Article

pubs.acs.org/EF

Effect of the Gas Composition and Gas/Oil Ratio on Asphaltene

Deposition
Ali A. AlHammadi,† Yi Chen,‡ Andrew Yen,§ Jianxin Wang,∥ Jefferson L. Creek,∥ Francisco M. Vargas,†
and Walter G. Chapman*,†
†
Department of Chemical and Biomolecular Engineering, Rice University, Houston, Texas 77005, United States
‡
Schlumberger, Houston Formation Evaluation Center, Sugar Land, Texas 77478, United States
§
Nalco Champion, An Ecolab Company, Sugar Land, Texas 77478, United States
∥
Chevron Energy Technology Company, Houston, Texas 77002, United States
*
S Supporting Information

ABSTRACT: Arterial deposition of asphaltene is a major flow assurance issue in pipelines and wellbores. The numerous
compounds constituting crude oils are mutually soluble at reservoir conditions, but precipitation can occur with changes in the
pressure, temperature, or composition. As the pressure and temperature change through the wellbore, asphaltenes can precipitate
and potentially deposit. Unfortunately, remediation by solvent soaks is expensive; hence, the need to forecast the potential risk of
asphaltene deposition. In this paper, a previously reported simulation tool, asphaltene deposition tool (ADEPT), is used to
predict the magnitude and location of asphaltene deposits in flow lines and wellbores. ADEPT is to be used to gauge the
frequency and location of deposits and how often intervention will be needed. The phase behavior of asphaltene is described by
the perturbed-chain statistical associating fluid theory equation of state, while the transport equations are coupled with kinetic
rates of precipitation, aggregation, and deposition. The transport model is simplified, resulting in dramatic speed up of the
simulator. This paper presents a field case as well as a simulation on the effect of different gases and gas/oil ratio on asphaltene
deposition.

■ INTRODUCTION
The deposition of asphaltene in the reservoir, wellbore, and
transportation pipelines has been a flow assurance concern.
Prediction of asphaltene precipitation is necessary yet not
sufficient for prediction of asphaltene deposition profiles. Large
capital and operating costs are associated with remediation of
these deposits, creating a need for better predictions to
minimize risks. Although thermodynamics enable us to identify
the most probable regions of deposition, the magnitude is still
of concern. The use of downhole deposition monitors is very
difficult as a result of high pressure levels and steep inclination
of some systems. In addition, these monitors might interfere
with hydrocarbons inducing deposition, causing blockage and
sometimes even the loss of the entire well.
Unlike wax and gas hydrates, the ability of asphaltene to
deposit even at high temperatures makes the problem even
harder. Moreover, paradoxes, such as deposition, are more
likely to occur in light undersaturated oils (less than 1 wt % low
asphaltene content) than in heavy oils (higher asphaltene
content), which shows that there are competing mechanisms.
In fact, Trbovich and King1 listed 11 causes of asphaltene
deposition, which are CO2 flooding, rich gas flooding, pH shift,
mixing of crude oil streams, incompatible organic chemicals,
simulation, shear, streaming potential, charged bare metal
surfaces, pressure drop, and temperature drop.
Unfortunately, there are only few publications focused on the
prediction of asphaltene deposition in wellbores or pipelines.
Most research on modeling particle deposition in pipe flows has
mainly focused on particle transport to the wall. Ramirez et al.2
modeled asphaltene deposition in a pipe based on the
hypothesis that asphaltene particle transport to the wall is by
molecular diffusion. Jamialahmadi et al.3,4 have developed a
mechanistic model, which was further applied to predict
asphaltene deposition for an Iranian oil field. Besides, Eskin et
al.5 have used particle flux mass transfer expressions for
turbulent flows to model the deposition process and have
shown a quantitative comparison between predicted and
measured deposition flux for a field case.
Nevertheless, a review of the existing literature reveals that
there is a lack of both qualitative and quantitative predictive
techniques to generate accurate and reasonable asphaltene
deposition rates and deposition profiles in comparison to both
laboratory- and field-scale data. This specific need stimulated
the development of the asphaltene deposition tool (ADEPT)
simulator.
The mechanism for ADEPT demonstrated in this paper was
proposed by Vargas et al.6 and later extended by Kurup et al.7
The transport of asphaltene in the wellbore is assumed to be a
multistep process, where phenomena, such as precipitation,
aggregation, diffusion, advection, and deposition, are the main
factors, as summarized in Figure 1. The model proposes that
small asphaltene-rich particles deposit, but larger asphaltene
aggregates do not deposit as a result of the flow field. Thus, the
Received: November 24, 2016
Revised: January 27, 2017
Published: January 30, 2017

© 2017 American Chemical Society 3610 DOI: 10.1021/acs.energyfuels.6b02313

Energy Fuels 2017, 31, 3610−3619
Energy & Fuels
Figure 1. Asphaltene phase transitions as the oil flows up the
wellbore.6 The black dots represent asphaltene particles resulting from
phase separation. As a result of flow forces, asphaltene will either meet
up with each other, forming a large enough aggregate, which is carried
by the flow or diffuses to the wall, building up a deposit.
rate of asphaltene precipitation is balanced by the deposition
rate and the rate of aggregation and advection.
However, the lack of field data makes validation of the
simulation difficult. Validating the model requires a combina-
tion of laboratory and field data. Results from the model can
then be compared to field experiences, as successfully done in a
previous case.8 Hence, a systematic data acquisition protocol
and a workflow of the simulator application to field scale are
required.
The objective of this paper is to compare ADEPT output to
field observations and to understand in a macroscopic view how
asphaltene is transported and deposited in the wellbore and, if
necessary, to revise the existing model, ADEPT.7
■
METHODOLOGY
An important tool to explain the phase behavior of asphaltene is the
solubility parameter (δ). For nonpolar fluids, the solubility parameter
can be readily calculated using the Hildebrand model, which provides
an estimation of the degree of material interactions. The cohesive
energy density is defined as the energy required to remove a molecule
from its neighboring molecules to an infinite distance, i.e., from the
real state to an ideal state.9 Therefore, the extent of miscibility of the
solvent and solute is directly linked to closeness of the solubility
parameters.
