SPE 116782

A Miscibility Scoping Study for Gas Injection into a High-Temperature

Volatile Oil Reservoir in the Cooper Basin, Australia
Peter Clark, University of Adelaide;∗ Serge Toulekima, Santos Limited; and Hemanta Sarma,
University of Adelaide

Copyright 2008, Society of Petroleum Engineers

This paper was prepared for presentation at the 2008 SPE Asia Pacific Oil & Gas Conference and Exhibition held in Perth, Australia, 20–22 October 2008.

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

As part of the Moomba Carbon Storage Project, a scoping study for the potential of miscible gas injection into
Tirrawarra, an Australian onshore high temperature volatile oil reservoir, has been completed. Twenty-six injectants
with varying compositions of CO2, CH4, C2H6, n-C5H12 and N2 are considered in the study. The volatile oil under
study has a CO2 content of approximately 20 mol%. High temperature reservoirs and volatile oils are two extremes
of a miscible gas injection system that have not commonly been studied in previous literature.

The minimum miscibility pressure (MMP) between oil and an injectant is a key parameter in a miscible gas injection
project, directly affecting project design. MMP is highly dependant on reservoir temperature as well as oil and
injectant compositions. Commonly, MMP calculations involve one of the following three methods: (i)
thermodynamic miscibility modeling, (ii) laboratory-based miscibility tests and simulations, (iii) use of correlations
in the literature. In this study an Equation-of-State is used to facilitate 1-D slim tube simulations, through which the
MMP between the volatile oil and various injectants is determined.

Applying commonly used MMP correlations to our study reservoir provided poor results – with some correlations
predicting MMP values at up to approximately 140% of those determined through 1-D slim tube simulation. This
was due to the inability of many correlations to account for the effect of volatile oil composition and high reservoir
temperatures. Our results show that MMP correlations can give misleading results if applied to reservoir data which
is dissimilar to that with which the correlation was developed.

The MMP with 100% CO2 injectant was calculated to be 2680 psig at 285°F. The highest MMP was obtained with
pure N2 injectant at greater than 3600 psig. Variation of injectant composition had clear effects on the MMP. The
increase of both N2 and CH4 concentrations in the injectant raised MMP by 25.6 psi/mol% and 15.2 psi/mol%,
respectively. Increasing the C2H6 and n-C5H12 concentrations decreased MMP by 6.4 psi/mol% and 34.4 psi/mol%,
respectively. The results of this study will help in developing an optimized injection strategy for the candidate
reservoir, in addition to providing insights into key design parameters for surface facilities.

Introduction

The primary aim of the project was to assess the potential of Tirrawarra oil to achieve miscibility with injection
gases of varying composition. Of specific interest was the determination of the minimum pressure at which the
Tirrawarra oil and the injection gases would become miscible at reservoir temperature (i.e. the minimum miscibility
pressure or MMP) and the effect of different common injection gas components on the MMP.

∗
Now with Santos Ltd.
2 SPE 116782

The Tirrawarra Field is located in the Cooper Basin, central Australia, 50km North of Moomba (Figure 1).
Tirrawarra Reservoir oil is a sweet undersaturated volatile oil with a high CO2 content (≈ 20 mol%). The potential of
the oil for miscibility with different injection gases was assessed through fluid characterization using the Peng-
Robinson Equation-of-State (PR-EoS). After characterization of the fluid model via calibration to laboratory
Pressure-Volume-Temperature (PVT) experiments, a series of simulated slim tube experiments were performed
using the fluid model and twenty-six different injection gas models, each with a different molecular composition.
Results from the slim tube simulations were analyzed using the ‘break-over point’ method as suggested by Danesh1
to determine the minimum miscibility pressure (MMP) between the reservoir oil and the injectant considered.
Simulation of a number of slim tube experiments with a number of different injection gas mixtures allowed clear
trends in the effect of injection gas composition on the MMP between the gas and oil to be observed.

The MMP obtained using slim tube simulation was compared to those estimated using various widely used MMP
correlations available in the literature. The most appropriate correlations for estimation of MMP between the
Tirrawarra oil and different injection mixtures were determined. Applying commonly used MMP correlations to our
reservoir fluid generally provided poor results – with some correlations predicting MMP values at up to
approximately 140% of those determined through the slim tube simulation. The most erroneous MMP values were
calculated when utilizing correlations which were developed using reservoir data sets that were most dissimilar to
the situation under study. Error was introduced due to the inability of these correlations to account for the effect of
volatile oil composition and high reservoir temperatures. Our results show that MMP correlations can give
misleading results if applied to reservoir data which is dissimilar to that with which the correlation was developed.

Theory & Definitions

Benefits of Miscible Gas Injection. The miscible gas injection process is an Enhanced Oil Recovery (EOR) method
that takes advantage of fluid thermodynamic properties to mix a displacing gas phase with a displaced oil phase in
order to create a single miscible displacement phase. Upon achievement of miscibility the interfacial tension (IFT)
between the displacing gas and the reservoir oil reduces to zero, resulting in a theoretically infinite local capillary
number (Nc). Nc is a dimensionless fluid dynamics term (see Equation 1 below) that describes the relative effect on
flow of viscous forces vs. capillary forces acting between two immiscible fluids at the pore scale (e.g. water/oil or
gas/oil under immiscible conditions). During a miscible gas flood the miscible phase essentially acts as a highly
efficient solvent when contacting pores filled with residual reservoir oil, improving microscopic flood efficiency.
Gas becoming single phase with oil also causes the oil phase to swell. This increase in oil volume causes expulsion
of some water from pores, in turn improving the relative permeability to the oil when flowing through a water-wet
porous medium.

Volumetric sweep efficiency is also improved in a miscible gas injection process as the viscous-to-gravity ratio (Rv/g
– Equation 2) increases the closer the displacing gas becomes to being miscible (as the solvent-oil density
difference is less than the initial gas-oil density difference). This leads viscous forces to have increased dominance
in the flow regime over gravity forces which in-turn leads to a more efficient vertical displacement profile.
vdisplacing μ displacing
Nc = ……………… (Equation 1)
σ displacing / displaced

vμo L
Rv / g = ……………………... (Equation 2)
ko gΔρh

MMP Determination from Slim Tube Experiment. The slim tube experiment is performed to determine the MMP
between a specific injection gas and reservoir oil at the prevailing reservoir temperature. By performing the
experiment at multiple pressures and measuring the ultimate oil recovery (usually truncated at around at 1 - 1.2 pore
volumes of gas injected) the MMP can be determined, as in Figure 2. During an immiscible gas injection scheme
the recovery of oil from the slim tube is expected to increase monotonically with increasing displacement pressure,
however above MMP this increase in recovery will be less per unit pressure increase (as the displacement, once
miscible, is already highly efficient for the reasons given above). The change in the incremental oil recovery per
pressure increase when the pressure exceeds MMP can be used to determine MMP using the ‘breakover method’. By
running numerous slim tube experiments above and below an estimated MMP, a graph similar in shape to Figure 2
will result. This is true of any injectant/oil system. A linear trend should be apparent in the data initially (far below
the MMP). Following a transition period, the data should again conform to a linear (approximately horizontal) trend
for pressures above MMP. By extrapolating both of these trends the MMP can be determined to occur at the
intersection of both linear trend lines.1
SPE 116782 3

