Urtec 2020 2501
Urtec 2020 2501
Urtec 2020 2501
This paper was prepared for presentation at the Unconventional Resources Technology Conference held in Austin, Texas, USA, 20-
22 July 2020.
The URTeC Technical Program Committee accepted this presentation on the basis of information contained in an abstract
submitted by the author(s). The contents of this paper have not been reviewed by URTeC and URTeC does not warrant the
accuracy, reliability, or timeliness of any information herein. All information is the responsibility of, and, is subject to corrections by
the author(s). Any person or entity that relies on any information obtained from this paper does so at their own risk. The information
herein does not necessarily reflect any position of URTeC. Any reproduction, distribution, or storage of any part of this paper by
anyone other than the author without the written consent of URTeC is prohibited.
Abstract
This paper presents an analytical model for full-flow-regime rate-transient analysis in unconventional
volatile oil reservoirs. The model allows us to forecast long-term production based on transient production
data and will be applicable in wells with substantial water production. We formulated and solved governing
non-linear partial differential equations (PDEs) with an inner boundary condition of constant BHP. By
defining pseudo-variables to transform the governing non-linear PDEs to linear forms, we were able to find
solutions of pseudopressure-normalized rate for oil, gas and water phases that describe all flow regimes
over the life of a multi-fractured horizontal well. For one-dimensional flow in closed reservoirs, our
analytical solutions that show the relationship between pseudopressure-normalized rate and dimensionless
time indicate a complicated decline with an exponential relation inside an infinite series. Our study also
highlighted that simply using uncorrected PVT data without removing separator effects can be in error.
More importantly, to provide an accurate pseudopressure calculation, we conducted numerous simulation
studies and proposed saturation-pressure (S-P) relations to enhance the accuracy of calculated
pseudopressures specifically for various volatile oils. Our analytical methods yielded reasonable
interpretations of not only simulated data but also actual field data. The solutions were validated through
comparisons with results from compositional simulation; the good agreements for both ordinary and near-
critical volatile oils verified the accuracy of our analytical method; notably, the validations were almost
exact during boundary-dominated flow. We also applied our methods to analyze production data from two
wells in shale oil reservoirs in the Midland Basin. These cases illustrated that our model can handle with
wells not only in transient (infinite-acting) flow but also in boundary-dominated flow.
Introduction
Background. Tight/shale oil development has increased significantly since 2010, driven by technological
improvements that have reduced drilling costs and improved hydraulic fracturing technology in major
tight/shale plays such as the Bakken, Eagle Ford, Permian Basin, and other resources through the world
(EIA 2017). The U.S. Energy Information Administration also reported that the total reserves of shale oils
worldwide are over three hundred billion barrels. In the next twenty years, successful development of these
reservoirs will still be crucial to maintain current oil supplies and further to achieve even greater production
and reserves levels. Because of the way hydrocarbons originated and accumulated in tight/shale rock
URTeC 2501
formations, light oils have been found to be common fluids in most of these reservoirs. In production of
these less-dense crude oils, we observe substantial gas flow, resulting in significant reduction of oil
production once reservoir pressure drops below bubblepoint pressure. The industry urgently needs an
accurate and practical method to forecast production, especially an analytical approach for field application
to unconventional volatile oil reservoirs.
Summary of Previous Work. Essentially all crude oil systems produce associated gas as pressure drops
to surface conditions, with the major difference being producing gas-oil ratios. Subsurface multiphase flow
requires an entirely different approach than single-phase flow in reservoir transient analysis. The essence
of the approach in published studies is related to reduction of nonlinear effects in analytical modeling.
Perrine (1956) and Martin (1959) introduced total mobility and compressibility parameters, as in single-
phase modeling, to analyze pressure transient responses for wells experiencing multi-phase flow. Many
more studies (Camacho and Raghavan 1989; Xu and Lee 1999; Heidari and Gerami 2011; Johnson and
Jamiolahmady 2018) continue to incorporate similar ideas by extending single-phase theory to multiphase
situations. Raghavan (1976) applied the pseudopressure concept to model pressure drawdown and buildup
tests in solution-gas-drive reservoirs. Jones and Raghavan (1988) used the Boltzmann transformation to
solve nonlinear flow equations; they suggested a pseudopressure function that incorporates the influence of
changes in relative permeability and fluid properties through a steady-state theory. Bøe et al. (1989) also
investigated pressure transient behavior for two-phase radial flow by employing the Boltzmann
transformation to PDEs; they derived theoretical saturation-pressure relations that can be used to calculate
pseudopressure functions in two limiting cases: initial flow and after long producing times. Later, Fevang
and Whitson (1996) further proposed a method that improves the reliability of using pseudopressure to
handle nonlinear effects in multiphase flow systems in conventional wells. They suggested splitting the
range of integration in an integral arising in the solution process into three regions to calculate multiphase
flow pseudopressure. The accuracy of their method has been repeatedly verified by many studies. Overall,
investigators have established that, for multiphase radial flow in conventional reservoirs, both the
pseudopressure concept and the Boltzmann transformation are justified to model well behavior. However,
given the differences in development of conventional and unconventional resources, many uncertainties
about the applicability of these methods to unconventional volatile oil reservoirs remain.
In unconventional plays, hydraulic fracturing must be employed to achieve economic production, and thus
investigations of multiphase flow behavior with hydraulic fractures has become our research focus.
