Interface-Capturing Methods For Two-Phase Ows: An Overview and Recent Developments
Interface-Capturing Methods For Two-Phase Ows: An Overview and Recent Developments
Interface-Capturing Methods For Two-Phase Ows: An Overview and Recent Developments
net/publication/322328389
CITATIONS READS
6 705
3 authors, including:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Suhas S Jain on 09 January 2018.
Two-phase flows
models
LBM SPH Two-fluid One-fluid
approaches
Interface-tracking Interface-capturing
classes
MAC Front-tracking VOF Level set Phase field CIP
Popinet (2018) observed that all well-known volumetric formulations, including the orig-
inal continuous surface force (CSF, Brackbill et al. 1992), ghost-fluid method (GFM,
Fedkiw et al. 1999) and smoothed Heaviside method (Sussman et al. 1994) can be writ-
ten in the form FST = σκδs n, which is commonly known as the CSF formulation,
where σ is the surface tension and δs is a numerical Dirac delta function specific to each
method. Therefore, a critical component of different surface tension calculation meth-
ods is the accuracy of calculating the normal vector (n) and curvature (κ). A common
approach has been to compute normal vectors from spatially differentiating a smooth
field (sometimes by a smoothening kernel), and subsequently taking the divergence of
the normal vectors to obtain curvatures. In some sharp interface approaches (e.g., VOF),
a smooth field is not readily available. Therefore, in recent years many new codes using
sharp-interface approaches have employed height-functions for more accurate estimation
of normals and/or curvatures (Sussman & Ohta 2009; Popinet 2009; Owkes & Desjardins
2015; Ivey & Moin 2015). Although height-functions can be very attractive in terms of
accuracy, convergence and momentum conservation properties, their accuracy and ro-
bustness suffers at low resolutions, specifically when κ∆x > 1/5 (Popinet 2018). Thus,
robust implementation of these methods is not straightforward and entails a smooth
transition to alternative techniques such as parabolic reconstruction of surface tension
(PROST), introduced in Renardy & Renardy (2002). The PROST method has been ex-
tended to unstructured meshes by Evrard et al. (2017). Their method can be interpreted
as a generalization of the height-function method, and it converges with the same order
of accuracy as height-function techniques. However, to date, it has only been applied
120 Mirjalili, Jain & Dodd
to two-dimensional unstructured meshes due to additional challenges and complexity
presented in three dimensions.
A discussion on numerical implementation of surface tension is incomplete without the
mention of spurious currents. These artificial flows can be detrimental in practical two-
phase flows, leading to artificial generation of kinetic energy and heat transfer (Hardt
& Wondra 2008; Gupta et al. 2009). Accurate direct numerical simulation (DNS) of
turbulent two-phase flows is only possible if the root-mean-squared (r.m.s.) velocity of
the spurious currents is negligible relative to the velocity fluctuations due to turbulence,
as demonstrated recently by Dodd & Ferrante (2016). Accurate curvature estimation
and well-balanced surface tension force discretization are required for reducing spurious
currents. A test case assessing the relative magnitude of these currents is therefore an
essential benchmark for the evaluation of surface tension implementation. Often times,
the magnitude of spurious currents in a static drop test case at a specific time is re-
ported (originally due to Gunstensen 1992). However, improvements to this measure are
necessary as shown by Magnini et al. (2016), for instance, by reporting the history or
time-averaged norm of spurious currents. In addition, these authors suggest examining
spurious currents for a moving drop.
Finally, much progress has been made in recent years towards the goal of discrete
conservation of energy for two-phase flows, particularly at high density ratios. For Carte-
sian staggered grids, a method that discretely conserves kinetic energy in the absence
of viscous dissipation and surface tension forces was presented in Fuster (2013). In gen-
eral, it is now well-established that for accurate and robust simulations at large density
ratios, momentum should be transported conservatively via Eq. (2.2) and also consis-
tently with respect to mass advection (Popinet 2009; Raessi & Pitsch 2012; Chenadec &
Pitsch 2013; Ivey et al. 2016). This correction has been shown to significantly improve
the energy conservation properties of two-phase flow solvers.
4. Volume-of-fluid method
In the class of volume-of-fluid (VOF) methods, the phase indicator function, H(x, t),
is defined as
(
1 if x is in the reference fluid,
H(x, t) = (4.1)
0 if x is in the other fluid.
In the absence of phase change, each fluid parcel retains its phase indicator value during
its motion. Therefore, the material derivative of H is zero, i.e.
