Mini-Project Report

Computational Modelling of Boiling of

a Liquid

Nikhil Mahalingesh
Niranjan E
Prajwal S
(Batch 16)
Fifth Semester (B.E)

Department of Mechanical Engineering

PESIT, Bangalore

Principal Guide:
Dr.K.S.Ravichandran
Chair Professor
Computational Fluid Dynamics
PES University
Table of Contents
1. Scope.............................................................................................................................. 1
2. Introdution.................................................................................................................... 1
3. Overview of the Project................................................................................................2
4. Review of literature...................................................................................................... 3
5. Geometric Modelling ................................................................................................... 4
6. Methodologies...............................................................................................................5
6.1. Monitors..................................................................................................................6
7. Mixture Model..............................................................................................................7
7.1. Heating Cycle..........................................................................................................7
7.2. Investigation Carried Out........................................................................................7
7.2.1. Condensation frequency...............................................................................7
7.2.2. Saturation temperature..................................................................................9
7.3. Cooling Cycle........................................................................................................15
8. VOF Model..................................................................................................................16
8.1. Heating Cycle........................................................................................................16
8.2. Cooling Cycle........................................................................................................18
9. Results and Discussions..............................................................................................19
10. Summary and Conclusions........................................................................................19
11. References...................................................................................................................20
 Page |1

1. Scope
In today’s world, many engineering systems of study are complex and non-linear, for which
simple, intuitive analytical solutions are not readily available. Deriving mathematical analytical
solutions to the problem can be tedious and overwhelming. Hence experimentation with the model
is done by computational modelling. Computational modelling uses mathematics, physics and
computer science with extensive computational resources to study the behavior of complex
systems by computer simulation. One such complex system of study is boiling of a liquid which
is a multiphase phenomenon. The project in question involves computational modelling of boiling
using ANSYS Fluent. The aim of this project is to simulate and monitor the behavior of
evaporation and condensation of water during repeated cycles of heating and cooling inside a
closed vessel.

2. Introduction
In ANSYS Fluent, there are three different approaches for simulating multiphase flow
phenomena i.e. the Volume of Fluid (VOF) model, the Mixture model and the Eulerian model [1].
The phase change method fundamentally depends on the properties and capabilities of its
underlying multiphase methods. The primary concern of this project is with the Mixture and the
VOF models and limiting the digression of the numerical solution from actual physics. Hence, the
fundamental purpose of this project is to explore and study the possible differences in the results
obtained from the VOF and Mixture models. The effects of Condensation frequency and Saturation
temperature on the numerical solution is also studied.

In order to simulate boiling of water one has to take into consideration the presence of liquid
and vapor states and the mass transfer between them. To achieve this, boiling is simulated inside
a simple closed 3D cylinder, half filled with liquid water. As the system is being heated, the
temperature of the liquid increases. This causes a change of phase of the liquid i.e. liquid-water
converts to water-vapor, thus increasing the pressure inside the cylinder. After critical conditions
are attained, the conditions of the system are sustained for a certain amount of time, before cooling.
 Page |2

3. Overview of the Project

 Page |3

4. Review of literature
Predicting precisely the highly transient interface phenomenon due to mass transfer coupled
with phase change due to evaporation is the major challenge encountered in modelling the boiling
phenomenon. Very little detailed research has been conducted on the subject matter due to its
undeniable complexity.

Reby Roy [2] has investigated higher self-pressurization in a 2D cylindrical vessel, three fourth
filled with water and one fourth filled with water vapor. The numerical simulations presented were
based on the VOF multiphase method in conjunction with the evaporation-condensation mass
transfer model. The study of self-pressurization rates with phase change was the main objective of
their project. Also the simulations included in that paper are for very short durations of time.

Film boiling process at the vicinity of a hot metal plate was studied by Timo Kuliju [3]. A VOF
model with realizable k-epsilon turbulence model was adopted for their study, in which the results
were found to correspond quite well the expected overall heat transfer coefficients.