Behmanesh et al. (2015) developed analytical methods for tight gas condensate reservoirs; their
methodology focused on the transient linear flow regime. Their study revealed that determination of
saturation-pressure relations in linear flow systems cannot simply use the theories developed for
conventional reservoirs. Zhang et al. (2015) transformed PDEs to ordinary differential equations (ODEs)
for linear flow in an infinite-acting gas condensate reservoir, and solved the equations using a Runge-Kutta
technique. The output of their methodology was a saturation-pressure relation and producing GOR.
Tabatabaie and Pooladi-Darvish (2017) extensively discussed behavior of multiphase linear flow in tight
oil reservoirs; they also employed the Boltzmann transformation to handle nonlinear equation (PDEs were
converted into ODEs). Wu et al. (2019) developed an approximate model by assuming that all pressure-
determined parameters are constant in each calculation step (to solve the problem of nonlinear effects in
modeling) and then applied it to production data analysis in multiphase reservoirs. However, solutions
adequate for early-time flow periods exhibit large errors once the pressure drawdown reaches the inter-
fracture closed boundaries (i.e., fracture interference) in a multi-fractured horizontal well. More
importantly, given that phase behavior of volatile oils differs from that of gas condensates, specific
investigations for volatile oil systems remained to be conducted. Luo et al. (2019) proposed a fully-
analytical methodology for production data analysis in low-permeability volatile oils; their model first
includes transient linear flow and boundary-dominated flow. But their saturation-pressure relationship
determination still followed methods developed for radial-flow systems, and thus requires further
justification. Their study also showed that more accurate pseudopressure calculations were needed.
URTeC 2501
In summary, our review of major studies of multiphase rate-transient analysis indicated that investigations
of specific methods to analyze and forecast production for volatile oil systems were very limited, especially
in unconventional reservoirs. However, many light oil resources have been developed (e.g., Bakken and
Eagle Ford); thus, published studies do not fully meet industry’s needs. Additionally, and more importantly,
full-flow-regime analysis should be highlighted and deserves more attention in future studies. A fact clearly
facing us is that increasing numbers of wells are about to enter boundary-dominated flow. Assuming endless
transient (infinite-acting) behavior in a multi-fractured horizontal well will inevitably lead to overestimates
of future production rates and ultimate recovery.
Objective. The objective of this paper is to extend our analytical model (Luo et al. 2019) for enhanced full-
flow-regime rate-transient analysis for unconventional volatile oil reservoirs. The enhanced model will be
able to forecast long-term production given only transient production data and will be applicable to wells
with substantial water production. Also, as a major motivation for this study, we intend to find a method to
determine an appropriate saturation-pressure relation for volatile oils, which should greatly improve the
accuracy of results in rate-transient analysis and production forecasting.
Methodology
Analytical Solution. For tight and shale reservoirs, linear flow in the stimulated formation around a
hydraulic fracture is dominant over the life of a fractured well, and approximate one-dimensional flow in a
closed region has been repeatedly found to reasonably represent the behavior of a multi-fractured horizontal
well. Additionally, we assume that the hydraulic fracture conductivity is infinite, and that the well is
producing at constant BHP (but we use material-balance time to deal with variable-BHP production). The
production rate from each symmetric drainage area is uniform; because of the symmetry, only a quadrant
of individual fracture-flow area is considered. The governing partial differential equation for the primary
phase can be written as follows:
∂ kro krg ∂p φ ∂ So Sg
+ Rv = + Rv
∂x µo Bo µ g Bg ∂x 0.00633k ∂t Bo (1)
Bg
with constant bottom-hole pressure and closed boundary conditions
p (=
x 0=) pwf (2)
∂p
(= ) 0
x L= (3)
∂x
Overall, for analytical modeling in multiphase flow systems, employing the Boltzmann transformation and
defining an appropriate multiphase pseudopressure are the only two ways used to linearize the governing
partial differential equation (the diffusivity equation), and linearization is a crucial issue in multiphase flow
modeling. Considering that the Boltzmann transformation has been established to be inapplicable with a
closed outer boundary condition, the only way to linearize the diffusivity equation for our study was to
introduce multiphase pseudopressures for each phase. The multiphase pseudopressure for the oil phase is
p k krg p
ptpo =∫pref µo Bo µ g Bg ∫pref λosc dp
ro
+ Rv dp = (4)
Transforming pressures into pseudopressures, the equations can be solved by the separation of variables
method (Luo et al. 2019). The solution for the primary phase is
URTeC 2501
(p tpoi − ptpo, wf )
=
141.2π L
qosc ∞ ( 2n − 1)2 π 2 ⋅ 0.00633kt (5)
khx f ∑ exp −
n =1 4φ L 2
⋅ ( ct / λ ) tp
Repeating the derivation used for the primary oil phase, we can readily obtain the solution for the gas phase.
The solution is essentially the same as Eq. 5. However, the dimensionless gas production rate should be
adapted to pseudopressures for the gas phase.
Pseudopressure Calculation. Having linearized the flow equations by introducing pseudopressures, we
need to calculate relation of pseudopressure and pressure accurately (a strong nonlinear relation as shown
in Fig. 1). From the solution we derived, we can readily infer that the calculation of pseudopressures will
have a decisive impact on oil and gas production rates. Fevang and Whitson (1996) suggested to calculate
pseudopressure for gas-condensate reservoirs by splitting the range of integration in the defining equation
(Eq. 4). This method is now widely used for conventional reservoirs (i.e., the radial flow systems).