Relevance of Slim Tube MMP. It must be stressed that the slim tube cannot be considered to represent the real flow
conditions of a gas injection into an oil reservoir; it is merely a well controlled experiment which allows us to gain
an idea of the phase behaviour and miscibility conditions that exist between the injected gas and reservoir fluids. In
fact the MMP measured through the slim tube experiment is often low, as effects such as dispersion and reservoir
heterogeneity have a negative effect on the process in real reservoirs. It is also important to note that water (connate
or otherwise) is always present in a real reservoir and will impact the miscibility process. The effect of water on the
miscibility process is not accounted for by the slim tube experiment. Even if miscibility is not actually achieved in
the reservoir when injecting at the slim tube MMP, the lowering of the inter-facial tension between the two fluids at
the flood front in a near-miscible flood will still provide a more efficient gas-flood displacement process than is
achievable via a totally immiscible flood.

MMP and the Importance of Bubble Point Pressure. There is potential for the MMP between an injection gas and
reservoir oil to be at a pressure lower than the bubble point pressure of the oil. If a slim tube experiment is conducted
at pressures below the bubble point pressure for the oil, free gas will exist within the slim tube. In this circumstance
injection gas will contact with liberated gas in the tube as well as with the oil. Detrimental experimental effects that
stem from injectant mixing with liberated gas are outlined by Yellig & Metcalfe3 including:

• Preferential displacement of liberated gas by injectant gas, reducing breakthrough time, reducing ultimate oil
recovery and increasing GOR at breakthrough.
• Mixing of liberated gas volatile components (e.g. Methane) with injectant gas. This tends to have a detrimental
effect on achievement of injectant-oil miscibility.

From this author’s research, no conclusive studies appear to have been conducted to examine the actual effects of
free gas saturation in the slim tube on the accuracy of MMP results. Dicharry et al.4 do address this issue in part in
their paper “Evaluation and Design of a CO2 Miscible Flood Project, SACROC Unit, Kelly-Snyder Field”. Dicharry
et al. conducted slim tube experiments specific to the oil of the Kelly-Snyder Field and determined that, whilst free
gas saturation was generally detrimental to the achievement of miscibility in the experiment, it was possible to
achieve and observe miscibility between their oil and a pure CO2 injection gas at pressures lower than the oil bubble
point. Oil recoveries in their experiments were comparable when miscible flow occurred above bubble point (93.7%)
and below bubble point (90.2%).

Equipment & Processes

Fluid Sampling. Proper reservoir fluid sampling is the first step towards any successful PVT study. Samples must
be taken in different ways depending upon reservoir and fluid characteristics, tools and expertise available. Data
requirements should also be a key consideration before undertaking any fluid sampling operation. The challenge that
must be overcome regardless of the sampling method employed is to capture a fluid sample that represents as closely
as possible the in-situ reservoir fluid. There are two main methods through which fluid sampling can be achieved –
subsurface sampling and surface recombination sampling. Whilst it is not the purpose of this paper to discuss the
issue of fluid sampling in-depth, the reader is referred to the reference list for a number of specialized papers
regarding this topic.6,7,8

PVT Experimentation. PVT experimentation is carried out on reservoir fluid in order to understand and quantify
the phase behaviour and properties of the reservoir fluid.1 The two major PVT experiments applicable to volatile oils
are the Constant Composition Expansion (CCE) Test and the Constant Volume Depletion (CVD) Test. The
Differential Liberation (DL) experiment, commonly performed on black oils, can also be performed on volatile oil
although it does not offer the advantage of maintaining a longer contact period between the gas and oil phases as in
the CVD test. Again it is not the purpose of this paper to provide an in depth assessment of the benefits and
drawbacks of individual PVT experiments and the reader is referred to the reference list for further resources
regarding this topic.1,9

Slim Tube Experiment. The slim tube (Figure 3) can be considered to be a 1-dimensional model reservoir. The
tube may be packed with glass beads or crushed sand, not to simulate the porous medium of the reservoir as may be
assumed, but instead to provide a medium for the mixing of the injectant and the oil in a multiple contact process.3
The test can also be run with no packing at all. The tube usually has a linear length of between 16 and 130 feet,
although the tube is generally coiled for space consumption reasons. Danesh1 suggests that a tube of length 40 feet is
usually sufficient for miscibility experiments.

The tube is initially saturated with either real or synthetic reservoir oil and is kept at reservoir temperature and at the
desired pressure of investigation. The oil is then displaced from the tube by injecting a volume of gas. The volume
4 SPE 116782

of gas injected is usually 1.2 pore volumes of the tube. The recovery at this point is termed the ultimate recovery, as
the incremental recovery at injected volumes above this is often minimal. The pressure differential across the tube is
quite small, and thus the process is assumed isobaric. The fluid produced from the outlet of the slim tube is flashed
to a separator at standard conditions where the rate of recovery, density and composition of the liquid phase is
measured. The gas phase from the separator is passed through a Gas Chromatograph to measure the gas
composition.

The breakthrough time for the injection gas is determined by monitoring any dramatic increase in the GOR of
produced fluids and/or by monitoring the produced gas composition. A sight glass at the tube outlet allows the
miscibility of the injected gas and tube oil to be visually verified. In a miscible process the colour of the produced
fluid at the outlet can be expected to change gradually from that of the oil to that of the injection gas as the miscible
front reaches the end of the tube. If the process is immiscible a two phase flow will be observed.1

Slim Tube Simulation. For the purposes of this project it was decided to initially conduct simulated slim tube
simulations. Simulation of the physical slim tube experiment has a number of advantages over actually performing
the experiment in the laboratory, primarily:

• Physical experiments can take a great deal of time, sometimes a number of days to perform a single run. This
can be expensive to the study group if purchasing third party laboratory time and results are returned slower.
Slim tube simulations can be run in a matter of minutes once a valid EoS is constructed.
• Long experimentation times tend to limit the number of cases that can be run (i.e. varying the pressure, varying
the injectant etc.). Simulation allows a great number of cases to be investigated quickly.
• The possibility of experimental error is eliminated.

Slim Tube Simulation Set-up and Assignment of Miscible Displacement Conditions. It was initially decided to
assign conditions required for miscible flow as 92% oil recovery at 1.2 slim tube pore volumes injected. This
assumption was in line with suggestions from previous authors.3,10,112

The assumption of miscible displacement at 92% recovery, 1.2 pore volumes gas injected was later revised however,
in favour of using the ‘break-over point’ method as suggested by Danesh1. The ‘break-over point’ method for MMP
determination was favoured because it is less ambiguous than assigning miscibility conditions to a recovery
percentage, the reasoning for this is as follows.

