18th Brazilian Congress of Thermal Sciences and Engineering

November 16-20, 2020 (Online)

ENC-2020-0505
MODEL VALIDATION OF A FALLING FILM EVAPORATION PROCESS
ON A VERTICAL TUBE USING SUBCOOLED WATER
M. C. Buss, J. C. Passos
Universidade Federal de Santa Catarina, LEPTEN-Laboratórios de Engenharia de Processos de Conversão e Tecnologia de Energia,
Departamento de Engenharia Mecânica, 88.040-970 – Florianópolis-SC
[email protected], [email protected]

S. C. Janssen, B. P. M. van Esch

Eindhoven University of Technology, Power and Flow Group, Department of Mechanical Engineering, Gemini-Zuid, Groene Loper,
15, 5612AZ, Eindhoven, The Netherlands.
[email protected], [email protected]

Abstract. Falling Film Evaporators (FFE) are used by a wide range of industrial sectors because of their high heat
transfer rate per unit of surface area while maintaining a relatively low temperature difference between surface and
liquid. This allows more compact system designs and thereby reduces material costs. Although they have been used for
decades the design of these evaporators is still mostly based on estimates and empirical data. The available literature is
reviewed as foundation for the mathematical model made for a vertical FFE design using subcooled water. Dimensionless
numbers are used to fundamentally compare its results with literature. Additionally, simulations are performed to validate
the model. An experimental setup of a vertical FFE was built in LEPTEN, at UFSC. An overview of its construction and
improvements is provided. This setup is used as a basis for the model, which means it can be used to validate the theory
in future works.
Keywords: Vertical Falling Film Evaporator, heat and mass transfer, Modelica modeling, subcooled water, evaporation

1. INTRODUCTION

Falling Film Evaporators (FFE) are used by a wide range of industrial sectors because of their high heat transfer rate
per unit of surface area while maintaining a relatively low temperature difference between surface and liquid. This allows
more compact system designs and thereby reduces material costs. FFE are used by the food sector (dairy products or
fruit juices for example) (Gourdon et al. (2015)), for water desalination, nuclear engineering (Huang et al. (2015)) among
others. Although they have been used for decades the design of these evaporators is still mostly based on estimates and
empirical data. Therefore, mathematical models should be created and validated to be able to better design a FFE. With
the aim of providing these models within an open source code this research uses Modelica as a programming language.

Modelica (Modelica Association (2007)) shows to be suitable software for modelling thermal systems because of its
component-based equation solver and its ability to implement physical equations independent from their numerical so-
lutions (Assaf et al. (2011)). de la Calle et al. (2012) successfully implemented a full horizontal FFE in Modelica with
submodels to describe the physical behaviour of each component. However, there is no implementation of a vertical
FFE model in Modelica yet. Since some industrial sectors use the evaporator in a vertical orientation it is important to
develop this mathematical model as well. This research aims to create this model and compare it with experimental results.

To obtain these results an experimental setup of a vertical FFE was built in the Laboratory of Energy Conversion
Engineering and Energy Technology, LEPTEN, at the Federal University of Santa Catarina, UFSC. An overview of its
construction and improvements is presented, following the works of de Souza (2018) and Jansen (2019) on aspects such
as the inflow distribution and direction, thermocouple placement for temperature measurements and modifications to in-
crease the surface area using grooves. Additionally the authors describe their work on the setup regarding improved
visualisation for future measurements.

Due to the Covid-19 outbreak, LEPTEN was closed so no experimental results could be obtained. The film thickness
is predicted with a simulation model. It is compared with experimental data of other authors and available theory.
M.C. Buss, S.C. Janssen, J.C. Passos, B.P.M. van Esch
Model Validation of a Falling Film Evaporation process on a vertical tube using subcooled water.

2. THEORY

The heat flow rate (q) for convection as described by equation 1 is a measure that helps to describe the performance
of a FFE by indicating the amount of heat exchange done by the FFE. It takes into account the heat transfer coefficient
(h), surface area (A), and the difference between the wall temperature (Tw ) and film temperature (Tf ).

q = hA∆T = hA(Tw − Tf ) (1)

In order to improve the performance, the heat flow rate should be increased as much as possible. There is a limit in
increasing ∆T because of temperature restrictions for the fluids used (milk, juice etc.). This work will mainly focus on
the evaluation of the convective heat transfer coefficient for a fixed area and a fixed heat flow by the heating element. The
variable h is dependent of the mass flow rate and thereby of the Reynolds number.
The Reynolds number is a dimensionless parameter used for evaluating the flow regime, usually having transition
values between the regimes (e.g. laminar to turbulent). Effectively it expresses the ratio between inertial and viscous
forces. It considers the mass flow rate (ṁ), dynamic viscosity(µ) and tube diameter (D). For a cylindrical thin film the
Reynolds number is defined as shown in equation 2. The film thickness defines the characteristic length and is included
within the mass flow rate.