At reservoir conditions, asphaltenes are generally stable. However,
as pressure and temperature decrease in the wellbore and pipeline or
the crude oil composition changes as a result of gas injection, the crude
oil becomes less compatible with asphaltene. This results in separation
of the heavy fraction in the form of a high asphaltene content liquid
phase as a result of the decrease in the solubility parameter of the oil.
This liquid phase, referred to as precipitate or primary particles, is
represented as small black circles in Figure 1. These particles will
encounter different forces that will direct the deposition process. As a
result of eddy diffusion in turbulent flows, these particles will either
meet up with other particles, forming aggregates, also known as
secondary particles, or meet up with the wall, in which they will adsorb
and, consequently, build up a deposit. However, as the temperature
and pressure continue to decrease, the bubble point is reached; the
light components leave the mixture, raising the solubility parameter of
the oil. Once the mixture has high enough solubility parameter,
asphaltenes become thermodynamically stable again. The pressure
when this occurs is known as the lower onset pressure. Consequently,
no more deposition is expected below this pressure; the precipitates
3611
Article
will dissolve back to the mixture, and the aggregates will start to
redissolve slowly. However, because some of these aggregates are large,
they might be very slow to completely dissolve. Such particles are
responsible for coke formation and fouling in refineries.10 The ability
to predict the phase behavior and deposition of asphaltene is therefore
of great significance.
In addition, advection plays a major role because it will carry both
the primary and secondary particles through the system to a
potentially stable region. In our model, we assume that, once the
aggregates are large enough, diffusion will be slow and advection will
move the aggregates through the system. Thus, the aggregates will not
contribute to deposition. Moreover, the shear force from advection can
lead to deformation or removal of the deposit, especially if the deposit
is assumed to be liquid-like.11
The following schematic (Figure 2) depicts the ADEPT simulator
structure. It consists of two modules: thermodynamic and deposition
Figure 2. Schematic of the ADEPT simulator structure. The simulator
consists of two modules: thermodynamic and deposition modules.
The thermodynamic module provides the location and amount
available to deposit, whereas the deposition module forecast the
magnitude of the deposit. Adapted from Vargas et al.6
modules. The thermodynamic module predicts when and how much
asphaltene will precipitate from crude oil. Next, the deposition module
calculates the rate of deposition, deposition profile, and pressure drop
as a result of the deposit.
The equilibrium concentration of asphaltene soluble in the crude oil
is the main output of the thermodynamic module and is a crucial input
to the deposition module. The phase equilibrium is calculated on the
basis of the wellbore and flow line PT trace and the fraction of
asphaltene precipitated. Consequently, accurate modeling of the phase
behavior of asphaltene is key to obtaining accurate representation of
the driving force of deposition. In the thermodynamic module, the
perturbed-chain statistical associating fluid theory equation of state
(PC-SAFT EOS)12 is used because it successfully describes asphaltene
stability.13 In this work, the PC-SAFT EOS is accessible with assistance
of commercial thermodynamic software, such as VLXE, Multiflash, and
PVTsim. An automatic PC-SAFT EOS characterization spreadsheet
was developed, and both PC-SAFT and cubic plus association (CPA)
EOS capabilities were tested.14 More details on the PC-SAFT
characterization can be found in the study by AlHammadi et al.14
Table 1 provides a summary of the oil properties for crude oil A,
Table 1. Oil Properties of Crude Oil Aa
saturates (wt %) 63.05
aromatics (wt %) 16.99
resins (wt %) 16.22
asphaltenes (wt %) 3.74
GOR (scf/stb) 669.00
API gravity (deg) 32.3
a
More data of the oil are available in the Supporting Information.
DOI: 10.1021/acs.energyfuels.6b02313
Energy Fuels 2017, 31, 3610−3619
Energy & Fuels Article

including saturates, aromatics, resins, and asphaltenes (SARA), gas/oil value of C0 is set to zero for simplicity. A non-zero initial condition
ratio (GOR), and American Petroleum Institute (API) gravity. SARA indicates the presence of a previous deposit in the tubings.
analysis is a technique that separates crude oil into four different The rate of precipitation is dependent upon the degree of
components based on polarity. In this study, a modified IP-143 supersaturation, which is the difference between the actual
method is used and the asphaltene fraction is defined as hexane concentration of asphaltene in the solution and the thermodynamic
insoluble. The rest of the fractions were determined by gas equilibrium concentration of asphaltenes in solution. The larger the
chromatography on the deasphaltic oil. The mass percent reported degree of supersaturation, the higher the value of the precipitation
were calculated on the basis of whole oil. kinetic rate. As more asphaltene is consumed through aggregation and
In the deposition module, a transient material balance equation was deposition, the actual concentration of asphaltene will decrease until
employed to describe the transport of the primary particles over a the degree of supersaturation reaches zero. After that point, the
control volume (Figure 1) of the wellbore or pipeline. The main mixture is stable again and is able to carry thermodynamically more
components of the model are composed of transport by advection and asphaltene particles. As a result, the system is undersaturated and the
dispersion and the three kinetic processes of precipitation, aggregation, precipitates will start to redissolve.