In the past, authors have justified the assignment of the term ‘miscible flow’ to flow in the slim tube which results in
oil recoveries of 90-95% based on the fact that these recovery percentages are commonly observed from miscible
displacement in a slim tube.3,10,11 Consequently the MMP of the oil/injectant pair would be determined as the first
pressure at which the assumed miscible recovery is observed. Miscible displacement may, however, result in many
different oil recoveries at 1.2 pore volumes injected, depending in part on the design of the slim tube and the
operating temperature. It is therefore more reasonable to determine that miscible flow is occurring in the slim tube
not simply by looking for a high recovery, but to look for a high recovery in conjunction with other miscibility
indicators. An industry standard approach to the determination of whether or not displacement in a slim tube is
miscible is still yet to be developed and a number of methods are commonly used, including but not limited to:

• 90-95% oil recovery at 1.2 pore volumes gas injected.

• Point of maximum curvature when oil recovery at 1.2 pore volumes gas injected is plotted versus displacement
pressure (the ‘break-over point method’ as in Figure 2) – The method chosen for this project.
• Analysis of composition of effluent stream from slim tube. Effluent from a miscible displacement will display
a smooth decreasing trend in the concentration of intermediate components in the effluent, whilst an
immiscible displacement will display a much sharper decrease in intermediate component concentration in the
effluent at injection gas break-through.

Data

Quality Control of PVT Data. A PVT study performed on fluid sampled from a Tirrawarra oil well in 1982 was
used as the basis for the fluid characterization and slim tube simulation detailed in this paper. The PVT study
consisted of a compositional analysis of a recombined oil sample to C11+, a CCE and CVD test, a separator test, and
an oil viscosity test. The reported Bo and Rs values were corrected using the results of the separator test.
SPE 116782 5

Fluid Data. The measured composition of the recombined reservoir fluid is tabled in Table 1. The oil density data
measured in the PVT study is shown in Figure 4 with the measured oil formation volume factor shown in Figure 5.
Solution GOR measurements are plotted in Figure 6. Measured oil viscosity data is plotted in Figure 7. In all cases
the initial EoS prediction based solely on composition and pseudo splitting is shown (as at end of Step 2 -below) as
well as the prediction from the matched, ungrouped EoS (as at end of Step 4 – below).

Results

Results of Fluid Characterization. Fluid characterization was carried out using the method as dot-pointed below.
Results and comments pertaining to each step are given underneath.

1. Enter all relevant PVT experiment data and oil compositional data into software package, including measured
molecular weight and specific gravity of the ‘C7+’ fraction.

2. Perform the steps detailed in Figure 8 to achieve an EoS which approximately matches the reported bubble
point pressure for the fluid.

The methodology shown in Figure 8 was followed. A match to within 5% of laboratory reported bubble point at
reservoir temperature was achieved by splitting the initial C7+ pseudo-component into four pseudo-components (C7-
C11, C12-C15, C16-C19, C20-C27+). The original C7+ pseudo component was split using a falling exponential splitting
method in which the distribution of component molecular percentages above C6 is assumed to follow a falling
exponential function.

3. Regress on the critical properties of oil components to achieve a match to lab PVT data

As anticipated this was quite a challenging step. The measured bubble point pressure, density of oil at bubble point,
CVD oil density data and separator test data were matched satisfactorily using a regression on the critical properties
of all components.

A regression on both the critical pressures and critical temperatures of all non-hydrocarbon components, pure
hydrocarbon and ‘pseudo-plus’ components yielded reasonable results using the regression method shown in Figure
9. It was found that repeating the regression method of Figure 9 with the non-hydrocarbon components excluded
from regression yielded better results. Throughout the regression process whilst the critical properties of the pure
components were regressed on in order to achieve a match to lab data, results were rejected if the regression method
caused the critical properties to deviate wildly from the true values and / or if the critical properties of the mixture
became inconsistent with reality (i.e. if a monotonic increase in Tc and monotonic decrease in Pc was not seen). The
laboratory data was satisfactorily matched using this method.

4. Regress on the critical volume of oil components to achieve a match to lab measured viscosity

The Lorenz-Bray-Clark viscosity model was used to match the lab measured viscosity via a regression on all
component critical volumes.

5. Group the oil components to minimize future computational expense

In this case, components with similar thermodynamic properties were grouped together, with CO2 remaining in a
separate group. It was considered important to preserve the critical properties of CO2 as much as possible in the
mathematical representation of our fluid (the EoS) as it is an important component controlling achievement of
miscibility between the Tirrawarra oil (approx 20 mol% CO2) and a planned CO2 based injection gas. The critical
properties of the groups as used by the EoS are shown in Table 2.

6. Repeat regression using Grouped Components

The regression on the grouped components was performed as in Figure 9 with the critical properties of all groups
regressed on at each regression stage. A good match to laboratory data was preserved. The final matches to oil
density, solution GOR and oil formation volume factor achieved with the grouped and ungrouped fluid models are
shown in Figure 10 to Figure 13 respectively. The match on oil viscosity (again using the Lorenz-Bray-Clark
model) is shown in Figure 14.
6 SPE 116782

Results of Slim Tube Simulation. It was necessary to conduct slim tube simulations below the oil bubble point in
order to determine the MMP between the oil and most of the injection gas mixtures considered. Although this was
initially done with some trepidation regarding the validity of the approach, results were encouraging (see “MMP and
the Importance of Bubble Point Pressure” above). No effect on recovery or achievement of miscible flow in the tube
was observed despite simulations being run at pressures below bubble point, and the simulated MMPs are in the
range of values expected. Results are summarized in Table 3.

Pure CO2 Gas Injection. By utilizing the break-over point analysis method, the slim tube simulations undertaken
assuming a pure CO2 injectant indicated an MMP with Tirrawarra oil of 2680 psig at 285°F (Figure 2).

In order to conduct a sense check on this MMP value, results were compared to those of Bon & Sarma12 who had
previously undertaken to determine the MMP between CO2 injection gas and an oil from a nearby oil reservoir, also
located in the Cooper Basin - dubbed ‘Field A’ in their report. The ‘Field A’ oil studied by Bon & Sarma was
notably more viscous than the Tirrawarra oil and contained approximately double the mole percentage of C7+
components. The experimentally determined MMP between CO2 and ‘Field A’ oil was 2795 psig.

It is reasonable to expect the MMP of ‘Field A’ oil with a CO2 gas injectant to be higher than that of Tirrawarra oil
due to the difference in the oil compositions. The proximity of the MMP values of Bon & Sarma and those of the
relevant correlations (discussed below) to the MMP determined from the method detailed in this report lends support
to the validity of the MMP determination method employed.

Although, as in this case, the MMP between an injection gas and Tirrawarra oil is evaluated to be below the bubble
point pressure of the oil, injection in the field should ideally be performed at or above the bubble point pressure.
This is to minimize the volume of liberated gas within the reservoir, due to its negative effects on oil quality and
miscible flow achievement.3

The ultimate recovery of oil in the slim tube simulation at 1.2 pore volumes injected, evaluated at bubble point
pressure, was 93-94% of slim tube pore volume when using pure CO2 injection gas.