4Γ 4ṁ
Re = = (2)
µ µπD
Transition values for Reynolds are widely available in literature, but mutual agreement between sources is rare. Some
authors provide discrete values, others calculate it based on dimensionless parameters, those being either Prandtl or
Kapitza. Some authors consider only the laminar and turbulent flow regimes, some take into account a wavy regime
between the two. It is also common to define a transition region between regimes. Several empirically obtained regime
transitions found in literature are listed in table 1.

Table 1: Reynolds transition values for a thin film according to literature.

Based on Prandtl number Based on Kapitza number
Authors Value Flow regime Authors Value Flow regime
56 Transition 1
Chun and Seban (1971) 5800/P r1.06 Turbulent
263 Wavy
= 1810 Ishigai et al. (1972)B
300 Transition 2
1600 Turbulent
Wilke (1962) apud Aviles 2460/P r0.65 Turbulent
(2007) = 1204 36 Transition 1
60 Wavy
Discrete value Al-Sibai (2004)B
1150 Transition 2
Authors Value Flow regime 3915 Turbulent
Ticona (2003) apud 25 Wavy
Teleken (2013) 1000 Turbulent
Nusselt (1916) apud
1600 Turbulent
Teleken (2013)
47 Wavy
Baehr and Stephan
611 Transition 2
(2006)AB
1600 Turbulent
A
Authors are using condensing films
B
Authors define Reynolds as Re = ṁ/ (µb) so the values have been multiplied by 4

The Prandtl number describes the ratio of momentum diffusivity over heat diffusivity and consists only of fluid
properties that are either temperature or pressure dependent (dynamic viscosity, specific heat capacity (cp ) and thermal
conductivity (k)). Equation 3 shows that relation.
ν (µ/ρ) µcp
Pr = = = (3)
α (k/ρcp ) k
Additionally, the Kapitza number expresses the ratio between the surface tension forces and viscous forces on the
 18th Brazilian Congress of Thermal Sciences and Engineering
November 16-20, 2020 (Online)

waves along the falling film. The Kapitza number is a function of fluid properties (surface tension (σ), density (ρ) and
kinematic viscosity (ν)) and gravity’s acceleration. The dimensionless number is thereby interesting to evaluate the film
uniformity for different fluids. Equation 4 is used to calculate it.

σ
Ka = (4)
ρg 1/3 ν 4/3
The film thickness (δ) and the convective heat transfer coefficient are inversely proportional according to the Nusselt
number (equation 7), which means that an increase in the film’s thickness causes a decrease in the heat transfer coefficient
and vice versa. So it’s desirable that the film remains as thin as possible. However when the film is too thin any perturba-
tion can severely compromise the film’s integrity and uniformity and thereby originate undesirable dry patches. Therefore
it is important to correctly model the film thickness.
In practice most falling films are outside the laminar regime, having waves and turbulence effects on the film. Beyond
that, even on laminar films thickness measurements are difficult to make. Therefore several empirical correlations are
available in literature to estimate the thickness. These correlations can be generalized by equation 5. The coefficients ci
have different values for each author as shown in table 2. Equation 5 shows that the film thickness tends to increase as Re
increases, as long as c3 is positive.

c2
v2

δ = c1 Rec3 Reynolds range: c4 (5)
g

Table 2: Coefficient values for film thickness calculation.

Authors c1 c2 c3 c4
1.4422B 1/3 1/3 Re<400
Stephan et al. (2010)A
0.4356 1/3 8/15 Re>400
Nusselt (1916), Chun and Seban 0.9086 1/3 1/3 Re<1600
(1971), Teleken (2013) 0.2628 1/3 1/3 Re>1600
A
Equation for turbulent regime is empirically obtained and agrees well
with correlations from 4 papers cited in this source.
B
Allowance for surface waves in the laminar regime can be made by
changing this value to 2.4 according to Kapitza (1948, 1949).

