water

Article
Numerical Study of Water-Oil Two-Phase Flow Evolution in a
Y-Junction Horizontal Pipeline
M. De la Cruz-Ávila 1, * , I. Carvajal-Mariscal 2 , Leonardo Di G. Sigalotti 3 and Jaime Klapp 1, *

1 Instituto Nacional de Investigaciones Nucleares (ININ), Carretera México Toluca km. 36.5, La Marquesa,
Ocoyoacac 52750, Mexico
2 Instituto Politécnico Nacional, ESIME-UPALM, Mexico City 07738, Mexico
3 Departamento de Ciencias Básicas, Universidad Autónoma Metropolitana-Azcapotzalco, Av. San Pablo 180,
Mexico City 02200, Mexico
* Correspondence: [email protected] (M.D.l.C.-Á.); [email protected] (J.K.);
Tel.: +52-(55)-53297200 (J.K.)

Abstract: The work aims to numerically evaluate different injection configurations for the analysis
of a two-phase flow behavior and evolution through a staggered Y-junction pipeline. To minimize
agglomeration between inlets, the injection zones have a separation distance, avoiding areas with
eddies or swirls owing to strong turbulence. Six input scenarios were examined accordingly with
injection system experimental data. Results show significant variations because the main fluid develops
a swirl over the pipe center. This is generated immediately after the phases’ supply zone due to the
oil-phase because it presents a partial pipe flooding, even in the water injection zone. Moreover, the
supply configuration has significant relevance to the main flow development. Accordingly, many flow
patterns can be achieved depending on the phases’ confluence coming from the supply system. The
interface velocities confirm the transition process and flow pattern development, which are driven
by the phases’ velocities describing the early stages of three flow patterns formed during the fluids’
Citation: De la Cruz-Ávila, M.; confluence. Finally, a substantial extent of the conjunction process points out that caution must be
Carvajal-Mariscal, I.; Sigalotti, L.D.G.; exercised during the injection supply system selection for this type of junction pipeline to achieve a
Klapp, J. Numerical Study of better, and smooth blend, with either narrow, medium, or wide emulsions.
Water-Oil Two-Phase Flow Evolution
in a Y-Junction Horizontal Pipeline. Keywords: numerical study; Y-junction pipe; two-phase flow; high-viscosity fluids; flow evolution;
Water 2022, 14, 3451. https://
water-oil flow
doi.org/10.3390/w14213451

Academic Editors: Gholamreza

Kefayati and Hasan Sajjadi
1. Introduction
Received: 29 September 2022
The two-phase oil-water flow is prevalent in a variety of industries, including bio-
Accepted: 26 October 2022
Published: 29 October 2022
chemical [1], energy, and petroleum extraction, transportation, and storage [2]. Because of
the turbulent momentum and interfacial tension [3] that effect each phase, oil and water
Publisher’s Note: MDPI stays neutral typically flow in the dispersed phase and continuous phase, respectively [3]. Significant
with regard to jurisdictional claims in variations in density and kinematic viscosity among both, oil and water, combined with
published maps and institutional affil- changes of the phase velocity, phase fraction, and pipe geometry, will influence flow condi-
iations.
tions, resulting in complicated and variable flow patterns [4]. Furthermore, the slip velocity
exists across the two phases [5]. As a result of the complicated flow structure and slip
phenomena, properly measuring the flow velocity is challenging [6,7].
Diverse researchers [6,8–11] have investigated oil and water flows extensively with
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
the purpose of characterizing the resultant flow patterns, pressure gradients, in situ phase
This article is an open access article
fractions, and interfacial phenomena under a range of different flow regimes, fluid char-
distributed under the terms and acteristics, and pipe designs. Unfortunately, the nature of these flows is still not totally
conditions of the Creative Commons understood because to the complicated flow dynamics seen in pipelines that involve more
Attribution (CC BY) license (https:// than liquid-liquid flows where the densities and viscosities of the two fluids are significantly
creativecommons.org/licenses/by/ different. Due to their impact on flow pattern transitions [6] and interface curvature [12], ad-
4.0/). ditional parameters such as interfacial forces, wetting properties [13], phase holdup [14–16],

Water 2022, 14, 3451. https://doi.org/10.3390/w14213451 https://www.mdpi.com/journal/water

Water 2022, 14, 3451 2 of 16

phase inversion [4], and the most interesting, miscible characteristic are important factors to
consider [17] in pipe junctions flows.
The T-junction [18–20] and Y-junction [16,21] are two popular pipeline supply methods
in industrial, chemical, and petroleum engineering [22,23]. There are many studies of
junction flow, including oil-water two-phase flow [24], thermal mixing [25], and turbulent
mixing [26], to which researchers are focusing due to variation rules in the various mixing
phases. Many studies have compared the properties of the Y-junction with the T-junction,
for example, Sierens and Verhelst [27] investigated the effect of injection parameters on
four types of junctions, T-junction, Y-junction, 45-degree junction, and 45-degree reverse
junction, and found that the Y-junction produced the highest power and 45-degree junction
produced the highest efficiency.
The turbulent flow in junction mixing is another significant research area. In these
earlier works, computational fluid dynamics (CFD) is a common non-intrusive method
for measuring properties and describing behaviours using numerical models [28]. The
k − ε model [29] and k − ω SST model [30,31] have been used to describe the turbulence in
numerical simulations. The SST model which is based on the k − ω, is used to model the
boundary layer outside, which results are more accurate.

1.1. Aim and Scope

Knowing how the phases’ conjunction develops from the supply system, substantial
improvements can be made to enhance the oil’s transport in pipes, as well as to improve the
techniques already used. As a result, the goal of this research is to use numerical simulations
to examine fluid-flow development in a two-phase system with an equal proportion of water
and oil, but distinct supply entry, minor fouling, and clogging in a Y-junction round horizontal
pipe. Moreover, it can help to better choose in advance the resolution of the holdup correlation
type based on the flow pattern. Last but not least, it can help to better propose the numerical
solution schemes not only based on the multiphase scheme, whether “non-miscible” or
“mixed” but with the consideration of selecting them based on the phase interpenetration.
One of the main purposes of this investigation is to confirm that the emulsification
process is not only driven by the interface or slip velocities at an early stage of the confluence
supply entrances but also by the two-phase flow pattern. The second is to verify that the
flow pattern, i.e., early stage of emulsions, develops according to the phases’ velocities
during the fluids’ confluence. Finally, to reveal if the supply configuration system has
significant relevance to the development of the main fluid flow and the substantial extent
of an emulsification process.

