See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/315978278

Numerical modelling of a direct contact condensation experiment using the

AIAD framework

Article  in  International Journal of Heat and Mass Transfer · August 2017

DOI: 10.1016/j.ijheatmasstransfer.2017.03.104

CITATIONS READS

5 401

3 authors:

Thomas Höhne Stasys Gasiunas

Helmholtz-Zentrum Dresden-Rossendorf Lithuanian Energy Institute
132 PUBLICATIONS   1,181 CITATIONS    16 PUBLICATIONS   19 CITATIONS   

SEE PROFILE SEE PROFILE

Marijus Šeporaitis
Lithuanian Energy Institute
28 PUBLICATIONS   35 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Investigation of coolant mixing at the ROCOM test facility View project

Two-phase flow View project

All content following this page was uploaded by Thomas Höhne on 23 January 2018.

The user has requested enhancement of the downloaded file.

 NUMERICAL MODELLING OF A DIRECT CONTACT CONDENSATION
EXPERIMENT USING THE AIAD FRAMEWORK

Thomas Höhne1, Stasys Gasiunas2, Marijus Šeporaitis2

1
Helmholtz-Zentrum Dresden-Rossendorf (HZDR) - Institute of Fluid Dynamics
P.O.Box 510119, D-01314 Dresden, Germany

2
Lithuanian Energy Institute (LEI), Kaunas, Lithuania

Abstract

The Lithuanian Energy Institute (LEI) test case deals with direct contact condensation (DCC) in the
two-phase stratified steam-water flow. The main goal of CFD simulations of these experiments is to
compute new models of heat and mass transport from saturated vapour to liquid over a free surface
and the temperature profiles across the liquid flow in a channel. Condensation occurs mainly on free
surfaces for instance at PTS scenarios. The knowledge of the accurate coolant temperature is
important for nuclear safety assessment.
Three different direct contact condensation models for the heat transfer within the AIAD framework
at the free surface were formulated and tested. The AIAD model describes a consistent set of model
correlations for the interfacial area density, the drag, the non-resolved disturbances of a free surface
and the turbulence damping the interface. The calculated surface temperature profiles agree well
with the experiment. Further model development should focus on “CFD grade” experimental data
and direct numerical simulations.

Keywords
CFD, CMFD, horizontal flow, PTS, PWR, AIAD, DCC, two-phase flow, two fluid flow.

1. Introduction
Condensation is a significant phenomenon in numerous engineering applications. Thermal phase
change processes are effective ways of heat removal, as the latent heat of condensation and boiling
provides high heat transfer. For designing heat exchangers the control of the heat transfer processes
is essential. Condensation occurs mainly on free surfaces. The gas-liquid interface depends on
whether the surface is wettable (film condensation) or not (drop-wise condensation). Direct contact
condensation occurs, if the vapour is in direct contact with the subcooled liquid. Contact
condensation on the free surfaces takes place for instance in pressurized thermal shock
(PTS)scenarios, when the injected cold water flows together with steam through the cold leg and the
other primary loop parts of pressurized water reactors (PWR). Accurate simulation of heating the
emergency core cooling water is important to control the effects of loss of coolant accidents.
The computational fluid dynamics (CFD) codes offer an effective and powerful way to simulate
industrial components. These codes solve the continuity equations in a three-dimensional domain.
Nevertheless, the 3D simulation of phase change heat transfer remains still a challenging task due to
the extensive computational time needed and the lack of “CFD grade” experiments.

Despite latest advances in the field of computational multi-fluid dynamics (CMFD), only very few
physical models dedicated to the problem of Direct Contact Condensation (DCC) in horizontal
stratified flow are available at the moment.

Two commonly used 1D correlations for heat and mass transfer during DCC in horizontal two phase
flows were derived from the experimental results in a horizontal pipe by Lim et al. (1984) and Kim et
al. (1985).

Celata et al. (1986) measured DCC on slowly moving subcooled water in a "pressurizer-like" geometry
and developed a special and limited set of integral correlations.

Chan and Yuen (1990) used the experimental device of Lim et al. (1984) and investigated the
influence of air on the DCC in the stratified horizontal flow.

Ramamurti and Kumar (2001) performed a DCC experiment on a thick layer of moving water in the
vessel with a stagnant vapour bubble and expressed the heat transfer coefficients in terms of Nusselt
number as a function of liquid Reynolds and Prandtl number and the rate of sub-cooling.

Widely used correlations are derived by Hughes and Duffey (1991). They introduced a "surface
renewal theory" for DCC in turbulent separated flow and developed a so-called "local" closure law for
description of the interphase heat and mass exchange.

Experiments and models of DCC in a rectangular duct and rectangular tank were later described by
Lorencez et al. (1997) and Mikielewicz et al. (1997), respectively. Especially Lorencez et al. (1997)
with their sophisticated experiment made comprehensive measurement of the turbulence near the
free surface and clarified the influence of the turbulence on the interfacial heat and mass transfer
coefficients.

Various methods for CMFD simulations of stratified two phase flows are described in the literature.

The one fluid approach with interface tracking is used for flow situations with large-scale interfaces
like film, annular or horizontal stratified flows. Usually the grid is fine enough for a localization of the
gas-liquid interface allowing a detailed resolution of the surface phenomena. Established methods
are straightforward interface tracking methods like surface-attached moving meshes or interface
capturing methods like the Volume of Fluid (VOF) (Scardovelli and Zaleski, 1999)- or the Level-Set-
method (Smereka, 2003) that are developed for the volume fraction advection step.

The assumption of the sharp-interface is not always appropriate (Anderson et al., 1998) as the
thickness of the interface may not be negligible comparing to the relevant scales especially near the
critical temperature. Anderson et al. (1998) present a review of the models and methods that can be
applied for simulations of diffuse-interfaces of finite thickness.

Olsson and Kreiss (2005) introduced a level set method in which the advection of the level-set
function is followed by an artificial compression step to ensure that the thickness of the interface
layer is preserved, inducing a volume conservation. Štrubelj et al. (2009) improved the two-fluid
2
model with a conservative level-set method proposed by Olsson and Kreiss (2005). Additionally the
model included a surface tension force based on the model proposed by Brackbill et al. (1992). The
interface sharpening method and the surface tension force were validated on several test cases
where viscosity was increased in order to achieve a damping of spurious currents. But up to now no
transitions between small- and large scale gas phases have been considered.

An advanced approach was introduced by Lakehal et al. (2003) based on pseudo-spectral DNS of
turbulent wavy flow at low Reynolds number but limited to a narrow range of flows with low Reynold
number and low subcooling rates. Lakehal and Labois (2011) used VLES within the TransAt code to
derive a heat transfer coefficient correlation at the liquid vapor interface for DCC. They used surface
divergence theory to define a correlation between the turbulence of the liquid and the heat and
mass transfer at the interface based on the direct numerical simulation.