In figure 1, the thickness equations are compared with experimental results by Zhang et al. (2000), who used an
optical-electronic method to measure the film thickness on a vertical tube falling film.

10-3
1.5 Nusselt, Chun, Teleken Zhang
[m]

Modelica laminar Stephan

Film thickness

0.5

0
1000 2000 3000 4000 5000 6000 7000 8000
Reynolds number [-]
Figure 1: Film thickness δ for different Reynolds numbers. Results by Zhang et al. (2000) are close to Stephan et al.
(2010). There is a discontinuity in the equations by Nusselt (1916), Chun and Seban (1971) and Teleken (2013). The
Modelica model is based on their laminar equation for the full Reynolds range.

Figure 1 shows a major difference between the two pieces of literature. In order to verify a suitable relation for the
thickness over the Reynolds number, additional experiments should be performed and fitted to a curve. That fit can be
used for the final Modelica model.
M.C. Buss, S.C. Janssen, J.C. Passos, B.P.M. van Esch
Model Validation of a Falling Film Evaporation process on a vertical tube using subcooled water.

Weber’s number expresses the ratio between inertial forces over surface tension. If the falling velocity gets too high,
the inertial forces become dominant resulting in an unstable film. Therefore the Weber number dictates a maximum flow
rate for a stable film based on the fluid properties and the film thickness. It is defined in equation 6.
ρν 2 δ
We = (6)
σ
When experimental results yield an appropriate relation (curve fit) for δ, the Weber number can be calculated over
the Reynolds domain. By comparing the uniformity during experiments and the related Weber numbers, a critical Weber
number can be obtained. It can be used in Modelica to find flow rate restrictions.
The Nusselt number is a dimensionless parameter describing the ratio between the convective and conductive heat
transfer in a fluid. The Nusselt number is stated in equation 7.
hδ N uk
Nu = =⇒ h = (7)
k δ
Many empirical correlations have been proposed to allow the determination of the Nusselt number with only the
knowledge of Re and P r as denoted by equation 8. These correlations are determined by the value of empirically
obtained constants (c1 , c2 and c3 ) and are valid for a certain condition (A flow regime or Reynolds number).

N u = c1 Rec2 P rc3 for Condition (8)

The simulations allow to calculate the values for the thickness and the heat transfer coefficient independently from the
Nusselts number. That means that the final results don’t rely on empirically obtained data. Among the literature several
constants have been proposed. Table 3 shows some of the correlations that were used to compare the data. It contains the
constants and the conditions to be valid, as defined by the authors.

Table 3: Coefficient values and their conditions to be valid for Nusselt cal-
culation
Authors c1 c2 c3 Condition
1.43 −0.33 0 LaminarA
Stephan et al. (2010) 0.0425 0.2 0.344 TransitionA
0.0136 0.4 0.344 Turbulent flowA
0.82 −0.22 0 Re < 1890
Chun and Seban (1971)
0.0038 0.4 0.65 Re > 1890
0.069 −0.33 0 Re < 155
Baehr and Stephan (2006)
0.0325 0.25 0.5 Re > 155
A
The authors state that the highest Nu value per Reynolds number of the
three equations should be used.

Pr = 3
Nusselt number [-]

0.6

0.4

0.2 Stephan et al.

Baehr and Stephan
Chun and Seban
0
0 1000 2000 3000 4000 5000 6000 7000 8000
Reynolds number [-]
Figure 2: Nusselt number for different Reynolds numbers according to Stephan et al. (2010), Baehr and Stephan (2006)
and Chun and Seban (1971) for water at 60◦ C and ambient pressure. The Nusselt number from Baehr and Stephan (2006)
shows a sudden jump at its Reynolds transition number, while the one from Stephan et al. (2010) shows a sharp transition
point.
 18th Brazilian Congress of Thermal Sciences and Engineering
November 16-20, 2020 (Online)

The convective heat transfer coefficient measures the effectiveness of the heat exchanger. By rewriting equation 1
equation 9 can be obtained.
q q
h= ; ∆T = Tw − Tf ; =⇒ h = (9)
A∆T A(Tw − Tf )
Also from inserting data from equations 5 and 8 into equation 7 it is possible to calculate h without having the
temperature data. With that figure 3 can be obtained from the available authors.