Pure Nitrogen Gas Injection. Poor results were obtained when pure Nitrogen gas was simulated to displace the
Tirrawarra oil in a slim tube. The MMP between pure Nitrogen injection gas and Tirrawarra oil was determined to
be 3600 psig at 285 °F (Figure 14) by the breakover point method. This value is considered to be unrealistically low
given the low recovery factor seen at 3600 psig, with the analysis in this case made difficult by the range of
pressures covered in the transition zone between strictly immiscible and strictly miscible flow. The extrapolation of
the correlation line through the MMP results from other CO2/Nitrogen injection gases versus a function (‘F’ – see
Figure 19) based on the critical temperatures of the injection gas pure components (Figure 19) indicates that a more
reasonable MMP of Tirrawarra oil with a pure Nitrogen injection gas to assume would be 5250 psig.

The Effect of Nitrogen in the Injection Gas. The effect of the addition of Nitrogen to an otherwise pure CO2
injection gas on the slim tube MMP can be seen in Figure 15. As can be seen in the chart, the addition of increasing
amounts of Nitrogen to the injectant caused a monotonic increase in the MMP. The average change in slim tube
MMP between simulated CO2/Nitrogen injectant and simulated Tirrawarra oil is +25.6 psi/mol% N2 over the range
0-30 mol% Nitrogen contamination in the base CO2 injectant.

The ultimate recovery of oil in the simulations at 1.2 pore volumes injected, evaluated at bubble point pressure, is
93-94% for Nitrogen contamination at or below 10 mol%, with recovery decreasing as shown in Figure 15 for
higher Nitrogen concentrations.

The Effect of Methane in the Injection Gas. The effect of the addition of Methane to an otherwise pure CO2
injection gas on the slim tube MMP can be seen in Figure 16. As can be seen in the chart, the addition of increasing
amounts of Methane to the injectant caused a monotonic increase in the MMP. The average change in slim tube
MMP between simulated CO2/Methane injectant and simulated Tirrawarra oil is +15.2 psi/mol% C1 over the range
0-30 mol% Methane in the injectant.

The ultimate recovery of oil in the simulations at 1.2 pore volumes injected, evaluated at bubble point pressure, is
93-94% for Methane contamination less than 20 mol%, with recovery decreasing as shown in Figure 16 for higher
Methane concentrations.

The Effect of Ethane in the Injection Gas. The effect of the addition of Ethane to an otherwise pure CO2 injection
gas on the slim tube MMP can be seen in Figure 17. As can be seen, the addition of increasing amounts of Ethane to
SPE 116782 7

the injectant caused a monotonic decrease in MMP. The average change in slim tube MMP between simulated
CO2/Ethane injectant and simulated Tirrawarra oil is -6.40 psi/mol% C2 over the range 0-30 mol% Ethane in the
injectant.

The ultimate recovery of oil in the simulations at 1.2 pore volumes injected, evaluated at bubble point pressure, is
93-94%.

The Effect of n-Pentane in the Injection Gas. The effect of the addition of n-Pentane to an otherwise pure CO2
injection gas on the slim tube MMP can be seen in Figure 18. As can be seen, the addition of increasing amounts of
n-Pentane to the injectant caused a monotonic decrease in MMP. The average change in slim tube MMP between
simulated CO2/n-Pentane injectant and simulated Tirrawarra oil is -34.4 psi/mol% C5 over the range 0-30 mol% n-
Pentane in the injectant.

The ultimate recovery of oil in the simulations at 1.2 pore volumes injected, evaluated at bubble point pressure, is
93-94%.

Correlation of MMP to Injection Gas Pure Component Critical Temperatures. In Figure 19 the strong
correlation between the critical temperature of the injection gas pure components and the simulated MMP value in
each case is evident. Extrapolating a linear correlation of the Nitrogen contaminant cases to the critical temperature
of Nitrogen (-232.4 °F) the MMP of a pure Nitrogen injection gas would be approximately 5250 psig. When
considering the large transition period between strictly immiscible and strictly miscible flow in the pure Nitrogen
gas stream cases (Figure 14) this value is considered feasible.

Results of MMP Estimation Using Established Correlations.

Literature Review. Numerous authors have undertaken to study the miscible displacement of oil by pure and
impure CO2 gas. From these studies have arisen a number of correlations that can be used to estimate the MMP
between a given oil and a given injection gas.

From a review of the literature regarding the subject1,3,10,11,13,14,15,16,17,18,19,20,21,22 it was apparent that two major types
of correlations are in use currently; those that predict the MMP between pure CO2 injectant and oil and those that
predict the change in MMP when a ‘contaminant gas’ is introduced into the CO2 injectant. A number of common
pure CO2 MMP correlations of use in the petroleum industry were applied to the Tirrawarra oil composition as
reported in the Tirrawarra PVT report. A summary of results can be seen in Table 5 and Figure 20. We will discuss
in more detail those correlations that we found best apply to the Tirrawarra oil, namely:

Pure CO2 Correlations

• Yuan et al.16 (2005) (inc. impure correction factor)
• Alston et al.21 (1985) (inc. impure correction factor)
Impure CO2 Correlations
• Sarma & Emera.22 (2006)
General Injection Gas Correlations
• Eakin & Mitch18 (1988) (inc. impure correction factor)

Specific equations for each correlation are tabulated in Table 4.

Pure CO2 Correlations

Yuan et al. (2005)

Yuan et al.16 developed a correlation for the MMP between oil and pure CO2 injectant by utilizing analysis of 70
MMPs calculated from Peng-Robinson EoS models for different oils, combined with 41 MMPs calculated from
physical slim tube experiments. Yuan et al. argue in their paper that whilst MMP between CO2 and reservoir oil
tends to increase with increasing reservoir temperature, this trend is not linear or even monotonically increasing as is
often assumed by other authors. Using data from an oil referred to in their paper as ‘Oil E’, Yuan et al. show the
trend of MMP vs. reservoir temperature is more likened to a concave downwards quadratic function, with the MMP
initially increasing with temperature, reaching some maximum value, and then decreasing with increasing
temperature. The trend does show that MMP tends to monotonically increase with temperature over the normal
range of reservoir temperatures, however.
8 SPE 116782

Yuan et al. also developed a correlation for use when the CO2 injectant is contaminated with Methane. They state
that the impure injectant correlation only provides reliable results when the concentration of Methane in the injection
gas is less than or equal to 40 mol%.

The Yuan et al. correlation is the only correlation that the authors were able to uncover in the literature that was
developed using reservoir temperature data that includes temperatures as high as that seen in Tirrawarra. The molar
percentage of C2-C6 in the Tirrawarra Oil is 30.63% - higher than the oils used to calibrate the Yuan et al. correlation
(10.36% – 24.70%). This may be a source of error when using this correlation in this instance.

Alston et al. (1985)

Alston et al.21 developed a correlation for pure CO2 MMP by studying a set of 30 pure CO2 and 38 impure CO2
miscible displacement data. Alston et al. also developed a correction factor (F), to be applied when the pure CO2
stream contained other gases, based on the critical temperature of the injection gas determined by the weight fraction
mixing rule.