1.2. Implementation
In this study, we present the results obtained numerically using the RANS technique
coupled with the Volume of Fluid (VOF) method, the appliance of the k − ω SST turbulence
model and high interface reconstruction scheme which have taken on greater relevance
and precision of results in the Computational Fluid Dynamics (CFD) study models for two-
phase systems. Through the use of the VOF multiphase model, the phases are modelled
for the representation of a flow in two phases of water-oil. It was possible to obtain the
interfacial or slip velocity, as well as the volumetric fraction of the phases during the
conjunction, showing that the fluid development process in general directly influences the
“mixing” of phases and the development of flow patterns.
When volume fraction or phase fraction is used in immiscible fluids, layers of fluid
are presented rather than bubble-like or drop-like structures, since most post-processing
programs assign a cut-off function or condition to those layers. If care is not taken in
advance to indicate which structures within the bulk fluid are considered as part of the
confluence interaction, it might not actually exist, which is why they are assumed as
misinterpretations. Therefore, an analysis of the strain rate must be made together with the
shear velocity [17], even when these structures exist in the form of layers, it is analyzed how
the bulk-fluid is deformed since the development of the flow pattern is not yet constituted.
 confluence interaction, it might not actually exist, which is why they are assumed as
misinterpretations. Therefore, an analysis of the strain rate must be made together with
the shear velocity [17], even when these structures exist in the form of layers, it is analyzed
how the bulk-fluid is deformed since the development of the flow pattern is not yet
Water 2022, 14, 3451 3 of 16
constituted.
With these results, an addition can be made to the numerical multiphase models for
the correct selection of some coupled model for mass transfer based on diffusion and slip
Withas
velocity, these
wellresults,
as the an addition
flow patterncan be made
based to the numerical
on specific multiphase
dimensionless numbers models
suchfor
as the
the
correct selection of some coupled model for mass transfer based on diffusion
Sherwood, Sh, Schmith, Sc, and multiphase Reynolds number, Reqp. With this informationand slip velocity,
as well
it is aspossible
also the flow to
pattern
selectbased on specific
the correct dimensionless
correlation for the numbers such
calculation ofas theholdup
the Sherwood, Sh,
of each
Schmith,
phase. Sc, and multiphase Reynolds number, Re qp . With this information it is also possible
to select the correct correlation for the calculation of the holdup of each phase.
In what follows, Section 2 briefly describes the experimental testing apparatus, which
In what follows, Section 2 briefly describes the experimental testing apparatus, which
specific details are presented in the respective citation while Section 3 exposes each and
specific details are presented in the respective citation while Section 3 exposes each and
every one of the particularities of the numerical methodology used. Section 4 presents the
every one of the particularities of the numerical methodology used. Section 4 presents
analysis of the results and the discussion of the findings obtained are carried out and
the analysis of the results and the discussion of the findings obtained are carried out and
Section 5 summarizes the relevant conclusions.
Section 5 summarizes the relevant conclusions.
2. Experimental
2. Experimental Setup
Setup
Figure 11 depicts
Figure depicts the
the experimental
experimental setup
setup pipeline
pipeline that
that is
is located
located at at the
the Engineering
Engineering
Institute-Universidad Nacional Autónoma de México (UNAM). This
Institute-Universidad Nacional Autónoma de México (UNAM). This facility was designed facility was designed
to evaluate the flow properties of liquid-liquid and liquid-air mixtures
to evaluate the flow properties of liquid-liquid and liquid-air mixtures constituted by constituted by
high-viscosity liquids. The P1 sensor data is used for this analysis where experimental
high-viscosity liquids. The P1 sensor data is used for this analysis where experimental rig is rig
is detailed
detailed described
described in [32–34].
in [32–34]. For theseForparticularly
these particularly
numerical numerical
simulations,simulations,
the geometrytheor
geometry or computational
computational domain were
domain were discretized discretized
explicitly on the explicitly
Y-junctiononsupply
the Y-junction
system tosupply
focus
system to focusresources
computational computational
under resources under
the evolution of the
mainevolution of main fluidThe
fluid development. development.
confluence
The confluence
system system
was selected andwas selected
marked as and marked
Injection as Injection
Point Point
illustrated illustrated
in the close-upinFigure
the close-
1B.
up Figure 1B. Further details are
Further details are described in [32]. described in [32].

Figure 1.
Figure 1. Top
Top view
view of
of the
the complete
complete experimental
experimental rig
rig setup
setup (A),
(A), injection
injection point
point close-up
close-up (B).
(B). (1),
(1), (2)
(2)
and (3) Phases Y-junction injection system. L is the characteristic length and D the characteristic
and (3) Phases Y-junction injection system. L is the characteristic length and D the characteristic
diameter of the pipe respectively for the dimensionless representative scale.
diameter of the pipe respectively for the dimensionless representative scale.

3. Numerical Setup
3.1.
3.1. Case
Case Study
Study
A staggered Y-junction nozzles with a distance of 25 cm between them and a 45◦ angled
junction was selected for the numerical domain model. All injectors have a pipe diameter,
D, of 7.62 cm and a total length, L, of 150 cm. ANSYS Fluent CFD commercial software [35]
in a Xeon 32 cores Workstation and two high-performance GPUs, Nvidia Quadro 6000 and
a Tesla C2075 to accelerate calculation was used to run two-phase water-oil 3D simulations.
For both fluids, the numerical domain is shown as a horizontal Y-junction pipe with two
confluence zones to the main pipeline. According to the research supply configurations [17],
the input mass flow is constant at 5 kg/s for both phases and is distributed symmetrically
Water 2022, 14, 3451 4 of 16

across the injection pipes. The continuous phase is marked as pure water fluid with the
corresponding thermophysical properties and the second phase or dispersed phase is
marked as oil fluid.
It is noteworthy that the second fluid, labelled as oil, has thermophysical proper-
ties set as the same as those of the fluid used in the experiments, which are displayed
elsewhere [34,36]. Glycerol is a very viscous and hygroscopic liquid; it absorbs moisture
from the air. Due to the fact that it is miscible with water in all proportions, it never be-
comes “insoluble” in water. This result in, to determine the general physical emulsification
process, it was not take into account the chemical dynamics [37,38] and the hydrogen
bonding velocity [39] during the conjunction. Therefore, even though the second phase has
the same thermophysical properties as glycerol, it was treated as an immiscible fluid. All
this with the main objective of focusing on the effects of its viscous and density properties
in the merely physical process of the bulk fluid evolution and development during the first
stages of supply into the pipe. This characterization of the phases allows for the application
of the VOF model.
To minimize agglomeration or collision between the inlets, the injection zones have a
separation distance, with the goal of reducing excessive turbulence and therefore avoiding
regions with eddies or swirls over the main channel. Six input cases or combinations were
examined using the phase procedure. Three cases for the incorporation of the low viscosity
fluid through two alternate supply zones and three other cases for the high viscosity fluid
respectively. The numerical simulations test matrix is listed in Table 1.

Table 1. Phases injection input scenarios.

Case Studies 1 2 3
Nozzle 1 2 3 1 2 3 1 2 3
Water X O O O X O O O X
Glycerol O X X X O X X X O
Case Studies 4 5 6
Nozzle 1 2 3 1 2 3 1 2 3
Water O X X X O X X X O
Glycerol X O O O X O O O X

3.2. Numerical Domain Details

The numerical domain mesh was created using a combination of advancing-front
meshing [40] and the Cut-Cell technique [41]. The advancing-front technique offers several
advantages over traditional grids, including the ability to alter the mesh density and
facilitate tessellation in geometrically complicated domains. The staggered nozzles, which
the Cut-Cell mesh must be fitted to, are the reason of the geometry’s refinement, particularly
at the Injection Point, as seen in Figure 2. When these two methods are used, a steady
increase of thin layers and cells from the pipe walls to the domain’s border is obtained,
allowing a good approximation of the physical phenomenon, especially when high-order
discretization schemes are applied.
Water 2022,
Water 2022, 14,
14, 3451
3451 55 of
of 16

Figure 2. Cut-Cell hexahedral mesh for the phase injection system.

system.

In the simulations,
In the simulations,adiabatic
adiabaticandand non-slip
non-slip enhanced
enhanced wallwall function
function to correct
to correct any
any mis-
miscalculation
calculation nearnear walls
walls treatment
treatment werewere employed
employed as boundary
as boundary conditions.
conditions. Because Because
a desire a
desire Y+ intypes
Y+ in these theseoftypes of two-phase
two-phase simulations,
simulations, which arewhich are mostly
mostly driven bydriven by the high-
the high-viscous
viscous flow, 1.5
flow, is from is from
to 2, a1.5 to 2, a generalized
generalized Y+ was
Y+ was taken taken
mostly mostlyas
desirable desirable
1 for theas 1 for the
initial-cell-
initial-cell-length of the constructed meshes whose outcome was to
length of the constructed meshes whose outcome was to get the correct length of theget the correct length
first
of the first cell at 1.6. This Y+ was taken into account also for the fastest phase,
cell at 1.6. This Y+ was taken into account also for the fastest phase, which in this case was which in
this case was water as the transport phase. The first grid length on the walls
water as the transport phase. The first grid length on the walls is 0.000235 m with an in- is 0.000235 m
with an increment of 20% to complete 10 layers. In each nozzle, a mass
crement of 20% to complete 10 layers. In each nozzle, a mass flow condition was employed flow condition
was employed
to inject to inject
the phases. Thethe phases. The
discharge discharge
occurs occurslimit
at the outlet at the
at outlet limit atconditions
atmospheric atmospheric of
conditions of 1 atm of pressure and
1 atm of pressure and 293 K of temperature. 293 K of temperature.