However, from our simulation results (also confirmed by Behmanesh et al. 2015), this empirical approach
will not achieve satisfactory accuracy for linear multiphase flow in unconventional reservoirs. The reason
is that it assumes a steady-state flow around wellbore, with short duration transient flow, and thus this
assumption during lengthy transient flow periods can cause large errors in most cases for unconventional
reservoirs. Therefore, to obtain more accurate solutions for either volatile oils or gas condensates in
unconventional formations, we need to analyze multiphase flow with a more appropriate approach. Based
on the definition of pseudopressure for the oil phase, the pseudopressure difference becomes
kro
pi krg
ptpoi − ptpo,=
wf ∫pwf µo Bo v µg Bg dp
+ R (6)
The key to calculating the pseudopressure difference is to correlate relative permeability and pressure;
saturation is intermediate in connecting the variables to each other. Thus, the issue is transformed to
determining the saturation-pressure relation.
6000
5000
Pseudopressure (psi)
4000
3000
2000
1000
0
0 1000 2000 3000 4000 5000 6000
Pressure (psi)
Fig. 1 — Comparison of ideal behavior (constant density and viscosity) of pseudopressure (dashed line)
and two-phase pseudopressure (solid line). This example is for a volatile oil with a bubblepoint pressure of
5704 psia and a GOR of 1500 scf/STB (Walsh and Lake 2003). Multiphase pseudopressure (and multiphase
pseudopressure difference) is a non-linear function of pressure.
Saturation-Pressure Relation. This section discusses published methods to determine saturation-pressure
relations. We summarize commonly used methods, including a steady-state flow path, constant-
URTeC 2501
composition expansion (CCE) path, constant volume depletion (CVD) path, and a tank-type model.
Additionally, and more importantly, we conducted numerous simulations to compare with results from the
commonly-used paths and to find the most appropriate relations for accurate pseudopressure calculation.
Steady-state flow means that the overall surface volumes of oil, gas, and water remain constant and
independent of position. It also implies that the gas-oil ratio remains constant during this flow regime. The
transient linear flow regime commonly occurs during the early-time flow period, and we observe (from
simulation) that the gas-oil ratio is constant at these times. The S-P relation for steady-state flow can be
obtained by Eq. 7, which indicates strong dependence on producing GOR.
krg R p − Rs µ g Bg
= ⋅ (7)
kro 1 − Rv R p µo Bo
The constant-composition expansion (CCE) path arose originally from laboratory tests. We can also derive
an equation to predict the results. The oil saturation in a three-phase flow system is
So =
(1 − Sw )
B ( R − Rs )
(8)
1 + g si
Bo (1 − Rsi Rv )
Constant volume depletion (CVD) is also an experimental process that is used to obtain PVT data for
volatile oils. However, in CVD experiments, in contrast to the CCE process, we remove partial gases after
each pressure decrement. Therefore, the process is similar to reservoir depletion once liquid production is
limited, because the cell volume in a CVD experiment is kept rigorously constant with deceasing pressure.
For volatile oils, completely following this method to determine the saturation-pressure relation may cause
errors. The data for this CVD process can be obtained from a laboratory PVT report.
Lastly, we consider the straightforward tank-type model originally proposed by Tarner (1944). The tank-
type path is derived from a macroscopic mass balance of stock-tank oil (Walsh and Lake 2003) and is
essentially a rearranged form of the material balance equation. The oil saturation is
Np Soi S gi Rvi φi R
1 − + − (1 − S w ) v
N p ,ult B Bgi φ Bg
So = oi
(9)
1 Rv
−
Bo Bg
Water Phase Saturation. For unconventional reservoirs with three-phase flow, connate water flows and
the saturation change is not negligible. However, the behavior of water saturation in unconventional
reservoirs is not as complicated as for hydrocarbons. Trend forecasting of water saturation behavior is
tractable based on collected production data. Water saturation can be determined for volatile oil reservoirs
by solving the stock-tank water macroscopic mass balance. This yields
φ Bw (WI − W p ) + We Bw S wi
=Sw i + (10)
φ V pi Bwi
The first term in brackets shows that the water saturation changes linearly with the net water entering the
reservoir by influx or injection. Without net injection or influx water, the water saturation will remain
approximately constant.
URTeC 2501
PVT Data Adjustment. PVT data are essential to complete pseudopressure calculations; the data are
usually obtained from lab tests. However, a lab test is not rigorously consistent with the complicated
production process from reservoir to separators. Our studies confirmed that lab test for volatile oils, such
as a differential vaporization test, does not represent the fully production process. The PVT properties
generated in a purely differential vaporization test require correction; the primary cause is the effect of the
separators. Recalling previous discussions about the path of produced fluid and the corresponding
vaporization process for a volatile oil from reservoir to surface, the process is basically a combination of
differential vaporization and flash separation.
To minimize the difference between lab tests and the production process, we adjust the lab-tested PVT data,
including solution and volume parameters. The adjustment will help enhance the accuracy of model
calculations when using this methodology. Amyx et al. (1960) were among the first to propose a PVT
adjustment method (ABW method), initially for black oils. The method is also applicable to partial
adjustments for volatile oils. However, Poettmann and Thompson (1986) pointed out that the method can
lead to negative dissolved gas-oil ratios at low pressures. This problem is commonly observed because the
method overcorrects oil-phase-related parameters and yields values that are slightly less than true values.
Hence, they provided a modified relation to fix the problem that arises with the ABW method. The corrected
relation (McCain 2002) for dissolved gas-oil ratio is
RsSb
Rs = RsD (11)
RsDb
The correction for Bo, based on Bofb and Bodb and the flash and the differential stock-tank-oil densities for
pressures above bubblepoint pressure, is
Bo = BoE BoSb (12)
1.3 600
Bo (RB/STB)
Rs (scf/STB)
1.2 400
1.1 200
1.0 0
0 1000 2000 3000 4000 5000 0 1000 2000 3000 4000 5000
Pressure (psi) Presseure (days)
Fig. 2 — Adjustments of oil FVF and producing GOR for an ordinary volatile oil to separator conditions.
