You are on page 1of 13

SPE 144847

Simulation of Two-Phase Flow in Carbon Dioxide Injection Wells


Lawrence J. Pekot, SPE, Pierre Petit, Yasmin Adushita, SPE, Stephanie Saunier, Schlumberger; and Rohan De
Silva, SPE, National Grid Carbon

Copyright 2011, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Offshore Europe Oil and Gas Conference and Exhibition held in Aberdeen, UK, 68 September 2011.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
To model a variety of potential operating conditions in pure carbon dioxide (CO2) injection wells we performed a multi-
phase transient well flow simulation study. A thermal reservoir simulator was also used to estimate the extent of reservoir
cooling and the variation of injectivity index to be expected from injection of cold CO2.
Depleted gas reservoirs are potentially attractive targets for CO2 but their low pore pressure results in low bottomhole
injection pressure and potentially two-phase flow regime in the wellbore. Other authors have noted the possible implications of
this condition; however, none have addressed the issue using transient flow simulation.
A vertical wellbore model was built in a multi-phase transient flow simulator, assuming representative Southern North Sea
conditions. To investigate wellbore profiles of pressure, temperature and CO2 liquid hold-up, parametric as well as thermal
reservoir simulations were performed. The latter simulations integrated the bottomhole conditions observed in the wellbore
model.
Results show that pure CO2 injected at the wellhead may vaporize or condense as it travels down the tubing, experiencing
continuous changes in pressure and temperature as dictated by its change in enthalpy. However, sudden vaporization or
condensation is not predicted by the simulator. Two-phase flow cases resulted in stable injection conditions. Well injectivity
index varied significantly with injection fluid temperature and pressure, but the extent of reservoir cooling away from the
wellbore was limited. This suggests that onerous processing to avoid a two-phase flow regime in CO2 injection wells, such as
pre-injection heating or downhole choking may not be necessary at the injection start-up into a depleted gas reservoir.

Introduction
A Challenge with Depleted Natural Gas Reservoirs
Long term storage of CO2 using depleted gas reservoirs (DGRs), as an action to reduce atmospheric greenhouse gas
emissions, is under worldwide study. DGRs offer several favorable characteristics for CO2 storage, including a caprock that
may be suitable to contain CO2, extensive data history, legacy infrastructure and they are often in reasonable proximity to large
CO2 sources. However, they also present challenges. Among them is the interaction between initially low reservoir pore
pressure and the phase behavior of CO2. This paper addresses some aspects of this particular challenge.
As shown on Figure 1, the phase behavior of CO2 is known to be complex and exhibits a critical point of 31.2C and 73.8
bara. DGRs may have a temperature below this critical temperature and are commonly abandoned at a pressure below this
critical pressure. Thus, CO2 injection would initially proceed in the vapor phase in the reservoir. However, CO2 may be
delivered to the wellhead via pipeline in a liquid, vapor or supercritical state. This implies that the injected CO2 may undergo
a phase transition and two-phase flow in the wellbore. This behavior is not fully or widely understood.
Historical Review
Several publications related to this subject have considered single phase operating conditions and raised concerns about
operating in two-phase flow conditions or investigated methods to circumvent the situation. One of the stated or implied
assumptions among these works has been that CO2 two-phase flow was to be avoided because it would lead to unstable
pressure, temperature and rate conditions and difficulty with injection well control. However, there are also publications that
specifically address two-phase wellbore conditions and their results are consistent with the results presented in this paper.
2 SPE 144847

An integrated asset model approach for planning a Southern North Sea (SNS) DGR injection project has been previously
described (Galic, et. al., 2009). This work traced the pressure and temperature evolution of CO2 from compressor discharge to
the reservoir, including seasonal temperature variations. However, a stated objective of the work was to avoid any liquid/vapor
phase change. This was accomplished primarily by maintaining the system pressure above 80 bara.