They found that the volatile fraction (Xv i.e. C1 and N2) in live oil tended to hinder the achievement of miscibility,
whilst the intermediate fraction (Xi i.e. C2-C4, H2S, CO2) tended to encourage miscibility with oil. The correlation of
Alston et.al correlates MMP with reservoir temperature and the molecular weight of the heaviest oil fraction (in this
case C5+). Alston et.al also included the ratio of volatile/intermediate components in the oil as a correction factor.

Errors may occur when using the Alston et.al correlation to predict Tirrawarra oil MMP as Tirrawarra reservoir is
some 38 °F hotter than the hottest reservoir studied by Alston et.al in the development of the correlation. Also, the
molecular weight of the C5+ fraction in the Tirrawarra oil sample is less than those considered by Alston et al. when
generating their correlation.

Impure CO2 Correlations

Emera & Sarma (2006). Emera & Sarma22 developed an equation to predict the MMP between oil and impure CO2
injectant. The authors report that the correlation can be used with non-CO2 concentrations in the injectant of 3-78
mol%, with a maximum N2 concentration of 20 mol%22. The correlation uses a genetic-algorithm based method to
predict MMP between an injection gas and oil, using the solvency of the injection gas in the oil, which is in turn a
function of the critical properties of the injection gas. The equation relies on the pure CO2/oil MMP being measured
through some other method. MFi values for different components in the injection gas are reported by the authors,
and are summarized in Table 4.

General Injection Gas Correlations

Eakin & Mitch (1988). Eakin & Mitch18 developed a correlation for the MMP between light and medium oils and
various injection gases including N2, H2S, CO2, Rich Natural Gas and Lean Natural Gas. Their correlation was based
on a data set of 102 light and medium oils. They report that they were able to correlate all of the data aside from five
outliers, all of which contained CO2 in the injectant.

A limitation of the Eakin & Mitch correlation is that it does not take into account the presence of some possible
injection gas components (e.g. SOx, O2) which may be present, particularly if the injection gas is flue gas. The
authors do not provide explicit data on the oil samples they used to develop their correlation, but do provide a chart
showing regions in which the MMPs of their samples plot on an axis of ln (Pr) vs. 1/Tr (reproduced schematically in
Figure 21). The results of applying the correlation to a displacement system comprising Tirrawarra oil and CO2
Injectant, as well as a system comprising Tirrawarra oil and Methane injectant can also be seen schematically on
Figure 21. The results from the Tirrawarra case have low 1/Tr values compared to the cases studied by the authors
when developing their correlation. This would indicate that the reservoir temperature in Tirrawarra is outside the
range of those tested by the authors, which could be a source of error when applying this correlation in this case.

Results of using Pure CO2 Correlations to Predict MMP for Tirrawarra Oil. The pure CO2 MMPs calculated
using various correlations published in the literature, including those described above, are summarized in Table 5
and Figure 20.

The correlation that was felt to be the most appropriate with which to predict pure CO2 MMP in Tirrawarra is that of
Yuan et al. The correlation approximated the pure CO2 MMP calculated using our EoS with an error of +10.99%,
which is on the order of errors seen when performing physical slim tube experiments.16 The simulated MMP fell
close to the error range specified by Yuan et al. at ±10.3% of their correlation value. The data used to develop the
SPE 116782 9

Yuan et al. correlation was also the most appropriate to the Tirrawarra oil and reservoir conditions.

The correlation of Eakin & Mitch matched the simulated pure CO2 MMP with an error of +14.27%.

The correlations which estimated the Tirrawarra EoS pure CO2 MMP poorly, i.e. Holm & Josendal, Yellig &
Metcalfe, Cronquist and Glaso, were all developed using oil data with a significantly higher mol% of the C7+
fraction than is present in the Tirrawarra oil. This may potentially account for the large over-estimation of pure CO2
MMP when these correlations are applied to the Tirrawarra oil data, as the correlations are developed for oils with a
generally lower API gravity and higher MMP with CO2 than Tirrawarra oil. In addition, the methods of Holm &
Josendal and Yellig & Metcalfe are graphical methods that required direct extrapolation of the original temperature
data beyond the original correlation data range in order to take account of the Tirrawarra reservoir temperature. This
extrapolation introduces significant error into the MMPs calculated using these methods, especially when the non-
linear behaviour of MMP with temperature as shown by Yuan et al. is considered.16

Results of using Other Correlations to Predict MMP for Tirrawarra Oil. The results of applying the Eakin &
Mitch and the Emera & Sarma correlations to the injection gas mixtures can be seen in Figure 22 and Figure 23
respectively.

Eakin & Mitch Correlation Results. The Eakin & Mitch correlation predicted both the value and relative change
in MMP when Methane was added to the injection gas reasonably well. Although inaccurate in regards to the actual
MMP values determined using simulation, the correlation predicted the relative change in MMP with increasing
amounts of N2 and C5 relatively well. The correlation was inaccurate in its prediction of MMP with addition of C2 to
the gas stream, indicating that MMP increases with increase in Ethane concentration. This is contrary to results seen
in this and other studies.

Emera & Sarma Correlation Results. The Emera & Sarma correlation predicted both the value and relative
change in MMP when Ethane was added to the injection gas stream extremely accurately. Results from the
correlation agreed very closely to those developed using the EoS in the case of Ethane presence in the injection gas.
The Emera & Sarma correlation underestimated the reduction in MMP due to n-Pentane addition to the injection gas
calculated by simulation of the n-Pentane cases. The correlation performed poorly in predicting MMP when
Methane was present in the gas stream, indicating a dramatic exponential rise in the MMP as Methane concentration
increased. This is not supported by simulation results. The correlation authors indicate that the correlation is to be
used for N2 concentrations in the injection gas of ≤ 8 mol%, so the correlation was of limited use in predicting
change in MMP with increase in N2 concentration.

Conclusions

Bearing in mind the limitations of this study and various assumptions made, the following conclusions have been
drawn:

• MMP was found to vary inversely to the variance in critical temperature of the injection gas mixture under
constant environmental conditions (i.e. invariant reservoir temperature, pressure and oil composition). The
addition of gases to the CO2 injection gas which increased the critical temperature of the injection gas (i.e. C2
and n-C5) reduced the MMP compared to the pure CO2 injection case. The addition of gases to the CO2 injection
gas which decreased the critical temperature of the injection gas (i.e. N2 and C1) increased the MMP compared to
the pure CO2 injection case.
• Although reservoir temperature has been shown by various authors to have a directly proportional relationship to
MMP over the normal range of reservoir temperatures16, it was found in this case that the volatile nature of
Tirrawarra oil helped to offset the negative effects on miscibility of the high reservoir temperature. It was
determined that injection at or around the bubble point pressure of the Tirrawarra oil would result in a miscible
or near miscible flood using a primarily CO2 based injection gas.
• The volatile nature of the oil in this study and the high prevailing reservoir temperature made it challenging to
find a suitable correlation to estimate MMP from established literature. It was observed that the correlations that
were calibrated using data most similar to the reservoir of interest provided the most reliable results.
• Whilst it is challenging to characterize a volatile oil using EoS, the EoS / slim tube simulation method employed
in this paper was found to provide good results in estimating MMP - allowing the authors to accurately answer
key questions regarding the miscible flood potential of the Tirrawarra reservoir oil within a tight designated
timeframe.
10 SPE 116782

Acknowledgements

The authors would like to thank Santos Ltd. for providing access to the necessary data and facilities to complete this
project. Special thanks also to Jan Bon Sr. and other staff of Petrolab for allowing access to their facilities and for
their helpful advice and support.