Fig. 2 illustrates the apparent difference in differential-vaporization test data and corrected data. Results
reveal that, with the separator effect, the disparity of both oil formation volume factor and solution GOR
URTeC 2501
grows from ambient condition to pressures near the bubblepoint. The errors in solution GOR values are up
to 8% in this case, and will be larger for near-critical volatile oils. This fact will cause errors in calculating
pseudopressures and further deviations in production forecasting.
Saturation-Pressure Relation Determination. Determination of the saturation-pressure relation is
essential for rate transient analysis and production forecasting in volatile oils. The four methods we
mentioned above are possible ways to determine the relation analytically. This section analyzes these
commonly-used methods and discusses them extensively.
We generated a synthetic dataset using a compositional simulator. The bubblepoint pressure of oil equaled
initial reservoir pressure, and water saturation varied. We calculated various S-P paths in a range of pressure
from initial reservoir pressure to bottom-hole pressure. Fig. 3 illustrates three calculated saturation-pressure
paths and two experiment-based results for a volatile oil. The results indicate a significant difference
between these paths, reemphasizing the importance of selecting a proper S-P relation. The oil saturation
along the steady-state path decreases sharply at the start and then slowly decreases from its value at initial
water saturation until it drops to approximately 41% PV at bottom-hole pressure (1000 psi). The path
depends strongly on producing GOR, and the computational process is tractable. Although saturation along
this path is overestimated for the entire production life of a well, many studies have examined the utility of
this path in certain applications during the early-time flow period, including both production-data and
pressure-transient analysis. The CCE path describes a saturation-pressure route determined from PVT
relations only. The relations shown in Fig. 3 differ significantly from those with the other two paths, but
they remain consistent with experimental results. The tank-type model describes the saturation-pressure
path of a subsurface fluid volume subjected to a continuously changing fluid leaving the volume. Along the
tank-type path, oil saturation monotonically decreases from 0.5 to approximately 0.34 at 1000 psi. The
calculations must be based on oil, gas, and water production data. In general, the predicted oil saturation is
lower than predicted along the steady-state path. The result from constant volume depletion, assuming gas
production only, shows higher oil saturation than the tank-type model and the CCE path. However, data for
this path is totally provided in PVT experiments, which means we do not need to make further calculations
to determine the S-P relation.
4000
Steady-state path Sw=50% PV
CCE path
3500 Tank-type model
CCE experiment
3000 CVD experiment
Pressure (psi)
2500
2000
1500
1000
0 10 20 30 40 50
Oil saturation (% PV)
Fig. 3 — Various S-P paths for a volatile oil. The figure compares the steady-state path, CCE path, CVD
path, tank-type model, and experimental data.
The S-P relations based on different assumptions exhibit apparently different paths with decreasing
saturations, from which we need to select the proper representation. Theoretical investigations often
overlook the fact that we not only have to consider the accuracy of the methodology with the selected S-P
relation, but also are required to consider the data necessary to calculate the S-P relation. Table 1 lists
required data for each type of path.
URTeC 2501
Production data is required in the tank-type model, but they are also the results of our production
forecasting, limiting the utility of this model. Producing GOR is required to determine steady-state path,
also limiting its utility. Therefore, the steady-state path and tank-type model cannot be used to complete
pseudopressure differences for the entire range from initial reservoir pressure to flowing bottom-hole
pressure. Additionally, the CCE and CVD path data, usually from PVT reports or further fluid property
simulation, are somewhat dependent on specific assumptions which may not match the reservoir depletion
process. Thus, we need to examine these paths further. We designed 15 simulation cases to examine the
characteristics of the saturation-pressure relation in volatile oil reservoirs estimated with the methods
discussed above. Basic information about these cases is provided in Table 2.
Table 1 — Required data for calculating all paths of saturation-pressure relation.
conditions approach the critical point on P-T diagram, the path predicted by the tank-type model gradually
deviates from the simulated data and large errors arise for the near-critical volatile oil as shown in Fig. 4(e).
Near-critical volatile oils are distinguished from ordinary oils by their sharply increasing GORs from the
start of production. In other words, a great deal of gas (vaporized oil) was released; meanwhile, oil
production is relatively limited, exhibiting an approximately constant volume depletion process in
laboratory tests. This raises another important observation, as displayed in Figs. 4(d) and 4(e), that CVD
paths are appropriate to predict saturation-pressure relation for near-critical oils.
4000 4000
Simulated Data Sw=50% PV Simulated Data Sw=50% PV
Steady-state path Steady-state path
3500 Tank-type model 3500 Tank-type model
CCE experiment CCE experiment
CVD experiment CVD experiment
Pressure (psi)
Pressure (psi)
3000 3000
2500 2500
2000 2000
1500 1500
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Oil saturation Oil saturation
(a) (b)
4000 4000
Simulated Data Sw=50% PV Simulated Data Sw=50% PV
Steady-state path Steady-state path
3500 Tank-type model 3500 Tank-type model
CCE experiment CCE experiment
CVD experiment CVD experiment
Pressure (psi)
Pressure (psi)
3000 3000
2500 2500
2000 2000
1500 1500
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Oil saturation Oil saturation
(c) (d)
4000 6000
Simulated Data Sw=50% PV
Steady-state path
Tank-type model 5000
3500
CCE experiment
Case 2 Case 3
CVD experiment 4000 Case 1 Case 4
0%
Pressure (psi)
Case 5
Pressure (psi)
3000
20% Critical point
3000
2500
40%
2000
2000
1000 60%
80%
1500 0
0.0 0.1 0.2 0.3 0.4 0.5 0 100 200 300 400 500 600 700 800 900
Oil saturation Temperature (°F)
(e) (f)
Fig. 4 — Saturation-Pressure paths for volatile oils at constant BHP conditions. This case compares steady-
state path, CCE path, CVD path, tank-type model, and simulation data. (a) Case 1, (b) case 2, (c) case 3,
(d) case 4, (e) case 5, (f) P-T diagram.