The Eulerian two-fluid model is most suited for small-scale dispersed flows like bubbly or droplet
flows. Such flow patterns are characterized by a scale of interfacial structures smaller than the used
grid size, therefore an averaged treatment is used and for each phase a corresponding set of
equations is solved. However, the Euler-Euler two-fluid models with appropriate algorithms for
tracking of the larger interfaces, might be an alternative to the pure interface tracking methods,
which fail when the surface characteristic scales become comparable or smaller than the grid size; a
discussion is given in Yadigaroglu (2005).

Moreover, already the simulation of adiabatic two-phase flows introduces difficulties. The VOF
method cannot simulate two-phase flows with high velocity differences between the phases.
Bartosiewicz et al. (2010) highlighted this issue in a simulation of slug formation in an air–water
channel.

Simulations of stratified flow with a 2D two-fluid model were further performed by Yao et al. (2003),
who made simulations of stratified flow with and without the condensation. CFD simulations of ECC
injection of subcooled water into horizontally stratified hot leg flow were performed by Coste (2011)
using two-fluid model with interfacial heat and mass transfer model based on surface renewal
concept.

Scheuerer et al. (2007) simulated DCC in LAKOON experimental device (Hein et al., 1995) and
achieved a good agreement between the measured and calculated condensation rate. The local
temperatures in simulation were underestimated.

The NEPTUNE-code works with a Large Interface Model (LIM) for stratified flows. This model locates
the free surface without any reconstruction in order to apply closure laws (Coste et al., 2012). For
this a refined gradient method is used which allows to detect stratified grid cells. The LIM-model has
been validated on several configurations (Bartosiewicz et al. 2010) and has also been compared with
other CFD-models within a benchmark (Lucas et al. 2011).

An alternative type of interface capturing method within the two-fluid model is implemented in the
CFX-code using a compressive advection discretization scheme which is applied to the volume
fraction equation (Zwart et al. 2007). This so-called Free Surface model has been used successfully
for the modelling of horizontally stratified pipe flows (Vallée et al. 2008). Nevertheless, it is not
appropriate to represent mixed flows so far as there is the need to identify the morphology of the
fluid phases.

3
An important aspect of the simulation is the evaluation of how the CFD methods treat turbulent
transport near the free surface, which primarily determines the inter-phase heat and mass transfer
predictions. Therefore, an Algebraic Interfacial Area Density (AIAD)-model was introduced by Höhne
and Vallée (2010). This model simulates the momentum exchange dependent on the morphological
form of the stratified flow pattern and is able to distinguish between bubbles, droplets and the free
surface using the liquid volume fraction values. It was implemented into the CFX-code and validated
against experimental data of counter-current flows in a hot leg model of a pressurized water reactor
(Höhne et al. 2011), in horizontal channels (Höhne and Mehlhoop, 2014) and at the Wenka test
facility (Porombka and Höhne, 2015). The AIAD-model represents a new alternative way to capture
the gas-liquid interface within the two-fluid model.

Štrubelj et al. (2010) used a two-fluid method to calculate DCC with the NEPTUNE_CFD code. They
assumed that the steam is at saturation condition; therefore, the energy equation of the vapour
phase was not solved.

Welch and Wilson (2000), too, assumed that the interface temperature is at the saturation
temperature of the pressure of the liquid phase. They developed a numerical VOF method, which
calculates the heat flux vectors on both sides of the interface, considering the jump term of the mass
conservation equation.

Sato and Niceno (2013) obtained the interfacial mass transfer rate directly from the heat flux balance
equation with the CFD code PSI-BOIL.

Thus, in order to fully understand and predict the DCC phenomenon, an experiment from LEI In
Kaunas Lithuania was utilized. Realizing this need, the AIAD concept was used and further developed
for flows with heat and mass transfer.

2. CFD simulation of Gas-Liquid two phase flows

2.1 Continuity, Momentum and Energy Equation
Multiphase CFD methods resolve the conservation equations for mass, momentum and energy and
they are distinguished by the different approaches and strategies used in describing the physical
closure models. One possibility is to use the two fluid Euler–Euler approach. This approach assumes
that at least two fluids are continuously penetrating each other. The volume fraction of the fluids in
each cell sums to unity. For each fluid, the full set of conservation equations is solved. Therefore,
each fluid has a different velocity field. The mechanisms of the interaction of the fluids are the
momentum transfer modelled by the flow resistance, the mass transfer modelled by phase change
and the energy transfer modelled by heat conduction.

In the calculation, we solved the conservation of mass (Eq. 1), momentum (Eq. 2) and energy
equations (Eq. 3) of the two-fluid model, which have the following form:

 k  k U k
 ( k  k U k )  K (1)
t

 k  k U k
 ( k  k U k U k )   k pk   k  k g   k ( v   kt )  U k ,i k  M i ,k   k   i (2)
t

4
 k  k H k D
 ( k  k H k U k )    k (q v  q kt )   k k p k  H ki k  q ki / Ls   k (3)
t Dt

where the subscript k denotes phase gas or liquid and i stands for the value at the interface, Ls
denotes the length scale at the interface,  is the density, U is the velocity vector, t is the time, p is
the pressure, g is the gravitational acceleration,  is the volume fraction,  is the shear stress (v is
the average viscous shear stress, t is the turbulent shear stress) i is the interfacial shear stress, k is
the mass generation, M i ,k the generalized interfacial drag, H the enthalpy, qki the interfacial heat
flux and  k the interfacial dissipation.

3. Algebraic Interfacial Area Density Model

The AIAD model was developed in close cooperation by ANSYS and HZDR and is described in Egorov
(2004) and Höhne (2014).

The basic idea of the AIAD model is:

 The interfacial area density (IAD) allows the detection of the morphological form and the
corresponding switch of each correlation from one object pair to another.
 It provides a law for the interfacial area density and the drag coefficient for a full range of
phase volume fractions from no liquid to no gas.
 The model improves the physical modelling in the asymptotic limits of bubbly and droplet
flows.
 The interfacial area density in the intermediate range is set to the interfacial area density for
free surface.
 The difference (flow behaviour, interfacial area) between a flat interface and a wavy
interface can be considered in the framework.