Figure 1 Pressure-Temperature Phase Diagram for Carbon Dioxide

Bottomhole injection pressure was not specifically discussed, but it is apparent that the initial bottomhole injection pressure
calculation result was approximately 140 bara against a depleted reservoir pore pressure of 27 bara. The wellbore hydraulic
calculations indicated a maximum pressure value in the middle of the tubing string followed by a reversal of pressure gradient
near the bottom of the well that was noted as due to a decrease in fluid density. It has also been suggested (Haigh, 2009) that
phase change in the wellbore could be avoided by choking and/or heating at the wellhead, changing the tubing to different
sizes during the project life to optimize wellbore pressure drop, or use of a downhole variable choke device to avoid two-phase
flow. The paper by Haigh was intended as a more wide-ranging discussion of several related CO2 injection topics and thus no
specific flow calculations were presented. Both of these papers referred to CO2 phase change in the well as unstable or may
lead to well instability. Another hypothetical case study (Hughes, 2009) recognizes the complexity of operating injectors
across the CO2 phase boundary. He notes that most conventional software designed to model hydrocarbon flow in pipes needs
to be used with caution when considering CO2 and that more rigorous CO2-specific software is needed. His calculations, using
a conventional software package, assume that CO2 in the well must be either in gaseous or liquid state and that a large range of
bottomhole pressure is not achievable between the two states. Two-phase flow is not considered. He concludes two possible
solutions. First, the problem can be avoided by heating the CO2 to above the critical temperature, although this is likely to be
economically prohibitive. Second, that without heating the use of a downhole flow control choke will be required.
However, there is also a different line of investigation. An early modeling study of two-phase CO2 injection was published
nearly thirty years ago (Cronshaw and Bolling, 1982). They observed the value of using a pressure-enthalpy chart to help
determine conditions in the well. Their simulation model, history matched to CO2 production data, produced both production
and injection well results with two-phase flow transitions extending 300 meters or more within the well. Their field data was
later revisited (Lu and Connell, 2007) with a model formulation that also considered impurities in the flow stream. This later
work validated Cronshaw and Bolling. Further, Lu and Connell created a model for a hypothetical injection case with a strong
geothermal gradient intended to induce gas, two-phase, liquid and supercritical conditions along the same wellbore. In their
well with a 1200 meter depth, pressure, density and fluid phase changed continuously along the well while the two-phase
SPE 144847 3

transition zone was calculated to extend 87 meters. Further interpretation of the above work (Paterson, et. al., 2008) concluded
that two-phase wellbore conditions do not prevent flow, but makes bottomhole calculations less straightforward and that a
more complete study should include a coupled wellbore-reservoir response.
Statoil considered two-phase operation of the Sleipner CO2 injection project (Baklid, et.al., 1996). Supercritical injection
was eventually chosen by Statoil, largely due to platform equipment placement constraints, not due to two-phase well flow
concerns.
Calculations of two-phase conditions in shut-in CO2 wells suggest a sharp fluid interface and change in density that can
cause a large change in bottomhole pressure with only a small change in wellhead pressure (Loizzo, et. al., 2009). Field data
from shut in CO2 observation wells at the Ketzin pilot injection project confirmed this (Henninges, et. al., 2011). Their data
includes recorded downhole video evidence showing condensation and vaporization occurring above and below a stationary
CO2 fluid interface in the static observation well. However, they also reported data from the same well showing a reversal of
the density gradient to a less dense fluid with greater depth, showing unexpected behavior even with shut-in observation wells.
Besides pure CO2 injection (including minor impurities) there is also documentation and modeling of two-phase flow
operations in Canadian acid gas injection wells (Stocker, 2006). Acid gas is primarily a mixture of CO2 and H2S.
From the above discussion it is clear to us that there is ample scope for additional work on this complex subject.
Objectives of this Work
This study investigated two primary concerns surrounding CO2 injection. First, dynamic wellbore simulation was used to
investigate flow regimes and estimate if two-phase flow can be sustained in CO2 injectors with conditions similar to DGRs in
the SNS. Second, thermal reservoir simulation was used to estimate possible effects of CO2 injection on reservoir injectivity.

Method
Dynamic wellbore simulation was carried out using the OLGA multiphase flow simulator with the Single Component
Module. The module allows for simulation of transient behavior of single components and their transition between liquid and
gas phase as well as their behavior beyond the critical point. The module is especially set up for steam and CO2 but other
single components can be entered for simulation. Use of the Single Component Module restricts the work to pure CO2
conditions. A single basic casing design for a vertical well was selected, intended to represent typical SNS conditions.
Numerous sensitivity cases were run to investigate variations in reservoir pressure, wellhead conditions and wellbore
equipment configurations.
Reservoir simulation calculations were carried out using an industry standard reservoir simulation software package with
the thermal modeling option. A single well radial model was constructed to represent an SNS DGR with reservoir properties
favorable for CO2 injection. A series of injection rates and bottomhole CO2 injection temperatures from the transient
simulation were used in the reservoir simulation to evaluate their effects on injectivity and reservoir temperature.
Assumptions
The set of assumptions have been selected to be representative of the depleted gas fields in the SNS. The subsurface
ambient temperature is based on a thermal gradient of 3.3C/100m with a temperature of 4C at wellhead and a reservoir
temperature of 100C. Table 1 describes the selected well design.