3.3. Numerical Domain Details

3.3. Numerical Domain Details
The numerical simulations were discretised as follows. For pressure spatial discretization
The numerical simulations were discretised as follows. For pressure spatial discreti-
the PRESTO Scheme (PREssure-STaggering-Option) [42], a third order MUSCL (Monotonic
zation the PRESTO Scheme (PREssure-STaggering-Option) [42], a third order MUSCL
Upstream-centred Scheme for Conservation Laws) [43] for momentum solution and the
(Monotonic Upstream-centred Scheme for Conservation Laws) [43] for momentum solu-
modified HRIC (High Resolution Interface Capturing) [44,45] for the volume fraction recon-
tion and the
struction. A modified
completely HRIC
linked(High Resolution
approach Interface
was also built Capturing) [44,45] for thesolution
for the pressure-velocity volume
fraction reconstruction. A completely linked approach was also built for the pressure-ve-
technique. The results demonstrate that, as compared to first-order techniques, the high-order
locity solution
MUSCL methodtechnique.
efficientlyThe results numerical
minimizes demonstrate that, asresulting
diffusion, compared to first-order
in superior tech-
two-phase
niques, the high-order MUSCL method efficiently minimizes numerical
flow resolution. Furthermore, when compared to other options, the PRESTO and HRIC with diffusion, result-
ing pressure-velocity
the in superior two-phase flow resolution.
fully coupled operating Furthermore, when compared
ensemble demonstrate greatertoconvergence.
other options,
the PRESTO and
In addition, HRIC
in this kindwith the pressure-velocity
of two-phase simulation, anfully coupled
adaptable step operating ensemble
size was included for
demonstrate greater convergence.
the numerical simulations to guarantee that the time-step was appropriate for the fluid flow
In addition,
development. Theininterval
this kindof of two-phase
time should be simulation,
short enoughan adaptable step size was included
to solve time-dependent features
for the numerical simulations to guarantee that the time-step was appropriate
and provide convergence in the timeframe given. The time step is calculated using forequation,
the fluid
flow development. The interval of time should be short enough to solve time-dependent
features and provide convergence in the Typical Cell Size
timeframe given. The time step is calculated us-
∆t = , (1)
ing equation, Characteristic Flow Velocity

and the values∆𝑡

considered 𝑇𝑦𝑝𝑖𝑐𝑎𝑙 𝐶𝑒𝑙𝑙 𝑆𝑖𝑧𝑒 have the order of 2.25 × 10−5 < ∆t < 5 × 10−3 .
in the simulations
= , (1)
𝐶ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐 𝐹𝑙𝑜𝑤 𝑉𝑒𝑙𝑜𝑐𝑖𝑡𝑦
and the values considered in the simulations have the order of 2.25 × 10 < ∆𝑡 <
5 × 10 .
Water 2022, 14, 3451 6 of 16

The most suited two-phase model for monitoring the surface of the two immiscible
fluids is the so-called Volume of Fluid (VOF, or surface-tracking approach), owing to the
hydrodynamics of the water-oil flow in this work.
The standing features of the VOF technique are:
• Accuracy in the prediction of two-phase flow. In this case, the high viscous water-oil
two-phase because it considers non-interpenetrating phases;
• Utilization of Third-order discretization schemes for phase tracking;
• The high phase reconstruction schemes for the volume fraction;
• More flexibility and efficiency than the finite-difference method. This includes dealing
with issues requiring extremely complex free surface configurations;
• It provides a simple and affordable way to monitor phases on three-dimensional grids;
• It runs in high-performance GPUs parallel processing.
In this technique, the phases are considered as continuous, which prevents them
from interpenetrating. Additionally, the phases in this approach are isothermal, transitory,
and do not include mass transfer or phase change. As a result, each of the fluids under
consideration has a single set of momentum equations, and the volume fraction of each
fluid in each computing cell is traced across the domain.
When utilizing the VOF model, there are a few things to keep in mind ahead of time
to ensure a good numerical description. That is, the volume fractions of all phases in each
control volume must amount to one. The fields for all variables and attributes are shared
by the phases and represent volume-averaged values as long as the volume fraction of
either of the phases is known per each location. As a result, the variables and features in
any particular cell are either solely reflective of one of the th-phases, or indicative of the
phases’ combination, depending on the volume fraction values,

0 → Cell is empty o f the qth f luid

 
αq = (2)
1 → Cell is f ull o f the qth f luid

0 < αq < 1 → The cell contains the inter f ace. (3)

Based on the local value of the phase αq the appropriate properties and variables will
be assigned to each control volume within the domain. In this study, the primary phase is
water, αw .

3.4. Governing Equations

The solution of the continuity equation for the phase volume fraction allows the
tracking of the interface between the phases, and is given by:

∂ρ →
∂t
+ ∇·ρ v = ∑ n Sn , z (4)

→
where ρ is the density, v the velocity vector, t the time and S = 0 due to the no mass transfer
assumption. For the interfacial tracking, oil as the secondary phase, α g , is achieved by
finding the solution of the Equation (1) for α g , thus,


∂ ρg αg →
+ ∇·ρ g α g v = 0. (5)
∂t
Therefore, from the aforementioned considerations, the volume fraction of αw is
computed from the relation αw + α g = 0.
Because the resultant velocity field is shared by all phases, just one momentum
equation is solved for the whole computational domain, which is determined by the
volume fractions of all phases through ρ and µ.

∂  → →T
 →→ →
ρ v + ∇· ρ v v = −∇ p + ∇·[µ(∇ v + ∇ v )] + ρg + F, (6)
∂t
Water 2022, 14, 3451 7 of 16

where, p, µ, g and F are the density, velocity, pressure in the flow field, viscosity, acceleration
due to gravity and the body force respectively. On the other hand, p and µ are estimated by
p p
using, ρ = ∑1 ρq αq and µ = ∑1 µq αq .

3.5. Interfacial and Surface Tension Treatment

The VOF approach constructs the interface between the fluids using a piecewise-linear
method. The interface between two fluids is considered to have a linear slope within each
cell. The system uses this linear form to compute fluid advection via the cell faces. The
initial stage of interface reconstruction uses the volume fraction and derivatives values
in each cell to determine the location of the linear interface respect to the centre of each
partially filled cell. The fluid advection across each face is then calculated using the
computed linear interface approximation and data about the normal and tangential velocity
profile on the face. Finally, the volume flux in each cell is computed using the balance of
fluxes obtained in the previous stage.
In addition, the VOF method takes into consideration surface tension at the phase
contact. The contact angle between the phases and the wall is specified in the model, and
the surface tension coefficient is assumed to be constant. The surface tension model use the
continuous surface force model [46] to do this. When surface tension is taken into account
in the VOF calculation, a source term, F, appears in the momentum equation at which the
pressure drop across the surface can be calculated using the surface tension coefficient, σ,
and the curvature of the surface can be estimated using the Young-Laplace equation and
two radii in orthogonal directions R1 and R2, defined as P2 − P1 = σ (1/R1 + 1/R2 ). As a
result, the pressure drop across the surface may be used to describe surface tension. The
source term for two-phase water-oil is then written as:

ρ∇α g
F = σκ 1
. (7)
2 ρ g − ρw

The interface curvature κ is defined in terms of the divergence of the unit normal, n̂,
as: κ = ∇·n̂, where n̂ = n/|n|. Here the surface normal is n = ∇αg . The surface curvature
is calculated based on the local gradient of the vector normal to the interface, defined as
the gradient of the volume fraction of oil αg .