DH ∂H
= + ∇ · (uH) − H∇ · u = 0. (4.2)
Dt ∂t
The volume fraction is defined as the spatial average of the phase indicator function (H)
in each computational cell (Ω),
Z
1
Ck (t) = H(x, t) dV, (4.3)
V Ω
where V is the volume of the k-th cell. We then integrate Eq. (4.2) over the computational
cell and use the definition introduced in Eq. (4.3) to obtain
Z Z
∂Ck (t)
V + (u · n)H(x, t) dS = H∇ · u dV. (4.4)
∂t ∂Ω Ω
A review of interface-capturing methods for two-phase flows 121
While in incompressible flows ∇ · u = 0, this term is retained in Eq. (4.4) for reasons
related to numerical implementation that will be explained later. In geometric VOF
methods an approximation to H is found geometrically (e.g., by a plane), whereas in
algebraic VOF methods H is represented by a function (e.g., polynomial or trigonometric
function).
4.1. Geometric VOF
Geometric VOF methods proceed in two steps. First, the interface is reconstructed in
each computational cell from the knowledge of the volume fraction field. Second, the
reconstructed interface is advected by computing the fluxed volume across each compu-
tational cell using geometric methods. In the following, we review the state of the art for
performing these two tasks.
4.1.1. Interface reconstruction
For the first task, the most widely used method is the piecewise linear interface calcu-
lation (PLIC) scheme (Debar 1974). In this approach, the interface is approximated in
each interfacial cell as a line in two dimensions or a plane in three dimensions as
n · x + α = 0, (4.5)
where α is the constant in the plane equation that is selected to enforce that the volume
cut by the interface is equal to Ck in the computational cell. The main choice left to
the user is how to compute the interface normal, n. Youngs’ method (Youngs 1982)
estimates n as a normalized gradient of Ck . This method is fast and performs best
at low resolutions, but at higher resolutions its accuracy decreases from second to first
order. A method that performs better at high resolutions is the centered-columns method
(Scardovelli & Zaleski 2003). An approach that combines the desirable properties at both
low and high resolutions is the mixed Youngs-centered (MYC) method (Aulisa et al.
2007). Pilliod & Puckett (2004) introduced the efficient least-squares volume-of-fluid
interface reconstruction algorithm (ELVIRA) method. On structured Cartesian meshes,
this approach uses a least-squares error minimization to select between normal vector
candidates. Its extension to three dimensional unstructured meshes is more complicated,
but has been accomplished by Jofre et al. (2014). ELVIRA is second-order accurate but
has high computational cost, especially in three dimensions due to an increased number
of normal candidates. For unstructured meshes and general polyhedral cells, Ivey &
Moin (2015) developed an embedded height-function framework to compute second-order
accurate normals. On Cartesian meshes, the MYC method appears to be the preferred
method because of its low cost and relatively high degree of accuracy on structured
meshes, and as such, it has been implemented in several free softwares (e.g., Popinet 2012,
2013; Arrufat et al. 2014). On unstructured meshes, the favored method for computing
interface normals is less clear.
After the normal n has been calculated, the next task is to compute α. For right
hexahedral cells, Scardovelli & Zaleski (2000) presented analytical methods for computing
α. For general polyhedra, a root-finding algorithm is typically employed to find α, for
which VOF-specific libraries exist, e.g., the VOFTools library by López et al. (2017).
4.1.2. Interface advection
The interface is advected by integrating Eq. (4.4) in time. The main decision left to the
user is on how the volume fluxes of the reference fluid are calculated. Two classes of advec-
tion algorithms have emerged: (i) split methods that rely on operator-splitting to perform
122 Mirjalili, Jain & Dodd
a series of one dimensional advections in each spatial dimension and (ii) unsplit methods,
which advect the interface in one step. Split methods are algorithmically straightforward
to implement compared to multidimensional (unsplit) schemes. Unsplit schemes involve
computation of complex flux polyhedra in computational cells containing the interface.
These computations are expensive and can lead to significant parallel load imbalance if
left untreated (Jofre et al. 2015). On the other hand, unsplit methods have the advan-
tage of only requiring one advection and reconstruction step and are more amenable for
general three-dimensional unstructured meshes. In addition, Pilliod & Puckett (2004)
showed that unsplit methods better resolve interfaces that are non-smooth (i.e. contain
corners). How this increased resolution capability translates into practical applications
(e.g., Prakash et al. 2016; Mortazavi et al. 2016) is an open question.