Table 1. Description of Well Tubulars

From To Nominal Diameter Wall thickness ID

m TVD m TVD Inches mm mm

0 100 Conductor 30 in 12.7 733.6

0 1000 13 3/8 in, 72 lb/ft 13.05 313.6

0 2000 9 5/8 in, 53.5 lb/ft 13.85 216.8

2000 2800 Liner 7 in 11.51 154.78

0 2800 Tubing 5 1/2 in 6.99 125.73

The study covers injection rates from 10kg/s, about 0.3 Mt CO2/year, to 50 kg/s, about 1.6 Mt CO2/year. This range was
selected to be representative of realistic and economic injection rates for CO2 storage projects in the reservoirs considered.
4 SPE 144847

The reservoir pressure of many DGRs in SNS is currently low to very low; but as CO2 is injected in the formation, the
pressure will gradually increase. Thus, a bottomhole pressure (BHP) range from 25 bara to 300 bara was evaluated. Well head
pressures from 25 bara up to 150 bara were considered.

Limitations
1. In a vertical well, for any cross section perpendicular to the direction of flow, the wellbore software assumes that
liquid and gas phases are in thermal and mechanical equilibrium (pressure and temperature are identical across a
given cross-section). However, when in operation CO2 phase behavior, and changing pressure and temperature due to
fluid flow down the wellbore, leads to a delay in establishing thermodynamic equilibrium. The software applies a user
specified relaxation time to model the impact of non-equilibrium thermodynamic conditions. Large values for
relaxation time allow larger non-equilibrium conditions to exist while very small times can lead to program
instabilities and may create non-physical results. Without physical data to help determine relaxation time, a sensitivity
study was performed and a constant relaxation time of one minute was applied to all the calculations.
2. The study was focused on the thermodynamic behavior of the CO2 in the wellbore and near wellbore area. Therefore,
other calculations associated with CO2 injection were not performed in the model, including the risk of hydrate
formation and chemical reactions of the CO2 with the rock or formation fluids, cement or other well materials.
3. Additionally, mechanical effects of the injection on the well and geomechanical effects in the near well bore area
were not part of the simulation performed.
4. Permeability and thickness of the reservoir is large compared to the range of injection rates, thus pressure drop and
temperature change across the perforations is minimal.
5. Finally, the effect of impurities on the phase envelope of the injected fluid was not considered.
Modeling of CO2 Flow in the Wellbore
The mechanism of the CO2 flow in the well is controlled by complex thermodynamic processes. This section aims to
provide some background on the thermodynamic behavior of pure component CO2.
Dynamic Single Component Module Simulation vs. Traditional Oil & Gas Software
Traditional Oil and Gas modeling software is normally used with complex mixtures of hydrocarbon components.
Considering a typical pressure-temperature chart, the phase envelope corresponds to the two-phase zone. In this area, there is
no discontinuity in the total enthalpy of the system. Considering the same chart but for a single component, the phase
envelope consists of a single line (refer to Figure 1). This line corresponds to a discontinuity in the enthalpy of the fluid. The
difference of enthalpy between liquid and gas at a given pressure and temperature corresponds to the latent heat (energy
required for a liquid system to vaporize). Traditional Oil & Gas industry modeling software does not take this parameter into
account. For this reason, they cannot correctly predict the phase transition for CO2.
In contrast, dynamic single component module simulation allows the software to take into account the latent heat, and thus
to predict the phase transition more accurately than a standard Oil & Gas industry modeling software.
The Role of Enthalpy
For a pure component in single phase, the pressure and the temperature fully describe the fluid conditions. In two-phase
flow at thermodynamic equilibrium, pressure and temperature are intrinsically related. The temperature at which the phase
transition occurs corresponds to a unique pressure. As a consequence, for a given pressure and temperature, there are an
infinite number of possible equilibrium states during phase transition. It is thus necessary to introduce an additional parameter
which will help to fully characterize the system. The parameter that we have chosen is the enthalpy.
SPE 144847 5