de Souza laminar HA Wilke turbulent

Heat transfer coefficient h [W/m2K]

10000 Modelica laminar de Souza Jansen

Chun and Seban
Pr = 3
8000

6000

4000

2000

0
0 1000 2000 3000 4000 5000 6000 7000
Reynolds number [-]
Figure 3: Heat transfer coefficient for different Reynolds numbers. All equations are plotted over the full domain to show
that the laminar equations of Chun and Seban (1971) are implemented on the Modelica model over the full Reynolds
range. de Souza (2018) uses the same equation for the turbulent regime as Chun and Seban (1971). Experimental results
from de Souza (2018) are close to the turbulent equation of Chun and Seban (1971), The results from Jansen (2019) are
closer to the theory of the Heat Atlas (HA) by Stephan et al. (2010). In experiments, Jansen (2019) used a grooved surface
instead of the smooth surface by de Souza (2018).

3. EXPERIMENTAL

In order to validate the simulations, experimental data are required. This data are obtained using the experimental
setup of a vertical FFE assembled in LEPTEN at the Mechanical Engineering Department of UFSC. The setup is shown
schematically in figure 4a.

(a) Schematic of the setup used: 1) Flowrate regulator by adjustable (b) Picture of most of the setup. In addition to figure 4a
tube, 2) Sponge disk and wire mesh to uniformly spread the water, lamps and a paper cover are added to fully diffusely light
3) Teflon cone to distribute the water over the stainless steel cylin- the FFE and thereby improve visualisation of the film.
der evenly. 4) Resistance wire used to heat the glass tube to pre-
vent condensation and thereby improve visualisation, 5) Variable
DC power supply, 6) Electrical cartridge resistance for heating the
FFE, 7) Power supply: JNG I1215426, 8) Thermal bath and pump.
Figure 4: Experimental setup in LEPTEN
M.C. Buss, S.C. Janssen, J.C. Passos, B.P.M. van Esch
Model Validation of a Falling Film Evaporation process on a vertical tube using subcooled water.

The outer metallic cylinder (D = 31.75mm, L=116mm) surface of the FFE itself has small laser machined grooves,
which increase the surface area by 3% when compared to a smooth tube. The grooves were an attempt by Jansen (2019)
to increase the wet contact area and thereby increase the amount of heat transfer according to equation 1. Additionally
he investigated varying the inflow angle using a funnel. That resulted in an improved distribution due to the tangential
flow but also caused splashes and the prototype was less reliable. The combination of a sponge disk and a wire mesh (as
mentioned in the caption of figure 4a) still provided the best result for the inflow.

Water is preheated to 60◦ C using a MQBMP-01 thermal bath from the company MicroQuimica Ltda.. The thermal
bath has a pump that is used for circulation through the system. The fluid is not saturated so there is no boiling (evapora-
tion only), preventing bubble generation that could cause hydrodynamic instability, interfering with the film’s uniformity.
An unstable film could lead to the formation of unwetted areas (dry patches) that reduce the heat transfer area with the
film. The mass flow rate is varied to evaluate heat transfer for different Reynolds numbers.

In order to collect temperature data, thermocouples attachment methods to the FFE have been evaluated. The method
to be adopted should take into account the fragile nature of the film’s uniformity and distort the measurements as little
as possible. de Souza (2018) tried using threadlocker adhesive to fix the thermocouples, but had some difficulties and
opted to use rivets instead. Jansen (2019) also looked into attachment methods and tried using epoxy glue, welding and
soldering, all of them without much success. His final solution was to use thermal tape, acknowledging that it was not an
optimal method and suggested that welding may be a viable alternative.
The use of a sealing silicon glue was tested and discarded due to a bad connection with the wall. The main options
considered were welding a thicker thermocouple and cyanoacrylate glue.
The entire cylinder is surrounded by a glass tube to reduce ambient interference. To prevent condensation and enable
visualisation, the tube is heated using a resistance wire. The wire is fixed with heat resistant tape in a way that does not
interfere with front side visualization.

According to Zhang et al. (2000) the measurement of the falling liquid film thickness is of critical importance but
complicated to be accurately obtained. Studies have been done to measure the thickness with probes that touch liquid’s
surface. This approach has been used by Härkonen (1994), whilst acknowledging that the method is invasive and could
lead to interference with the quantity to be measured. Therefore Zhang et al. (2000) devised a method using lasers to
measure the thickness without interfering with the film’s flow. Aviles (2007) stated a table with an overview of multiple
intrusive and non-intrusive methods. For this experimental setup, it is intended to use an optical method in the future to
visualise and measure the film thickness.