As previously mentioned, split advection schemes use operator-splitting to decompose
Eq. (4.4) into a series of one dimensional advections. Because terms like ∂u/∂x are
generally non-zero, the right-hand side in Eq. (4.4) is retained. This is in contrast to
unsplit schemes which omit the last term due incompressibility (∇·u). In two dimensions,
the first split advection scheme to satisfy the boundedness condition, 0 ≤ Ck ≤ 1, and
mass conservation in incompressible flow is known as the Eulerian implicit - Lagrangian
explicit (EI-LE) scheme (Tryggvason et al. 2011). Aulisa et al. (2007) extended the EI-LE
scheme to three dimensions by creating EILE-3D. EILE-3D uses a series of three EI-LE
advection steps, one in each spatial direction. This discretely mass-conserving scheme
satisfies the boundedness condition, but has the disadvantage of requiring six advection
and interface reconstruction steps as well as the requirement to compute three two-
dimensional divergence-free velocity fields. Later, Weymouth & Yue (2010) developed
a split advection scheme that consists of three Lagrangian explicit (LE) steps that are
both bounded (0 ≤ Ck ≤ 1) and discretely conservative. The only drawback is that it
introduces an additional CFL restriction to guarantee boundedness of Ck . However, this
restriction is relatively benign. The Weymouth & Yue method is implemented in the free
software Basilisk (Popinet 2013) and PARIS Simulator (Arrufat et al. 2014). Another
method to extend the EI-LE advection scheme to three dimensions was given by Baraldi
et al. (2014). This approach introduces an algebraic step in the third dimension that
ensures mass is discretely conserved. The downside to this is that the algebraic step can
produce overshoots (Ck > 1) or undershoots (Ck < 0), and therefore a redistribution
algorithm is required to ensure boundedness. Further tests of the split advection schemes
which include a cost comparison would be beneficial.
Unsplit VOF advection methods compute and transport the fluxed volume in one step.
Rider & Kothe (1998) proposed one of the first unsplit VOF advection schemes in the
framework of PLIC interface reconstruction. Face centered velocity components were
used to construct trapezoidal flux regions. These flux regions can potentially overlap,
which can ultimately lead to the condition 0 ≤ Ck ≤ 1 not being satisfied exactly.
López et al. (2004) developed a two-dimensional unsplit geometric advection scheme
which used cell vertex velocities instead of edge velocities. This change resulted in no
overlap of flux regions which improved the mass conservation compared to Rider & Kothe
(1998). Owkes & Desjardins (2014) extended the work of López et al. (2004), yielding
a bounded, mass-conserving three-dimensional unsplit VOF advection algorithm. Their
scheme addressed outstanding conservation issues in three-dimensional unsplit methods,
namely how to address flux hexahedra that overlap during the advection step. Jofre
et al. (2014) extended the work of Owkes & Desjardins (2014) to unstructured meshes
and non-convex polyhedra. Ivey & Moin (2017) further advanced the computation of
A review of interface-capturing methods for two-phase flows 123
flux polyhedra on unstructured meshes to yield a discretely conservative and bounded
unsplit advection scheme. It should be noted that all the state of the art unsplit VOF
advection methods use cell vertex velocities to form the flux polyhedra as opposed to edge
velocities. A possible departure from this framework is given by Roenby et al. (2016).
Although the performance and accuracy of this method has not yet been compared to
conventional geometric VOF methods, it offers a potentially more cost-effective approach
to VOF advection on unstructured meshes because it does not require PLIC interface
reconstruction and complex geometric computations of flux volumes.
It is important to note that in VOF advection schemes, discrete mass conservation
is only guaranteed if ∇ · u = 0 is satisfied discretely in every grid cell. In practice,
the condition ∇ · u = 0 is typically only satisfied up to some user-specified tolerance
through the iterative solution of the Poisson equation for pressure. To obtain discrete
conservation of mass to machine precision in every grid cell cost effectively, the Poisson
equation should be solved directly instead of iteratively (Dodd & Ferrante 2014).
5. Level-set methods
Level set methods were first developed and used in the context of computer graphics
and image processing (Sethian 1999). They were extended to the case of two-phase flows
by Sussman et al. (1994). A traditional level set function is a signed-distance function
φ(x, t), that represents the shortest distance to the interface. These methods solve the
advection equation
∂φ
+ u · n|∇φ| = 0, (5.1)
∂t
to update φ at every time step and are commonly discretized with a Hamilton-Jacobi
weighted essentially non-oscillating (HJ-WENO) type scheme in space and a total vari-
ation diminishing Runge-Kutta (TVD-RK) scheme in time. Sadly, φ loses its signed-
distance property after the advection step, and hence needs to be reinitialized.
Level set methods offer many advantages over other methods such as: (1) accurate
computation of normals and curvature using n = ∇φ/|∇φ| and κ = ∇ · n, due to
smoothness of the φ field and (2) straightforward extensions to Cartesian adaptive mesh
refinement (AMR). However, the biggest disadvantage of these methods is the fact that
mass of each phase is not conserved, a crucial requirement for numerical modeling of
realistic two-phase flows. While there are multiple fixes to limit the mass loss as described
in Section 5.2, this problem cannot be eliminated completely.