and deposition. The mathematical model developed by Kurup et al.7,8 Preventing deposition has been a subject of extensive research given
is the proposed competition between aggregation and deposition; it
might be better to enhance aggregation and, thus, limit the amount of
∂C ∂C 1 ∂ 2C asphaltene available for deposition. However, such an approach might
=− + + rp − Daagg C 2 − DadC
∂t ∂Z Pe ∂Z2 (1) raise issues beyond the wellbore; large aggregates can cause coke
where C is the dimensionless concentration of precipitated asphaltene formation and fouling in the refineries.10 In addition, it is reasonable to
assume that the smaller aggregates are easier to redissolve, as shown by
particles, t is dimensionless time, Z is the dimensionless axial
coordinate, Pe is the Peclet number, Da is the Damköhler number Boek et al.15 Aske et al. have demonstrated that only partial
(agg and d stand for aggregation and deposition, respectively), and rp redissolution of the aggregates is observed and the rest takes a long
time to redissolve.16 They also showed that redissolution is typically
is the rate of precipitation. Notice that this equation is dimensionless,
and the following dimensionless parameters can be defined: very slow and depends upon the physical state of the system. Schabron
and Rovani have found that aggregation can be completely reversible
UZL using different solvents of high polarity.17 In addition, Pan and
Pe = Firoozabadi were able to illustrate that insoluble asphaltene particles
Daxial (2a)
can be readily redissolved using ultrasonic waves, which breaks the
kaggL aggregates into smaller particles and, thus, makes it easy for them to
Daagg = redissolve.18 Therefore, the rate of precipitation and dissolution can be
UZ (2b) written as a conditional function as follows:7
kdL rp = Da p(Cf − Ceq) if Cf > Ceq
Dad = (5a)
UZ (2c)
z rp = rdiss = − kdissDa pC if Cf < Ceq (5b)
Z=
L (2d)
where Dap is the Damköhler number of precipitation, kdiss is the
C′ dissolution factor, and Cf and Ceq are the dimensionless actual and
C= maximum predicted equilibrium concentrations of asphaltene in the oil
C0 (2e) phase, respectively. For simplicity, the dissolution factor is assumed to
t ′UZ be 1. The precipitation kinetic constant is fit to laboratory data. The
t= experimental procedure and the fitting are described in later sections.
L (2f)
An important observation is that both precipitation and deposition
where UZ is the average velocity in the axial direction, L is the total rates are modeled to be pseudo-first-order for simplicity. The orders
axial length of the pipe, Daxial is the axial dispersion coefficient, kagg and can be readily modified as more data become available from the field.
kd are the aggregation and deposition rate constants, respectively, z is On the other hand, aggregation is assumed to be second-order
the axial length coordinate, C′ is the precipitated asphaltene following the Smoluchowski theory, which states that, when two
concentration, and C0 is the initial concentration of asphaltene in particles are within a distance R, they will stick to each other and the
the oil. The axial dispersion coefficient can be defined as new formed entity will continue to move as one bigger particle.19
The model assumes that the asphaltene-rich phase must precipitate
UZ 2R2 before forming a deposit. Further, it assumes that a competition exists
Daxial = Dm +
48Dm (3) between aggregation and deposition. Because large aggregates are
drawn to the center of the flow field, they are unavailable for
where Dm is the diffusion coefficient of the particle and R is the radius deposition. We will assume steady state. The asphaltene deposition
of the pipeline. Equation 1 is subjected to the following boundary and process is generally a very slow process. Although the deposit is
initial conditions: building up in the pipe, the effect on the flow rate is slowly changing
C(t = 0, Z) = C0 and minimal at early stages of deposition.
(4a)
As demonstrated by Figure 3 for crude oil B, the difference between
C(t , Z = 0) = C0 steady and non-steady states is not significant over a simulated period
(4b)
of a couple of weeks. Moreover, as a result of the number of iterations,
∂C the non-steady-state calculation becomes computationally expensive,
|(Z = 1) = 0 especially as the deposit buildup increases. Simplifying the equation
∂Z (4c)
can speed up the simulator dramatically, enabling quick sensitivity
Equation 4a is an initial condition that specifies that the concentration studies.7 The kinetic constants are from Kurup et al.8 and are
of asphaltene nanoaggregates initially is C0 throughout the wellbore. summarized in Table 2.
The second equation (eq 4b) is a boundary condition specifying that, For crude systems in this and previous studies, the Peclet number is
at any time, the concentration of nanoaggregates in the entrance of the found to be large. The large Peclet number causes the computation to
wellbore is C0. The last equation (eq 4c) is a boundary condition be slow and less stable. We have found that, in several cases, the term
specifying that, in the vicinity of the exit of the wellbore, there is no including the reciprocal of the Peclet number is very small and may be
change in the concentration of nanoaggregates. In this example, the neglected. In the following, we neglect the axial dispersion term. It is

3612 DOI: 10.1021/acs.energyfuels.6b02313

Energy Fuels 2017, 31, 3610−3619
Energy & Fuels
Figure 3. Asphaltene deposition in steady and non-steady states over a
4 week period. The difference is not significant. Consequently, for a
quick design estimate, the steady-state model is recommended. The
kinetic parameters used in this case are from Kurup et al.8 and are
summarized in Table 2.
Table 2. Kinetic Parameters Obtained from the Batch and
Capillary Laboratory Experiments for Crude Oils A and Ba
crude oil kp (s−1) kag (s−1) kd (s−1)
A 8.05 × 10 −4
7.61 × 10 −4
2.17 × 10−5
B 1.32 × 10−3 7.29 × 10−5 2.47 × 10−6
a
The kinetic constants for crude oil B are from Kurup et al.8 The
laminar boundary layer was used.
worth noting that the Damköhler numbers for the precipitation,
aggregation, and deposition are generally very small, and thus, the
system is reaction-limited. Such small values indicate that, although
asphaltene is available to deposit, the small rate of deposition limits the
amount of asphaltene deposited. This further simplifies the equation
and allows for quick sensitivity analysis calculations without sacrificing
accuracy. It is worth noting that, for the capillary experiments, it might
be better to include the full form as a result of the lower range of the
Peclet number.
With negligence of the axial dispersion term, the deposit buildup is
dependent upon the precipitation rate and aggregation and deposition
kinetic rates. The concentration of precipitated asphaltene available for
deposition is calculated from the following equations:
∂C
= Da p(Cf − Ceq) − Daagg C 2 − DadC if Cf > Ceq
∂Z
∂C
= − kdissDa pC − Daagg C 2 − DadC if Cf < Ceq
∂Z
Unfortunately, despite the simplification, no analytical expression is
obtained for the first equation, while the second equation can be
solved analytically. At steady state, as a result of the continuous oil
flow, the temperature, pressure, and oil composition can be assumed
to be a function of the position and not time, as shown by Figure 1.