Nomenclature

a1-a10 = Coefficients in Yuan et al. pure and impure CO2 correlations

Bo = Oil Formation Volume Factor (rb/stb)
co = Undersaturated oil compressibility, 1/psi
C = Separator test shrinkage factor, STB/rbb
Ci = Hydrocarbon component with i carbon atoms per molecule
Ci+ = Hydrocarbon components with i and greater carbon atoms per molecule
Δ = “Difference Between”
EOR = Enhanced Oil Recovery
EOS = Equation of State
F = Coefficient in Alston et al. impure-CO2 MMP correlation
GOR = Gas-Oil Ratio
g = Acceleration due to Gravity
h = Formation Height
m = Coefficient used in Yuan et al. impure-CO2 MMP correlation
MCi+ = Molecular weight of hydrocarbon components with i and greater carbon atoms per molecule
MFi = Emera & Sarma coefficient for gas component i
MMP = Minimum Miscibility Pressure
MMPPURE = MMP between pure CO2 injection gas and oil
MMPIMPURE = MMP between CO2 injection gas containing impurities and oil
μ = Viscosity
Nc = Capillary Number
ω = EOS acentric factor
P = Pressure, psi
Pc = Critical pressure, psia
PC = Critical pressure, MPa
PC,CO2 = Critical pressure of CO2, MPa
PCO2 = Molar percentage of CO2 in injection gas, %
PCM = Molar average pseudocritical pressure of injection gas, MPa
PCW = Weight average pseudocritical pressure, MPa
PC2-C6 = Molar percentage of C2-C6 in oil, %
Pr = Reduced Pressure
Pr, impure = Reduced pressure of impure CO2 injection gas
Pr, pure = Reduced pressure of pure CO2 injection gas
PR = Peng-Robinson (EOS)
PVT = Pressure, Volume, Temperature
ρ = Density
R = Stock Tank Gas-Oil Ratio, scf/stb
Rs = Solution Gas-Oil Ratio, scf/stb
Rsep = Separator Gas-Oil Ratio, scf/sep.bbl
Rv/g = Viscous-to-Gravity Ratio
S = Separator-Stock Tank shrinkage factor, stb/sep.bbl
σ = Interfacial Tension (IFT)
Sor = Residual oil saturation, fraction
SRK = Soave-Redlich-Kwong (EOS)
T = Reservoir/PV Cell temperature, °F
Tc = Critical temperature, °R
TC = Critical temperature, °C
TcK = Critical temperature, °K
TcM = Molar average pseudocritical temperature of Injection Gas, °F
TcMK = Molar average pseudocritical temperature of Injection Gas, °K
Tcw = Weight average pseudocritical temperature of Injection Gas, °F
SPE 116782 11

TCW = Weight Average pseudocritical temperature using the MFi, °C

Tr = Reduced temperature
V = Volume
ν = Velocity
wi = Weight fraction of component i in fluid
xi = Mole fraction of gas component i
Xv = Molar fraction of C1 and N2 in reservoir oil
Xi = Molar fraction of C2-C4, H2S and CO2 in reservoir oil
yi = Mole fraction of gas component i

Subscripts
b = “At bubble point pressure”
d = “Obtained from differential pressure experiment”
f = “Following flash to stock tank conditions”
g = gas
i = Component i, injection
o = oil
w = water

SI Metric Conversion Factors

bbl x 1.589874 E-01 = m3
ft3 x 2.831685 E-02 = m3
scf/stb x 5.6146 E+00 = m3/ m3
cP x 1.0* E+03 = Pa.s
ft x 3.048* E-01 = m
in. x 2.54* E+00 = cm
(°F – 32)/1.8 E+00 = °C
°F + 459.67 E+00 = R
psi x 6.894757 E+00 = kPa
* Conversion factors are exact

References

1. Danesh, Ali (1998). “PVT and Phase Behaviour of Petroleum Reservoir Fluids”. Amsterdam; New York : Elsevier

2. Sarma, H.K. (2007). Enhanced & Improved Oil Recovery Methods (Academic Course Notes). Australian School of Petroleum,
Adelaide.

3. Yellig, W.F. Metcalfe, R.S. (1980). “Determination and Prediction of CO2 Minimum Miscibility Pressure”. JPT, January. 1980.
pp. 160-168

4. Dicharry, R.M. Perryman, T.L., Ronquille, J.D. (1973). “Evaluation and Design of a CO2 Miscible Flood Project, SACROC Unit,
Kelly-Snyder Field”. SPE Paper 4083.

5. Ahmed, T. (2007) “On Equations of State”, SPE Paper 107331, Presented at the 2007 SPE Latin American and Caribbean
Conference on Petroleum Engineering, Buenos Aires, Argentina, 15-18 April 2007.

6. N.R. Nagarajan, M.M. Honarpour, K. Sampath. (2006). “Reservoir Fluid Sampling and Characterization – Key to Efficient
Reservoir Management”. JPT, August 2007. pp. 80-91

7. H.A. El-Banbi, W.D. McCain. (2001). “Sampling Volatile Oil Wells”. SPE Paper 67232, Presented at the SPE Production and
Operations Symposium, Oklahoma City, U.S.A, 24-27 March 2001

8. J. Bon, H.K. Sarma, J.T. Rodriguez, J.G. Bon. (2007). “Reservoir Fluid Sampling Revisited – A Practical Perspective”. SPEREE
December 2007, pp. 589-596.

9. L.P. Dake. (1978). “Developments in Petroleum Science 8 - Fundamentals of Reservoir Engineering”. Ed. 19. Amsterdam:
Elsevier Science.

10. Graue, D.J. Zana, E.T. (1981). “Study of a Possible CO2 Flood in Rangely Field”. SPE Paper 7060-PA, July. 1981.

11. Jacobson, H.A. (1972) “Acid Gases and Their Contribution to Miscibility”, JCPT, 57-59 (April-May). 1972.
12 SPE 116782

12. Bon, J. Sarma, H.K. (2004) “A Technical Evaluation of a CO2 Flood for EOR Benefits in the Cooper Basin, South Australia”.
SPE Paper 88451. Presented at the SPE Asia-Pacific Oil and Gas Conference and Exhibition, Perth, Australia, 18-20 October
2004.

13. Enick, R.M. Holder, G.D. Morsi, B.I. (1988). “A Thermodynamic Correlation for Minimum Miscibility Pressure in CO2
Flooding of Petroleum Reservoirs”. SPE Paper 14518. University of Pittsburgh.

14. Sebastian, H.M. Wenger, R.S. Renner, T.A. (1985). “Correlation of Minimum Miscibility Pressure for Impure CO2 Streams”.
SPE Paper 12648. Amoco Production Company.