5.1. Reinitialization
The conventional approach to reinitialize φ is to compute the interface location (φ = 0)
and to recompute the distance to the interface at all other points. However, this method
was found to be too expensive with a complexity of O(N 3 ) and was found to distort the
interface. Hence, Sussman et al. (1994) proposed a PDE based reinitialization method
φτ = sgn(φ0 )(|∇φ| − 1) = 0, (5.2)
A review of interface-capturing methods for two-phase flows 125
where sgn is a smoothed signum function. Reinitialization of φ can also be performed by
solving the Eikonal equation of the form
||∇φ|| = 1. (5.3)
Fast numerical methods such as the fast-marching method (FMM) (Chopp 2001) and the
fast-sweeping method (FSM) (Zhao 2005) have been developed to solve this equation.
FMM uses a heap structure to store and order the grid points where the Eikonal equation
is solved and has a complexity of O(N log N ), whereas the FSM uses Gauss-Seidel iter-
ations with alternating sweeping orderings and hence the number of sweeps is bounded:
2n for Rn with a complexity of O(N ). These methods seem attractive due to their low
cost and reduced mass-loss error with the use of higher-order extensions; however, there
is an inherent difficulty in scalable parallel implementation of these methods due to load-
balancing issues and the requirement of high number of iterations in the regions where
the characteristics cross the sub-domain multiple times.
7. Hybrid methods
There are various hybrid methods available in the literature, such as coupled-level-
set and volume-of-fluid method (CLSVOF) (Sussman & Puckett 2000), front-tracking
coupled with VOF (Aulisa et al. 2007), and phase field coupled with VOF (Liu & Yu
2016). Of all these methods, CLSVOF has been quite successful since it takes advan-
tage of level-sets in accurate computation of normals and curvature, and also improves
A review of interface-capturing methods for two-phase flows 129
mass conservation 33 33 3 33 33 33
momentum and kinetic-energy conservation 3 7 7 3 3 3
ease of programming 3 3 3 7 3 33
normals and curvature 3 3 33 33 3 3
cost and scalability 3 33 3 7 33 33
unstructured Meshes 33 33 3 3 3 3
overall accuracy 33 3 3 33 3 3
overall potential 33 3 7 3 3 33
mass conservation by utilizing the VOF framework. There have been a few variations of
CLSVOF that achieve discrete mass conservation (e.g. Tomar et al. 2005). Overall, the
accuracy of this method is typically higher than a level-set or a VOF method, although
the cost is also higher. This method also has issues in achieving parallel scalability due to
the inherent difficulty in achieving balanced loads in the presence of operations required
for both VOF and level-set methods.
8. Concluding remarks
In this report, we have provided an overview of the most popular interface-capturing
methods, with emphasis on modern developments. In the spirit of comparison, we have
summarized the status of these methods in Table 1. The reader should be able to extract
the rationale behind the ratings in this table from discussions presented in Sections
4-7. We found that the biggest challenge remaining in the field is in developing mass,
momentum and total energy conserving schemes with reasonable cost and scalability (see
Table 1). Indeed, these are required for realistic simulations of two-phase flows, especially
in turbulent conditions. Moreover, by considering other relevant factors, particularly
accuracy and modular adaptability to problems with practical significance (including
compressible flows, unstructured meshes, scalar transport, and surfactant transport),
we conclude in Table 1 that VOF and phase-field methods can be considered the most
promising classes for future investment.
On a separate note, while reviewing the literature for this report, we observed a conspic-
uous scarcity of benchmarks and test-cases that could help to evaluate the performance
of new schemes in problems and conditions pertinent to practical applications. Firstly,
we recommend the design of several test cases at realistic non-dimensional numbers in
which the accuracy and robustness of the fully coupled solver (interface capturing scheme
and momentum equation implementation) can be assessed. Secondly, when performing
DNS or LES of two-phase turbulent flows, it is inevitable that some features (e.g., lig-
130 Mirjalili, Jain & Dodd
aments, films, and droplet fragments) are left under-resolved. It is important that the
under-resolved features do not cause the numerical simulation to diverge or introduce
artificial effects. Yet in papers introducing numerical methods we commonly see a focus
on convergence rates and accuracy at rather high resolutions. Through assessment and
presentation of the performance of solvers at low resolutions, this deficiency should also
be addressed in future work. Hopefully, these improved benchmarks will motivate the
emergence of methods that are more appropriate for tackling difficult, realistic problems
in two-phase flows.
As a final remark, we believe that additional work on multi-physics aspects of two-phase
flows would be beneficial. A number of engineering applications require consideration of
complex phenomena beyond the purely capillary effects reviewed here. For instance, the
utilization of electrostatic effects in spray atomizers represents an interesting method to
control the atomization of liquids (Gomez & Tang 1994; Tang & Gomez 1994; de la Mora
2007). Similarly, the interaction of acoustic waves with droplets, bubbles and interfaces in
general is a problem of relevance for the acoustic traceability of ships in ocean (Trevorrow
et al. 1994), for biomedical applications (Adami et al. 2016), and for fuel atomization
in jet engines (Gajan et al. 2016) and hydrocarbon-fueled scramjets (Urzay 2017). The
treatment of these advanced problems is in its infancy and requires creative extensions
of the existing frameworks reviewed in this report.