To learn more about simulating change of phase by evaporation, Moon Soo lee’s [4] paper on
two phase simulation methods proved helpful. The development history and the numerical
simulation of multiphase and phase changing flow is discussed with a background review on
multiphase methods.
 Page |4

5. Geometric Modelling
The 3D cylinder modelled using ICEM CFD is of the following specifications:

Diameter 100mm
Height 100mm
Type of element Tetra/Mixed
Number of elements 10868
Number of fluid cell zones 2

The cylinder has two fluid zones separated by an interface, hence the whole cylinder is divided
into two equal halves: the lower half corresponding to liquid-water and the upper half
corresponding to water-vapor. While meshing the geometry, the lower half was meshed first
corresponding to liquid-water as the primary phase and subsequently the geometry was translated
to the upper half to mesh the upper volume which corresponds to the water-vapor, the secondary
phase. The lower and upper halves of the mesh are interconnected by introducing an interface
between them. Tetra/mixed elements were used for the volume mesh and the number of elements
was limited to about 10,000 so that running calculations is feasible on laptop computers.
 Page |5

6. Methodologies
The following methodologies were adopted after the finalized mesh file was read in ANSYS
Fluent:

Solver Pressure-Based, Transient

Models Standard k-ε, Standard Wall Functions, Multiphase – Mixture and VOF
Phases Primary – Liquid Water
Secondary – Water Vapor
Phase Interaction Mass Transfer from water phase to vapor phase
Mass transfer mechanism: Evaporation-Condensation
Boundary Conditions Bottom: Uniform wall temperature of 1173 K
Top: Wall with Mixed thermal conditions of 350 K free
Heating cycle stream temperature
Vapor-wall: Wall with Mixed thermal conditions of 350 K
free stream temperature
Water-wall: Uniform wall temperature of 1173 K
Bottom: Wall with Mixed thermal conditions of 300 K free
stream temperature
Top: Wall with Mixed thermal conditions of 300 K free
Cooling cycle stream temperature
Vapor-wall: Wall with Mixed thermal conditions of 300 K
free stream temperature
Water-wall: Wall with Mixed thermal conditions of 300 K
free stream temperature
Solution Methods Scheme: SIMPLE
Gradient: Green-Gauss Cell Based
Pressure: PRESTO!
Transient formulation: First-Order Implicit
 Page |6

A pressure based solver was chosen with transient flow conditions and the effect of normal
gravitational acceleration was considered. The k-epsilon turbulence model with standard wall
functions was used due to their proven accuracies in solving multiphase problems. For the vapor
material, the density was specified as ideal-gas. The density of liquid-water was assigned as
boussinesq. In phase interactions, the ‘evaporation-condensation’ mass transfer mechanism was
defined from liquid-water to water-vapor. First order implicit transient formulation was used. First
order upwind discretization scheme was used for the momentum equation, density, volume
fraction, energy, turbulent kinetic energy and turbulent dissipation rate. In pressure velocity
coupling, SIMPLE coupling scheme was used.

To simulate heating, the bottom wall and the walls surrounding the water phase was assigned
a uniform temperature of 1173 K. The remaining boundaries were assigned the mixed thermal
boundary conditions with free stream temperature of 350 K.

The convergence criterion is based on the residual value of the calculated variables. In the
present calculations, the threshold values were set to a 1/1000th of the initial residual value of each
variable, except the residual value of energy which was set to a millionth.

6.1. Monitors
A number of surface and volume monitors were set up to plot and record the data obtained
from the calculations.

 Surface monitors: area-weighted average pressure; area-weighted average temperature and

total surface heat flux from the water wall.
 Volume monitors: maximum and minimum temperature; maximum and minimum
pressure; mass of water, vapor and mixture; enthalpies of water, vapor and mixture;
average temperatures of water, vapor and mixture; average pressure of water, vapor and
mixture; pv and mrt.

The monitors were closely examined at regular intervals to check whether the model behaved
as expected throughout the entire simulation. Custom monitors to record PV and mRT values
against flow time were set up to verify the ideal gas law for vapor phase at every point during the
simulation.
 Page |7

7. Mixture Model
The Mixture model is a simplified multiphase model that can be used in different ways. The
mixture model can model n phases (fluid or particulate) by solving the momentum, continuity, and
energy equations for the mixture, the volume fraction equations for the secondary phases, and
algebraic expressions for the relative velocities. The Mixture model is a good substitute for the full
Eulerian multiphase model in several cases.