The cases 1-5 we discussed earlier show the saturation-pressure paths for wells produced with the same
bottom-hole pressure and delivered important observations in determination of S-P relation. We further
examine these observations based on same vapor percentage assumptions at wellbore with another five
URTeC 2501
cases. From Fig. 5(a), the tank-type model again performs well for ordinary volatile oils, suggesting that
the tank-type model in determination of S-P relation is not affected by well production conditions.
Furthermore, the trend of the resulting S-P path during late-time period keeps highly consistency with that
of CVD path. This observation inspired us to provide a solution for correction of S-P paths in Figs. 5 (b),
5(c) and 5(d) based on CVD path. Again, for a near-critical volatile oil (as shown in Fig. 5(e)), the S-P
relation determined by CCE matches simulated data for a near-critical volatile oil. However, under small
pressure drawdown in case 9 (compared with case 4), the CVD path is not accurate. Therefore, we further
need to discuss the effect of varying pressure drawdown on the accuracy of CVD-based S-P relation.
4000 4000
Simulated Data Sw=50% PV Simulated Data Sw=50% PV
Steady-state path Steady-state path
3500 Tank-type model 3500 Tank-type model
CCE experiment CCE experiment
CVD experiment CVD experiment
Pressure (psi)
Pressure (psi)
3000 3000
2500 2500
2000 2000
1500 1500
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Oil saturation Oil saturation
(a) (b)
4000 4000
Simulated Data Sw=50% PV Simulated Data Sw=50% PV
Steady-state path Steady-state path
3500 Tank-type model 3500 Tank-type model
CCE experiment CCE experiment
CVD experiment CVD experiment
Pressure (psi)
Pressure (psi)
3000 3000
2500 2500
2000 2000
1500 1500
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Oil saturation Oil saturation
(c) (d)
4000 6000
Simulated Data Sw=50% PV
Steady-state path
Tank-type model 5000
3500
CCE experiment
Case 7 Case 8
CVD experiment 4000 Case 6 Case 9
0%
Case 10
Pressure (psi)
Pressure (psi)
3000
20% Critical point
3000
2500
40%
2000
2000
1000 60%
80%
1500 0
0.0 0.1 0.2 0.3 0.4 0.5 0 100 200 300 400 500 600 700 800 900
Oil saturation Temperature (°F)
(e) (f)
Fig. 5 — Saturation-Pressure paths for volatile oils at constant vapor percentage conditions at bottom hole.
This case compares steady-state path, CCE path, CVD path, tank-type model, and simulation data. (a) Case
6, (b) case 7, (c) case 8, (d) case 9, (e) case 10, (f) P-T diagram.
The results in previous cases (4, 5, 10) repeatedly demonstrate that the CVD path can accurately
characterize the saturation-pressure relation for near-critical volatile oils. We next examine near-critical
URTeC 2501
volatile oils under varying flowing bottom-hole pressure. The results displayed in Fig. 6 indicate that the
CVD-determined relation is not reliable at large levels of BHPs (small pressure drawdowns); the reason is
that, at such conditions, producing GOR is not large (produced gas is not dominant) although reservoir
conditions are close to the critical point and thus very sensitive to a small pressure drawdown. Therefore,
for most operating conditions, the CVD path, for which laboratory data are available, provides an accurate
way to describe the saturation-pressure relation for a near-critical volatile oil. We should, however, be on
the alert for wells operating with high flowing BHP (small drawdowns) and consider a modification to this
procedure.
4000 4000
Simulated Data Sw=50% PV Simulated Data Sw=50% PV
Steady-state path Steady-state path
3500 Tank-type model 3500 Tank-type model
CCE experiment CCE experiment
CVD experiment CVD experiment
Pressure (psi)
Pressure (psi)
3000 3000
2500 2500
2000 2000
1500 1500
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Oil saturation Oil saturation
(a) (b)
4000 4000
Simulated Data Sw=50% PV Simulated Data Sw=50% PV
Steady-state path Steady-state path
3500 Tank-type model 3500 Tank-type model
CCE experiment CCE experiment
CVD experiment CVD experiment
Pressure (psi)
Pressure (psi)
3000 3000
2500 2500
2000 2000
1500 1500
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Oil saturation Oil saturation
(c) (d)
4000 6000
Simulated Data Sw=50% PV
Steady-state path
Tank-type model 5000
3500
CCE experiment
CVD experiment 4000 0%
Pressure (psi)
Pressure (psi)
3000
20% Critical point
3000 Case 11
2500 Case 12
40%
2000 Case 13
Case 14
2000
1000 60%
Case 15
80%
1500 0
0.0 0.1 0.2 0.3 0.4 0.5 0 100 200 300 400 500 600 700 800 900
Oil saturation Temperature (°F)
(e) (f)
Fig. 6 — Saturation-Pressure paths for volatile oils at constant reservoir temperature conditions. This case
compares the steady-state path, CCE path, CVD path, tank-type model, and simulation data. (a) Case 11,
(b) case 12, (c) case 13, (d) case 14, (e) case 15, (f) P-T diagram.