Acknowledgments
This investigation was funded by ONR, Grant #N00014-15-1-2726.
REFERENCES
Adami, S., Kaiser, J., Adams, N. & Bermejo-Moreno, I. 2016 Numerical model-
ing of shock waves in biomedicine. Proceedings of the Summer Program, Center for
Turbulence Research, Stanford University, pp. 15–24.
Anderson, D., McFadden, G. & Wheeler, A. 1998 Diffuse-interface methods in
fluid mechanics. Annu. Rev. Fluid Mech. 30, 139–165.
Arrufat, T. J., Dabiri, S., Fuster, D., Ling, Y., Malan, L., Scardovelli, R.,
Tryggvason, G., Yecko, P. & Zaleski, S. 2014 The PARIS-Simulator code.
http://www.ida.upmc.fr/∼zaleski/paris/index.html.
Aulisa, E., Manservisi, S., Scardovelli, R. & Zaleski, S. 2007 Interface recon-
struction with least-squares fit and split advection in three-dimensional cartesian
geometry. J. Comput. Phys. 225, 2301–2319.
Badalassi, V., Ceniceros, H. & Banerjee, S. 2003 Computation of multiphase
systems with phase field models. J. Comput. Phys. 190, 371–397.
Balczar, N., Jofre, L., Lehmkuhl, O., Castro, J. & Rigola, J. 2014 A finite-
volume/level-set method for simulating two-phase flows on unstructured grids. Int.
J. Multiphase Flow 64, 55–72.
Baraldi, A., Dodd, M. & Ferrante, A. 2014 A mass-conserving volume-of-fluid
method: volume tracking and droplet surface-tension in incompressible isotropic tur-
bulence. Comput. Fluids 96, 322–337.
Brackbill, J., Kothe, D. & Zemach, C. 1992 A continuum method for modeling
surface tension. J. Comput. Phys. 100, 335–354.
Chenadec, V. L. & Pitsch, H. 2013 A monotonicity preserving conservative sharp
A review of interface-capturing methods for two-phase flows 131
interface flow solver for high density ratio two-phase flows. J. Comput. Phys. 249,
185–203.
du Chéné, A., Min, C. & Gibou, F. 2008 Second-order accurate computation of
curvatures in a level set framework using novel high-order reinitialization schemes.
J. Sci. Comput. 35, 114–131.
Chiodi, R. & Desjardins, O. 2017 A reformulation of the conservative level set reini-
tialization equation for accurate and robust simulation of complex multiphase flows.
J. Comput. Phys. 343, 186–200.
Chiu, P. H. & Lin, Y. T. 2011 A conservative phase field method for solving incom-
pressible two-phase flows. J. Comput. Phys. pp. 185–204.
Chopp, D. L. 2001 Some improvements of the fast marching method. SIAM J. Sci.
Comput. 23, 230–244.
Clift, R., Grace, J. R. & Weber, M. E. 2005 Bubbles, Drops, and Particles. Courier
Corporation.
Darwish, M. & Moukalled, F. 2006 Convective schemes for capturing interfaces of
free-surface flows on unstructured grids. Numer. Heat Tr. B-Fund. 49, 19–42.
Debar, R. 1974 Fundamentals of the KRAKEN code. LLNL Tech. Rep. #UCIR-760.
Desjardins, O., Moureau, V. & Pitsch, H. 2008 An accurate conservative level
set/ghost fluid method for simulating turbulent atomization. J. Comput. Phys. 227,
8395–8416.
Ding, H., Spelt, P. D. & Shu, C. 2007 Diffuse interface model for incompressible
two-phase flows with large density ratios. J. Comput. Phys. 226, 2078–2095.
Dodd, M. S. & Ferrante, A. 2014 A fast pressure-correction method for incompress-
ible two-fluid flows. J. Comput. Phys. 273, 416–434.
Dodd, M. S. & Ferrante, A. 2016 On the interaction of Taylor length scale size
droplets and isotropic turbulence. J. Fluid Mech. 806, 356–412.
Dong, S. & Shen, J. 2012 A time-stepping scheme involving constant coefficient matri-
ces for phase-field simulations of two-phase incompressible flows with large density
ratios. J. Comput. Phys. 231, 5788–5804.
Enright, D., Fedkiw, R., Ferziger, J. & Mitchell, I. 2002 A hybrid particle level
set method for improved interface capturing. J. Comput. Phys. 183, 83–116.
Evrard, F., Denner, F. & van Wachem, B. 2017 Estimation of curvature from
volume fractions using parabolic reconstruction on two-dimensional unstructured
meshes. J. Comput. Phys. 351, 271–294.
Fedkiw, R. P., Aslam, T., Merriman, B. & Osher, S. 1999 A non-oscillatory
Eulerian approach to interfaces in multimaterial flows (the ghost fluid method). J.