Figure 2 Pressure-Enthalpy Diagram for Carbon Dioxide

All the transformations in the wellbore can be drawn on a pressure-enthalpy diagram, Figure 2, as long as the system is in
thermodynamic equilibrium. From this diagram, one can deduce that two-phase flow is possible and is thermodynamically
stable. Two-phase conditions persist throughout the area under the solid black curve.
The pressure-enthalpy diagram presented in Figure 2 is valid when the system has reached a thermodynamic equilibrium (a
system has reached an equilibrium when the entropy of the system is maximum).
Equilibrium in the well
The pressure and the temperature of the flowing fluid evolve along the well path. The conditions of thermodynamic
equilibrium for the fluid are thus continually changing along its journey. A way to show that a fluid is, or is not, at equilibrium
is to compare the calculated Gibbs energy (a.k.a. Free Enthalpy) for each phase.
The Gibbs energy is defined as follows:

g=hTs

Where h is the enthalpy in kJ/kg, T is the temperature in Kelvin and s is the entropy in kJ/K kg.
If two phases coexist in a fluid and the phases are at thermodynamic equilibrium, the Gibbs energy of the liquid is equal to
the Gibbs energy of the gas. In conditions of disequilibrium, the more stable phase is the phase with the lower Gibbs energy. If
both phases are present and the modeling software calculates different Gibbs energy for the two phases, the system is not in
thermodynamic equilibrium and phase transition is in progress.
It is important to note that phase transition and thermodynamic disequilibrium of CO2 does not necessarily imply unstable
or unsustainable rate, temperature or pressure conditions during injection operations.

Model Cases
Numerous simulations, both in the wellbore and the reservoir, were run but we present here only a few groups of cases to
highlight some critical points about CO2 flow.
Three cases describe single phase injection
Five cases describe variation in injection rate
Six cases describe variation in reservoir pressure
A matrix of cases is shown with fixed wellhead pressure (WHP) and BHP and the calculated injection rate
Four cases describe development of the cool CO2 front in the reservoir
A matrix of cases describes the changing injectivity index for CO2 versus T and P
6 SPE 144847

Model Results
CO2 Flow in the Well Bore
Figure 3 shows the calculated evolution of the pressure and enthalpy for three cases of single phase injection in the well
bore. The first case corresponds to an injection rate of 50 kg/s associated with a reservoir pressure of 300 bara. The
corresponding well head pressure is 60 bara. Keeping in mind that the wellhead temperature is constant and set at 4C, this
indicates that the CO2 is liquid. Due to the hydrostatic pressure in the well, the pressure increases and exceeds 74 bara, the
critical pressure of CO2. Above this pressure, the fluid is in either liquid or supercritical phase and reaches the reservoir as a
dense fluid. The two other cases correspond to an injection rate of 10 kg/s and an associated reservoir pressure of 25 or 50
bara. In both of these cases, the calculated pressure-enthalpy traverse demonstrates that the CO2 is injected as a gas phase. For
those three cases, the software predicts that there is no two-phase flow, which is validated by the pressure-enthalpy diagram as
the calculated results do not enter the two-phase zone.

300 bara, 50 kg/s

50 bara, 10 kg/s

25 bara, 10 kg/s

Figure 3- Pressure-Enthalpy Diagram with Calculations Shown for CO2 Injection as a Single Phase

Figure 4 compares five different mass injection rates into a reservoir with a pressure of 100 bara. All the cases start with
this same wellhead temperature, 4C. The software calculates that for all five cases, the required minimum well head pressure
is equal to 39 bara. The pressure-enthalpy diagram similarly indicates that for a temperature of 4C, the two-phase transition
occurs at a constant pressure of 39 bara. Thus, all five cases start with two-phase fluid at the wellhead and the only way to
differentiate these cases is their initial enthalpy, which is associated with a given liquid mass fraction. The pressure in the well
increases with depth due to the hydrostatic pressure. Since the reservoir pressure is higher than the critical pressure, the fluid
eventually leaves the two-phase zone and will be injected as a dense phase in the reservoir. Injection can also be achieved at
higher wellhead pressure.
SPE 144847 7

50 kg/s 40 kg/s

30 kg/s

20 kg/s

10 kg/s

Figure 4- Pressure-Enthalpy Diagram with Calculations Shown for Injection Rate 10-50kg/s, Reservoir Pressure 100, bara