The approach used in the AIAD model is to define blending functions depending on the volume
fraction that enable switching between the morphologies of dispersed droplets, dispersed bubbles
and the free surface. Based on these blending functions, different equations for the interfacial area
density and the drag coefficient can be applied according to the local morphology. The blending
functions are defined as Eq. (4) and Eq. (5) for droplet and bubble region, respectively and Eq. (6) for
the free surface.
−1
𝑓𝐷 = [1 + 𝑒 𝑎𝐷 (𝛼𝐿−𝛼𝐷,𝑙𝑖𝑚𝑖𝑡 ) ] (4)
−1
𝑓𝐵 = [1 + 𝑒 𝑎𝐵 (𝛼𝐺−𝛼𝐵,𝑙𝑖𝑚𝑖𝑡 ) ] (5)
𝑓𝐹𝑆 = 1 − 𝑓𝐷 − 𝑓𝐵 (6)

With 𝑎𝐷 and 𝑎𝐵 being the blending coefficients for droplets and bubbles, respectively and 𝛼𝐷,𝑙𝑖𝑚𝑖𝑡
and 𝛼𝐵,𝑙𝑖𝑚𝑖𝑡 the volume fraction limiters. In the simulations presented here, values of 𝑎𝐷 = 𝑎𝐵 = 70
and 𝛼𝐷,𝑙𝑖𝑚𝑖𝑡 = 𝛼𝐵,𝑙𝑖𝑚𝑖𝑡 = 0.3 were used. For all model coefficients same values were used as in
previous studies (Höhne 2010, 2011 and 2013). They were chosen independent of the actual
geometry and flow regime and no tuning of the AIAD model was done for the work presented here.
The threshold value 𝛼𝐵,𝑙𝑖𝑚𝑖𝑡 = 0.3 is a critical volume fraction before the coalescence rate increases
sharply and is verified by experiments in both vertical and horizontal flows. Published data agree that

5
bubbly flow rarely exceeds a gaseous volume fraction of about 0.25-0.35 when the transition to
resolved structures occurs (Griffith and Wallis 1961, Taitel et al. 1980, Murzyn and Chanson 2009).
Parameter studies also indicated that the model is not very sensitive towards a change of the
blending function parameters.

For simplicity bubbles and droplets are for now assumed to be of spherical shape, with a constant
diameter of 𝑑𝐵 and 𝑑𝐷 , respectively. Non-drag forces in the regions of dispersed flow are neglected.
The resulting formulation for the interfacial area density for droplets 𝐴𝐷 is given by

6𝛼𝐿
𝐴𝐷 = (7)
𝑑𝐷

in which 𝛼𝐿 is the volume fraction of the liquid phase. The IAD for bubbles 𝐴𝐵 is formulated
analogous. The IAD of the free surface 𝐴𝐹𝑆 is defined as the magnitude of the gradient of the liquid
volume fraction 𝛼𝐿 , as given in Eq. (8), with 𝒏 being the normal vector of the free surface.

𝜕𝛼𝐿
𝐴𝐹𝑆 = |∇𝛼𝐿 | = (8)
𝜕𝑛

The local interfacial area density 𝐴 is then calculated as the sum of 𝐴𝑗 , weighted by the blending
functions 𝑓𝑗 :

𝐴 = ∑ 𝑓𝑗 𝐴𝑗 , 𝑗 = 𝐹𝑆, 𝐵, 𝐷 (9)
𝑗

3.1 Modelling the free surface drag

In the general case of a two-phase flow, there is a velocity difference between the fluids, which is
commonly called slip velocity. In contradiction to most Volume of Fluid (VOF) methods, where only
one velocity field is present, within the multi-fluid framework each phase has its own velocity and
turbulence model. Thereby a drag force is induced at the phase boundary that is acting on both
phases. The drag force can be correlated with the slip velocity 𝑼, the fluid density 𝜌, the surface area
𝑎 and the dimensionless drag coefficient 𝐶𝐷 . For geometry independent modeling the drag force is
expressed as the volumetric force density 𝐹𝐷 and 𝑎 is then replaced by the area density. The
magnitude of the drag force density is given by Eq. (10).

1 2
|𝐹𝐷 | = 𝐶𝐷 𝐴 𝜌| U | (10)
2

For a dispersed flow, the density of the continuous phase is used in Eq. (10). In case of a free surface
the phase averaged density is used, i.e.

𝜌 = 𝛼𝐺 𝜌𝐺 + 𝛼𝐿 𝜌𝐿 (11)

with 𝜌𝐺 and 𝜌𝐿 being the density of the gas and the liquid, respectively.

In simulations of free surface flows, Eq. (10) does not represent a realistic physical model. It is
reasonable to expect that the velocities of both fluids in the vicinity of the interface are rather

6
similar. In Höhne (2010), it is assumed that the shear stress near the surface behaves like a wall shear
stress on both sides of the interface in order to reduce the velocity differences of both phases. It is
supposed that the morphology region “free surface” is acting like a wall and a wall like shear stress is
introduced at the free surface, which influences the loss of gas velocity. The components of the
normal vector n⃗ at the free surface are taken from the gradients of the void fraction.

To use these directions of the normal vectors the gradients of gas/liquid velocities, which are used to
calculate the wall shear stress onto the free surface, are weighted with the components of the
normal vector. Three different drag coefficients can be applied:

1. 𝐶𝐷,𝑏𝑢𝑏𝑏 (for bubbly regions): Drag coefficient is assumed constant and equal to 0.44
2. 𝐶𝐷,𝑑𝑟𝑜𝑝 (for droplet regions): Drag coefficient is assumed constant and equal to 0.44
3. 𝐶𝐷,𝑓𝑠 (for free surface region): The drag is ruled by viscous surface stresses and the
formulation as proposed by Höhne and Mehlhoop (2014) is used.

(12)
2[𝛼𝐿 𝜏𝑤,𝐿 + 𝛼𝐺 𝜏𝑤,𝐺 ]
𝐶𝐷,𝐹𝑆 = 𝑚𝑎𝑥 [0.01, 2 ]
𝜌𝐿 U 𝑠𝑙𝑖𝑝

3.2 Turbulence Modelling

Fluid dependent turbulence option has been used for the continuous liquid respective gaseous
phase. In terms of turbulence treatment, the k-ω approach is used for both continuous phases. One
of the advantages of the k-ω model over the k-ε is the treatment when in low Reynolds numbers for
a position close to the wall (Menter, 1994).

3.2.1 Turbulence Damping

Without any special treatment of the free surface, the high velocity gradients at the free surface,
especially in the gaseous phase, generate levels of turbulence that are too high throughout the two-
phase flow when using differential eddy viscosity models like the k-ε or the k-ω model. Therefore, a
certain amount of damping of turbulence is necessary in the region of the interface, because the
mesh is too coarse to resolve the velocity gradient in the gas phase at the interface.

For the two-fluid formulation, Egorov (2004) proposed a symmetric damping procedure. This
procedure provides a solid wall-like damping of turbulence in both gas and liquid phases. More
information can be found in Höhne and Mehlhoop (2014).