7.1. Heating Cycle

The methodologies mentioned in section 6 were applied and solution was initialized. Before
the calculations were started, it was assumed that the upper half of the vessel is completely filled
with vapor and the lower half is completely filled with water. The calculations were started with a
small time step size of 1e-06 seconds for 100 time steps. A maximum of 100 iterations per time
step was provided so that better solution convergence is achieved in each time step. As the solution
converged faster, the time step size was gradually increased while carefully observing the
variations of residuals.

Some abnormalities were observed with the mass of water-vapor as flow time increased. The
mass of vapor decreased gradually by an appreciable amount during the initial stages of heating.

7.2. Investigations Carried Out

7.2.1. Condensation frequency

It is known that during the heating cycle, there is no condensation of the fluid. But in the
numerical simulation of boiling, it was observed that the mass of vapor decreased initially by
approximately 3e-06 kg as shown in Figure 1 implying that condensation had taken place. To limit
the deviation of the numerical solution from the known physics some variations were conducted
with Condensation frequency which is specified in mass transfer mechanisms between the two
phases. The default value for condensation frequency is 0.1.
 Page |8

Figure 1

Condensation frequency = 0.1 (default)

0.0006
Volume Integral density-vapour (kg)

0.0005

0.0004

0.0003

0.0002

0.0001

0
0 1 2 3 4 5 6 7
Flow Time (seconds)

To comprehend how the value of condensation frequency affects the final results of the model
and also to study how the numerical solution could possibly deviate from physical aspect, the
condensation frequency was set to 1.

Figure 2

Condensation frequency = 1.0

0.0004
Volume Integral density-vapour (kg)

0.0003

0.0002

0.0001

0
0 1 2 3 4 5 6 7
Flow time (seconds)
 Page |9

During this time, it was observed that the mass of vapor decreased from an initial value of
2.8689e-04 kg to 2.3784e-04 kg in 4.43 seconds of flow time (approximately by 5e-05 kg), after
which the vapor mass increased steadily.

Then the condensation frequency was set to 0 as shown in Figure 3 to see how it compares
with the results obtained from condensation frequency of 1 and 0.1.

Figure 3

Condensation Frequency = 0
Volume Integral density-vapour (kg)

0.0005

0.0004

0.0003

0.0002

0.0001

0
0 1 2 3 4 5 6 7
Flow Time (seconds)

The mass of vapor decreased from an initial value of 2.869e-4 kg to 2.8547e-4 kg in 0.298
seconds of flow time after which it increased rapidly. The decrement observed in this case is very
minute and also the drop in vapor mass persisted for a smaller amount of time. Thus reducing the
deviation of the numerical solution from the physical aspect. For all further calculations of heating
cycle, the condensation frequency was set to 0.

7.2.2. Saturation temperature

The saturation temperature of the fluid was set as a constant (373.15 K) independent of the
absolute pressure inside the cylinder at first. As seen in Figures 4, 5, 6 there is a rapid increase in
vapor mass and pressure after the temperature inside the vessel crosses the saturation temperature
of water (about 6.3 seconds).
 P a g e | 10

Figure 4

Saturation Temperature kept constant

4.50E-03
Volume Integral density-vapour (kg)

4.00E-03

3.50E-03

3.00E-03

2.50E-03

2.00E-03

1.50E-03

1.00E-03

5.00E-04

0.00E+00
0 2 4 6 8 10
Flow Time (seconds)

Figure 5

Saturation Temperature kept Constant

450
Volume-Average Static Temperature (K)

400

350

300

250

200

150

100

50

0
0 1 2 3 4 5 6 7 8 9 10
Flow Time (seconds)
 P a g e | 11

Figure 6

Saturation Temperature kept Constant

Volume-Average Absolute Pressure (Pa)

2500000

2000000

1500000

1000000

500000

0
0 2 4 6 8 10
Flow Time (seconds)

Since the system under study i.e., the cylinder is a closed vessel and there is constant
evaporation from the surface. Thus causing the absolute pressure inside the cylinder to increase.
As a result, the saturation temperature of the fluid also increases accordingly.