Through numerous simulations and comparisons, we conclude that CVD paths are appropriate to represent
saturation-pressure relations for near-critical volatile oils, and corrections for other volatile oils are required,
URTeC 2501
as discussed below. Based on observations of these cases, quadratic polynomial functions can be employed
to fit the CVD-based or CVD-corrected relations. The general representation is
p ( So ) =pi − a ( So − Soi )
2
(14)
We determined that this quadratic function is appropriate for matching actual saturation-pressure paths and
CVD path for volatile oils. Further, the correction of the CVD path for some specific situations depends
uniquely on the coefficient (a) in Eq. 14. Based on our preliminary conclusions from the simulated cases,
the correction of coefficient a should be conducted for ordinary volatile oil with high producing GOR and
near-critical oils with small pressure drawdown. From simulation results, when saturations determined by
tank-type path is overall large than CVD path, the correction is usually needed. Therefore, we can actually
observe the tank-type path and the CVD path to judge if the correction is required. In summary, we
recommend that saturation-pressure relations should be determined as indicated in Table 3.
Table 3 — Summary of saturation-pressure relation determination for volatile oils
Case Study
In this section, we present four cases to illustrate application of the results of our study to rate transient
analysis and production forecasting in volatile oil reservoirs. We first validate our methods with two
synthetic cases, an ordinary volatile oil and a near-critical volatile oil. The initial conditions of these
reservoirs and major properties of these volatile oils are summarized in Table 4. We then built a one-
dimensional reservoir model and output oil and gas production data generated with a commercial reservoir
simulator. The values of parameters we employed in the simulation are listed in Table 5.
Table 4 — Basic reservoir parameters used in simulator for synthetic cases.
Initial oil saturation 0.55
Initial water saturation 0.45
Reservoir length 100 ft
Reservoir width 50 ft
Net pay 100 ft
Porosity 0.07
Permeability 800 nd
Rock compressibility 10-6 psi-1
URTeC 2501
Table 5 — Initial conditions for volatile oil reservoirs and producing BHPs
Fluid Type Pi / psi Pwf / psi Ti/°F
Ordinary volatile oil 4735 2000 250
Near-critical volatile oil 4576 2000 600
Synthetic Case Study 1: Ordinary Volatile Oil. This case study briefly illustrates a workflow for rate
transient analysis of a typical volatile oil well. The initial pressure is bubblepoint pressure, and the initial
conditions shown in Table 5 indicate that this volatile oil is not a near-critical volatile oil. Therefore,
determining a saturation-pressure relation based on a CVD-modified path is suggested. The producing GOR
profile is displayed in Fig. 7.
6000 100000
60000
Rp (ft3/bbl)
3000
20%
40000
2000 40%
60% 20000
1000
80%
0
0 100 200 300 400 500 600 700 800 0 2000 4000 6000 8000 10000
Temperature (°F) Production Time (days)
(a) (b)
Fig. 7 — P-T diagram (a) and producing gas-oil ratio profile (b) for synthetic cases. The producing GOR
remains constant for almost a month (2000 scf/STB) and then increases to large values as the reservoir
depletes.
The producing GOR is constant at the start of production (one-month duration in this case) and increases
as the reservoir depletes. Additionally, the increasing trend and level of the producing GOR are noteworthy.
Thus, the correction should be based on both CCE and CVD paths (similar to case 3 in the S-P relation
determination). The S-P relation, determined by fitting laboratory CVD data with Eq. 14, is
p ( So ) =
4735 − 5.19* ( So − 55 )
2
(15)
The corrected path and its comparisons with the original CVD path and simulation result are shown in Fig.
8.
URTeC 2501
5000
CCE path
CVD path
4500 Tank-type model
Corrected path (a=-5.1)
4000
Pressure (psi)
3500
3000
2500
2000
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Oil saturation
Fig. 8 —S-P paths including corrected path based on fit of laboratory CVD data.
Using the corrected saturation-pressure relation based on the fit of laboratory CVD data in our
pseudopressure calculation, we verified the proposed model and associated methodologies by comparing
to data from a compositional simulator. Fig. 9 displays the results, showing a good agreement between our
solution and simulated data.
101 105
Simulated Data Simulated Data
Predicted Data Predicted Data
4
100 10
10-1 103
qo (STB/day)
qg (scf/day)
10-2 102
10-3 101
10-4 0 100 0
10 101 102 103 104 105 10 101 102 103 104 105
Production Time (days) Production Time (days)
(a) (b)
Fig. 9—Production rate predictions for synthetic case 1. Log-log plots show both simulated data generated
by compositional simulator and the results calculated by our model: (a) oil production rate, and (b) gas
production rate.
Synthetic Case Study 2: Near-Critical Saturated Volatile Oil. We next discuss a case with a near-critical
volatile oil. The pressure drawdown from reservoir to wellbore is relatively large. Therefore, following the
observations in our earlier discussions of the saturation-pressure relation determination, in this case we use
the original CVD path from the PVT report without any correction scheme. Fig. 10 compares the predicted
and simulated data, which verifies the accuracy of our methodology for determining the saturation-pressure
relation for near-critical volatile oils.
URTeC 2501
101 105
Simulated Data Simulated Data
Predicted Data Predicted Data
100 104
10-1 103
qo (STB/day)
qg (scf/day)
10-2 102
10-3 101
10-4 0 100 0
10 101 102 103 104 105 10 101 102 103 104 105
Production Time (days) Production Time (days)
(a) (b)
Fig. 10 — Production rate predictions for synthetic case 2. Log-log plots show both simulated data
generated by compositional simulator and the results calculated by our model: (a) oil production rate, and
(b) gas production rate.