3.2.2 Sub-grid wave turbulence

Small waves created by Kelvin-Helmholtz instabilities that are smaller than the grid size are neglected
in traditional two phase flow CFD simulations, but the influence on the turbulence kinetic energy of
the liquid side can be significantly large.

The consequence of the specific turbulent kinetic energy kSWT is prescribed in a source term where
the gradients of the local velocities and the liquid density are present and which is added to the total
turbulent kinetic energy k (Wilcox, 1994). More information about the sub-grid wave turbulence is
found in Höhne and Mehlhoop (2014).

7
4. Modelling the heat and mass transfer

4.1 Interphase Heat Transfer Models

Due to thermal non-equilibrium, heat transfer occurs across the phase interface. It is described in
terms of an overall heat transfer coefficient ℎ𝐿𝐺 , which is the amount of heat energy crossing a unit
area per unit time per unit temperature difference between the phases. Rate of heat transfer 𝑄𝐿𝐺
per unit time across a phase boundary of interfacial area per unit volume 𝐴𝐷 , from liquid to gas, is :

𝑄𝐿𝐺 = ℎ𝐿𝐺 𝐴𝐷 (𝑇𝐺 − 𝑇𝐿 ) (13)

Or, this can be rewritten as:

(ℎ) (14)
𝑄𝐿𝐺 = 𝑐𝐿𝐺 (𝑇𝐺 − 𝑇𝐿 )

(ℎ)
𝑐𝐿𝐺 = ℎ𝐿𝐺 𝐴𝐷 for the particle model that we have used in the case.

Hence, the interfacial area per unit volume and the heat transfer coefficient ℎ𝐿𝐺 needs to be known.

𝜆𝑁𝑢
We know that ℎ = 𝑑

In the particle model, 𝜆 is taken to be the thermal conductivity of the continuous phase and the
length scale 𝑑 is the mean diameter of the dispersed phase, that is:

𝜆𝐿 𝑁𝑢𝐿𝐺 (15)
ℎ𝐿𝐺 = 𝑑𝐺

In our case of heat transfer between liquid and gas, the use of overall heat transfer coefficient is not
sufficient to model the interphase heat transfer process. This model considers separate heat transfer
process on each side of the phase interface. This is achieved by using two heat transfer coefficients
defined on each side of the phase interface.

The sensible heat flux to liquid from the interface is given as:

𝑞𝐿 = ℎ𝐿 (𝑇𝑠 − 𝑇𝐿 ) (16)

Similarly, the sensible heat flux to gas from the interface:

𝑞𝐺 = ℎ𝐺 (𝑇𝑠 − 𝑇𝐺 ) (17)

The fluid specific Nusselt number is given by:

ℎ𝐿 𝑑𝐿𝐺 (18)
𝑁𝑢𝐿 =
𝜆𝐿

Phase transition:

The interphase mass transfer is determined from the total heat balance:
Total heat flux to phase L from the interface:

𝑄𝐿 = 𝑞𝐿 + 𝑚𝐿𝐺
̇ 𝐻𝐿𝑆 (19)

8
Total heat flux to phase G from the interface:

𝑄𝐺 = 𝑞𝐺 + 𝑚𝐿𝐺
̇ 𝐻𝐿𝑆 (20)

HLS and HGS represent interfacial values of enthalpy carried into and out of the phases du to phase
change. The total heat balance 𝑄𝐺 + 𝑄𝐿 = 0 now determines the interphase mass flux:

 LG  QL  QG  H GS  H LS 
m (21)

4.2 Modelling the Direct Contact Condensation

No universal correlations for the inter-phase momentum and heat transport at free surfaces are
available so far because of the diversity of horizontal flow patterns. A very limited number of
correlations, published in the literature, are typically calibrated for one-dimensional thermo-
hydraulic codes. Their direct implementation in three-dimensional CFD codes results in either grid-
dependent models or in introducing fitting parameters which have to be re-calibrated for each new
flow condition. Their applicability to a general situation, with its typically complex multi-dimensional
transient flow pattern, may bring additional uncertainties rather than increase the simulation
accuracy.

The turbulence treatment is a critical issue in selecting a modelling approach for the interfacial
momentum and heat transfer at a free surface.

Namely, the profiles of the statistically averaged velocity, temperature, and turbulence variables are
to be smooth throughout the interface zone. Besides, the velocity and the temperature slip between
the phases can only be present to account for the unresolved interface disturbances (sub-grid wave
turbulence).

Using the turbulence treatment (turbulence damping at the interface, sub-grid wave turbulence) and
drag closure law from the AIAD framework, the following model setups were utilized and compared
with the experiment:

 Egorov DCC model including the Ranz Marshall correlation

 Modified Hughes-Duffey (HD) model including the Ranz Marshall correlation
 Adapted Coste DCC model.

4.2.1 Egorov DCC model

The low Reynolds number model is taken from Egorov (2004). For spherical droplets it is advisable to
set the Ranz Marshall (1952) correlation for the Droplet regime. The Ranz Marshall correlation is
given as:

𝑁𝑢 = 2 + 0.6𝑅𝑒 0.5 𝑃𝑟 0.3 (22)

It is based on boundary layer theory for steady flow past a spherical particle.

The following formulations are used:

𝑁𝑢𝐺 = ∑𝑗 𝑓𝑗 𝑁𝑢𝐺 𝑗 , 𝑗 = 𝐹𝑆, 𝐵, 𝐷 (23)

9
 𝑁𝑢𝐺,𝐵 = 6.58

0.5
𝑁𝑢𝐺,𝐷 = 2 + 0.6 ∙ 𝑅𝑒𝑆𝑙𝑖𝑝 ∙ 𝑃𝑟𝐺0.3 (24)

𝑁𝑢𝐺,𝐹𝑆 = 400

𝑁𝑢𝐿 = ∑𝑗 𝑓𝑗 𝑁𝑢𝐿 𝑗 , 𝑗 = 𝐹𝑆, 𝐵, 𝐷 (25)

𝑁𝑢𝐿,𝐵 = 6.58

𝑁𝑢𝐿,𝐷 = 2

𝑁𝑢𝐿,𝐹𝑆 = 400

with 𝑅𝑒𝑠𝑙𝑖𝑝 as the slip velocity Reynolds number:

𝑈𝑠𝑙𝑖𝑝 ∙𝑑𝐷
𝑅𝑒𝑠𝑙𝑖𝑝 = (26)
𝜐𝐺
and the Prandtl number 𝑃𝑟𝐺 :

𝜂𝐺 ∙𝑐𝑝,𝐺
𝑃𝑟𝐺 = 𝜆𝐿
(27)

As the droplet temperature increases to the saturation temperature, the Nusselt number approaches
its asymptotic value of 6.58. For 𝑁𝑢𝐿,𝐹𝑆 and 𝑁𝑢𝐺,𝐹𝑆 a high value was suggested by Egorov (2004).
Calculation of the HTC from experimental findings in Gasiunas et al. (2012) showed, that 𝑁𝑢𝐹𝑆 = 400
is sufficient high enough.