Comput. Phys. 152, 457–492.
Fernández. de la Mora, J. 2007 The fluid dynamics of Taylor cones. Annu. Rev.
Fluid Mech. 39, 217–243.
François, M. M., Cummins, S. J., Dendy, E. D., Kothe, D. B., Sicilian, J. M.
& Williams, M. W. 2006 A balanced-force algorithm for continuous and sharp
interfacial surface tension models within a volume tracking framework. J. Comput.
Phys. 213, 141–173.
Fuster, D. 2013 An energy preserving formulation for the simulation of multiphase
turbulent flows. J. Comput. Phys. 235, 114–128.
Gajan, P., Simon, F., Orain, M. & Bodoc, V. 2016 Investigation and modeling of
combustion instabilities in aero engines. AerospaceLab AL11–09.
132 Mirjalili, Jain & Dodd
Gibou, F., Fedkiw, R. & Osher, S. 2018 A review of level-set methods and some
recent applications. J. Comput. Phys. 353, 82–109.
Gomez, A. & Tang, K. 1994 Charge and fission of droplets in electrostatic sprays.
Phys. Fluids 6, 404–414.
Gunstensen, A. K. 1992 Lattice-Boltzmann Studies of Multiphase Flow Through
Porous Media. Ph.D. Thesis, MIT.
Gupta, R., Fletcher, D. F. & Haynes, B. S. 2009 On the CFD modelling of Taylor
flow in microchannels. Chem. Eng. Sci. 64, 2941–2950.
Hardt, S. & Wondra, F. 2008 Evaporation model for interfacial flows based on a
continuum-field representation of the source terms. J. Comput. Phys. 227, 5871–
5895.
Herrmann, M. 2008 A balanced force refined level set grid method for two-phase flows
on unstructured flow solver grids. J. Comput. Phys. 227, 2674–2706.
Heyns, J. A., Malan, A., Harms, T. & Oxtoby, O. F. 2013 Development of a
compressive surface capturing formulation for modelling free-surface flow by using
the volume-of-fluid approach. Int. J. Numer. Methods Fluids 71, 788–804.
Hirt, C. W. & Nichols, B. D. 1981 Volume of fluid (VOF) method for the dynamics
of free boundaries. J. Comput. Phys. 39, 201–225.
Ii, S., Xie, B. & Xiao, F. 2014 An interface capturing method with a continuous
function: The thinc method on unstructured triangular and tetrahedral meshes. J.
Comput. Phys. 259, 260–269.
Ishii, M. & Hibiki, T. 2010 Thermo-fluid dynamics of two-phase flow . Springer Science
& Business Media.
Ivey, C. & Moin, P. 2015 Accurate interface normal and curvature estimates on
three-dimensional unstructured non-convex polyhedral meshes. J. Comput. Phys.
300, 365–386.
Ivey, C., Moin, P., Bose, S. & Kim, D. 2016 Geometric volume-of-fluid framework
for the simulation of two-phase flows on unstructured grids. In Proceedings of the
31st Symposium on Naval Hydrodynamics.
Ivey, C. B. & Moin, P. 2017 Conservative and bounded volume-of-fluid advection on
unstructured grids. J. Comput. Phys. 350, 387–419.
Jacqmin, D. 1999 Calculation of two-phase Navier-Stokes flows using phase-field mod-
eling. J. Comput. Phys. 155, 96–127.
Jofre, L., Borrell, R., Lehmkuhl, O. & Oliva, A. 2015 Parallel load balanc-
ing strategy for Volume-of-Fluid methods on 3-D unstructured meshes. J. Comput.
Phys. 282, 269–288.
Jofre, L., Lehmkuhl, O., Castro, J. & Oliva, A. 2014 A 3-D Volume-of-Fluid
advection method based on cell-vertex velocities for unstructured meshes. Comput.
Fluids 94, 14–29.
Kim, D. 2011 Direct numerical simulation of two-phase flow with application to air layer
drag reduction. Ph.D. Thesis, Stanford University.
Kim, J., Lee, S. & Choi, Y. 2014 A conservative Allen-Cahn equation with a space-
time dependent Lagrange multiplier. Int. J. Eng. Sci. 84, 11–17.
Li, Y., Choi, J.-I. & Kim, J. 2016 A phase-field fluid modeling and computation with
interfacial profile correction term. Commun. Nonlinear. Sci. Numer. Simul. 30, 84–
100.
A review of interface-capturing methods for two-phase flows 133
Liu, Y. & Yu, X. 2016 A coupled phase–field and volume-of-fluid method for accurate
representation of limiting water wave deformation. J. Comput. Phys. 321, 459–475.
López, J., Hernández, J., Gómez, P. & Faura, F. 2004 A volume of fluid method
based on multidimensional advection and spline interface reconstruction. J. Comput.