4. Modeling

Within Modelica there are several native libraries available. Three of them were mainly used. Modelica.Fluid was
used to implement a fluid flow into and out of the model, as well as setting the ambient temperature and pressure. Model-
ica.Media was incorporated to model the working fluid and make it possible to switch to other fluids. Modelica.Thermal
was the last one to be implemented, it provides the heat source for the system. Those three libraries work with the FFE
block created for this work.
The models’ assumptions are laminar flow, no shear stress on the liquid-vapour interface, axisymmetric flow, neg-
ligible heat transference from the surrounding vapour to the film and conduction is the main heat transfer phenomena,
whilst advection is negligible because the film is so thin. Also, due to complexity, wavy effects are not considered. The
hydrodynamic equations from Nusselt (1916) model on the film thickness are used because they are based on the same
assumptions as mentioned above. The only modification is that the film isn’t at saturation temperature, therefore the film’s
temperature is not constant, but increases over the vertical distance traveled. Because of this complication a numerical
approach to solve the equations is chosen. For any given height, the temperature is averaged over the radial distance
spanning the film, making the temperature a function of the height of the cylinder only.
The setup is assumed axisymmetric and thereby modeled 2D considering a vertical and radial direction. Using the
2D Navier-Stokes equation in the vertical direction (equation 10) and applying the assumptions above, equation 11 is
obtained. Applying the boundary conditions returns the velocity profile in the y direction, as shown in equation 12.
 18th Brazilian Congress of Thermal Sciences and Engineering
November 16-20, 2020 (Online)

∂2u ∂2u
   
z ∂u ∂u ∂u ∂p
y ρ +u +v =− + µl + 2 + ρl g (10)
∂t ∂x ∂y ∂x ∂x2 ∂y
x
L   2
1 ∂p y
u= − ρl g + c1 y + c2
µl ∂x 2
(
@ y = 0 : u = 0 ⇒ c2 = 0 ∂p
dx and = ρv g (11)
@ y = δ : ∂u
∂y = 0
∂x

y2
 
g
u(y) = (ρl − ρv ) δy − (12)
µl 2
Z Z δ
ρl bg
Figure 5: Schematic drawing of the models’ con- ṁ = ρl udA = ρl b udy =⇒ ṁ = (ρl − ρv ) δ 3 (13)
trol volume (dashed lines), assuming the film is A 0 3µl
thin enough in comparison to the tube diameter
to be flattened

The mass flow rate can be obtained by using equation 12 on the upper surface of the control volume shown in figure 5.
The result, equation 13, is the same as the one obtained by Nusselt and is one of the equations used on the Modelica model.
An equation for the turbulent regime was not obtained and the attempts to implement a finite elements on Modelica were
not successful. Figure 6 shows the energy balance on a slice of the control volume. Equation 14 is the energy balance
used by the Modelica model.

δ
y y

x ṁcp Tf x ṁcp Tf
L
Tw

Tf
q”dA dx hlv dṁ q”dA hlv dṁ

T∞

(ṁ − dṁ) cp (Tf + dTf ) (ṁ − dṁ) cp (Tf + dTf )

(a) Spatial coordinates for the energy balance. (b) Temperature profile on the control volume.
Figure 6: Energy balance drawings

ṁcp Tf + q”dA = (ṁ + dṁ) cp (Tf + dTf ) + hlv dṁ (14)

The last main equation used by the model is equation 1. Having those three equations (1, 13 and 14) and the proper
parameters the model is set to solve for thickness, evaporation rate and film temperature. The parameters used are the tube
perimeter (b), length (L), Reynolds number or mass flow rate, ambient pressure and temperature, fluid entrance temper-
ature, heat flow rate and working fluid. Using the equations mentioned and these input parameters, Modelica calculates
and provides the required output variables.

Data on horizontal FFEs was found in literature. For comparison, the Modelica model was adjusted. The definition
of the Reynolds number was changed to half of equation 2, the length of the tube in the model was considered half the
circumference of the horizontal tube and the total area was multiplied by two (two tube halves). Value b is the length of
the horizontal tube.
M.C. Buss, S.C. Janssen, J.C. Passos, B.P.M. van Esch
Model Validation of a Falling Film Evaporation process on a vertical tube using subcooled water.