Therefore, for accurate results, the variation of saturation temperature as a function of absolute
pressure obtained from thermodynamic steam tables was incorporated for the working range of the
system i.e. from 1 bar to 220 bar.

For better accuracy, the relationship between saturation temperature and absolute pressure was
defined using a 7th order polynomial using the data points. The coefficients of the equation
obtained from MATLAB were incorporated in Fluent. But the underlying problem of higher order
polynomials i.e., wiggles between the interpolated data points, resulted in poor results from the
simulation.

Hence, a piecewise linear variation of saturation temperature with absolute pressure was
chosen. 31 data points were selected to represent the dependence. This resulted in better results
from the model as shown in Figures 7,8,9. To elaborate, the convergence of the solution in each
time step improved i.e. the solution took lesser number of iterations to converge in each time step.
 P a g e | 12

Figure 7

Piecewise linear variation

4.50E-03
Volume Integral density-vapour (kg)

4.00E-03

3.50E-03

3.00E-03

2.50E-03

2.00E-03

1.50E-03

1.00E-03

5.00E-04

0.00E+00
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Flow Time (seconds)

Figure 8

Piecewise-linear variation
450
Volume-Average Static Temperature (K)

400

350

300

250

200

150

100

50

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13

Flow Time (seconds)

 P a g e | 13

Figure 9

Piecewise linear variation

Volume-Average Absolute Pressure (Pa)

2500000

2000000

1500000

1000000

500000

0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36

Flow Time (seconds)

It is evident from the Figures 4,5,7,8 that to attain the same values of vapor mass (4e-03 kg)
and pressure (2000000 Pa), time taken by the model when saturation temperature was kept constant
was very less (around 9 seconds) compared to when a piecewise linear variation was used (where
the time taken was about 35seconds).

Henceforth, the piecewise linear variation of Saturation temperature is incorporated for further
simulations since better results were obtained from it.

As a result of the investigations carried out, changes were made to the initial model.
Calculations were run for the complete heating cycle, i.e. till the absolute pressure inside the simple
cylinder reached 220 bar. The time taken for the complete heating cycle was 135.198 seconds. The
final mixture temperature attained was 653.5 K.

The variations of Static Temperature and Absolute Pressure with Flow Time have been plotted
for the complete heating cycle in Figures 10, 11.
 P a g e | 14

Figure 10

Pressure versus time

25000000

20000000
Absolute Pressure (Pa)

15000000

10000000

5000000

0
0 20 40 60 80 100 120 140 160
Flow Time (seconds)

Figure 11

Temperature versus time

700

600

500
Temperature (K)

400

300

200

100

0
0 20 40 60 80 100 120 140
Flow Time (seconds)
 P a g e | 15

7.3. Cooling Cycle

When the heating cycle was completed, the cooling cycle was started with changes to boundary
conditions. All boundaries were set to mixed thermal conditions with a free stream temperature of
300 K no shell conduction. The condensation frequency was set back to its default value of 0.1 so
that condensation takes place inside the cylinder. After the heating was stopped, it was observed
that the absolute pressure was still increasing. The pressure increased to 239 bar from 220 bar after
25 seconds of cooling and then the pressure started decreasing at a very slow rate i.e. the pressure
decreased by 1 bar after 3 seconds of flow time which can be visualized in Figure 12.

Figure 12

Cooling Cycle - Mixture Model

Volume-Average Absolute Pressure (Pa)

30000000

25000000

20000000

15000000

10000000

5000000

0
135 160 185 210 235 260 285 310 335 360 385 410

Flow Time (seconds)

 P a g e | 16

8. VOF Model
Volume of Fluid method is the most well-known and widely used direct multiphase models.
The VOF model is designed for two or more immiscible fluids in which the position of the interface
between the fluids is of interest. In this model, a single set of momentum equations is shared by
the fluids, and the volume fraction of each of the fluids in each computational cell is tracked
throughout the domain.