4.2.3 Modified Hughes-Duffey (HD) model

The Hughes and Duffey (1991) model uses the surface renewal theory. The surface renewal theory
predicts the condensation mass transfer from the liquid turbulent properties and the renewal period
of the eddies. The theory suggests that turbulent eddies transport bulk liquid, which has lower
temperature, to the surface, and removes saturated condensate from the interface. The local
temperature gradient across the interface thus increases, which enhances the condensation
compared to laminar heat conduction. The renewal period controls the contact time between the
bulk liquid and the saturated vapour; consequently, this is the driving force for the condensation
mass transfer. The renewal theory calculates the effect of the reduced interfacial shear and the eddy
diffusivity distribution on heat and mass transfer by directly deriving them from the momentum
transfer.

This model is called Hughes-Duffey (HD) model in the present paper:

𝑁𝑢 = 1.13 ∙ 𝑅𝑒𝐿,𝑇𝑢𝑟𝑏 ∙ 𝑃𝑟 0.5 (28)

where Pr is Prandtl number and ReTurb turbulent Reynolds number of liquid.

Nevertheless, the turbulent Reynolds number is actually calculated from the local parameters as:

10
 𝑈𝑇𝑢𝑟𝑏 ∙𝐿𝑇𝑢𝑟𝑏
𝑅𝑒𝐿,𝑇𝑢𝑟𝑏 = 𝜐𝐿
(29)

where νL is liquid viscosity, LTurb is length scale and UTurb velocity scale, which are calculated from
liquid turbulence kinetic energy kL and turbulence dissipation εL:

1/4 1/2
𝑈𝑇𝑢𝑟𝑏 = 𝐶𝜇 𝑘𝐿 , 𝐶𝜇 = 0.09 (30)

3/2
𝑘𝐿
𝐿 𝑇𝑢𝑟𝑏 = 𝐶𝜇 𝜀𝐿
(31)

With Eq. (27) the Nusselt numbers are the following:

𝑁𝑢𝐺,𝐵 = 6.58

0.5
𝑁𝑢𝐺,𝐷 = 2 + 0.6 ∙ 𝑅𝑒𝑆𝑙𝑖𝑝 ∙ 𝑃𝑟𝐺0.3 (32)

𝑁𝑢𝐺,𝐹𝑆 = 1.13 ∙ 𝑅𝑒𝐺,𝑇𝑢𝑟𝑏 ∙ 𝑃𝑟𝐺0.5 (33)

With Eq. (29) the Nusselt numbers are the following:

𝑁𝑢𝐿,𝐵 = 6.58

𝑁𝑢𝐿,𝐷 = 2

𝑁𝑢𝐿,𝐹𝑆 = 1.13 ∙ 𝑅𝑒𝐿,𝑇𝑢𝑟𝑏 ∙ 𝑃𝑟𝐿0.5 (34)

And with Eq. (31) the Prandtl numbers are the following:

𝜂𝐺𝑎𝑠 ∙𝑐𝑝,𝐺
𝑃𝑟𝐺 = 𝜆𝐺
(35)

𝜂𝐿 ∙𝑐𝑝,𝐿
𝑃𝑟𝐿 = 𝜆𝐿
(36)

4.2.4 Adapted Coste DCC model

The heat transfer coefficient model proposed by Coste (see Lucas et al. 2011) was utilized and
adapted to the AIAD framework. The turbulent Reynolds number is calculated from turbulence
quantities similar as in HD model. The heat transfer coefficient model is similar to HD model and
comes from surface renewal model with renewal frequency, which is defined with Kolmogorov
length scale and turbulent velocity.

With Eq. (27) the Nusselt numbers are the following:

𝑁𝑢𝐿,𝐵 = 6.58

11
 𝑁𝑢𝐿,𝐷 = 2

0.875
𝑁𝑢𝐿,𝐹𝑆 = 2.7 ∙ 𝑅𝑒𝑇𝑢𝑟𝑏 ∙ 𝑃𝑟𝐿0.5 (37)

With ReTurb and PrL according to Eq. (33) resp. Eq. (40). In the paper of (Coste 2004) he was not using
his model for the gas side for the validation of his model. Therefore, it is also neglected in this case.

In this case the gas temperature is fixed at the saturation temperature value, corresponding to the
given pressure level. It also means that, for the calculation of the condensation rate, zero heat flux is
assumed on the gas side of the interface.

5. Experimental facility
The test facility was designed and constructed for DCC in co-current steam and water flow research.
The interest is to evaluate interfacial shear, steam condensation influence to the water turbulence,
local HTC. Fig. 5.1 shows general view of the test section and Fig. 5.2 shows schematic of the
experimental facility, which is designed such that optical non-disruptive together with disruptive
traditional measurements of water temperature and velocity could be performed along the channel.
The test section has a rectangular cross-section with a length of 1.2 m, a width of 0.02 m, and a
height of 0.1 m. The test section is made of stainless steel 0.01 m thick and it is thermo insulated. For
this measurement setup the two borosilicate-glass windows (0.4 m x 0.09 m) are installed in the side
wall.

Fig. 5.1. Photo of experimental facility.

12
Fig. 5.2. Schematic of the test section: 1 – steam inlet pipe; 2 – rectangular test section; 3 –
thermocouples; 4 – flow equalizer; 5 – vortex type steam flow meter; 6 – electric valve for steam
flow adjustment; 7 – steam separator; 8 – electromagnetic water flow meter; 9 –water supply
adjustment valve; 10 – sliding assembly of the thermocouples.

Steam is produced by an electric steam boiler and the flow rate is measured by vortex-type flow
meter Endress+Hauser W72 (5). Superheated steam is supplied through 0.027 m diameter, 2 m long
horizontal pipe (1), connected to the test section through flow equaliser (4). It is a ceramic comb-like
structure, made of rectangular 0.001 m width/height and 0.095 m long mini channels, which
eliminates large steam eddies and also forms rectangular initial velocity profile.