3.6. Turbulence Model

If one or more of the phases are in the turbulent regime, turbulence models must be
included in numerical simulations of two-phase flow. In this work, the k − ω Shear Stress
Transport (SST) turbulence model [30], which is based on the Reynolds Averaged Navier-
Stokes (RANS) technique, was used since it has certain benefits when it comes to high
viscosity water-oil flow [47]. Specifically, it is used to make it easier to estimate the onset
and degree of flow separation under adverse pressure gradients by including transport
effects into the eddy-viscosity approximation over k − ε turbulence models. In comparison
to the k − ε model it also provides certain advantages in free-stream independence. The
model is given by,
 
∂ ∂ ∂ ∂κ
(ρκ ) + (ρκvi ) = Γκ eκ − Yκ + Sκ ,
+G (8)
∂t ∂xi ∂x j ∂xi
!
∂ ∂ ∂ ∂ω
(ρω ) + (ρωvi ) = Γω + Gω − Yω + Dω + Sω . (9)
∂t ∂xi ∂x j ∂x j

The term Geκ represents the production of turbulence kinetic energy due to mean
velocity gradients; Gω the generation of ω: Γκ and Γω the effective diffusivity of κ and ω,
respectively; Yκ and Yω the dissipation of κ and ω due to turbulence, respectively; Dω the
cross-diffusion term; Sκ and Sω are user-defined source terms. The effective diffusivities
for k − ω are the same as in standard k − ω model.
Water 2022,
Water 2022,14,
14,3451
3451 88 of 16
of 16

3.7. Sensitivity
3.7. Sensitivity Analysis
Analysis and
and Validation
Validation
Four mesh
Four mesh variants
variantswere
werecreated
createdininorder
ordertotoget
getnumerical
numerical results
results that
that were
were notnot reli-
reliant
ant on the mesh. The variations of the outcomes for each mesh version are shown
on the mesh. The variations of the outcomes for each mesh version are shown in Figure 3. in Figure
3. The
The data
data forfor
thethe sensitivity
sensitivity analysis
analysis were
were collected
collected from
from thethe data
data retrieved
retrieved using
using a cen-
a central
tral line
line or central
or central marker
marker ofpressure
of the the pressure gradient
gradient along along the calculation
the calculation domain.
domain.

Figure 3.
Figure 3. Pressure plot,
plot, values
values along
along the
the central
central marker.
marker. Experimental
Experimental data
data obtained
obtainedfrom
from[33].
[33].

Mesh-A significantly underestimates pressure values, suggesting suggesting that the represen-
tation of flow,
flow, i.e.,
i.e., hydrodynamics,
hydrodynamics,was wasaccurately
accuratelysimulated,
simulated, but butthethe precision
precision of the
of the re-
results is questionable when compared to Meshes -B, -C, and
sults is questionable when compared to Meshes -B, -C, and -D. Mesh-B, on the other hand,-D. Mesh-B, on the other
hand, continues
continues to underestimate
to underestimate pressurepressure
values values when compared
when compared to real
to real data. data. Despite
Despite the fact
the fact that the number of cells is slightly over three times
that the number of cells is slightly over three times that of Mesh-A, 44% less that of Mesh-A, 44%thanlessMesh-
than
Mesh-C,
C, and 54% andless
54% lessMesh-D,
than than Mesh-D, the outcomes
the outcomes are stillare
farstill
fromfarthose
fromobtained
those obtained in the
in the exper-
experimental
imental data.data. MeshMeshC, forC,example,
for example, displays
displays a region
a region withwith values
values thatthat
areare already
already near
near to
to the experimental ones, but with a 22% of variance, which is still
the experimental ones, but with a 22% of variance, which is still unsatisfactory owing to unsatisfactory owing
to
thethe needed
needed accuracy
accuracy of the
of the experimental
experimental findings.
findings. Finally,
Finally, Mesh-DMesh-D exhibits
exhibits a 9.98%a 9.98%
dis-
discrepancy
crepancy from the experimental values essentially in the area of the injectors, but as
from the experimental values essentially in the area of the injectors, but as the
the
flow
flow develops,
develops, it it stabilizes
stabilizes with
with aa 0.23%
0.23% deviation
deviation before
before 11.811.8L/D.𝐿/𝐷. More
More refined
refined meshes
meshes
beyond
beyond this thiscells
cellsnumber
number of of
Mesh-D
Mesh-D are are
not possible owing
not possible to computing
owing to computingresources, which
resources,
might
whichresult
mightin result
unattainable time convergence,
in unattainable forcing the use
time convergence, of LES
forcing therefined
use of meshes, which
LES refined
need
meshes, which need a different kind of model and discretization. Therefore, Mesh-Dsince
a different kind of model and discretization. Therefore, Mesh-D was selected was
the highest
selected precision
since of theprecision
the highest findings ofis required,
the findingsas well as the fastest
is required, as well possible convergence
as the fastest possi-
time, to be able totime,
ble convergence characterize theto
to be able development
characterizeofthe thedevelopment
main flow in of thethe
early
mainemulsification
flow in the
procedure. For these types of mixing simulations, equilibrium, or stability, is reached with
early emulsification procedure. For these types of mixing simulations, equilibrium, or sta-
a residual convergence of 1 × 10−5 in 15 s of flow. For−5 this reason and mostly due to the
bility, is reached with a residual convergence of 1 × 10 in 15 s of flow. For this reason and
confluence, the development of the flow pattern is not yet constituted.
mostly due to the confluence, the development of the flow pattern is not yet constituted.
It is worth noting that the deviations from estimated behavior between meshes are
It is worth noting that the deviations from estimated behavior between meshes are
ascribed to modelling considerations including the inherent irreversibilities within the
ascribed to modelling considerations including the inherent irreversibilities within the
process; however, these are deemed not significant based on the strict parametrization
process; however, these are deemed not significant based on the strict parametrization
conducted and thorough meshing procedure, selection of the discretization schemes, the
conducted and thorough meshing procedure, selection of the discretization schemes, the
functions for the treatment of the phases, as well as the boundary conditions. Since the
functions for the treatment of the phases, as well as the boundary conditions. Since the
development of the flow pattern is not yet constituted in the initial stage of the pipeline,
development of the flow pattern is not yet constituted in the initial stage of the pipeline,
fluctuations in the results are expected, resulting in these deviations. However, it should be
fluctuations
noted that thesein the
areresults are deviations
not major expected, resulting in these
but are within thedeviations.
parametersHowever, it should
for the simulations.
Water 2022, 14, 3451 9 of 16

Water 2022, 14, 3451 be noted that these are not major deviations but are within the parameters for the simula-
9 of 16
tions.

4.
4. Results
Results and
and Analysis
Analysis
Water
Water is considered the
is considered the transport
transport phase
phase or or continuous
continuous phase phase inin this
this water-oil
water-oil two-two-
phase
phase research, whereas oil is the dispersed phase. The variables of the continuous phase
research, whereas oil is the dispersed phase. The variables of the continuous phase
inside
inside the
the VOF
VOF modelmodel will
will determine
determine the the fluid’s
fluid’s develop
develop assessment.
assessment. HavingHaving stated
stated so,
so,
the
the calculation and data acquired from the behavior throughout the continuous phase’s
calculation and data acquired from the behavior throughout the continuous phase’s
calculation domainwill
calculation domain willbebeused
usedtotoanalyze
analyze thethe evolution
evolution of the
of the water-oil
water-oil flowflow develop-
development.
ment. The results
The results will bewill be analyzed
analyzed by means by means
of slices ofonslices on XY-plane
XY-plane along
along the the Z axis.
Z axis.
The Figure 4 illustrates
illustrates the phases
phases confluence
confluence in the numerical
numerical domain. The The evolution
evolution
of the main fluid stands out
stands out from case to case in particular, on how the interaction of the
phases develops. In this figure,
develops. In this figure, it can be seen that for the cases of double water injection
with apparent lateral thrust, in specific Cases 1 and 3, an an envelope
envelope of the the continuous
continuous phase
phase
is presented
presentedoveroverthethedispersed
dispersedphase.
phase.That That is,is,
when
when thethe
injection of water
injection of watersupplied
suppliedby the
by
main pipeline
the main pipelineand and
the lateral nozzle,
the lateral either
nozzle, rightright
either or left
or(1,left2 (1,
or 2, 3),2,during
2 or the first
3), during the mo-
first
ments
moments 3.933.93
< 𝐿</ 𝐷 L< / D15.72 the water
< 15.72 surrounds
the water surroundsthe oil.
the This observation
oil. This observationis notis un-
not
known
unknown as as
it bears
it bearssimilarity
similarity totothethewavy
wavystratified
stratifiedregime
regimeaccordingly
accordingly to to Angeli
Angeli and
Hewitt [11]. This
Hewitt This result
result entails
entails characteristics
characteristics for for the
the possible
possible development
development of of narrow
emulsions [6,48].
emulsions [6,48]. Additionally,
Additionally, aa slight
slight turn
turn of of the main flow can be seen in the opposite
direction of the injection nozzle.