As a result of the steady-state case, one can conclude that the actual
concentration of asphaltene entering the unstable section is relatively
constant and is depleted by aggregation and deposition. As previously
stated, this depletion is controlled by the equilibrium concentration of
asphaltenes (Ceq). The concentration of asphaltene dissolved in the oil
Cf depends upon the rate of precipitation with the driving force (Cf −
Ceq).
∂Cf
= − Da p(Cf − Ceq)
∂Z
Despite the simplicity of the deposition model, it has shown potential
in modeling the asphaltene deposition profile. However, the
deposition profile depends upon the kinetic constants of deposition,
Article
precipitation, and aggregation of asphaltene. As mentioned earlier, it is
difficult to obtain the kinetic constants from limited field data directly;
thus, laboratory experiments become the only feasible method to
acquire them.
The precipitation and aggregation kinetic constants can be extracted
from laboratory asphaltene batch experiments. The experiment is
performed by preparing a series of mixtures each with the same oil/n-
alkane precipitant ratio. The mixtures are left in an oven at a given
temperature to age, and at different times, one of the mixtures is
removed and filtered to separate asphaltene precipitates. The
precipitates obtained at different aging time are then weighed, yielding
a precipitated mass as a function of time. The equations describing the
precipitation and aggregation rates of asphaltene are
dCf
= − k p(Cf − Ceq)
dt ′ (8a)
dCagg
= kaggC 2
dt ′ (8b)
dC
= k p(Cf − Ceq) − kaggC 2
dt ′ (8c)
On the basis of these models, the kinetic constants for precipitation
and aggregation can be simply extracted by fitting the constant to
match the precipitated amounts, as shown in Figure 4. The kinetic
parameters are summarized in Table 2 for crude oil A. The properties
of oil A are summarized in Table 1.
Figure 4. Precipitated mass of asphaltene through a 0.2 μm filter as a
function of the aging time for crude oil A. These data are used to
determine the kinetic constants of precipiation and aggregation. A total
of 1 mL of oil is mixed with 0.92 mL of heptane. The kinetic constants
of precipitation and aggregation are summarized in Table 1. See Table
(6a) 1 for crude properties. The y axis is dimensionless because it shows the
ratio between the amount precipitated to the initial amount of
asphaltene present in the oil.
(6b)
The data in Figure 4 were obtained for crude oil A at the oil/
heptane ratio of 52:48 in a 70 °C oven (1 mL of oil with 0.92 mL of
heptane). At different times, the samples were removed from the oven
and filtered using a 0.2 μm filter. We assume that these filtered
asphaltenes are aggregates.
Once the precipitation and aggregation rates are obtained, the
model can be used to match the capillary deposition experiments. A
key concept of the capillary scale experiment is that not only must the
pressure drop with time be followed but the profile must also be
determined as well as the mass flux of deposition. The experimental
setup is a pressure- and temperature-controlled capillary viscometer.
The fluid composition is controlled and adjusted by two different
pumps, one containing n-alkane and the second containing the oil, and
(7)
the complete mixing by an ultrasonic bath occurs prior to entering the
capillary.20 Under normal conditions, where no deposition occurs, the
pressure drop of a fluid flowing through a capillary tube can be simply
calculated using the Hagen−Poiseuille equation as follows:
3613 DOI: 10.1021/acs.energyfuels.6b02313
Energy Fuels 2017, 31, 3610−3619
Energy & Fuels
Figure 5. Schematic representation of the capillary deposition experiment.20
8μQ L
ΔP0 =
π r0 4 (9)
where μ is the fluid viscosity, Q is the flow rate, L is the length of the
tube, and r0 is the capillary tube radius. In the previous equation, the
radius of the capillary is fixed. However, this is surely not the case
when asphaltene starts to deposit. If the deposit is gradual and
uniform, the equation can be modified to account for a decreasing
radius with time, as shown in the following equation:
8μQ
L
ΔP(t ) =
π [r0 − Δr(t )]4 (10)
The relative change of the pressure drop will simply be
ΔP(t ) − ΔP0 [r − Δr(t )]−4 − r0−4
= 0
ΔP0 r0−4 (11)
If the thickness of the deposited layer is small compared to the
capillary radius, then the previous equation can be approximated.21
ΔP(t ) − ΔP0 Δr(t )
≈4
ΔP0 r0 (12)
In addition, the last equation is valid only if the flow rate and fluid
viscosity values are constants. A schematic representation of the setup
is shown in Figure 5.
The pressure transducer was used to measure the pressure drop
across the capillary tube, which was simultaneously recorded using a
computer. To ensure isothermal conditions, the capillary tube is
immersed in either a water bath or an oven. In addition, the total flow
was maintained within the laminar regime because it is believed that
there exists a laminar boundary layer in the pipelines that dictates the
deposition process.
Prior to each test, toluene is pumped through the capillary tube to
measure its actual radius and ensure that no deposition is present. At
the end of each run, nitrogen was pumped through the capillary to
gradually displace any remaining liquid in the capillary tube. An
electronic balance can be then used to record the effluent weight with
time, which is later used to calculate the deposition thickness. Another
method is to use some viscous immiscible fluids to be injected and to
measure the pressure drop using the transducer. This is a more
effective method because it can determine more easily the local
effective capillary radius, especially in case of non-uniform deposits.7
The asphaltene content in crude oils is generally low but still may
pose problems. For instance, Hassi-Messaoud oil contains as low as
0.1%, yet asphaltene deposits in the wellbore.22 In addition, the
asphaltene deposition rate is considerably low. Consequently, in the
3614
Article
capillary viscometer, a large amount of sample is used (1 L at least) to
obtain significant amounts of deposits, and therefore, it works as a
single-pass system. The ability of this experiment to mimic the laminar
boundary layer near the wall, its high surface area, and the relatively
low expense compared to other methods makes this experiment an
attractive choice to study deposition. The capillary experiment yields
the deposition thickness as a function of the axial length. This is later
converted to deposition flux. For this case unfortunately, only the
pressure drop of the capillary was reported by the operating company,
which was converted to a deposition flux. The capillary deposition rate
constant kd is obtained by matching the peak of the deposition flux.