15. Glaso, O. (1985). “Generalized Minimum Miscibility Pressure Correlation”. SPE Paper 12893. Petroleum Technology Research
Inst.

16. Yuan, H. et al. (2004). “Improved MMP Correlations for CO2 Floods Using Analytical Gas Flooding Theory”. SPE Paper
89359. Presented at the 2004 SPE Symposium on Improved Oil Recovery, Tulsa, Oklahoma. 17-21 April 2004.

17. Bon, J. Sarma, H.K. (2005). “An Investigation of Minimum Miscibility Pressure for CO2-Rich Injection Gases with Pentanes-
Plus Fraction”. SPE Paper 97536. Presented at the SPE International Improved Oil Recovery Conference in Asia-Pacific, Kuala
Lumpur, Malaysia. 5-6 December 2005.

18. Eakin, B.E. Mitch, F.J. (1988). “Measurement and Correlation of Minimum Miscibility Pressures of Reservoir Oils”. SPE Paper
18065. Presented at the 63rd SPE Annual Technical Conference and Exhibition, Houston, Texas. October 2-5. 1988.

19. Benham, A.L. et al. (1959). “Miscible Fluid Displacement – Prediction of Miscibility”. SPE Paper 1484-G. Presented at AIChE-
SPE Joint Symposium, San Francisco, December 6-9, 1959.

20. Kuo, S.S. (1985). “Prediction of Miscibility for the Enriched-Gas Drive Process”. SPE Paper 14152. Presented at the 60th SPE
Annual Technical Conference and Exhibition, Las Vegas, Nevada. September 22-25 1985.

21. Alston, R.B. Kokolis, G.P. James, C.F. (1985). “CO2 Minimum Miscibility Pressure, A Correlation for Impure CO2 Streams and
Live Oil Systems” SPE J., 268-274 (April 1985).

22. Emera, M.K. Sarma, H.K. (2005). “A Reliable Correlation to Predict the Change in Minimum Miscibility Pressure when CO2 is
Diluted With Other Gases”. SPE Paper 93478. Presented at the 2005 SPE Middle East Oil and Gas Show and Conference,
Manama, Bahrain. 12-15 March 2005.

Tables

Table 1 - Composition of recombined reservoir fluid

Table 2 – Regressed Properties of Grouped Fluid Components

SPE 116782 13

Table 3 - Calculated MMP between Tirrawarra Oil and 26 separate injection gases using slim tube simulation
14 SPE 116782

Correlation Equation Equation Data Range

Author (Pure CO2 Injectant) (Impure CO2 Injectant)
Yuan et al. MMPPURE = a1 + a2 M C 7 + + a3 PC 2−C 6 + MMPPURE Pure:
= 1.0 + m( PCO 2 − 1)
MMPIMPURE 71 ≤ T ≤ 300 °F
⎛ ⎞
⎜ a + a M + a PC 2−C 6 ⎟T + Where: 139 ≤ MC7+ ≤ 319
⎜ 4 5 C7+ 6 M 2 ⎟ m = a + a M + a PC + 2% ≤ PC2-C6 ≤ 40.03%
⎝ C7+ ⎠ 1 2 C 7 + 3 2 − C 6
2
(a7 + a8MC 7 + + a9 MC 7 + + a10 PC 2 −C 6 )T 2
⎛ PC ⎞ Impure:
⎜ a + a M + a 2 − C 6 ⎟T +
Where, ⎜ 4 5 C 7 + 6
M
2 ⎟ 110 ≤ T ≤ 300 °F
a1 = -1.4634 E+03 a6 = 8.1661 E+03 ⎝ C 7 + ⎠
a2 = 0.6612 E+01 a7 = -1.2258 E-01
2 140.6 ≤ MC7+ ≤ 245.4
a3 = -4.4979 E+01
a4 = 0.2139 E+01
a8 = 1.2283 E-03
a9 = -
(a 7 + a 8 M C 7 + + a 9 M C 7 + + a 10 F R )T 2
10.36%≤PC2-C6≤ 24.70%
a1 = -6.5996 E-02 a6 = -2.7344 E-02
4.0152 E-06
a5 = 1.1667 E-01 a10 = - a2 = -1.5246 E-04 a7 = -2.6953 E-06 0%≤ mol% C1 in
9.2577 E-04 a3 = 1.3807 E-03 a8 = 1.7279 E-08 injection gas ≤ 40%
a4 = 6.2384 E-04 a9 = -3.1436 E-11
a5 = -6.7725 E-07 a10 = -1.9566 E-08

Alston et MMPPURE = 8.78 x10 −4 (T )1.06 M C1.578+ ( X v / X i ) 0.136 MMPIMPURE = F ( MMPPURE ) 90 ≤ T ≤ 243 °F
169.2 ≤ MC5+ ≤ 302.5
al. ⎛ 87.8 ⎞
1.935⎜⎜ ⎟
⎟ 0.141 ≤ Xv/Xi ≤ 13.61
⎛ 87.8 ⎞ ⎝ Tcw ⎠
F = ⎜⎜ ⎟⎟ 70.7 ≤ Tcw ≤ 115.7
⎝ Tcw ⎠ Tc of H2S and C2 in
n
Tcw = ∑ wi Tc i − 459.7 Injection Gas altered to
i
125.33 °F

Emera & - Pr , impure ⎛ T ⎞ ⎛ ⎞

2
105.44 ≤ T ≤ 233.96 °F
= 3 . 406 + 5 . 786 ⎜ CW ⎟ − 23 . 0 ⎜ T CW ⎟ + 166.2 ≤ MC5+ ≤ 267.5
Sarma Pr , pure ⎜ T C , CO ⎟ ⎜ T C , CO ⎟
⎝ 2 ⎠ ⎝ 2 ⎠ 0 ≤ Xv ≤ 0.486
3 4
⎛ T ⎞ ⎛ T ⎞ 0.0131 ≤ Xi ≤ 0.403
20 . 48 ⎜ CW ⎟ − 5 . 7 ⎜ CW ⎟
15.332 ≤ TCW ≤ 131.18
⎜ T C , CO ⎟ ⎜ T C , CO ⎟
⎝ 2 ⎠ ⎝ 2 ⎠ (in °F)
MMP IMPURE 764.35 ≤ PCW ≤ 1174.81
P r , impure =
PCW (in psia)
n
PCW = ∑ i =1
w i PCi

MMP PURE
P r , pure =
PC , CO 2
n
T CW = ∑ i =1
MF i w i T Ci

Eakin & - ⎛ 0.06912 ⎞ ⎛ 0.005955M C 7 + ⎞

ln( Pr ,impure ) = ⎜⎜ 0.1697 − ⎟⎟ y C1 ( M C 7 + ) 0.5 + ⎜⎜ 2.3855 − ⎟⎟ y C 2 −C 4
⎝ Tr ⎠ ⎝ Tr ⎠
Mitch
⎛ 3
⎞
⎛ 0.01023 ⎞ ⎜ 0.0005899( M C 7 + ) 2 ⎟
+ ⎜⎜ 0.1776 − ⎟⎟ y N 2 ( M C 7 + ) 0.5 + ⎜ 0.01221( M C 7 + ) − ⎟ y CO 2
⎝ Tr ⎠ ⎜ Tr ⎟
⎝ ⎠
⎛ 101.429 0.00375( M C 7 + ) ⎞
+ ⎜⎜ + ⎟⎟ y H 2 S
⎝ M C 7+ Tr ⎠
Where;
MMPIMPURE n
Pr ,impure = PcM = ∑ xi PCi
PcM .......... i =1