Field Case Study 1: Production Forecasting. This case examines production data from a well in a shale
reservoir. The well was produced for only about one year. We plot oil and gas production vs. material
balance time (Blasingame et al., 1991) profiles in Fig. 11, all showing apparent linear flow. Following our
methodologies, we conducted rate transient analysis from which we obtained a reasonable estimate of
average permeability in the stimulated region of volatile-oil shale well. This result confirmed the
applicability of our model and associated methodology. In addition, in comparison with existing models,
what makes our model special is that long-term production behavior can be predicted from available
production data still in transient linear flow. Using models that include only transient linear flow will
produce large errors in production forecasting. Fig. 11 indicates that the well production will exhibit the
response of boundary-dominated flow after around 2400 days of material balance time. At that time, the
pressure drawdown will propagate to the no-flow boundary located at the point of fracture interference at
the middle of two neighboring clusters, and production will begin to decline more rapidly. This superiority
of our multiphase flow model can thus help us build a long-term perspective for shales development.
105 109
Simulated Data Simulated Data
Predicted Data Predicted Data
104 108
103 107
qo (STB/day)
qg (scf/day)
102 106
101 105
100 -1 104 -1
10 100 101 102 103 104 105 10 100 101 102 103 104 105
Material Balance Time (days) Material Balance Time (days)
(a) (b)
Fig. 11 — Production data analysis and forecasting for field case 1: (a) oil production rate, and (b) gas
production rate.
Field Case Study 2: Full-Flow-Regime Analysis. This field case shows a well that has been producing oil
and gas for over four years. The oil and gas production profiles are displayed in Fig. 12. By employing our
URTeC 2501
model, we smoothly handled with the production data containing both transient linear flow and boundary-
dominated flow. The production profiles clearly illustrate that production was beginning to decline,
indicating the well was experiencing the boundary-dominated flow regime. Moreover, due to incorporating
all flow regimes of a multi-fractured horizontal well in our analytical model, our results can also tell the
time at which fracture interference occurred and an estimate of cluster spacing. The field case further
verified our applicability of our proposed methodologies.
105 108
Simulated Data Simulated Data
Predicted Data Predicted Data
104 107
103 106
qo (STB/day)
qg (scf/day)
102 105
101 104
100 -1 103 -1
10 100 101 102 103 104 105 10 100 101 102 103 104 105
Material Balance Time (days) Material Balance Time (days)
(a) (b)
Fig. 12 — Full-flow-regime production data analysis for field case 2: (a) oil production rate, and (b) gas
production rate.
Conclusions
This paper focused on an analytical method for rate transient analysis in unconventional volatile oil
reservoirs. We proposed a pseudopressure-based solution associated with a significantly improved
saturation-pressure relation which allows us to deal with the issue of nonlinear differential equations
modeling multiphase flow. The conclusions of this study are summarized as follows.
1. With appropriate definition of pseudopressure, we can linearize the governing partial-differential
equations in our model. The resulting solution allows full-flow-regime analysis over the life of a multi-
fractured horizontal well in low-permeability volatile oil reservoirs.
2. We highlighted the importance of PVT adjustments to prepare data for rate transient analysis in volatile
oil reservoirs. Due to effects of separator conditions, using PVT properties of volatile oils obtained in lab
tests can cause errors on the order of 8% for solution GOR, with the error varying in different fluids.
3. Multiphase pseudopressure calculation in unconventional reservoirs is much more complicated than for
radial-flow systems. The results are essentially dominated by the relationship between pressure and relative
permeability for each phase. Relative permeability curves (usually determined by history matching) and an
appropriate saturation-pressure relation should be bundled to complete this process.
4. Based on evidence from a comprehensive simulation study, we proposed an accurate method to determine
saturation-pressure relations for unconventional volatile oils reservoirs. Near-critical volatile oils produced
with typical pressure drawdowns can simply use an S-P relation determined directly from a laboratory CVD
report to calculate pseudopressures, however, for other volatile oils, we should follow our suggestions for
correction of laboratory CVD data.
5. Synthetic and field case studies were designed to show the applicability of our methodology. Synthetic
cases, including individual analysis for ordinary and near-critical volatile oils, demonstrated the accuracy
of our model and verified the feasibility of our assumptions. Field cases demonstrated that our model can
URTeC 2501
predict long-term production rates and allows us to interpret production data of wells experiencing
boundary-dominated flow.
Nomenclature
Bg = gas formation volume factor, ft3/scf
Bo = oil formation volume factor, RB/STB
Bw = water formation volume factor, RB/STB
ct = total compressibility,1/psi
h = formation thickness, ft
k = matrix permeability, md
krg = gas relative permeability, fraction
kro = oil relative permeability, fraction
L = half distance of fracture spacing, ft
N = original oil in place, STB
Np = cumulative oil production, STB
p = pressure, psi
pb = bubblepoint pressure, psi
pi = initial pressure, psi
pref = reference pressure, psi
ptpo = three-phase pseudopressure for oil phase, psi/cp
pwf = bottom-hole pressure, psi
qg = gas production rate, scf/day
qo = oil production rate, STB/day
Rp = producing gas oil ratio, scf/STB
Rs = solution gas oil ratio, scf/STB
Rv = solution oil gas ratio, STB/scf
Sg = gas saturation, fraction
So = oil saturation, fraction
Sw = water saturation, fraction
t = production time, days
Vp = pore volume
Wp = cumulative water production, STB
xf = fracture half-length, ft
Greek Letters
λosc = component mobility for oil phase,1/cp
μg = gas viscosity, cp
μo = oil viscosity, cp
ϕ = porosity, fraction
Subscripts
D = differential vaporization
Db = differential vaporization test at bubblepoint pressure
URTeC 2501
g = gas
i = initial
o = oil
Sb = separator test at bubblepoint pressure
sc = standard condition
tp = three-phase
ult = ultimate
wf = bottom-hole
References
Behmanesh, H., Hamdi, H., and Clarkson, C.R., 2015. Production Data Analysis of Tight Gas Condensate
Reservoirs. Journal of Natural Gas Science and Engineering, 22, 22–34.
http://dx.doi.org/10.1016/j.jngse.2014.11.005.