The cases shown in Figure 5 consider the same mass rate of 20 kg/s, but different reservoir pressures. For all these cases, the
wellhead temperature is again set at 4C and the calculated required wellhead pressure is again 39 bara, which means that the
fluid starts in the wellbore as a two-phase flow.
When injected into a reservoir at a pressure of only 25 bara, the pressure in the well decreases with depth and the
temperature slowly rises with depth but then falls as the fluid continues to vaporize. However, the fluid travels the entire
tubing length and arrives at the perforations as two-phase flow, as shown in Figure 6a. For the 50 bara case, well temperature
and pressure slowly increases with depth, but again the fluid travels the entire length of the tubing in two-phase conditions.
Higher reservoir pressure cases eventually transition out of two-phase and into dense phase conditions, as shown in Figure 6b
for a reservoir pressure of 100 bara. The gas and liquid densities converge slowly over the first 2100 meters of the well and
then converge more rapidly to complete the transition to single phase supercritical state at a depth of approximately 2250
meters. There is no sudden wellbore pressure or temperature surge or jump associated with this transition. Even the 300 bara
case is achieved by starting with a two-phase wellhead fluid of 39 bara and 4C, although the fluid almost immediately
completes a transition into the liquid phase. Note that this does not imply that wellhead pressure must be 39 bara to achieve
injection.
8 SPE 144847

300 bara

200 bara

100 bara

50 bara

25 bara

Figure 5 Pressure-Enthalpy Diagram with Calculations Shown for a Constant Injection Rate of 20 kg/s and Different Reservoir
Pressures

Figure 6a Pressure, Temperature and Phase Densities vs. Depth for an Injection Rate of 20 kg/s and a Reservoir Pressure of 25
bara

Figure 6b Pressure, Temperature and Phase Densities vs. Depth for an Injection Rate of 20 kg/s and a Reservoir Pressure of 100
bara

Results presented Table 2a and Table 2b are based on a given WHP and BHP. The software then calculates the mass rate
and bottomhole temperature (BHT) required to meet those boundary pressures. The wellhead temperature is again set equal to
4C. It is important to recall that the calculated rates shown in Table 2 are based on the generalised conditions deployed in this
study only and should not be assumed as typically applicable to any specific SNS injection project. However, they do suggest
that a wide range of injection rates and BHP may be achievable if two-phase flow in the wells is allowed.
SPE 144847 9

With these conditions and a WHP of 25 bara and a reservoir pressure of 50 bara and higher, CO2 cannot be injected into the
well. As such, those cases were not considered further.
For a WHP and a reservoir pressure of 25 bara, CO2 is injected as a gas. For a WHP of 60 bara and higher, CO2 is injected
as liquid or dense phase, as indicated in Table 2a.

Table 2a- Mass rate and Corresponding Wellhead Phase for different WHP, Reservoir Pressure and a
constant WHT of 4C
Wellhead Pressure (bara)
Reservoir
Pressure Mass Rate (kg/s) WH phase
(bara)
25 60 100 150 25 60 100 150
25 7 144 151 163 Gas Liquid Dense Dense
50 132 145 159 Liquid Dense Dense
75 126 140 154 Liquid Dense Dense
100 121 135 150 Liquid Dense Dense
200 93 110 127 Liquid Dense Dense

Table 2b- Bottomhole Pressure and Temperature for different WHP, Reservoir Pressure and a constant
WHT of 4C
Wellhead Pressure (bara)
Reservoir
Pressure Bottomhole Temperature (C) Bottomhole Pressure (bara)
(bara)
25 60 100 150 25 60 100 150
25 112 10 9 11 25 36 36 37
50 16 15 14 59 60 62
75 18 17 16 85 86 87
100 20 19 18 109 110 111
200 24 21 22 207 210 211

CO2 Flow in the Near Well Bore Area


Reservoir temperature change
The wellbore simulations undertaken indicate a range of BHT between 0 to 60C for various BHP and injection rates,
lower than the reservoir temperature of 100C. A near wellbore radial model was constructed to estimate propagation of the
cool front and the injectivity variation as an impact of cold CO2 injection during a 30 year injection period. Table 3 presents a
list of cases that were evaluated. The reservoir model has a radius of 1000 meters and a thickness of 200 meters.