Figure
Figure 4.
4. Water-oil and oil-water
Water-oil and oil-water flows
flows development
development by
by means
means of
of phase
phase fraction.
fraction.

In an
aninitial
initialassessment,
assessment,it can be be
it can said thatthat
said there is anisappearance
there an appearanceof drops at the water
of drops at the
oil
waterinterface coincident
oil interface withwith
coincident the interface
the interfacewavywavy development,
development, andanda apossibly
possibly narrow
narrow
emulsion, which could be attributed to the interfacial interfacial velocity [6,49]. However,
However, because of
the numerical domain total length selected, at the first instance, to study thethe
the numerical domain total length selected, at the first instance, to study confluence
confluence in
ainY-junction,
a Y-junction, caution
caution should
should bebeconsidered
consideredononthe theanalysis
analysisofofthe
the interface
interface because
because the
development of a specific or complete interface is still evolving.
The lateral confluence in which the phases phases of CaseCase 22 develop,
develop, stabilizes
stabilizes the
the main
main flow
flow
confines it
and confines it towards
towards the the centre
centre of of the
the pipe from 11.79
pipe from 𝐿/𝐷 until the exit boundary. In
11.79 L/D
contrast, it resembles
resembles aa developing
developingannular
annularflow.flow.However,
However,according
according toto
thethe
relationship
relationshipof
surface velocities of the flow pattern map, this flow pattern developed
of surface velocities of the flow pattern map, this flow pattern developed at velocities at velocities greater
than 10 than
greater m/s approximately. In this case,
10 m/s approximately. a major
In this case, development and behaviour
a major development of the wavy
and behaviour of
stratified
the wavy flow couldflow
stratified be achieved
could be in lengthsin
achieved greater
lengths than L/D.19.65 𝐿/𝐷.
19.65than
greater
When the oil confluence is lateral from 1.96 L/D or 3.93 L/D respectively, the develop-
ment of the main fluid does not differ much between Cases 4 and 6, but rather by the fluid
load inclined towards one side of the main pipeline and opposite to the injection boundary.
Still, this behaviour is similar to the wavy stratified flow pattern as in every case as noticed
Water 2022, 14, 3451 10 of 16

When the oil confluence is lateral from 1.96 𝐿/𝐷 or 3.93 𝐿/𝐷 respectively, the de-
Water 2022, 14, 3451 velopment of the main fluid does not differ much between Cases 4 and 6, but rather 10 ofby
16
the fluid load inclined towards one side of the main pipeline and opposite to the injection
boundary. Still, this behaviour is similar to the wavy stratified flow pattern as in every
case as noticed
by Angeli by Angeli
and Hewitt and
[11]. Hewitt
Main fluid[11]. Main fluidmust
development development
continuemustin order continue in order
to achieve the
to achieve
real the real
or complete flowor complete
pattern and flow pattern
thereby and thereby
determine if thedetermine if the emulsion
emulsion generated gener-
is narrow or
ated
mediumis narrow
[6,11,48].or medium [6,11,48].
The Case
Case55isisthe themost
mostcomplicated
complicated development
development achieved
achievedwithin this numerical
within this numerical sim-
ulation
simulationbecause
becauseit encloses all theallcharacteristics
it encloses the characteristicsof the ofcases
theaforementioned.
cases aforementioned. First, during
First,
the confluence,
during the confluence,the dispersed
the dispersed phase slightly
phase surrounds
slightly surrounds thethecontinuous
continuousphase, phase, from
3.93
3.93 𝐿/𝐷
L/D to to 5.89
5.89 𝐿/𝐷.
L/D.Immediately
Immediately after, thethe
after, oiloil
flow
flowcomes
comes from thethe
from left left
sideside
injector (3),
injector
(3), destabilizing
destabilizing the theflow flow
andand causing
causing a secondzone
a second zoneofofenvelopment
envelopmentfrom from 5.89
5.89 𝐿/𝐷L/D to
13.7 𝐿/𝐷. As
13.7 L/D. Asthe theflow
flow continues
continues to to evolve,
evolve, thethe flow
flow pattern
pattern could
could bebe considered
considered as wavy
stratified with
with thethepossibility
possibilityofofgenerating
generatinga anarrow
narrow emulsion
emulsion [6,48]. However,
[6,48]. However, thetheinterface
inter-
that that
face develops
developsis notiscompletely
not completely defined, at least
defined, it is not
at least it isappreciated
not appreciatedthrough the slices
through used
the slices
so far. For all of the above, it is necessary to analyse the velocity
used so far. For all of the above, it is necessary to analyse the velocity contours and cor- contours and corroborate
that the interfacial
roborate velocity effectively
that the interfacial describesdescribes
velocity effectively the aforementioned flow pattern.
the aforementioned flow pattern.
Figure55the
In Figure theoil-water
oil-water flow
flow development
development by by
means means of velocity
of velocity contours
contours of XYofmul-XY
multiple
tiple slices
slices alongalongthe the Z axis
Z axis is depicted.
is depicted. In In this
this analysis,the
analysis, themain
mainfluid
fluidvelocity
velocity contours
contours
cannot show
cannot show the the interface
interface oror the
the interface
interface velocity
velocity values.
values. However,
However, it it is
is possible
possible to to show
show
the velocity values where the phases
the velocity values where the phases are located. are located.

Figure 5. Oil-water flow development: velocity contours of multiple slices along the Z axis.
Figure 5. Oil-water flow development: velocity contours of multiple slices along the Z axis.

In
In Case 1, as
Case 1, as same
same inin Case
Case 3,3, after
afterthe
theconfluence
confluencejunction
junctionL/D𝐿/𝐷==6.88
6.88namely
namely injec-
injector
tor 3, the velocity contour in every slice is between 𝑣 ≈ 1.55 and ≈
3, the velocity contour in every slice is between v ≈ 1.55 and ≈ 2.33 m/s and hardly 2.33 m/s and hardly
recognisable
recognisable phase
phase velocity
velocity indicating
indicating thatthat phases
phases are
are moving
moving together
together with
with almost
almost the the
same
same velocity. Despite that, an interface velocity analysis must be performed in order
velocity. Despite that, an interface velocity analysis must be performed in order to to
determine
determine the the phases
phases velocity
velocity apart
apart from
from each other. WhenWhen water
water is supply
supply by the later
injectors
injectors it thrusts the oil performing a subtle blend, which, in many cases, the resulting
emulsion is medium [6,11,48].
In Cases 2, 44 and
and 6,6, the
thephases
phasesvelocities
velocitiesare
arewell
welldefined.
defined.For
Foreach
eachcase,
case, the
the oiloil that
that is
is moving slower, approximately 𝑣 ≈ 1.6 m/s, than the water 𝑣 ≈
moving slower, approximately v g ≈ 1.6 m/s, than the water vw ≈ 2.33 m/s precipitates to2.33 m/s precipi-
tates to the bottom
the bottom of the
of the pipe pipe forming
forming a basin as a basin as explained
explained Hasson etHasson et Specifically,
al. [50]. al. [50]. Specifically,
for Case
2 oil-water,
for and for the
Case 2 oil-water, andother twoother
for the Cases,twoof water-oil. Additionally,
Cases, of water-oil. Trallero etTrallero
Additionally, al. [6] andet
Al-Hadhrami
al.[6] et al. [49] describe
and Al-Hadhrami how
et al. [49] surface
describe velocity
how determines
surface the flow pattern,
velocity determines andpat-
the flow the
development
tern, of some typeofofsome
and the development emulsion,
type ofbeemulsion,
it narrow,bemedium or medium
it narrow, wide. They detailThey
or wide. that,
under a low water-oil surface velocity ratio, the flow is gravity dominated and the phases
are totally segregated, where the interface is smooth. This fluid characteristic is particularly
of the stratified wavy flow pattern. However, the wavy behavior is normal to the lateral
injectors and in the radial direction, and not according to the flow pater map that is the
Water 2022, 14, 3451 11 of 16