The field value of the deposition kinetic constant can be obtained
by scaling the laboratory parameter to field boundary layer thickness.
Three different boundary layers are present in the wellbore, which are
the momentum, mass-transfer, and laminar boundary layers. The
momentum boundary layer can be calculated using Prandtl boundary
layer theory.23
δmom = 62.7DtRe−7/8 (13)
23
The mass-transfer boundary layer is calculated as
Dt
δmass =
Sh (14)
5
The laminar boundary layer can be calculated as follows:
ρf
δ laminar = 5
μf u* (15)
where Dt is the diameter of the pipeline, Re is the Reynolds number, ρf
and μf are the density and viscosity of the fluid, respectively, u* is the
friction velocity, and Sh is the Sherwood number. A summary of the
kinetic constants is provided in Table 2. The deposition kinetic is
scaled on the basis of the boundary layers
2δ ⌀
kd = kd(cap)
R ⌀+1 (16)
where
⌀=
Dm 1
δ 2 kd(cap) (17)
A quick look on the parameters from these two cases illustrates that
the kinetic parameters are within a similar range. The kinetic
coefficient of precipitation for both cases is quite close. Moreover,
the ratio for the aggregation and deposition kinetic constants between
crude oils A and B seems consistent.
DOI: 10.1021/acs.energyfuels.6b02313
Energy Fuels 2017, 31, 3610−3619
Energy & Fuels
Because the three kinetic parameters are crucial for a quantitative
prediction, a sensitivity analysis was carried out to understand their
effects on asphaltene deposition and pressure drop. It was found that
the magnitude of their effect varies from case to case but the overall
trend is similar. The kinetic rate of precipitation can typically
determine the shape of the deposition profile. A higher precipitation
constant means a steeper profile with a sharp peak. However, its effect
on the pressure drop is not as pronounced. This is because, after a
certain value of the precipitation constant, the system seems to show
instantaneous precipitation. A more reasonable value for the
precipitation rate is to be fast compared to deposition and aggregation
but not instantaneous. This is more representative of the polydisperse
nature of asphaltene.
The aggregation kinetic rate is another important parameter. A
higher aggregation rate means less asphaltene available for deposition.
Therefore, a higher value of the aggregation rate is equivalent to a
lower asphaltene deposition flux and, hence, deposition. However, the
dependence is almost linear, with a slight negative slope. The change
in the pressure drop is also very slight.
In comparison, the kinetic rate of deposition seems to have the
most effect. The effect on the asphaltene deposition and pressure drop
seems to be exponential. An increase in the kinetic rate of deposition
means that the system is able to deposit asphaltene much faster.
Because the system seems to be reaction-limited, an increase of the
deposition rate is critical and can have pronounced consequences on
the deposition thickness and pressure drop.
■ RESULTS
Accurate modeling of the phase behavior is essential to
represent the driving force of precipitation. The PC-SAFT EOS
is used to model the system and calculate the equilibrium
concentration of asphaltene soluble in the crude oil. The
difference between the actual asphaltene concentration present
in the system and the equilibrium concentration drives the
deposition behavior. Figure 6 represents the phase behavior of
Figure 6. Asphaltene phase behavior of the system. The PT trace
demonstrates that, at the start, the system is stable but, at one point
along the wellbore, the onset condition is passed and a liquid−liquid
phase separation occurs. This continues until the bubble curve is
reached.
the crude oil at a GOR of 669 scf/stb. Data points for the
bubble point and the asphaltene onset pressure are matched by
PC-SAFT (curves). Also shown is the pressure−temperature
trace from reservoir conditions up the wellbore. Asphaltenes are
unstable below the onset pressure. As the pressure continues to
decrease, the bubble curve is reached and gas starts to leave the
mixture. The largest driving force for asphaltene deposition is at
the bubble point pressure. Once the pressure is below the
bubble curve, the solubility of the oil raises again, leading to
asphaltene becoming stable again at the lower onset.
3615
Article
The input to the deposition simulator includes the pressure
and temperature profile, all kinetic constants, the asphaltene
equilibrium concentration from the thermodynamic module,
and operation conditions, such as pipeline length and diameter,
flow rate, and initial dimensionless asphaltene concentration in
the oil phase. The deposition simulator uses this information to
predict the profile of asphaltene deposition along the wellbore
or pipeline. Figure 7 illustrates the deposition thickness along
the wellbore length. No deposition occurs before the onset
condition is reached.
Figure 7. Deposition thickness along the wellbore length. No
deposition occurs when the system is stable, i.e., above the onset
condition.
The traditional way to determine the deposition profile is to
use calipers. However, sometimes this option is not possible,
especially in subsea pipelines; alternatively, the pressure drop is
measured continuously. The occurrence and magnitude of the
additional frictional pressure drop indicate the presence and
thickness of the asphaltene deposit. In this field case, there are
no direct deposition profile measurements to compare against
but a frictional pressure drop was measured in the first 14 days
after pipeline wash. To compare to this value, the frictional
pressure drop was calculated through the Darcy−Weisbach
formula24 and shown in Table 3.
Table 3. Frictional Pressure Drop Prediction of Crude Oil A
Using the Darcy−Weisbach Formulaa
GOR (scf/stb) frictional pressure drop (psi)
669 122.8
field data ≈ 140 psi
a
The roughness value used is 0.18, which was calculated from the
Colebrook−White equation.26 The laminar boundary layer was used
for this calculation.