T + 460 n
Tr = TcM = ∑ xi TCi
TcM + 460 .......... i =1

Table 4 – Details of Correlation Methods considered to model MMP in Tirrawarra Oil

SPE 116782 15

Method Pure CO2 MMP (psig) Error (%)

Initial Pressure* 4300* -
Tirrawarra EoS Bubble Point* 2995* -
Slim Tube Simulation 2680 0.00%
Cronquist (1978) 3821 42.56%
Holm & Josendal (1974) 3700 38.06%
Glaso (1985) 3449 28.71%
Yellig & Metcalfe (1980) 3335 24.44%
Eakin & Mitch (1988) 3062 14.27%
Yuan et.al (2005) 2975 10.99%
Alston et al. (1985) 2222 -17.09%
Table 5 - Results of MMP Correlations applied to Tirrawarra Oil Data (Pure CO2) injection - Error quoted as % of Slim Tube
Simulation pure CO2 injection MMP)
* This is a reference value only, not indicative of a MMP pressure
16 SPE 116782

Figures

Figure 1– Tirrawarra Location Map

Figure 2 - "Breakover Point" Method for Determination of MMP from Slim Tube Experiment

Figure 3 - Typical Slim Tube Experiment Setup (after Sarma)

SPE 116782 17

Figure 4 - Oil Density measured from separate PVT studies and a matched and un-matched PR-EoS

Figure 5 - Oil Formation Volume Factor measured from separate PVT studies and a matched and un-matched PR-EoS

Figure 6 - Solution Gas-Oil Ratio measured from separate PVT studies and a matched and un-matched PR-EoS
18 SPE 116782

Figure 7 - Oil Viscosity Data measured from separate PVT studies and a matched and un-matched PR-EoS

Figure 8 - Pseudo-Splitting Methodology

SPE 116782 19

Figure 9 - Regression Method followed for Fluid Characterization

Figure 10-Comparison of match to Oil Density data between Grouped and Ungrouped Component EoS
20 SPE 116782

Figure 11 - Comparison of match to Oil Formation Volume Factor between Grouped and Ungrouped Component EoS

Figure 12 - Comparison of match to Solution GOR data between Grouped and Ungrouped Component EoS

Figure 13- Comparison of match to Oil Viscosity data between Grouped and Ungrouped Component EoS
SPE 116782 21

Figure 14 - MMP determination for pure Nitrogen injection gas case

MMP Determination for CO2/N2 Gases

Slim Tube Simulation, 285 Degrees Fahrenheit
100 Bubble
Point

95

90
Ultimate Recovery at 1.2 PV Injected

85

80
Pure CO2
75 5% N2
10% N2
15% N2
70
20% N2
25% N2
65
30% N2

60

55

50
2200 2300 2400 2500 2600 2700 2800 2900 3000 3100 3200 3300 3400 3500 3600 3700
Pressure (psig)

Figure 15 – Detailed effect of addition of Nitrogen gas on Slim Tube MMP and Recovery at 1.2 PV injected

MMP Determination for CO2/CH4 Gases

Slim Tube Simulation, 285 Degrees Fahrenheit
100
Pb

95
Ultimate Recovery at 1.2 PV Injected

90

Pure CO2
85
5% C1

10% C1

80 15% C1

20% C1

75 25% C1

30% C1

70
2500 2600 2700 2800 2900 3000 3100 3200 3300
Pressure (psig)

Figure 16 – Detailed effect of addition of Methane gas on Slim Tube MMP and Recovery at 1.2 PV injected
22 SPE 116782

MMP Determination for CO2/C2 Gases

Slim Tube Simulation, 285 Degrees Fahrenheit Bubble
Point
100

95

90
Ultimate Recovery at 1.2 PV Injected

85

80
Pure CO2
5% C2
75
10% C2
15% C2
70
20% C2
25% C2
65
30% C2

60

55

50
2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000
Pressure (psig)

Figure 17 – Detailed effect of addition of Ethane gas on Slim Tube MMP and Recovery at 1.2 PV injected

MMP Determination for CO2/nC5 Gases

Slim Tube Simulation, 285 Degrees Fahrenheit Bubble
Point
100

90

80
Ultimate Recovery at 1.2 PV Injected

Pure CO2
70

60 5% nC5

50 10% nC5

40 15% nC5

30 20% nC5

20 25% nC5

10 30% nC5

0
1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000 3100 3200
Pressure (psig)

Figure 18 – Detailed Effect of addition of n-Pentane gas on Slim Tube MMP and Recovery at 1.2 PV injected

Figure 19 - Correlation of Simulated Slim Tube MMPs to Injection Gas Pure Component Critical Temperature Function ‘F’
SPE 116782 23

Figure 20 - Results of Pure CO2 Injection Gas Correlations applied to Tirrawarra Oil Composition

Figure 21 - Schematic of Eakin & Mitch18 Chart Showing areas in which MMPs between Light Oils and various Injectants
plotted vs. Results from Correlation applied to Tirrawarra Oil
4500

4000

3500
MMP between Injection Gas and Tirrwarra Oil

Ethane
3000
Bubble
Point n-Pentane
2500

Methane

2000
Nitrogen

1500
Eakin & Mitch - Ethane

1000 Eakin & Mitch - n-Pentane

Eakin & Mitch - Methane

500

Eakin & Mitch - Nitrogen

0
95/5 90/10 85/15 80/20 75/25 70/30
Injection Gas Mix (CO2 mol% / Secondary Gas mol%)

Figure 22 - Accuracy of Eakin & Mitch Correlation in Predicting Change in MMP with various Gas Mixtures
24 SPE 116782

7000

6000

MMP between Injection Gas and Tirrwarra Oil

5000
Ethane

n-Pentane
4000

Methane

3000 Bubble Nitrogen

Point

Emera & Sarma - Ethane

2000
Emera & Sarma - n-
Pentane

1000 Emera & Sarma - Methane

Emera & Sarma - Nitrogen

0
95/5 90/10 85/15 80/20 75/25 70/30
Injection Gas Mix (CO2 mol% / Secondary Gas mol%)

Figure 23 - Accuracy of Emera & Sarma Correlation in Predicting Change in MMP with various Gas Mixtures

1 Danesh, Ali (1998). “PV T and Phase Behaviou r of Petroleum Reservoir Flu ids”. Amsterdam; New York : Elsevier
2 Jacobson, H.A. (197 2) “Acid Gases and Their Co ntribution to M iscibility”, J CPT, 57-59 (April- May ). 1972.