Table 3- List of Cases in Near Wellbore Model Simulation


Case Reservoir Pressure, Bottomhole Injection CO2 Injection Rate, Phase
bar Temperature, C kg/s
1 25 47 10 Gas

2 100 56 10 Dense

3 100 45 30 Dense

4 100 38 50 Dense

All cases demonstrate that despite the CO2 plume sweeping the entire near wellbore area, the cool front propagates only
within a limited radius from the wellbore. In Case1, shown in Figure 7, where the CO2 is injected as gas, the cold temperature
extends about 150 m and the radius does not vary with depth. In Cases 2, 3, and 4, where the CO2 is injected as dense phase,
the cool front propagates up to a radius of approximately 150 m, 300 m, and 350 m, respectively, indicating that the cooling
radius is unsurprisingly dependent on the injection rate and the injected fluid temperature. Case 3 is shown in Figure 8. As
cooling occurs due to flow of cold CO2 into the reservoir, the conductive heat exchange with the reservoir rocks rapidly heats
up the injected stream before it moves further into the reservoir.
10 SPE 144847

However, in Cases 2, 3, and 4, the simulations indicate a separation in temperature and pressure fronts, which does not
appear in the gas phase injection case. The cool front radius is larger at the bottom of the reservoir than at the top as the
injected CO2, which is colder and denser than the existing fluid in the formation and segregates downwards.

)
!"#$"%&'(%")*%+,-."/)0"1)2
) )

Figure 7 Temperature Profile, Case1, Gas Phase CO2 Injection, year 30.

)) )) )
!"#$"%&'(%")*%+,-." ) ) ) ) 234)5"67-'8/)91:#;
)
) ) ) ) ) )) <=)>"7)*)?
))) )) ) ) )@@)A&%))&60)%&'")+,);@)91:7B
) ) ) ))

Figure. 8 Left: Temperature Profile, Case 3. Right: CO2 Density Profile, Case 3, Dense Phase CO2 Injection.

Injectivity Index change


The change of CO2 injectivity index with temperature and pressure is a complex behavior which depends on its variations
of density and viscosity.
Regardless of fluid phase, CO2 density decreases with higher temperature, thus reducing the injectivity index due to a
higher volumetric rate. In liquid and supercritical phases, CO2 viscosity decreases with higher temperature, thus increasing the
injectivity index as fluid mobility is higher. However, in gas phase the viscosity weakly increases with higher temperature and
therefore acts to reduce the injectivity index. As a net result, in liquid and supercritical phases the impacts of density and
viscosity change on injectivity index act to cancel each other, while in gas phase the two effects amplify each other.
The impact of pressure on injectivity index could also be described in a similar manner. Regardless of the fluid phase, CO2
density and viscosity increases with higher pressure, and therefore their impacts on injectivity index act to offset each other.
During the phase transition from gas to dense phase, the increase of density is more rapid than the increase in viscosity.
As illustrated in Figure 9, the well injectivity index varies more strongly with pressure than with temperature. During the
phase transition from gas to dense phase, pressure increases cause a rapid increase in density and the injectivity index may
increase by as much as 400%. Also while CO2 is in the gas phase, the injectivity index tends to decrease slowly with
temperature increase. As the CO2 exceeds supercritical conditions, injectivity index increases with both temperature and
pressure. Higher BHT now increases the injectivity index, as the CO2 viscosity decreases. Overall, the injectivity index might
vary by approximately 30% with changes to the BHT between 5 to 100C.
SPE 144847 11

These injectivity index variations will change the bottomhole and well head pressure requirements, and thus must be taken
into account in the detailed design of injection scenarios.
Hydrate Formation
Cold CO2 injection has also raised concerns about possible hydrate formation in the wellbore and sand face. However,
studies (Chapoy et al., 2009) have shown that for a pure CO2 stream with less than 250 ppm water content, hydrate formation
is unlikely at temperatures of -2 to 30C and at pressures up to 200 bar, which is the range of bottomhole operating conditions
studied in this evaluation. An even lower water content is required to avoid hydrate formation if impurities are present. This
signifies that dehydration of the injected stream is important for the range of bottomhole pressure and temperatures estimated
in this study. Connate water vaporization in the near wellbore area might also increase the water content of the CO2 stream as
it flows into the reservoir. The CO2-water phase partition experimental work that has been done (Spycher and Pruess, 2005)
indicates that under equilibrium conditions, the water mole fraction in the injected CO2 at 30C and between 25 to100 bara
could be up to approximately 2000 to 4000 ppm. In a gas reservoir, however, water vaporization into the CO2 rich phase may
be lower if the quantity of water in the near wellbore area is limited and not replenished by capillary forces. Thus, the injected
stream may be heated up by the reservoir before it gets saturated with water.