axial direction normal to the main pipeline inlet boundary. In this specific study, the fluid
flow development resembles more the annular flow pattern formation described by Hasson
et al. [50] with the possibility of developing a medium to wide emulsion [6,11,48]. In Case
5, the development behavior is rather different, specifically, from 5.89 L/D until 15.72 L/D
where the confluence states a different and chaotic blend.
Now, when analyzing the velocity of the interface, information is obtained about the
Water 2022, 14, 3451 12 of 16
development and evolution of the flow pattern that must be formed under this confluence
configuration or Y-junction supply system proposed. Figure 6 presents the complete
interface of the water-oil flow development which was obtained by means of iso-surface
represents the oil-water interphase, specifically meaning that oil surrounding water and
where water has a 50% mass fraction and 50% of oil mass fraction [51] like in Figure 4.
an annular flow pattern might be developing [50]. However, there is no conclusive infor-
Additionally,
mation threeisperspective
about what happening withviewsthewere selected
water-oil to vice
and/or facilitate
versa the complete observation
emulsification, or
of the development of the interface: Isometric, YZ plane and XZ plane
the specific flow pattern development, at least not according to data described ofby
the entire internal
Nädler
computational
and Mewes [48]. In domain. On itthe
spite of this, canother hand,that
be ensured thedue
velocity contours
to the chaotic were placed
development and over the
iso-surface
the proportionsto perceive the phases’
of the phases, interaction
a wide emulsion over
begins to this thin[6,49].
develop layer.

Figure
Figure6.6.Iso-surface of the
Iso-surface water-oil
of the interface
water-oil with the
interface respectively
with velocityvelocity
the respectively contours contours
with different
with different
viewing planes: Isometric, YZ and XZ planes.
viewing planes: Isometric, YZ and XZ planes.

Figure 7 shows the velocity scatterplots of the interface which illustrate the velocity
of the main fluid mixture development. By contrasting the results obtained in this work
against the flow pattern maps, the early stages of emulsification are observed. The distri-
bution of the oil volume fraction in a cross section in the pipeline for the mixing velocity
Water 2022, 14, 3451 12 of 16

In Case 1, the images show that the path of this interface follows, has a S shape,
starting from 5.89 L/D until the outlet boundary (XZ plane). On the other hand, in the
confluence zone, specifically at 5.89 L/D, a slightly thrust is exerted on the main fluid
where the interface curves (YZ plane). As the fluid evolution continues, a wave that
ranges from 7.86 L/D to 15.72 L/D confirms the wavy stratified pattern. At 12.78 L/D
the aforementioned oil-water basin is better perceived (isometric). All this trajectory
development reveals a delicate twist that the iso-surface undergoes, indicating that a
fluid swirl is caused by the lateral confluence thrust. And finally, the velocity contours
indicate that both phases move along the numerical domain in almost constant velocity of
v ≈ 2.33 m/s.
In the same way, the Case 2 show the thrusting zone but in a subtle way than Case
1. Contrary to Case 1, the interface fulfils the numerical domain reaching the wall’s edge
(XZ plane) immediately. The S-shaped path is less noticeable, which better explains the
stability of the main fluid such that mixing occurs at the centre of the pipeline. The wavy
behaviour starts at 5.89L/D but unlike the previous case, the oil remains in the lower zone
of the pipe (YZ plane). Here, there are no traces of the slight swirl of the main fluid as in
the previous case, although it also still shows the basin that is produced from oil-water.
The velocity contours show an increase in the interface velocity, which reaffirms a velocity
of the phases v < 3 m/s. The water that is located in the upper part moves faster than the
oil located in the lower area of the pipe. Despite this, there is an impulse that transmits the
water to the oil originating the stratified wavy flow pattern [11], and probably a narrow
emulsion [6].
The development and evolution of the conjunction in Case 3 is very similar to what
happens in Case 1. However, in this Case, the fluid load hangs towards the right side of
the pipe, also causing the main fluid to swirl. While for Case 1 the swirl does clockwise, in
Case 3 does counter clockwise. The fluid path is mirror-like S-shape, 7.86 L/D (XZ plane)
because the supply system is very much like the opposite of Case 1. The wavy behaviour
starts at 11.79L/D and the water surround oil without presenting the basin of the previous
two cases. The interphase velocity contours show that velocity between phases is slightly
lower 1.55 m/s < v < 1.94 m/s.
The Case 4 and 6, has a similar development an evolution than Case 2 with two main
differences. One is that the swirl is higher perceptible, counter clockwise at ≈ 10.8 L/D and
clockwise ≈ 9.82 L/D respectively and two; the surrounding basin is water-oil like noticed
Nädler and Mewes [48]. These Cases also present an interface velocity of v < 3 m/s.
Despite all these cases’ apparently stable water-oil conjunction flow development
and evolution, Case 5 is rather chaotic. Two iso-surfaces were created, which ensemble
represents the oil-water interphase, specifically meaning that oil surrounding water and an
annular flow pattern might be developing [50]. However, there is no conclusive informa-
tion about what is happening with the water-oil and/or vice versa emulsification, or the
specific flow pattern development, at least not according to data described by Nädler and
Mewes [48]. In spite of this, it can be ensured that due to the chaotic development and the
proportions of the phases, a wide emulsion begins to develop [6,49].
Figure 7 shows the velocity scatterplots of the interface which illustrate the veloc-
ity of the main fluid mixture development. By contrasting the results obtained in this
work against the flow pattern maps, the early stages of emulsification are observed. The
distribution of the oil volume fraction in a cross section in the pipeline for the mixing
velocity (main fluid flow) v > 1.7 m/s, and the oil and water inlet volume fraction at 50%
respectively, describes a concave shape, which in this work is described as basin, behaviour
and development also described by Angeli and Hewitt [11]. This occurs in almost all cases
with the exception of Case 3 and 5 in which the shape of the interface is complex.
Water 2022, 14, 3451 13 of 16

(main fluid flow) 𝑣 > 1.7 m/s, and the oil and water inlet volume fraction at 50% respec-
tively, describes a concave shape, which in this work is described as basin, behaviour and
Water 2022, 14, 3451 13 of 16
development also described by Angeli and Hewitt [11]. This occurs in almost all cases
with the exception of Case 3 and 5 in which the shape of the interface is complex.

Figure7.7.Interface
Figure Interfacevelocity
velocityscatterplots.
scatterplots. The
The iso-surface
iso-surface thatthat represents
represents the interface
the interface is projected
is projected over
over a 2d scatterplot.
a 2d scatterplot.