5. RESULTS

This section first shows the behaviour of the modeled FFE over its length in figure 7. Additionally the model outcomes
are compared to data available in literature.

(a) Heat transfer coefficient along the x axis, showing an increase (b) Temperatures along the x axis. The ambient temperature T∞
as predicted by equation 9, considering constant heat flow and the is constant, matching the boundary condition. The temperature
increase in temperature shown in figure 7b. difference between wall and fluid is positive, resulting in heating
of the fluid. Due to the constant ambient temperature and the in-
creased fluid temperature, evaporation takes place (increasing over
the length).

(c) Mass flow rate (liquid) along the x axis is decreasing due to the (d) Film thickness along the x axis is decreasing due to evapora-
evaporated liquid. tion.

Figure 7: Results from the simulations in Modelica. The parameters are those of the LEPTEN experimental setup and Re
is set to 4780 at the fluid’s entrance. Additionally the heat supply was set to 284.6 W .

Figure 7 shows the expected behaviour from a FFE suggesting the model is functioning in an appropriate manner. The
absolute values should now be compared to literature and experiments for complete validation.

Narváez-Romo and Simões-Moreira (2017) obtained experimental data on a horizontal tube FFE which is compared
to a horizontal version of our FFE model in figure 8.
It is noted that the Modelica model has a higher slope of temperature along the traveled distance, even though the
average remains close on both data sets. This could be due to the increased influence of gravity in the vertical tube as
opposed to the horizontal one. The average could make it so the effect of the curvature evens out along the upper and
lower halves. From the comparison in figure 8 it is interesting to note that the Modelica model always returns a higher
value than the ones computed by the authors.
 18th Brazilian Congress of Thermal Sciences and Engineering
November 16-20, 2020 (Online)

Figure 8: Comparison between average temperatures on the simulated model and experimental data on a horizontal tube
for different Reynolds numbers. Experimental data was available for both 50 and 70 degrees Celsius. Thereby there were
two simulations performed in Modelica using these inflow temperatures.

6. DISCUSSION

There is an experimental difference between the works of Jansen (2019) and de Souza (2018), noted both on the setup
and acquired data. Whilst de Souza used a smooth vertical tube Jansen made several longitudinal grooves on the tube in
an attempt to increase surface area and thereby increase the heat transfer. The effective area increase was 3% at most,
showing little significance. Therefore the grooves were not considered on the final model. The large difference between
experimental results obtained by both authors on similar setups is mostly unexplained. Another interest on the grooved
surfaces was to increase film stability and prevent dry patches. Unfortunately the model does not take into account film
stability and experimental data couldn’t be obtained, therefore it was not feasible to evaluate this possible effect.
From figure 3 it can be observed that the model behaves like the laminar one from Chun and Seban (1971) on the
full regime despite the water is not at saturation temperature. The correct correlation for the turbulent regime should be
empirically determined. Unfortunately due to the Covid-19 pandemic these experiments were interrupted.

For future works it is recommended to:

• Finish setting up the high speed camera at LEPTEN for measurements on film thickness and stability;
• Attach thermocouples to the tube so that the model can be evaluated for its accuracy;
• Develop an approach with less assumptions to solve the problem, so that the turbulent regime can be better repre-
sented;
• Evaluate the impact of the grooves on film stability.
7. ACKNOWLEDGEMENTS

Thanks to professor Simões-Moreira and PhD student Narváez-Romo from Poli-USP for their availability and coop-
eration. Thanks to CNPQ and PIBIC for their support to the scientific community.