#
@(''%1A: BCD(E#545#F BC/(G.%3$;</<=,+

;88888%9 >88888%
3$ ; </<=,+

!"##"$%&"'(%)($*(+,#-+(.%
#"
/(01

:88888%9 ;88888%
3$ ; </<=,+
>88888%9 ?88888%
3$ ; </<=,+
788888%9
:88888%
3$ ; </<=,+
!""
$# %# !"" $"" &""

2(3(+4"5+%6+(33-+(.%!,+

Figure 9 Well Injectivity Index Variation with Pressure and Temperature, CO2 Injection.

Discussion
The use of DGRs in the SNS for CO2 storage has received considerable attention in the last several years. Although the
process appeared simple at first glance, more recent work has shown that several unresolved technical concerns remain.
Several of these concerns have focused on the operational performance effects of CO2 phase behavior, both in the wellbore
and in the reservoir, especially at pressure below the critical pressure of CO2. The work presented here agrees with previous
work that CO2 phase behavior must be evaluated seriously and is not straightforward for a DGR project. However, it also
suggests that it is not necessarily a project killer. Nor does CO2 phase behavior in the wells or the reservoir necessarily place
absolute operational requirements or limitations on the project, such as heating of the CO2 , use of downhole chokes or
maintenance of single phase flow; all of which are expensive and/or inefficient.
The simulation work presented suggests that stable operation of two-phase CO2 injection wells may be achievable in
commercial scale projects. Phase transition in the wells is calculated to occur over a considerable vertical distance, or even the
entire well length, without severe or sudden pressure, temperature or rate instabilities. CO2 phase change in the wells may
actually function as part of a self-stabilizing process in response to changes in wellhead or reservoir conditions. A similar
observation has been made previously by others (Baklid, et. al., 1996).
Reservoir pressure below the CO2 critical pressure of 74 bara may persist for several years at the start of a DGR injection
project. During that time, CO2 injectivity index should increase significantly due to its change in density and viscosity. This
will have a continuing impact on bottomhole injection conditions and subsequently on wellbore hydraulics. The need for a
reliable coupled solution between reservoir and well for CO2 injection is clear.
12 SPE 144847

Much remains to be done to confirm these results. Many uncertainties remain to be explored further, including the software
limitations previously listed and aspects not considered by this study such as variable/transient operating conditions of start-
up, shut-down, turn-down and lower injectivity reservoir conditions. Moreover, corroboration of the broad outcome of these
studies using an analytically different approach (e.g. computational fluid dynamics and separately determined PVT data)
would provide further confidence to the outcomes of this study. Two-phase flow field data for direct comparison with
calculations is very limited and needs to be fully utilized to build a stronger body of evidence.
Commercial operation of a large scale two-phase CO2 injection project may require integrated data collection and a
sophisticated process control system to ensure optimal performance. However, compared to the cost of heating CO2, extra
compression or frequent workovers to continually optimize well completions, the smart management of a two-phase injection
system may be not only possible, but superior.

Conclusions
Wellbore Modeling Conclusions
Traditional software packages used in the oil and gas industry are not adequate to model two-phase CO2 flow conditions in
the wellbore and can lead to misleading conclusions about CO2 injection well performance. Although verification work is
needed, wellbore transient modeling software provides a better picture of complex CO2 flow behavior than traditional oil and
gas software. This picture suggests that stable two-phase flow conditions can be sustained over long transitions in the well,
even extending over the entire length of the well. Phase transitions appear to occur without severe or sudden changes in
temperature, pressure or rate. However, continuous changes, and even reversals, in pressure, temperature and density gradients
may accompany the flow. Thermodynamic equilibrium may or may not be maintained in the well during these transitions. The
possibility of sustained, stable two-phase CO2 injection led to the execution of a set of several simulation cases that show a
wide range of injection rates and bottomhole pressures may be achievable, without major gaps in the range of performance. An
implication of this is that two-phase flow must be considered as an operating possibility and that expensive operating options
designed to allow only single phase flow may not be necessary.
Injectivity Conclusions
During a 30 year injection life, reservoir cooling effects were shown to extend from 150 m to 350 m out from the wellbore.
Using a homogeneous radial thermal model, the cool front was seen to be wider at the base of the reservoir where the coldest
and most dense fluid would settle. In a stratified reservoir, this effect may be substantially different.
Pressure is the most important factor in CO2 injectivity index variation because of the sensitivity of CO2 density with
pressure, especially when operating in gas phase conditions.
Temperature change is a less significant factor for variation of injectivity index. However, due to the large temperature
range that may occur with CO2 injection, temperature effects on injectivity index are not negligible.
The reservoir model in this study represents a high injectivity index formation. In case of a lower injectivity index and low
reservoir pressure, the impact of BHT variation on injectivity would be more significant; BHP would be higher and more wells
would be required, higher pressure drop across perforations would lead to additional cooling, multiphase flow into the
reservoir might lead to a lower phase relative permeability, and the colder temperature in the near wellbore area might have
geomechanical effects that reduce the fracture pressure.
Hydrate formation would appear to not be a concern over the range of conditions used for this study. However, operating
conditions for each project are unique and must be fully investigated.
The changing conditions encountered by CO2 injection in DGRs, both in the well and the reservoir implies that a rigorous
integrated modeling approach is needed for detailed injection design.