Amyx, J., Bass, D. and Whiting, R.L., 1960. Petroleum reservoir engineering physical properties.
Blasingame, T., McCray, T., and Lee, W. 1991. Decline Curve Analysis for Variable Pressure
Drop/Variable Flowrate Systems. Presented at the SPE Gas Technology Symposium, Houston, Texas, 22-
24 January. SPE-21513-MS. https://doi.org/10.2118/21513-MS.
Bøe, A., Skjaeveland, S. M., and Whitson, C. H. 1989. Two-phase Pressure Test Analysis. SPE Form Eval
4 (4): 604-610. SPE-10224-PA. http://dx.doi.org/10.2118/10224-PA.
Camacho-V., R., and Raghavan, R. 1989. Performance of Wells in Solution-Gas-Drive Reservoirs. SPE
Form Eval 4(04): 611 - 620. SPE-16745-PA. https://doi.org/10.2118/16745-PA.
Energy Information Agency, 2017. Tight oil expected to make up most of U.S. oil production increase
through 2040. https://www.eia.gov/todayinenergy/detail.php?id=41033.
Fevang, Ø., and Whitson, C. 1996. Modeling Gas-Condensate Well Deliverability. SPE Res Eng 11(04):
221 - 230. SPE-30714-PA. https://doi.org/10.2118/30714-PA.
Johnson, C. and Jamiolahmady, M., 2018. Production Data Analysis in Gas Condensate Reservoirs Using
Type Curves and the Equivalent Single Phase Approach. J. Pet. Sci. Eng. 171 592-604.
http://dx.doi.org/10.1016/j.petrol.2018.07.064.
Jones, J., and Raghavan, R. 1988. Interpretation of Flowing Well Response in Gas-Condensate Wells
(includes associated papers 19014 and 19216). SPE Form Eval 3(03): 578-594. SPE-14204-PA.
https://doi.org/10.2118/14204-PA.
Luo, L., Cheng, S., and Lee, J., 2019. Analytical Model for Rate Transient Analysis in Low-Permeability
Volatile Oil Reservoirs. Presented at SPE Annual Technical Conference and Exhibition, Calgary, Canada,
30 September- 2 October, SPE-195900-MS. http://dx.doi.org/10.2118/195900-MS.
Martin, J.C. 1959. Simplified Equations of Flow in Gas Drive Reservoirs and the Theoretical Foundations
of Multiphase Pressure Buildup Analyses, Trans., AIME 116, 309-11.
McCain Jr, W.D., 2002. Analysis of black oil PVT reports revisited. Presented at SPE Annual Technical
Conference and Exhibition, San Antonio, Texas, 29 September–2 October. https://doi.org/10.2118/77386-
MS.
Perrine, R. L. 1956. Analysis of Pressure-Buildup Curves. Drilling and Production Practice.
URTeC 2501
Poettmann, F.H. and Thompson, R.S., 1986. Discussion of Engineering Applications of Phase-Behavior of
Crude-Oil and Condensate Systems. Journal of Petroleum Technology 38 (12): 1263-1263.
https://doi.org/10.2118/15835-PA.
Raghavan, R. 1976. Well Test Analysis: Wells Producing by Solution Gas Drive. SPE J 16 (4): 196-208.
SPE-5588-PA. http://dx.doi.org/10.2118/5588-PA.
Sureshjani, M. H., and Gerami, S. 2011. A New Model for Modern Production-Decline Analysis of
Gas/Condensate Reservoirs. J. Can. Pet. Technol. 50(7/8): 14 - 23. SPE-149709-PA.
https://doi.org/10.2118/149709-PA.
Tabatabaie, S.H., and Pooladi-Darvish, M., 2017. Multiphase Linear Flow in Tight Oil Reservoirs. SPE
Res. Eval. & Eng., 20(01), 184-196. SPE-180932-PA. http://dx.doi.org/10.2118/180932-PA.
Tarner, J., 1944. How Different Size Gas Caps and Pressure Maintenance Programs Affect Amount of
Recoverable Oil. Oil Weekly, 144 (2) 32-34.
Walsh, M. P. and Lake, L. W. 2003. A Generalized Approach to Primary Hydrocarbon Recovery.
Amsterdam: Elsevier.
Wu, Y., Cheng, L., Huang, S., et al. 2019. An Approximate Semianalytical Method for Two-Phase Flow
Analysis of Liquid-Rich Shale Gas and Tight Light-Oil Wells. J. Pet. Sci. Eng. 176, 562-572.
https://doi.org/10.1016/j.petrol.2019.01.085.
Xu, S. and Lee, W.J., 1999. Gas Condensate Well Test Analysis Using a Single-Phase Analogy. SPE
Western Regional Meeting. http://dx.doi.org/10.2118/55992-MS.
Zhang, M., Becker, M.D., and Ayala, L.F., 2016. A Similarity Method Approach for Early-Transient
Multiphase Flow Analysis of Liquid-Rich Unconventional Gas Reservoirs. Journal of Natural Gas Science
and Engineering, 28, 572–586. http://dx.doi.org/10.1016/j.jngse.2015.11.044.