The operator reported that the frictional pressure drop as a
result of deposit is approximately 140 psi in the 2 weeks after a
wellbore wash. It is found that the predicted results are very
close to the field data measured. This demonstrates that the
ADEPT simulator is able to predict the pressure drop as a result
of asphaltene deposition quantitatively based on laboratory
experimental results.
Effects of the GOR on Asphaltene Deposition in the
Wellbore. Because the GOR can initially decrease with oil
production or increase as a result of a change in the production
horizon, the asphaltene stability and deposition in the wellbore
can be affected accordingly. Thus, one wellbore (crude oil A)
from the Gulf of Mexico having an asphaltene deposition
DOI: 10.1021/acs.energyfuels.6b02313
Energy Fuels 2017, 31, 3610−3619
Energy & Fuels
problem and a decreasing GOR was studied through the
ADEPT simulator to understand the effects of the GOR on the
asphaltene phase behavior and deposition.
In this case, the original GOR (zero flash) of the reservoir
fluid is 669 scf/stb. The operator reported that GOR decreased
about 60 scf/stb over 4 months. GOR would increase to some
extent as a result of change in the production horizon. Hence,
we consider two other GORs, 549 and 1000 scf/stb, set as the
desired conditions.
Figure 8 depicts the asphaltene phase behavior predicted at
different GORs. The straight solid line corresponds to the PT
Figure 8. Asphaltene phase behavior at different GORs (■, onset
pressure tested; ▲, bubble pressure tested; black line, PT trace; blue
line, asphaltene upper onset pressure; red line, bubble pressure; green
line, asphaltene lower onset pressure; solid line, predictions at GOR =
669 scf/stb; dashed line, predictions at GOR = 1000 scf/stb; and
dotted line, predictions at GOR = 549 scf/stb).
profile in the wellbore. The square and triangle points are
experimental results for asphaltene precipitation and bubble
points, respectively. The blue, red, and green curves are
prediction results. It is found that both the asphaltene onset
pressure and bubble pressure increase significantly with an
increasing GOR but the effect on the lower onset pressure
curve is minimal. This is because gas components are
precipitants for asphaltene; thus, the oil with a higher GOR
Figure 9. Solubility parameter of the crude oil as a function of the pressure at a fixed temperature (T = 190 °F). A higher GOR reaches the
asphaltene instability solubility at a higher pressure, causing asphaltene to be unstable over a wider range of pressures. The GOR unit is scf/stb.
3616
Article
becomes unstable at a higher pressure. It means even less
depressurization (high pressure) can result in asphaltene
precipitation for a higher GOR system. With regard to the
bubble point shift, it can be explained by the gas dissolution
capability. A higher GOR system with more gases needs more
pressure to keep gas components dissolved in the oil phase. For
the unchanged lower onset curves, the reason is that the oil
composition with different GORs are almost the same when the
gas evolves below the bubble point. Those light gases escape
from the oil phase, making it gradually a good solvent for
asphaltenes. Consequently, the precipitated asphaltene can
redissolve back to the oil. Note that, only in the system with
GOR of 1000 scf/stb, the oil is unstable to asphaltene initially.
Such a system is likely to have asphaltene deposition in the
reservoir.
As shown by Figure 9, at higher GOR, the solubility
parameter of the oil is lower and reaches the asphaltene
instability solubility at a much higher pressure, causing the
asphaltene to phase separate (upper onset). This continues
until the bubble point is reached, at which the gases escape,
causing the solubility to rise. The lower onset is reached when
the solubility parameter is above the asphaltene instability
solubility. Notice that the upper onset for different GORs is at
various pressures compared to the lower onset, which is
relatively at a much narrower range, as further explained by
Figure 9.
Figure 10 shows the deposit thickness predicted with respect
to different GORs in 2 weeks (14 days). The two dashed curves
(GOR = 669 and 549 scf/stb) show a similar trend with the
solid curve (GOR = 1000 scf/stb) but are still in the initial part
of stage I. Even though the decrease in the deposition for both
GORs is not shown, it is expected that it will appear beyond the
wellbore at some downstream location under a much lower
pressure.
In this specific case, the magnitudes of deposit thickness in
the three systems were found to be similar. In other words, the
deposit thickness is not sensitive to the GOR. The appearance
of the deposit shifts to the lower location (higher pressure)
with an increasing GOR. This is because oil having a higher
GOR can become unstable even at a high pressure; in turn,
asphaltene starts to precipitate and deposit at that location.
According to the discussion above, it can be concluded that
the GOR can significantly affect the asphaltene phase behavior
DOI: 10.1021/acs.energyfuels.6b02313
Energy Fuels 2017, 31, 3610−3619
Energy & Fuels
Figure 10. Asphaltene deposit thickness at different GORs. A higher
GOR leads to a higher deposition thickness.
and change the deposit location strongly. Such effects should be
taken into account before the GOR is increased.
Additionally, the deposition flux was predicted, and the result
is shown in Figure 11. The curve indicates that the deposition
rate grows fast initially and then it slows until it reaches the
maximum (stage I). After that, it starts decreasing slightly until
a critical point (stage II) and then drops sharply (stage III).
This curve shows the typical trend of the deposition rate along
the wellbore or pipeline as long as the pressure of the well
bottom or pipeline inlet is in the asphaltene instability region.
This trend can be explained through the asphaltene
precipitation driving force (Figure 12) and kinetic effects.
The deposition rate is a function of the concentration of
asphaltene primary particles (generated from precipitation),
which is strongly dependent upon precipitation (primary
particle formation) and aggregation and deposition (primary
particle consumption). The difference between the amount
precipitated and consumed is what drives an increase in the
deposition amount.
In stage I, asphaltenes precipitate right after the oil enters the
pipeline according to phase behavior prediction results.