Forv𝑣>>11m/s,
For m/s,ititstarts
starts the
the wavy
wavy stratified
stratified flow
flow pattern
pattern and
and narrow
narrow emulsion
emulsion devel-
devel-
oping. For 𝑣 = 1.5 m/s, there is no slug flow transition, while for
oping. For v = 1.5 m/s, there is no slug flow transition, while for high v = 2 m/s, thehigh 𝑣 = 2 m/s, the
smoothdispersed
smooth dispersed droplet
droplet flow
flow pattern
pattern starts
starts with
with aa narrow
narrow to
to medium
medium emulsion
emulsion growth.
growth.
Thisbehaviour
This behaviour isis achieved
achieved by by Cases
Cases 11 andand 3. 3. However,
However, 44 Cases have v𝑣 >
Cases have > 2.5
2.5m/s
m/s which
which
implies that according to Trallero et al. [6], Angeli & Hewitt [11] and Vielma
implies that according to Trallero et al. [6], Angeli & Hewitt [11] and Vielma [51] the annular [51] the an-
nular flow pattern is developing. In this way, it is confirmed that the basin
flow pattern is developing. In this way, it is confirmed that the basin formed is due to the formed is due
development of an annular flow, specifically case 2, 4, and 6 and medium to wide emulsions
might be starting to form. For case 5, highly prone to developing wide emulsions [6,11,48].
Water 2022, 14, 3451 14 of 16

5. Conclusions
Different injection configurations were numerically analysed in order to describe the
evolution and development of the water-oil flow pattern in this study. In a horizontal Y-
junction pipeline, six distinct phase supply arrangements were described. Through the
analysis of the fraction of the phases, as well as the slip velocities, different behaviours were
observed. For this reason, the results demonstrate meaningful differences in pattern flow
development, attributable mostly to the phase loaded to one or the other side of the pipeline.
Cases 2 and 5 indicate a more complex structure due to relative restriction to the
pipeline centre, which is caused by the phases’ supply arrangement. The blending process
in Case 3 is the smoothest because the continuous phases gradually direct the primary fluid
to the pipeline centre. This behaviour is followed by Case 1. The rest of the cases, have a
similar development mainly because the high viscous fluid was precipitated to the pipeline
bottom modifying the early stage of the flow pattern formation and evolution.
The interfaces velocities confirm the mixture process and the flow pattern development
which is driven by the phases’ velocities. This is determined precisely when velocities are
1 m/s < v < 3 m/s describing the early stages of the three flow patterns formed during
the fluids’ confluence obtained with these numerical simulations. In addition, a slight swirl
was detected directly attributable to the staggered Y-junction supply system.
Therefore, the supply configuration has a significant relevance on the development of
the main fluid flow and substantial extent on the emulsification and the fluid flow pattern
development. Finally, care must exercise during the supply system in a Y-junction pipeline
to achieve better and smooth blend turning the emulsification process in order to obtain
either narrow, medium or wide emulsion.

Author Contributions: Conceptualization, M.D.l.C.-Á.; methodology, M.D.l.C.-Á.; software, M.D.l.C.-

Á. and I.C.-M.; validation, M.D.l.C.-Á. and L.D.G.S.; formal analysis, M.D.l.C.-Á.; investigation,
M.D.l.C.-Á. and I.C.-M.; resources M.D.l.C.-Á.; data curation, M.D.l.C.-Á.; writing—original draft
preparation, M.D.l.C.-Á.; writing—review and editing, M.D.l.C.-Á., I.C.-M. and L.D.G.S.; visualiza-
tion, M.D.l.C.-Á.; supervision, I.C.-M. and J.K.; project administration, J.K.; funding acquisition, J.K.
All authors have read and agreed to the published version of the manuscript.
Funding: This research work described in this paper was funded by European Union’s Horizon 2020
Programme under the ENERXICO Project (Grant Agreement No. 828947) and under the Mexican
CONACYT-SENER-Hidrocarburos (Grant Agreement No. B-S-69926).
Data Availability Statement: Not applicable.
Acknowledgments: The authors of this paper are particularly grateful to Laboratory of Applied
Thermal and Hydraulic Engineering (LABINTHAP) from the National Polytechnic Institute of Mexico
for the contributions made and the computational resources provided, in collaboration with the
Engineering Institute of the National Autonomous University of Mexico, IINGEN-UNAM.
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design
of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or
in the decision to publish the results.

References
1. Kowalska, G.; Baj, T.; Kowalski, R.; Szymańska, J. Optimization of Glycerol–Water Extraction of Selected Bioactive Compounds
from Peppermint and Common Nettle. Antioxidants 2021, 10, 817. [CrossRef] [PubMed]
2. Bertero, L.; DiLullo, A.; Lentini, A.; Terzi, L. Innovative way to produce and transport heavy oil through dispersion in water:
Laboratory study and field test results. In Proceedings of the SPE Annual Technical Conference and Exhibition, New Orleans, LA,
USA, 25–28 September 1994; pp. 283–295. [CrossRef]
3. Brauner, N.; Maron, D.M. Flow pattern transitions in two-phase liquid-liquid flow in horizontal tubes. Int. J. Multiph. Flow 1992,
18, 123–140. [CrossRef]
4. Brauner, N.; Ullmann, A. Modeling of phase inversion phenomenon in two-phase pipe flows. Int. J. Multiph. Flow 2002, 28,
1177–1204. [CrossRef]
5. Shi, H.; Holmes, J.A.; Durlofsky, L.J.; Aziz, K.; Diaz, L.R.; Alkaya, B.; Oddie, G. Drift-Flux Modeling of Two-Phase Flow in
Wellbores. SPE J. 2005, 10, 24–33. [CrossRef]
Water 2022, 14, 3451 15 of 16