8. REFERENCES

Al-Sibai, F., 2004. Experimentelle Untersuchung der Strömungscharakteristik und des Wärmeübergangs bei welligen
Rieselfilmen. Ph.D. thesis, RWTH, Aachen, Germany.
Assaf, K., Zoughaib, A. and Clodic, D., 2011. “Modelica-based modelling and simulation of dry-expansion shell-and-
tube evaporators working with alternative refrigerant mixtures”. In International Journal of Refrigeration. Elsevier,
Vol. 34, pp. 1471–1482. ISSN 01407007. doi:10.1016/j.ijrefrig.2011.03.010.
Aviles, M.L., 2007. Experiments on falling film evaporation of a water-ethylene glycol mixture on a surface with longitu-
dinal grooves. Ph.D. thesis, Technischen Universitat Berlin, Berlin, Germany.
Baehr, H.D. and Stephan, K., 2006. Wärme- und Stoffübertragung. Springer-Verlag, Berlin, 5th edition.
Chun, K.R. and Seban, R.A., 1971. “Heat transfer to evaporating liquid films”. Journal of Heat Transfer, Vol. 93, pp. 391
– 396.
de la Calle, A., Yebra, L.J. and Dormido, S., 2012. “Modeling of a falling film evaporator”. In Proceedings of the 9th
International Modelica Conference. Munich, Germany.
de Souza, J.N., 2018. Análise experimental do processo de evaporação em película descendente sobre tubo vertical.
Master’s thesis, Universidade Federal de Santa Catarina, Florianopolis, Brasil.
Gourdon, M., Innings, F., Jongsma, A. and Vamling, L., 2015. “Qualitative investigation of the flow behaviour during
falling film evaporation of a dairy product”. Experimental Thermal and Fluid Science, Vol. 60, pp. 9 – 19. doi:
https://doi.org/10.1016/j.expthermflusci.2014.07.017.
M.C. Buss, S.C. Janssen, J.C. Passos, B.P.M. van Esch
Model Validation of a Falling Film Evaporation process on a vertical tube using subcooled water.

Huang, X.G., Yang, Y.H. and Hu, P., 2015. “Experimental study of falling film evaporation in large scale rectangular
channel”. Annals of Nuclear Energy, Vol. 76, pp. 237 – 242. doi:https://doi.org/10.1016/j.anucene.2014.09.053.
Härkonen, M., 1994. “Heat transfer to evaporating falling films of water and sugar/water solution”. Wärme- und Stof-
fübertragung, Vol. 29, pp. 349 – 355.
Ishigai, S., Nakanisi, S., Koizumi, T. and Oyabi, Z., 1972. “Hydrodinamics and heat transfer of vertical falling liquid
films”. Bull JSME, Vol. 15, pp. 594–602. doi:https://doi.org/10.1299/jsme1958.15.594.
Jansen, T.D., 2019. Water Falling Film Evaporation on a Vertical Tube. Master’s thesis, Eindhoven University of Tech-
nology, Eindhoven, The Netherlands.
Kapitza, P.L., 1948, 1949. “Wellenstromung der dunnen schichten einer viskosen flussigkeit”. J Exp Theoret Phys
(UdSSR, p. 18(1):3–18; 19(2):105–120.
Modelica Association, 2007. “Modelica specification”. Modelica. 24 june 2020 <http://www.modelica.org/documents>.
Narváez-Romo, B. and Simões-Moreira, J.R., 2017. “Falling liquid film evaporation in subcooled and satu-
rated water over horizontal heated tubes”. Heat Transfer Engineering, Vol. 38:3, pp. 361 – 376. doi:
https://doi.org/10.1080/01457632.2016.1189275.
Nusselt, W., 1916. Die Oberflachenkondensation des Wasserdampfes. Z. Vereinsdeutscher Ininuere, 60th edition.
Stephan, P., Kabelac, S., Kind, M., Martin, H., Mewes, D. and Schaber, K., 2010. VDI Heat Atlas. Springer-Verlag Berlin
Heidelberg, 2nd edition. doi:https://doi.org/10.1007/978-3-540-77877-6.
Teleken, J.G., 2013. Modelagem matemática e análise fluidodinâmica do processo de destilação por filme líquido de-
scendente. Ph.D. thesis, Universidade Federal de Santa Catarina, Florianopolis, Brasil.
Ticona, E.M., 2003. Determinação experimental do coeficiente de troca de calor em um gerador de "pasta de gelo".
Master’s thesis, Pontificia Universidade Catolica do Rio de Janeiro, Rio de Janeiro, Brasil.
Wilke, W., 1962. “Wärmeübergang an rieselfilme”. VDI-Forschungsheft, Vol. 490.
Zhang, J.T., Wang, B.X. and Peng, X.F., 2000. “Falling liquid film thickness measurement by an optical-electronic
method”. Review of Scientific Instruments, Vol. 71, pp. 1883 – 1886. doi:https://doi.org/10.1063/1.1150557.

9. RESPONSIBILITY NOTICE

The authors are solely responsible for the printed material included in this paper.