Moreover, the largest precipitation driving force results in the
fastest increase rate of the asphaltene precipitated amount and
deposition flux. Later in the stage I, with the consumption of
the asphaltene concentration in the oil phase, the precipitation
Figure 11. Asphaltene deposition flux prediction at GOR = 1000 scf/stb (simulation). The deposition flux goes through different stages.
3617
Article
Figure 12. Precipitation driving force in the system with GOR = 1000
scf/stb (simulation). A large driving force initially leads to a large
deposition flux. The driving force quickly weakened. CF and CEQ are
Cf and Ceq as defined in eqs 5a and 5b, respectively. The Ceq value is
0.367.
driving force gradually decreases to some extent and leads to
the deposition flux increasing slower, even though the particle
amount accumulates. The precipitation driving force continues
to decrease to an even lower level right below the bubble point.
Thus, the slight decrease of the deposition rate in stage II can
be attributed to the limited accumulation of asphaltene particles
as a result of the co-effects of all processes. The starting point of
stage III is the beginning of the redissolution based on our
model. That means no precipitation occurs any more.
Meanwhile, the particle redissolving causes the primary particle
concentration to decrease immediately. Therefore, the
deposition rate decreases very fast in stage III.
Effects of Gas Injection on Asphaltene Deposition in
the Wellbore. Another interesting issue is the effect of gas
injections on asphaltene deposition. A touch on this issue is
given briefly in the previous section, where adding gas to the
system results in a higher GOR. Consequently, asphaltene
becomes less stable, leading to a higher deposition range. On
this section, however, the effect of individual gas components is
examined. Surely, adding gas to the oil will result in a higher
GOR. Nonetheless, the increase in the GOR is not comparably
DOI: 10.1021/acs.energyfuels.6b02313
Energy Fuels 2017, 31, 3610−3619
Energy & Fuels
large, and the most prevalent cause of the deposition is the
interaction of these components with the oil and the
consequent effect on asphaltene phase behavior. Figures 13
and 14 illustrate the effect of methane and nitrogen on
asphaltene phase behavior and deposition, respectively.
Figure 13. Asphaltene phase behavior as a result of 20% methane and
nitrogen injection at GOR = 669 scf/stb. The lower onset pressure is
very close to the bubble curve. Thus, the bubble curve serves as an
indication of the proximity of dissolution (■, onset pressure tested; ▲,
bubble pressure tested; blue line, asphaltene upper onset pressure; red
line, bubble pressure; solid line, predictions at no gas injection; dashed
line, predictions at 20 mol % methane injection; and dotted line,
prediction at 20 mol % nitrogen injection).
Figure 14. Asphaltene deposition thickness as a result of different gas
injections. Nitrogen injection leads to a large deposition thickness over
narrower lengths.
The deposition thickness of the various injections reflects the
phase behavior and position of the PT trace. A larger instability
region leads to a larger deposition area, which is the case of
methane. Nitrogen, on the other hand, gives a narrower but
larger deposition thickness. This is because the oil with
nitrogen injection has a lower equilibrium concentration,
resulting in a fast depletion of asphaltene-precipitated particles
over a smaller area.25 It is worth mentioning that, in all of these
simulations, it was assumed that asphaltenes were stable prior
to the start of the simulation.
■ CONCLUSION
Our prediction technique, the ADEPT simulator, can help the
oil industry to not only identify the conditions that can cause
asphaltene problems but also improve production planning and
3618
Article
wellbore or pipeline design. Furthermore, with laboratory data,
ADEPT can help distinguish the crude oil that will have
deposition issues from those only having a precipitation
phenomenon during production operations.
ADEPT would be applied to estimate the frequency of
remediation for a given well or development during the
economic evaluation of whether or not to go forward with a
prospective project. If it requires wellbore intervention too
frequently, the project could become uneconomical.
In this paper, a summary of the procedure of application of
the ADEPT simulator to field cases is presented through a
wellbore case from the Gulf of Mexico. On the basis of the
current research, with full reliance on the laboratory
experimental measurements, the current ADEPT simulator is
able to provide a good quantitative agreement with field
observations.
This application of the ADEPT simulator should, therefore,
be followed by benchmarking against field examples with
known fluid properties and deposition problems. The simplified
mathematical model gives the same accuracy with a first-order
ordinary differential equation (ODE) as opposed to the second-
order partial differential equation (PDE) and is much faster
computationally. Moreover, an analytical expression for the
actual concentration and dissolution of asphaltene can be
obtained.
In addition, the GOR shows significant effects on the
asphaltene phase behavior and deposition location in the
wellbore generally. Higher GOR and gas injections (methane
and nitrogen) lead to a higher asphaltene deposition amount
and an earlier deposition as a result of a shift in onset
conditions.
■
*
ASSOCIATED CONTENT
S Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.energy-
fuels.6b02313.
Compositional analysis of the crude oil used in this study
(PDF)
■ AUTHOR INFORMATION
Corresponding Author
*E-mail: wgchap@rice.edu.
ORCID
Ali A. AlHammadi: 0000-0002-2747-2492
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
Ali A. AlHammadi gratefully acknowledges the Abu Dhabi
National Oil Company (ADNOC) support through the Ph.D.
scholarship. Yi Chen and Walter G. Chapman acknowledge
financial support from the DeepStar consortium. The authors
are thankful to the DeepStar consortium for granting the
approval to present this paper. The authors thank Andrea
Gutierrez, Sai Panuganti, Le Wang, and Mohammed I. L. Abu
Taqiya for their helpful discussions.
■ REFERENCES
(1) Trbovich, M.; King, G. Asphaltene deposit removal: Long-lasting
treatment with a co-solvent. Proceedings of the SPE International
DOI: 10.1021/acs.energyfuels.6b02313
Energy Fuels 2017, 31, 3610−3619
Energy & Fuels Article

Symposium on Oilfield Chemistry; Anaheim, CA, Feb 20−22, 1991; (23) Deen, W. M. Analysis of Transport Phenomena (Topics in
DOI: 10.2118/21038-MS. Chemical Engineering); Oxford University Press: New York, 1998; Vol.