Phys. 195, 718–742.
López, J., Hernández, J., Gómez, P. & Faura, F. 2017 VOFTools–A software
package of calculation tools for volume of fluid methods using general convex grids.
Comput. Phys. Commun. 223, 45–54.
Magnini, M., Pulvirenti, B. & Thome, J. 2016 Characterization of the velocity
fields generated by flow initialization in the CFD simulation of multiphase flows.
Appl. Math. Modell. 40, 6811–6830.
McCaslin, J. O. & Desjardins, O. 2014 A localized re-initialization equation for the
conservative level set method. J. Comput. Phys. 262, 408–426.
McKee, S., Tomé, M. F., Ferreira, V. G., Cuminato, J. A., Castelo, A., Sousa,
F. & Mangiavacchi, N. 2008 The MAC method. Comput. Fluids 37, 907–930.
Melville, W. K. 1996 The role of surface-wave breaking in air-sea interaction. Annu.
Rev. Fluid Mech. 28, 279–321.
Mirjalili, S., Ivey, C. B. & Mani, A. 2016 Cost and accuracy comparison between
the diffuse interface method and the geometric volume of fluid method for simulating
two-phase flows. Bull. of the Am. Physical Soc. 61, BAPS.2016.DFD.R29.8.
Mirjalili, S., Ivey, C. B. & Mani, A. 2017 Boundedness proof for a conservative
phase-field method discretized by central finite differences. Bull. of the Am. Physical
Soc. 62, BAPS.2017.DFD.Q7.12.
Mortazavi, M., Le Chenadec, V., Moin, P. & Mani, A. 2016 Direct numerical
simulation of a turbulent hydraulic jump: turbulence statistics and air entrainment.
J. Fluid Mech. 797, 60–94.
Muzaferija, S., Peric, M., Sames, P. & Schelin, T. 1998 A two-fluid Navier-
Stokes solver to simulate water entry. In Proceedings of the 22nd Symposium on
Naval Hydrodynamics.
Olsson, E. & Kreiss, G. 2005 A conservative level set method for two phase flow. J.
Comput. Phys. 210, 225–246.
Olsson, E., Kreiss, G. & Zahedi, S. 2007 A conservative level set method for two
phase flow II. J. Comput. Phys. 225, 785–807.
Owkes, M. & Desjardins, O. 2014 A computational framework for conservative, three-
dimensional, unsplit, geometric transport with application to the volume-of-fluid
(VOF) method. J. Comput. Phys. 270, 587–612.
Owkes, M. & Desjardins, O. 2015 A mesh-decoupled height function method for
computing interface curvature. J. Comput. Phys. 281, 285–300.
Pilliod, J. E. & Puckett, E. G. 2004 Second-order accurate volume-of-fluid algo-
rithms for tracking material interfaces. J. Comput. Phys. 199, 465–502.
Plateau, J. 1873 Experimental and theoretical statics of liquids subject to molecular
forces only. Gauthier-Villars, Trübner and Co.
Popinet, S. 2009 An accurate adaptive solver for surface-tension-driven interfacial flows.
J. Comput. Phys. 228, 5838–5866.
Popinet, S. 2012 The Gerris flow solver. http://gfs.sourceforge.net/.
Popinet, S. 2013 The Basilisk code. http://basilisk.fr/.
134 Mirjalili, Jain & Dodd
Popinet, S. 2018 Numerical models of surface tension. Annu. Rev. Fluid Mech. 50,
49–75.
Prakash, S. R., Jain, S. S., Tomar, G., Ravikrishna, R. V. & Raghunandan,
B. N. 2016 Computational study of liquid jet breakup in swirling cross flow. In Proc.
Annu. Conf. Inst. Liq. Atom. Spray Syst. As., 18th, ILASS As.. Chennai, India.
Prosperetti, A. & Tryggvason, G. 2009 Computational methods for multiphase
flow . Cambridge University Press.
Raessi, M. & Pitsch, H. 2012 Consistent mass and momentum transport for simulating
incompressible flows with large density ratios using the level set method. Comput.
Fluids 63, 70–81.
Renardy, Y. & Renardy, M. 2002 Prost: A parabolic reconstruction of surface tension
for the volume-of-fluid method. J. Comput. Phys. 183, 400–421.
Rider, W. J. & Kothe, D. B. 1998 Reconstructing volume tracking. J. Comput. Phys.
141, 112–152.
Roenby, J., Bredmose, H. & Jasak, H. 2016 A computational method for sharp
interface advection. Roy. Soc. Open Sci. 3, 160405.
Russo, G. & Smereka, P. 2000 A remark on computing distance functions. J. Comput.
Phys. 163, 51–67.
Sato, Y., Sadatomi, M. & Sekoguchi, K. 1981 Momentum and heat transfer in
two-phase bubble flow-I. Theory. Int. J. Multiphase Flow 7, 167–177.