Filtered tap water is supplied at the beginning of the test section through another ceramic comb in
order to avoid big eddies. Right after the combs there is placed 0,163 m long metal plate between
steam and water to get some velocity profile with boundary layer which helps to avoid initial
disturbance of water surface. The water level of 0.025 m is maintained using the weir with adjustable
height at the end of the channel and by slight degree (up to ±1ᵒ) of channel inclination. The water
flow rate is controlled by a valve in accordance with readings of an electromagnetic water flow meter
ISOIL ML201 (8). At the end of the channel, the water and steam is freely discharged to the
atmosphere. The density of the steam and water was estimated from the temperature and pressure
(atmospheric).

5.1 Water temperature measurement

The water temperature was measured using sliding assembly (10), made of three K-type
thermocouples, placed at different heights, specified in Fig. 5.3.

Fig. 5.3. Sliding assembly of thermocouples.

13
Thermocouples measure water temperature at any distance within the x=0.163 – 1.123 m of test
section. The 0.02 m interval was selected for water temperature longitudinal profile measurement.
Figure 5.4 shows water longitudinal temperature profiles measured along the channel at three
vertical heights from the bottom: 0.002 m measured with thermocouple T1; 0.0125 m with T2 and
0.023 m with T3. The steam inlet velocity is 6 m/s and water inlet velocity is 0.028 m/s.

Fig. 5.4. Time averaged water temperature along the channel (steam 6 m/s, cold water 0.028 m/s)

5.2 Flow regime

Measurements are performed at different steam (4, 6, 8 m/s) and water (0, 0.028, 0.055 m/s) inlet
velocities. Fig. 5.5 shows free surface of the hot water pictures at two windows of the test section
when steam inlet velocity is 6 m/s and no condensation is present. The first window begins at test
section distance 0.16 m and ends at 0.56 m, the second begins at 0.63 m and ends at 1.03 m. It is
clearly seen that in the first half of the channel is stratified smooth flow regime and in the second
half of the channel it becomes a little wavy with amplitude of 0.0015 m and wavelength of 0.03 m.
According to the flow pattern map of Mandhane (1974), flow regime is in the transition between
stratified and wavy. Next Fig. 5.6 shows free surface of water when steam inlet velocity is 6 m/s and
cold water supply velocity is 0.028 m/s. As steam condenses, its mass flow rate decreases along the
channel (about 20%). It influences the free surface of the water and water waves get even smaller
(0.001 m in amplitude and 0.017m in wavelength). According to the von Arx (1974) it is capillary
waves.

In this paper, for condensing two-phase flow modelling, the 6 m/s of steam with 0.028 m/s of cold
water inlet velocity combination at ambient pressure is selected.

Fig. 5.5 Free surface shape at steam supply 6 m/s and water supply 0 m/s (no condensation).

14
Fig. 5.6 Free surface shape at steam supply 6 m/s and water supply 0.028 m/s.

6. Numerical Setup
Gas (continuous gas) is taken as saturated steam at 373 K. Properties of dry steam at saturation
temperature have been taken from the steam tables. Water is taken as a liquid. The water is initially
at temperature 293 K. The inlet boundary conditions are the steam flow rate 32.4 m3/h and the
water flow rate 49.7 l/h.

The walls are set adiabatic. It is worth noticing that, at the walls, a no-slip boundary condition is
applied for both continuous phases (i.e., gas and liquid). The boundary condition at the outlet is set
as a pressure boundary condition. The water level at the initial state is 25 mm. The steam cross
section is 20x75 mm.

For the boiling modelling the reader should refer to the section 4.

According to the experimental setup this flow is considered to be stationary. The selected case
corresponds to the stratified smooth flow regime. Only capillary waves were observed. The Reynolds
number of steam was in the range of 9000. The inlet was supplied with the saturated dry steam at
the temperature corresponding to the measured pressure. The channel walls are thermally insulated.
The following assumptions allow one to consider steam above the free surface as isothermal gas with
the known constant values of all properties (density, specific enthalpy, specific heat capacity,
dynamic viscosity):
 No inlet steam superheat
 Adiabatic upper and side walls
 No homogeneous condensation (no wet steam)
Due to the carefully designed inlet and outlet of water the free surface is plane and horizontal
everywhere. Reduction of the steam flow rate inside the channel is significant, since a big portion of
inlet steam is condensed in the measured flow section. During the condensation the specific heat of
phase transition is released and fully consumed for heating up the initially sub-cooled water. The
correspondent temperature gradient, normal to the free surface, develops in water. Therefore the
condensation rate is limited by the transport rate of heat from the free surface to the bulk water
flow.

Since steam is isothermal, no energy equation and thus no thermal boundary conditions for the
vapour phase are necessary. The steam properties (temperature, density, molecular transport
coefficients) correspond to the saturated dry steam at the given pressure.

Separation of phases in this flow is provided by the gravity force.

15
6.1 Geometry and Mesh
The simulations are performed using a geometric representation of the experiment. The channel
cross section is 20x100mm, the channel is 1200 mm long. The inlet blade is modelled, also the weir at
the water outlet. Medium turbulence intensity (5%) was assumed at the inlet. The outlet has a
pressure boundary condition including hydrostatic pressure.

The geometry has been discretized using the ICEM software. The resulting mesh is made of
approximately 1.3 Mio hexahedral cells. A grid resolution study was conducted to ensure that
convergence with respect to the spatial resolution has been achieved. One of the side walls was
modelled with symmetry boundary condition.

Fig. 6.1 Outline of computational domain with Fig. 6.2 Zoom of the grid
boundary conditions

The simulations were run on the Linux-Cluster hydra4. It consists of two heads and 72 compute
nodes. Each node has 2 Intel CPUs (Xeon 8-Core or 16-Core). In general hydra consists of more than
1500 CPU-cores and 8,5 TB of main memory. The network is a 1GbE Ethernet and additionally an
InfiniBand fabric with a bandwidth of 40 Gbit/s, which qualifies hydra for sequential, coarse
granularity parallel and massively parallel jobs. The operating system is Ubuntu Linux. The theoretical
general peak performance (single precision) amounts to 70,5 TFlop/s.

ANSYS CFX is an element-based finite-volume method with second-order discretisation schemes in

space and time. It uses a coupled algebraic multigrid algorithm to solve the linear systems arising
from discretisation. The discretisation schemes and the multigrid solver are scalably parallelized. The
following table shows the numerical scheme used in the case:

Table 6.1 Solver setup

Advection scheme Option High Resolution

Convergence control Timescale control 15000 Iterations
Convergence criteria Residual type RMS
Residual target 1e-04

16
The simulations of AIAD applied to phase transfer case were run on 1 core and 16 processors for
around 2 days (50 hours).

6.2 Overview of the settings and models used in the AIAD framework
The tables 6.2 and 6.3 show the overview on the different settings for the two phases respective for
the phase pairs. All the physical sub-models were programmed using the command language CCL and
no additional User Fortran routines were necessary.