Acknowledgement
The authors express their thanks to Schlumberger Limited and National Grid Carbon for their permission and support to
publish this work.

References
Baklid, A., Korbol, R., and Owen, G. 1996. Sleipner Vest CO2 Injection Into a Shallow Underground Aquifer. Paper SPE 36600, Presented
at the SPE Annual Technical Conference, Denver, 6-9 October
Chapoy, A., Burgass R., and Tohidi, B. 2009. Effect of Common Impurities on the Phase Behaviour of Carbon Dioxide Rich Systems:
Minimizing the Risk of Hydrate Formation and Two-Phase Flow. Paper SPE 123778, Presented at the SPE Offshore Europe Oil and
Gas Conference, Aberdeen, 8-11 September
Cronshaw, M.B. and Bolling, J.D. 1982. A Numerical Model of the Non-Isothermal Flow of Carbon Dioxide in Wellbores. Paper SPE 10735
presented at the SPE California Regional Meeting , San Francisco, 24-26 March
Galic, H., Cawley, S. and Bishop, S. 2009. CO2 Injection Into Depleted Gas Reservoirs. Paper SPE 123788, Presented at the SPE Offshore
Europe Oil and Gas Conference, Aberdeen, 8-11 September
Haigh, M.J. 2009. Well Design Differentiators for CO2 Sequestration in Depleted Reservoirs. Paper SPE 124274, Presented at the SPE
Offshore Europe Oil and Gas Conference, Aberdeen, 8-11 September
Henninges, J., Liebscher, A., Bannach, A., Brandt, W., Hurter, S., Kohler, S., Moller, F. and CO2 SINK Group, 2010. P-T-! and two-phase
fluid conditions with inverted density profile in observation wells at the CO2 storage site at Ketzin (Germany). Energy Procedia 4
(2011) 6085-6090
SPE 144847 13

Hughes, D.S. 2009. Carbon storage in depleted gas fields: Key challenges. Energy Procedia 1 (2009) 3007-3014
Loizzo, M., Lecampion, B., Berard, T., Harichandran, A. and Jammes, L. 2009. Reusing O&G Depleted Reservoirs for CO2 Storage: Pros
and Cons. Paper SPE 124317, Presented at the SPE Offshore Europe Oil and Gas Conference, Aberdeen, 8-11 September
Lu, M. and Connell, L.D., 2007. Non-isothermal flow of carbon dioxide in injection wells during geological storage. International Journal of
Greenhouse Gas Control 2 (2008) 248-258
Paterson, L., Lu, M., Connell, L.D. and Ennis-King, J. 2008. Numerical Modeling of Pressure and Temperature Profiles Including Phase
Transitions in Carbon Dioxide Wells. Paper SPE 115846, Presented at the SPE Annual Technical Conference, Denver, 21-24
September
Spycher, N. and Pruess K. 2005. CO2-H2O Mixtures in the Geological Sequestration of CO2. II. Partitioning in Chloride Brines at 12 to
100C and up to 600 bar". Geochim. and Cosmochim. Acta, (69), No. 13, 3009-3020
Stocker, R., 2006. Acid Gas Injection-Compositional Modeling for Flow of Acid Gases in Wellbores. Fekete Instructional Video, Fekete
Reservoir Engineering Software and Services

You might also like