IR point detectors. The alarm level assumed was 20%
tors are used for the monitoring of heavier, more toxic methane LEL for low level alarm and 60% methane LEL
hydrocarbon gas components and are recommended for for high alarm level. These values are in accordance with
heavy gas monitoring around condensate pumps and in the the Fire and Gas Detection Safety Philosophy of the
wellbay of the platform. The analysis was thus based on platform.
these two types of infrared flammable gas detectors. The Point detectors are least sensitive to methane and are
order of detection is the order in which the hazardous com- therefore calibrated to this gas. This ensures that any
ponents (HC, CO2 and H2S) in the gas stream are detected. It released process gas is not overestimated. The mechanism
is the order in which the individual gas components reach of operation of point detectors is also additive. Table 2,
their low and high alarm levels. These low and high alarm obtained courtesy of Groveley4 Detection, gives the cross
levels are presented in Table 1. sensitivities of the Simtronics GD10P IR point detector to
other hydrocarbon components.
The result for the produced gas stream (export com-
OPEN PATH DETECTORS pressor outlet stream), selected to represent the open path
Open-path detectors typically consist of a radiation source detection analysis and the result for the MP separator oil
and a physically separate remote detector. The detector flash gas stream, selected to represent the point detection
measures the average concentration of gas along the path analysis are presented in this section as examples. The
of the beam. The unit of measurement is concentration results present the detection profile of the areas covered by
multiplied by path length (% LEL  m or ppm  m). This both flammable detector types. They are useful for evaluating
makes it impossible to distinguish whether a reading is the suitability of the detectors over the field life. The order of
due to a high concentration along a small part of the beam detection was determined at low and high level alarm for all
or a lower concentration distributed over a longer length. streams in the process. The stream composition data used in
The open-path detector used for the analysis was the the analysis was obtained from the Heat and Mass Balance
Dräger Polytron open path gas detector and the alarm levels Report and was based on the 2010 HP Ops data.
used for the analysis was 1 LELm for low level alarm and The determination of the order of detection enabled
3 LELm for high level alarm in accordance with Fire and the gas concentrations in air at low and high-level alarm
Gas Detection Safety Philosophy of the platform. levels to be established. The gas concentrations in air at
Infrared detectors are calibrated for the hydrocarbon low and high level alarm levels are a prerequisite in evalu-
gas they are least sensitive to. IR open path detectors ating the toxic properties of the gas at the point of detection.
operate at or around 2.1 mm (these are the most modern The results present the order of detection of selected streams
open path devices) and are least sensitive to propane gas. for both open path and point flammable detectors, and the
They are therefore calibrated for propane. The detector concentration of the detected gas components at the first
will alarm when the infrared light absorbed by the detected gas.

Table 1. Low and high alarm levels

Gas component Low level alarm High level alarm

Hydrocarbons Open path 1 LELm 3 LELm

Point 20% LEL Methane 60% LEL Methane
CO2 0.5% (5000 ppm) 1.5% (15,000 ppm)
H2 S 0.0005% (5 ppm) 0.001% (10 ppm)

453
SYMPOSIUM SERIES NO. 156 Hazards XXII # 2011 IChemE

Table 2. Simtronics4 GD10P IR point detector sensitivities L ¼ corresponding threshold limit of the toxic component
(toxic component impairment criterion)
Sensitivity low Sensitivity
level alarm high level Oxygen depletion excluded (relevant for immediate
Component (%LFL) alarm (%LFL) fatality analysis only).
If the total dose exceeds unity, then the impairment
Methane 20 60 level is deemed to be exceeded.
Ethane 12 36 The minimum gas concentrations corresponding to a
Propane 4 12 toxic load of 1 were determined using the model and were
i-Butane 5 15 based on four impairment criteria. These impairment cri-
n-Butane 5 15 teria values were sourced from the toxicity report. These
i-Pentane 10 30 values are listed in Table 3.
n-Pentane 10 30 The results from the toxicity analysis are based on
n-Hexane 16 48 the 15 minute short term exposure limit (STEL) because of
its relevance to the time frame for detection and escape.
The toxic load corresponding to the short term exposure
TOXICITY ANALYSIS limit (STEL) is the total toxic load received over a 15
Toxicity is the degree to which a product can cause personal minute exposure duration. The results presented by the
injury or illness when inhaled, swallowed, or absorbed toxic load profile Figure 3 give the toxic load received over
through the skin. The increase in CO2 content of the pro- a 1 and 5 minute exposure duration, assuming a linear corre-
duced gas over the field life has the potential to significantly lation of toxic load with time. It is assumed that the average
alter the toxic risk profile of the platform. It was thus necess- time frame from the detection of a gas leak to the initiation of
ary to investigate the properties of the toxic components in a response by the control room is about 5 minutes.
the process streams (HC, H2S and CO2) with a view to The toxic load profile Figure 3 for each stream show
assessing the toxic potential of these components in the that the toxic load received in 5 minutes is more than the
event of a loss of containment. toxic load corresponding to STEL with the absence of an
The toxicity analysis enabled an evaluation of the early warning system and the toxic load received in 5
toxic properties of the released gas at low and high alarm minutes is less that the toxic load corresponding to
levels. The assessment was useful in determining whether STEL with the presence of an early warning system. The
the concentrations of the detected gas at the low and high change in toxic load observed from Figure 3 over the field
alarm levels are within the limits prescribed by the HSE life is a result of the changing composition of toxic com-
workplace exposure limits3. ponents within the stream. This observation justifies the
Other objectives derived from the analysis include: need for an early warning system to facilitate a quick
escape of personnel from the hazardous area at low level
. the determination of the variation in toxicity of process alarm. The results confirm that the current detection levels
streams over field life. are appropriate for the monitoring of gas releases, subject
. the determination of the largest contributor to the tox- to the implementation of an early warning system.
icity amongst the individual components. Figure 3 showing the contributions to the toxic
load show that the largest contributors to the toxic load