Table 6.2: Overview on the different settings for the two phases

Liquid Steam

Morphology continuous fluid continuous fluid

Turbulence model k- k-

Turbulence Damping AIAD AIAD

Sub-grid wave turbulence AIAD none

Table 6.3: Overview on the different settings for the two phase pairs

Fluid Pair Gas | Liquid

Interphase transfer model Mixture Model
Momentum transfer AIAD (Höhne, 2014)
Mass transfer Thermal phase change according to chapter 4

Heat transfer Two resistance according to chapter 4

7. Results
As an example to illustrate the flow and temperature behaviour the next figures present the results
of the HD condensation model. The resulting free surface shape, shown in Fig. 1, is flat and
horizontal.

The steam superficial velocity shown as streamlines is depicted in Fig. 7.2 and at the symmetry plane
in Fig. 7.4. Reduction of the steam flow rate inside the channel is significant, since a big portion of
inlet steam is condensed on the interface. During the condensation the specific heat of phase
transition is released and fully consumed for heating up the initially sub-cooled water. It is visible on
the temperature distribution in a symmetry plane in Fig. 7.5.

The water superficial velocity shown as streamlines is depicted in Fig. 7.3 and at the symmetry plane
in Fig. 7.6. After the inlet plane there is a larger recirculation area, were the fluid is mixed. Higher
velocities are noticed at the free surface, initiated by the steam flow. The water is transported to the
outlet gap.

17
Fig. 7.1 Steam volume fraction in the symmetry plane of the channel

Fig. 7.2 Steam superficial velocity shown as Fig. 7.3 Water superficial velocity shown as
streamlines streamlines

Fig. 7.4 Steam superficial velocity shown in the symmetry plane of the channel

18
Fig. 7.5 Water temperature in the symmetry plane of the channel

Fig. 7.6 Water superficial velocity shown in the symmetry plane of the channel

The temperature profiles at the horizontal measurement plane just below the water surface are
shown in Fig. 7.7 for different condensation models. There is a constant rise of water temperature
towards the outlet of the channel due to heat and mass transfer from the steam side. The effect of
the latent heat release of condensation is tested by calculating different direct contact condensation
models. All the results of different DCC models show reasonable good agreement with the
experiment. Vertical temperature profile at three horizontal positions (x=0.25 m, x=0.6 m and x=1 m)
at the Fig. 7.8 show, that the water temperature is constantly rising towards the outlet of the
channel and the vertical temperature difference in the water is rather low. This is because of the
turbulent mixing of the water due to recirculation areas. Fig. 7.9 shows the calculated vertical
Steam/Water HTC at position x=1 m for the HD condensation model. The highest values occur at the
free surface.

Fig. 7.7 Temperature profile of different models Fig. 7.8 Vertical temperature profile at 3
at the measurement position T3 horizontal positions (HD model)

19
Fig. 7.9 Vertical Steam/Water Heat transfer Fig. 7.10 Vertical Steam/Water Interphase mass
coefficient at position x=1 m (HD model) transfer rate at position x=1 m (HD model)

Because of the AIAD framework the different HTC are possible for different flow morphologies. Fig.
7.10 shows the vertical Steam/Water interphase mass transfer rate at position x=1 m of the HD
condensation model. It is evident that almost all the interphase mass transfer is taking place at the
free surface.

In future it is planned to measure interfacial shear, water turbulence and velocity. Without these
data there is still a risk of error compensation in the direct contact condensation models.

Summary and Outlook

Direct Contact Condensation in stratified steam-water flow was calculated with ANSYS CFX using the
AIAD model. A few consistent sets of model correlations for the heat transfer within the AIAD
framework at the free surface were formulated and tested. The calculated temperature profile
agrees well with the experiment. Further model development should focus on the turbulence
treatment of the free surface. For direct contact condensation model validation also experiments at
different flow conditions are necessary. In future velocity and turbulence profile will be measured.
Also the use of direct numerical simulations are planned to further improve the model correlations.

Nomenclature
a surface area (m2)

𝑎𝐷 , 𝑎𝐵 blending coefficients for droplets and bubbles

A interfacial area density (1/m)

𝐶𝐷 drag coefficient

d diameter (m)

𝑓𝑗 blending function

𝐹𝐷 drag force (N)

g gravitational acceleration (m/s2)

h heat transfer coefficient, HTC (W/m2 K)

20
H specific enthalpy (J/kg)

L length scale at the interface (m)

m mass flux due to phase transition (kg/s)

⃗
n normal vector at the free surface

p pressure (Pa)

q heat flux (W/m²)

Q rate of heat transfer (W)

t time (s)

U velocity (m/s)

x axial distance from the water inlet (m)

y vertical distance from the bottom of the channel (m)

z horizontal cross-stream distance from the wall of the channel (m)

 volume fraction

 mass generation

 interfacial dissipation

 density (kg/m3)

𝜆 thermal conductivity (W/m K)

 shear stress (Pa)

Subscripts

B bubble
D drag
FS free surface
G gas
i interface
k phase gas or liquid
L liquid
Turb turbulent
W wall

Literature
Anderson, D. M., McFadden, G. B., Wheeler, A. A. “Diffuse-interface methods in fluid mechanics”
Annual Review of Fluid Mechanics, 30, 139–65 (1998).

21
Bartosiewicz, Y.; Seynhaeve, J.-M.; Vallee, C.; Höhne, T.; Laviéville, J.
Modeling free surface flows relevant to a PTS scenario: comparison between experimental data and
three RANS based CFD-codes - Comments on the CFD-experiment integration and best practice
guideline, Nuclear Engineering and Design 240(2010), 2375-2381

Brackbill, J., D.B. Kothe, and C. Zemach, A continuum method for modeling surface tension. Journal of
computational physics, 1992. 100(2): p. 335-354.

Celata, G. P., Cumo, M. Farello, G.E., Focardi, G., “Direct contact condensation of steam on slowly
moving water”, Nuclear Engineering and Design, 96, 21-31 (1986). CFX-10 Documentation, Solver
Theory, Multiphase Flow Theory, ANSYS, 2006.

Chan, T. S., Yuen, M. C., “The effect of air on condensation of stratified horizontal concurrent steam
water-flow”, Journal of heat transfer, 112, 1092-1095. (1990).

Coste, P., 2004, Computational simulation of multi-D liquid –vapour thermal shock with
condensation, 5th International Conference on Multiphase Flow (ICMF'04), paper no. 420

Coste et al., 2012, Validation of the large interface method of NEPTUNE_CFD 1.0.8 for pressurized
thermal shock applications, Nuclear Engineering and Design, 253, pp. 296-310

Egorov., Y., Validation of CFD codes with PTS relevant test cases. Technical Report EVOL-ECORA-D07,
ANSYS, Germany, 2004.