The initializations adopted in the Mixture model were retained for the VOF model so that
comparisons can be drawn with the results of the Mixture model. Calculations were run with the
same conditions of heating and cooling.

8.1. Heating Cycle

Very little change in heating cycle was observed between the VOF and the Mixture methods.
The model reached critical pressure of 220 bar after 147 seconds of flow time. The temperature
inside the cylinder reached 650 K. It was observed that the convergence of the solution took more
number of iterations when the VOF model was used compared to the Mixture model.

The observed reading can be visualized in figures 13 and 14.

 P a g e | 17

Figure 13

Pressure versus time

25000000
Volume-Average Absolute Pressure (Pa)

20000000

15000000

10000000

5000000

0
0 20 40 60 80 100 120 140 160
Flow Time (seconds)

Figure 14

Temperature versus time

Volume-Average Static Temperature (K)

700

600

500

400

300

200

100

0
0 20 40 60 80 100 120 140 160
Flow Time (seconds)
 P a g e | 18

8.2. Cooling cycle

Again, very little change in cooling cycle was observed between the VOF and the Mixture
methods. The absolute pressure inside the cylinder increased even after heating was stopped, like
the Mixture model, before dropping down gradually. The rate of decrease of absolute pressure
inside the cylinder is very slow, in a similar manner as the Mixture model. The slow decline is
pictured in figure 15.

Figure 15

Cooling cycle - VOF Model

Volume-Average Absolute Pressure (Pa)

30000000

25000000

20000000

15000000

10000000

5000000

0
145 170 195 220 245 270 295 320 345
Flow Time (seconds)
 P a g e | 19

9. Results and discussions

Sensible heating takes place during the initial stages, since the liquid is in sub-cooled state.
Once the temperature attains Saturation temperature anywhere in the liquid domain, phase change
starts to occur. As a result, vapor fraction of water-vapor increases. Hence, the pressure inside the
vessel increases since it is enclosed.

With time, the liquid near the wall attains saturation temperature and vaporization takes place
near the wall and the vapor rises to the liquid vapor interface, thus increasing turbulence and
convection in the liquid zone. As a result, heat transfer to the liquid zone increases and phase
change rate also increases.

The mass transfer rate will be in a slow pace during the initial stages. With time, the bulk of
the fluid reaches saturation temperature and thermal stratification increases and proportionally the
mass transfer rate also increases.

All the calculations run were found to obey the ideal gas law at every point during the
simulation which is validated using the monitors.

10. Summary and conclusions

Numerical simulation and experiments were conducted to investigate the boiling of water
inside a closed vessel. A comparison is provided between the Mixture and the VOF models for the
process of boiling inside a closed cylinder.

It was observed that the convergence of the solution took more number of iterations when the
VOF model was used compared to the Mixture model. But the total flow time for the complete
heating cycle were almost the same for both the multiphase models.

As far as the cooling cycle is concerned, both the models showed similar results. But there is
scope for exploring other possibilities for cooling of water inside a closed vessel in ANSYS Fluent.

Both the Mixture and the VOF multiphase methods in conjunction with evaporation-
condensation mass transfer mechanisms is a useful tool in studying boiling. The results obtained
from both the multiphase models are encouraging. The predictions using this methodology needs
to be checked with experiments.
 P a g e | 20

11. References
[1] Fluent 16.1 User’s Guide, 2015.

[2] Hari Krishna Raj, Jinan S, Dr. K E Reby Roy, numerical prediction of higher self-pressurization
in a typical storage tank, International Journal of Engineering Sciences and Technology (IJEST)

[3] Timo Kulji, Esa Muurinen, Riitta Keiski, Film Boiling studies with CFD, Energy Research,
university of Oulu.

[4] Moon Soo Lee, Vikrant Aute, Amir Riaz, Reinhard Radermacher, Areview on direct two-
phase, phase change flow simulation methods and their applications, International Refrigeration
and Air Conditioning Conference 2012.