6. Trallero, J.L.; Sarica, C.; Brill, J.P. A Study of Oil/Water Flow Patterns in Horizontal Pipes. SPE Prod. Facil. 1997, 12, 165–172.
[CrossRef]
7. Rodriguez, I.H.; Yamaguti, H.K.; de Castro, M.S.; Da Silva, M.J.; Rodriguez, O.M.H. Slip ratio in dispersed viscous oil–water pipe
flow. Exp. Therm. Fluid Sci. 2011, 35, 11–19. [CrossRef]
8. Russell, T.W.F.; Hodgson, G.W.; Govier, G.W. Horizontal pipeline flow of mixtures of oil and water. Can. J. Chem. Eng. 1959, 37,
9–17. [CrossRef]
9. Charles, M.E.; Govier, G.W.; Hodgson, G.W. The horizontal pipeline flow of equal density oil-water mixtures. Can. J. Chem. Eng.
1961, 39, 27–36. [CrossRef]
10. Arirachakaran, S.; Oglesby, K.D.; Malinowsky, M.S.; Shoham, O.; Brill, J.P. An analysis of oil/water flow phenomena in horizontal
pipes. In Proceedings of the SPE Production Operations Symposium, Oklahoma City, OK, USA, 13–14 March 1989; pp. 155–167.
[CrossRef]
11. Angeli, P.; Hewitt, G. Flow structure in horizontal oil–water flow. Int. J. Multiph. Flow 2000, 26, 1117–1140. [CrossRef]
12. Chen, S.H.; Choi, S.M. Measurement and Interpretation of Curvatures of the Oil–Water Interface in Isometric Bicontinuous
Microemulsions. J. Appl. Crystallogr. 1997, 30, 755–760. [CrossRef]
13. Kee, K.E.; Babic, M.; Richter, S.; Paolinelli, L.; Li, W.; Nesic, S. Flow patterns and water wetting in gas-oil-water three-phase flo. In
Proceedings of the CORROSION 2015, Dallas, TX, USA, 15–19 March 2015; pp. 1–16.
14. Colombo, L.P.M.; Guilizzoni, M.; Sotgia, G.M.; Bortolotti, S.; Pavan, L. Measurement of the oil holdup for a two-phase oil-water
flow through a sudden contraction in a horizontal pipe. J. Physics Conf. Ser. 2014, 501, 012015. [CrossRef]
15. Edomwonyi-Otu, L.C.; Angeli, P. Pressure drop and holdup predictions in horizontal oil–water flows for curved and wavy
interfaces. Chem. Eng. Res. Des. 2015, 93, 55–65. [CrossRef]
16. Ren, G.; Ge, D.; Li, P.; Chen, X.; Zhang, X.; Lu, X.; Sun, K.; Fang, R.; Mi, L.; Su, F. The Flow Pattern Transition and Water Holdup
of Gas–Liquid Flow in the Horizontal and Vertical Sections of a Continuous Transportation Pipe. Water 2021, 13, 2077. [CrossRef]
17. De la Cruz-Ávila, M.; Carvajal-Mariscal, I.; Klapp, J.; Guzmán, J.E.V. Numerical Study of Multiphase Water–Glycerol Emulsifica-
tion Process in a Y-Junction Horizontal Pipeline. Energies 2022, 15, 2723. [CrossRef]
18. Hosseini, S.M.; Yuki, K.; Hashizume, H. Experimental Investigation of Flow Field Structure in Mixing Tee. J. Fluids Eng. 2009, 131, 051103.
[CrossRef]
19. Yang, L.; Wang, J.; Ma, Y.; Liu, S.; Tang, J.; Zhu, Y. Oil-Water-Gas Three-Phase Separation in Multitube T-Junction Separators.
Water 2019, 11, 2655. [CrossRef]
20. Grbčić, L.; Kranjčević, L.; Družeta, S.; Lučin, I. Efficient Double-Tee Junction Mixing Assessment by Machine Learning. Water
2020, 12, 238. [CrossRef]
21. Steegmans, M.L.J.; Schroën, K.G.P.H.; Boom, R.M. Microfluidic Y-junctions: A robust emulsification system with regard to
junction design. AIChE J. 2010, 56, 1946–1949. [CrossRef]
22. Adeosun, J.T.; Lawal, A. Numerical and experimental studies of mixing characteristics in a T-junction microchannel using
residence-time distribution. Chem. Eng. Sci. 2009, 64, 2422–2432. [CrossRef]
23. Naik-Nimbalkar, V.; Patwardhan, A.; Banerjee, I.; Padmakumar, G.; Vaidyanathan, G. Thermal mixing in T-junctions. Chem. Eng.
Sci. 2010, 65, 5901–5911. [CrossRef]
24. Wang, L.-Y.; Wu, Y.-X.; Zheng, Z.-C.; Guo, J.; Zhang, J.; Tang, C. Oil-Water two-Phase Flow Inside T-Junction. J. Hydrodyn. 2008,
20, 147–153. [CrossRef]
25. Sökmen, C.N. Effect of property variations on the mixing of laminar supercritical water streams in a T-junction. Int. Commun.
Heat Mass Transf. 2011, 38, 85–92. [CrossRef]
26. Dehbi, A.; de Crécy, F. Validation of the Langevin particle dispersion model against experiments on turbulent mixing in a
T-junction. Powder Technol. 2011, 206, 312–321. [CrossRef]
27. Sierens, R.; Verhelst, S. Influence of the Injection Parameters on the Efficiency and Power Output of a Hydrogen Fueled Engine. J.
Eng. Gas Turbines Power 2003, 125, 444–449. [CrossRef]
28. Burlutskii, E. CFD study of oil-in-water two-phase flow in horizontal and vertical pipes. J. Pet. Sci. Eng. 2018, 162, 524–531.
[CrossRef]
29. Jones, W.; Launder, B. The prediction of laminarization with a two-equation model of turbulence. Int. J. Heat Mass Transf. 1972, 15,
301–314. [CrossRef]
30. Menter, F.R. Two-equation eddy-viscosity turbulence models for engineering applications. AIAA J. 1994, 32, 1598–1605. [CrossRef]
31. Frank, T.; Lifante, C.; Prasser, H.-M.; Menter, F. Simulation of turbulent and thermal mixing in T-junctions using URANS and
scale-resolving turbulence models in ANSYS CFX. Nucl. Eng. Des. 2010, 240, 2313–2328. [CrossRef]
32. Noguera, J.F.; Torres, L.; Verde, C.; Guzman, E.; Sanjuan, M. Model for the flow of a water-glycerol mixture in horizontal pipelines.
In Proceedings of the 2019 4th Conference on Control and Fault Tolerant Systems (SysTol), Casablanca, Morocco, 18–20 September
2019; pp. 117–122. [CrossRef]
33. Noguera-Polania, J.F.; Hernández-García, J.; Galaviz-López, D.F.; Torres, L.; Guzmán, J.; Sanjuán-Mejía, M.E.; Jiménez-Cabas, J.
Dataset on water–glycerol flow in a horizontal pipeline with and without leaks. Data Brief 2020, 31, 105950. [CrossRef]
34. Torres, L.; Noguera, J.; Guzmán-Vázquez, J.E.; Hernández, J.; Sanjuan, M.; Palacio-Pérez, A. Pressure Signal Analysis for the
Characterization of High-Viscosity Two-Phase Flows in Horizontal Pipes. J. Mar. Sci. Eng. 2020, 8, 1000. [CrossRef]
35. ANSYS Fluent. ANSYS Fluent User’s Guide; ANSYS, Inc.: Canonsburg, PA, USA, 2013.
Water 2022, 14, 3451 16 of 16

36. Peters, F.; Arabali, D. Interfacial tension between oil and water measured with a modified contour method. Colloids Surfaces A:
Physicochem. Eng. Asp. 2013, 426, 1–5. [CrossRef]
37. Hayashi, Y.; Puzenko, A.; Feldman, Y. Slow and fast dynamics in glycerol–water mixtures. J. Non-Crystalline Solids 2006, 352,
4696–4703. [CrossRef]
38. Egorov, A.V.; Lyubartsev, A.P.; Laaksonen, A. Molecular Dynamics Simulation Study of Glycerol–Water Liquid Mixtures. J. Phys.
Chem. B 2011, 115, 14572–14581. [CrossRef]
39. Dashnau, J.L.; Nucci, N.V.; Sharp, K.A.; Vanderkooi, J.M. Hydrogen Bonding and the Cryoprotective Properties of Glycerol/Water
Mixtures. J. Phys. Chem. B 2006, 110, 13670–13677. [CrossRef]
40. Seveno, E. Towards an adaptive advancing front method. In 6th International Meshing Roundtable; 1997; pp. 349–362. Available
online: https://scholar.google.com.hk/scholar?hl=en&as_sdt=0%2C5&q=Seveno%2C+E.+Towards+an+adaptive+advancing+
front+method.+In+6th+International+Meshing+Roundtable%3B+Publisher%3A+Location%2C+1997%2C+p.+349%E2%80%93
62.&btnG= (accessed on 25 October 2022).
41. Ingram, D.; Causon, D.; Mingham, C. Developments in Cartesian cut cell methods. Math. Comput. Simul. 2003, 61, 561–572.
[CrossRef]
42. Patankar, S.V. Numerical Heat Transfer and Fluid Flow; Hemisphere Publishing Corporation: New York, NY, USA, 1980. [CrossRef]
43. van Leer, B. Towards the ultimate conservative difference scheme. V. A second-order sequel to Godunov’s method. J. Comput.
Phys. 1979, 32, 101–136. [CrossRef]
44. Muzaferija, S.; Peric, M.; Sames, P.; Schellin, T. A two-fluid Navier-Stokes solver to simulate water entry. In Proceedings of 22nd
Symposium on Naval Architecture, 1999; The National Academy Press: Washington, DC, USA, 1999; pp. 638–651.
45. Waclawczyk, T.; Koronowicz, T. Comparison of cicsam and hric high-resolution schemes for interface capturing. J. Theor. Appl.
Mech. 2008, 46, 325–345.
46. Brackbill, J.; Kothe, D.; Zemach, C. A continuum method for modeling surface tension. J. Comput. Phys. 1992, 100, 335–354.
[CrossRef]
47. Alagbe, S.O. Experimental and Numerical Investigation of High Viscosity Oil-Based Multiphase Flows. Ph.D. Thesis, Cranfield
University, Cranfiled, UK, 2013.
48. Nädler, M.; Mewes, D. Flow induced emulsification in the flow of two immiscible liquids in horizontal pipes. Int. J. Multiph. Flow
1997, 23, 55–68. [CrossRef]
49. Al-Hadhrami, L.M.; Shaahid, S.M.; Tunde, L.O.; Al-Sarkhi, A. Experimental Study on the Flow Regimes and Pressure Gradients
of Air-Oil-Water Three-Phase Flow in Horizontal Pipes. Sci. World J. 2014, 2014, 1–11. [CrossRef]
50. Hasson, D.; Mann, U.; Nir, A. Annular flow of two immiscible liquids. Can. J. Chem. Eng. 1970, 48, 514–520. [CrossRef]
51. Vielma, M.; Atmaca, S.; Sarica, C.; Zhang, H.Q.; Al-Sarkhi, A.S. Characterization of oil water flows in inclined pipes. In
Proceedings of the SPE Annual Technical Conference and Exhibition, Denver, CO, USA, 21–24 September 2008; pp. 1273–1286.
[CrossRef]