Scardovelli, R. & Zaleski, S. 2000 Analytical relations connecting linear interfaces
and volume fractions in rectangular grids. J. Comput. Phys. 164, 228–237.
Scardovelli, R. & Zaleski, S. 2003 Interface reconstruction with least-square fit and
split eulerian-lagrangian advection. Int. J. Numer. Methods Fluids 41, 251–274.
Sethian, J. A. 1999 Level set methods and fast marching methods: evolving interfaces
in computational geometry, fluid mechanics, computer vision, and materials science.
Cambridge University Press.
Shaw, R. A. 2003 Particle-turbulence interactions in atmospheric clouds. Annu. Rev.
Fluid Mech. 35, 183–227.
Shen, J. & Yang, X. 2010 A phase-field model and its numerical approximation for
two-phase incompressible flows with different densities and viscosities. SIAM J. Sci.
Comput. 32, 1159–1179.
Sirignano, W. A. 1999 Fluid dynamics and transport of droplets and sprays. Cambridge
University Press.
Sun, Y. & Beckermann, C. 2007 Sharp interface tracking using the phase-field equa-
tion. J. Comput. Phys. 220, 626–653.
Sussman, M. & Ohta, M. 2009 A stable and efficient method for treating surface
tension in incompressible two-phase flow. SIAM J. Sci. Comput. 31, 2447–2471.
Sussman, M. & Puckett, E. G. 2000 A coupled level set and volume-of-fluid method
for computing 3D and axisymmetric incompressible two-phase flows. J. Comput.
Phys. 162, 301–337.
Sussman, M., Smereka, P. & Osher, S. 1994 A level set approach for computing
solutions to incompressible two-phase flow. J. Comput. Phys. 114, 146–159.
Tang, K. & Gomez, A. 1994 On the structure of an electrostatic spray of monodisperse
droplets. Phys. Fluids 6, 2317–2332.
Tomar, G., Biswas, G., Sharma, A. & Agrawal, A. 2005 Numerical simulation of
A review of interface-capturing methods for two-phase flows 135
bubble growth in film boiling using a coupled level-set and volume-of-fluid method.
Phys. Fluids 17, 112103.
Trevorrow, M. V., Vagle, S. & Farmer, D. M. 1994 Acoustical measurements of
microbubbles within ship wakes. J. Acoust. Soc. Am. 95, 1922–1930.
Tryggvason, G., Scardovelli, R. & Zaleski, S. 2011 Direct numerical simulations
of gas–liquid multiphase flows. Cambridge University Press.
Ubbink, O. & Issa, R. 1999 A method for capturing sharp fluid interfaces on arbitrary
meshes. J. Comput. Phys. 153, 26–50.
Urzay, J. 2017 Supersonic combustion in air-breathing propulsion systems for hyper-
sonic flight. Annu. Rev. Fluid Mech. 50, 593-627.
Waclawczyk, T. 2015 A consistent solution of the re-initialization equation in the
conservative level-set method. J. Comput. Phys. 299, 487–525.
Wang, Y., Shu, C., Shao, J., Wu, J. & Niu, X. 2015 A mass-conserved diffuse
interface method and its application for incompressible multiphase flows with large
density ratio. J. Comput. Phys. 290, 336–351.
Weymouth, G. D. & Yue, D. K.-P. 2010 Conservative volume-of-fluid method for
free-surface simulations on cartesian-grids. J. Comput. Phys. 229, 2853–2865.
Xiao, F., Honma, Y. & Kono, T. 2005 A simple algebraic interface capturing scheme
using hyperbolic tangent function. Int. J. Numer. Methods Fluids 48, 1023–1040.
Xie, B., Ii, S. & Xiao, F. 2014 An efficient and accurate algebraic interface captur-
ing method for unstructured grids in 2 and 3 dimensions: The thinc method with
quadratic surface representation. Int. J. Numer. Methods Fluids 76, 1025–1042.
Yang, X., Feng, J. J., Liu, C. & Shen, J. 2006 Numerical simulations of jet pinching-
off and drop formation using an energetic variational phase-field method. J. Comput.
Phys. 218, 417–428.
Youngs, D. L. 1982 Time-dependent multi-material flow with large fluid distortion.
Numer. Methods Fluid Dyn. 1, 41–51.
Yue, P., Zhou, C. & Feng, J. J. 2007 Spontaneous shrinkage of drops and mass
conservation in phase-field simulations. J. Comput. Phys. 223, 1–9.
Zhang, D., Jiang, C., Liang, D., Chen, Z., Yang, Y. & Shi, Y. 2014 A refined
volume-of-fluid algorithm for capturing sharp fluid interfaces on arbitrary meshes.
J. Comput. Phys. 274, 709–736.
Zhao, H. 2005 A fast sweeping method for Eikonal equations. Math Comput. 74, 603–
627.