(2) Ramirez-Jaramillo, E.; Lira-Galeana, C.; Manero, O. Modeling 3.
asphaltene deposition in production pipelines. Energy Fuels 2006, 20 (24) Whitaker, S. Introduction to Fluid Mechanics; Krieger Publishing
(3), 1184−1196. Company: Malabar, FL, 1968.
(3) Jamialahmadi, M.; Soltani, B.; Müller-Steinhagen, H.; Rashtchian, (25) Juyal, P.; McKenna, A. M.; Fan, T.; Cao, T.; Rueda-Velásquez,
D. Measurement and prediction of the rate of deposition of flocculated R. I.; Fitzsimmons, J. E.; Yen, A.; Rodgers, R. P.; Wang, J.; Buckley, J.
asphaltene particles from oil. Int. J. Heat Mass Transfer 2009, 52 (19), S.; Gray, M. R.; Allenson, S. J.; Creek, J. Joint industrial case study for
4624−4634. asphaltene deposition. Energy Fuels 2013, 27 (4), 1899−1908.
(4) Soltani Soulgani, B.; Rashtchian, D.; Tohidi, B.; Jamialahmadi, M. (26) Colebrook, C.; White, C. Experiments with fluid friction in
A novel method for mitigation of asphaltene deposition in the roughened pipes. Proc. R. Soc. London, Ser. A 1937, 161, 367−381.
wellstring. Iran. J. Chem. Chem. Eng. 2010, 29 (2), 131−142.
(5) Eskin, D.; Ratulowski, J.; Akbarzadeh, K.; Pan, S. Modelling
asphaltene deposition in turbulent pipeline flows. Can. J. Chem. Eng.
2011, 89 (3), 421−441.
(6) Vargas, F. M.; Creek, J. L.; Chapman, W. G. On the development
of an asphaltene deposition simulator. Energy Fuels 2010, 24 (4),
2294−2299.
(7) Kurup, A. S.; Vargas, F. M.; Wang, J.; Buckley, J.; Creek, J. L.;
Subramani, H. J.; Chapman, W. G. Development and application of an
asphaltene deposition tool (ADEPT) for well bores. Energy Fuels 2011,
25 (10), 4506−4516.
(8) Kurup, A. S.; Wang, J.; Subramani, H. J.; Buckley, J.; Creek, J. L.;
Chapman, W. G. Revisiting asphaltene deposition tool (ADEPT):
Field application. Energy Fuels 2012, 26 (9), 5702−5710.
(9) (a) Hildebrand, J.; Scott, R. The Solubility of Nonclectrolytes, 3rd
ed.; Reinhold: New York 1949; Vol. 488. (b) Hildebrand, J. H.; Scott,
R. L. Regular Solutions; Prentice-Hall: Upper Saddle River, NJ, 1962.
(10) Wiehe, I. A. Process Chemistry of Petroleum Macromolecules; CRC
Press: Boca Raton, FL, 2008.
(11) Vargas, F. M. Modeling of asphaltene precipitation and arterial
deposition. Ph.D Dissertation, Rice University, Houston, TX, 2009.
(12) Gross, J.; Sadowski, G. Perturbed-chain SAFT: An equation of
state based on a perturbation theory for chain molecules. Ind. Eng.
Chem. Res. 2001, 40 (4), 1244−1260.
(13) Panuganti, S. R.; Vargas, F. M.; Gonzalez, D. L.; Kurup, A. S.;
Chapman, W. G. PC-SAFT characterization of crude oils and
modeling of asphaltene phase behavior. Fuel 2012, 93, 658−669.
(14) AlHammadi, A. A.; Vargas, F. M.; Chapman, W. G. Comparison
of cubic-plus-association and perturbed-chain statistical associating
fluid theory methods for modeling asphaltene phase behavior and
pressure−volume−temperature properties. Energy Fuels 2015, 29 (5),
2864−2875.
(15) Boek, E. S.; Ladva, H. K.; Crawshaw, J. P.; Padding, J. T.
Deposition of colloidal asphaltene in capillary flow: Experiments and
mesoscopic simulation. Energy Fuels 2008, 22 (2), 805−813.
(16) Aske, N.; Kallevik, H.; Johnsen, E. E.; Sjöblom, J. Asphaltene
aggregation from crude oils and model systems studied by high-
pressure NIR spectroscopy. Energy Fuels 2002, 16 (5), 1287−1295.
(17) Schabron, J. F.; Rovani, J. F. On-column precipitation and re-
dissolution of asphaltenes in petroleum residua. Fuel 2008, 87 (2),
165−176.
(18) Pan, H.; Firoozabadi, A. Thermodynamic micellization model
for asphaltene precipitation inhibition. AIChE J. 2000, 46 (2), 416−
426.
(19) Lang, R.; Xanh, N. X. Smoluchowski’s theory of coagulation in
colloids holds rigorously in the Boltzmann-Grad-limit. Zeitschrift für
Wahrscheinlichkeitstheorie und Verwandte Gebiete 1980, 54 (3), 227−
280.
(20) Wang, J.; Buckley, J. S.; Creek, J. L. Asphaltene deposition on
metallic surfaces. J. Dispersion Sci. Technol. 2004, 25 (3), 287−298.
(21) Broseta, D.; Robin, M.; Savvidis, T.; Féjean, C.; Durandeau, M.;
Zhou, H. Detection of asphaltene deposition by capillary flow
measurements. Proceedings of the SPE/DOE Improved Oil Recovery
Symposium; Tulsa, OK, April 3−5, 2000; DOI: 10.2118/59294-MS.
(22) Ali, M.; Islam, M. The effect of asphaltene precipitation on
carbonate-rock permeability: An experimental and numerical ap-
proach. SPE Prod. Facil. 1998, 13 (03), 178−183.

3619 DOI: 10.1021/acs.energyfuels.6b02313

Energy Fuels 2017, 31, 3610−3619