COMBINED TOXIC EFFECT MODEL Table 3. Impairment criteria values

The combined toxic effect model is used to determine the
gas concentration corresponding to a toxic load of 1. A Impairment criteria (%)
toxic load of 1 represents the toxicity of a gas mixture
that corresponds to the toxicity at the reference impairment Low High level
level. The model considers the additive effects of all level alarm alarm (15
toxic components in the gas and could also be modified (8 hour minutes Immediate
to account for the oxygen depletion effect (relevant Component LTEL) STEL) Escape fatality
for immediate impairment criteria only) of the gas thus
predicting the overall gas concentration that leads to H2S 0.0005 0.001 0.06 0.06
impairment. The combined toxic effect model is represented CO2 0.5 1.5 5 10
by the Equation (1) below. Methane 0.1 0.15 0.45 30
Ethane 0.1 0.15 0.45 13
C1 C2 C3 Cn Propane 0.1 0.15 0.45 4.7
þ þ  ¼1 (1) Butane 0.06 0.075 0.2 1.7
L1 L2 L3 Ln
Propane 0.06 0.075 0.3 0.8
Where, Hexane 0.002 0.05 0.3 0.3
O2 13
C ¼ concentration of the toxic component in air

454
SYMPOSIUM SERIES NO. 156 Hazards XXII # 2011 IChemE

% Contribution to toxic load Variation in toxic load over field life relative to 2010
120%

100%
% Contribution to toxic load

80%

60%

40%

20%

0%
HC1
HC2
HC3
HC4
HC5
HC6

HC1
HC2
HC3
HC4
HC5
HC6

HC1
HC2
HC3
HC4
HC5
HC6

HC1
HC2
HC3
HC4
HC5
HC6

HC1
HC2
HC3
HC4
HC5
HC6
O2

O2

O2

O2

O2
H2S

H2S

H2S

H2S

H2S
CO2

CO2

CO2

CO2

CO2
2010 2015 2018 2023 2030
Year of operation

Figure 3. % Contribution of individual components to the toxic load corresponding the immediate impairment criteria and variation
in toxic load over field life (produced gas stream)

corresponding to the short term exposure limit for all the detector types over the life of the field. The analysis thus
selected streams are hydrocarbons. This observation holds confirmed the need for hydrocarbon and CO2 detectors.
true for most of the streams in the process based on the The analysis also presented the possibility of detecting
STEL criteria. H2S by inference subject to a detailed analysis of the recov-
Figure 3 shows a steady increase in the contribution to ered flare gas and fuel gas reject CO2 streams as well as the a
the toxic load over field life by the CO2. It also shows a few other hydrocarbon process streams and produced water
decrease in the contribution to the toxic load by the hydro- streams with concerns regarding H2S detection levels.
carbons but shows the toxic load remaining fairly constant. The toxicity analysis results confirm that toxic
This observation is explained by the increase in toxic load hazards are generally unchanging over the life of the field.
resulting from the CO2 increase being offset by the toxic This conclusion ignores synergistic effects of hyperven-
load reduction due to the decrease in hydrocarbon content. tilation due to CO2 inhalation. A study into this effect is
This trend is mirrored in all the streams with increasing recommended as a detailed design activity.
CO2 content.
The results from the toxicity analysis show that toxic
hazards are generally unchanging over the field life (ignor- ACKNOWLEDGEMENTS
ing synergistic effects of hyperventilation due to CO2 inha- The author thanks Mr John Rowlands for his permission to
lation). This is contrary to the conclusions drawn from publish this paper. I am also grateful to Mr Ian Herbert for
earlier studies into the effect of increasing CO2 on the his assistance during the course of preparing this paper.
toxic hazards. Toxic hazards were generally expected to
increase steadily over the field life due to increasing CO2.
REFERENCES
1. Technical reports.
2. In house Modelling.
CONCLUSIONS 3. Health and Safety Executive. Workplace Exposure Limits.
During the CO2 EOR deign stage; the safety engineer played Document Number EH40/2005.
a key role in focusing the design team scrutiny on the 4. Groveley Detection Technical Note, GD10P Cross Sensi-
options, which were available to them to deliver an inher- tivities.
ently safer design. 5. Flammability Characteristics of Combustible Gases
The flammability analysis confirmed the need for and Vapours, ISA The Instrumentation, Systems, and
hydrocarbon and CO2 detectors in production manifolds, Automation Society ANSI/ISA –TR12.13.01–1999.
export gas compressor outlet stream, HP separator, MP 6. BS 5306-4: 2001, Fire extinguishing installations and
and LP separator gas streams over the life of the field. equipment on premises – Part 4: Specification for carbon
This is due to the presence of both flammable and high dioxide systems.
CO2 content streams in these modules. 7. ISO 10156: 1996 Gases and gas mixtures – Determination
Results from the detection analysis indicate the preva- of fire potential and oxidizing ability for the selection of
lence of hydrocarbons and CO2 detectors over H2S for both cylinder valve outlets.

455