Gasiunas et al., Interfacial Shear of Co-Current Steam–Water Flow Estimation, Heat Transfer
Research 43(5), 1–18 (2012)

Griffith, P., “Screening reactor steam/water piping systems for water hammer”, NUREG/CR-6519,
U.S. Nuclear Regulatory Commission (1996).

Hein, D., Ruile, H., Karl, J., “Kühlmittelerwärmung bei Direktkontaktkondensation an horizontalen
Schichten und vertikalen Streifen zur Quantifizierung des druckbelasteten Thermoschocks“, BMFT-
Forschungsvorhaben 1500906, Abschlussbericht, Lehrstuhl für Thermische Kraftanlagen, TU
München, Germany. (1995).

Höhne, T. and C. Vallée, Experiments and numerical simulations of horizontal two-phase flow
regimes using an interfacial area density model. The Journal of Computational Multiphase Flows,
2010. 2(3): p. 131-143.

Höhne, T. and J.-P. Mehlhoop, Validation of closure models for interfacial drag and turbulence in
numerical simulations of horizontal stratified gas–liquid flows. International Journal of Multiphase
Flow, 2014. 62(0): p. 1-16.

Höhne, T.; Deendarlianto; Lucas, D., 2011. Numerical simulations of counter-current two-phase flow
experiments in a PWR hot leg model using an area density model. Int. J. Heat Fluid Flow 32, Issue 5,
1047-1056.

Hughes, E. D., Duffey, R. B. “Direct contact condensation and momentum-transfer in turbulent

separated flows”, International Journal of Multiphase flow 17, 599-619. (1991).

22
Kim, H. J., Lee, S. C., Bankoff, S. G. “Heat transfer and interfacial drag in countercurrent steam-water
stratified flow”, International Journal of Multiphase Flow 11, 593-606. (1985).

Lakehal, D. and Labois, M., A New Modeling Strategy for Phase-change Heat transfer in Turbulent
Interfacial Two-phase Flows. Internal Journal of Multiphase Flow, 37:627–639, 2011.

Lakehal, D., Fulgosi, M., Yadigaroglu, G., Banerjee, S., “Direct numerical simulation of turbulent heat
transfer across a mobile, sheared gas-liquid interface”, Journal of heat transfer, 125, 1129-1139
(2003).

Lim,J. S., Tankin, R. S., Yuen, M. C., “Condensation measurement of horizontal cocurrent steam-water
flow”, Journal of Heat Transfer, 106, 425-432 (1984).

Lucas, D., Coste, P., Höhne, T., Lakehal, D., Bartosiewicz, Y., Bestion, D., Scheuerer, M., Galassi, M. C.,
CFD modeling of free surface flow with and without condensation, Multiphase Science and
Technology 23(2011), 253-342

Mandhane, J. M., Gregory, G. A., and Aziz, K., "A Flow Pattern Map for Gas-Liquid Flow in Horizontal
Pipes" Int. J. Multiphase Flow, 1, 537-553 (1974).

Menter, F. R. “Two-Equation Eddy-Viscosity Turbulence Models for Engineering Applications,” AIAA

Journal, vol. 32, no. 8, pp. 1598-1605, 1994

Mikielewicz, J., Trela, M., Ihnatowicz, E., “A theoretical and experimental investigation of direct
contact condensation on a liquid layer”, Experimental Thermal and fluid Science 15, 221-227 (1997).

Olsson, E. and Kreiss, G., 2005. A conservative level set method for two phase flow. J. Comput. Phys.
210, 225-246.

Porombka, P., Höhne, T., Drag and turbulence modelling for free surface flows within the two-fluid
Euler-Euler framework, Chemical Engineering Science 134(2015), 348-359

Ranz, W. and W. Marshall, Evaporation from drops. Chem. Eng. Prog, 1952. 48(3): p. 141-146.

Sato, Y. and Niceno, B.. A sharp-interface phase change model for a mass-conservative interface
tracking method. Journal of Computational Physics, 249:127–161, 2013.

Scardovelli, R., Zaleski, S., “Direct numerical simulation of free-surface and interfacial flow”, Annual
Review of Fluid Mechanics, 31, 567-603 (1999).

Scheuerer, M., Galassi, M. C., Coste, P., D’Auria, F., “Numerical simulation of free surface flows with
heat and mass transfer”, The 12th International Topical Meeting on Nuclear Reactor Thermal
Hydraulics (NURETH-12), Pittsburg, Pennsylvania, U.S.A., September 30-October 4 (2007).

Smereka, P., Sethian, J.A., “Level set methods for fluid interfaces”, Annual Review of Fluid Mechanics
35, 341-372 (2003).

Štrubelj, L., Ézsöl, G. and Tiselj, I., Direct Contact Condensation Induced Transition from Stratified to
Slug Flow. Nuclear Engineering and Design, 240:266–274, 2010.

23
 Štrubelj, L.; Tiselj, I.; Mavko, B., 2009. Simulations of free surface flows with implementation of
surface tension and interface sharpening in the two-fluid model. Int. J. Heat Fluid Flow 30, 741-750.

Taitel, Y., Dukler, A.E., “Model for predicting flow regime transitions in horizontal and near horizontal
gas-liquid flow”, AICHE Journal, 22, 47-55 (1976).

Vallée, C.; Höhne, T.; Prasser, H.-M.; Sühnel, T., 2008. Experimental investigation and CFD simulation
of horizontal stratified two-phase flow phenomena. Nucl. Eng. Des. 238, 637-646.

Von Arx, W.S., “Oceanus”, Energy: Natural limits and abundances, 17, 2-13 (1974).

Welch, S. W. J. and Wilson, J., A Volume of Fluid Based Method for Fluid Flows with Phase Change.
Journal of Computational Physics, 160:662–682, 2000.

Wilcox, D. C. 1994. Turbulence modelling for CFD. La Cañada, California: DCW Industries Inc.

Yadigaroglu, G., “Computational fluid dynamics for nuclear applications: from CFD to multi-scale
CMFD”, Nuclear Engineering and Design, 235, 153-164 (2005).

Yao, W., Coste, P., Bestion, D., Boucker, M., “Two-phase pressurized thermal shock investigations
using a 3D two-fluid modeling of stratified flow with condensation”, The 10th international Topical
Meeting on Nuclear Reactor Thermal Hydraulics (NURETH-10), (2003).

Zwart, P. J.; Burns, A. D.; Galpin, P. F., 2007. Coupled Algebraic Multigrid for Free Surface Flow
Simulations. Proc. 26th Int. Conf. on offshore mechanics and arctic engineering, OMAE2007-29080.

24

View publication stats