Ahmmed - 2009 - Numerical Study of Cavitating and Noncavitating Flow Around 2D Hydrofoil

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/215687748
Numerical Study of Cavitating and Noncavitating Flow Around 2D Hydrofoil
Thesis · January 2009
CITATIONS READS
0 473
1 author:
Mohammad Shakil Ahmmed

The University of Queensland
21 PUBLICATIONS 63 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Computational modelling of the heat transfer phenomena in the process of photovoltaic solar cell manufacturing View project
CFD Modelling of Anaerobic Biogas Digestor View project
All content following this page was uploaded by Mohammad Shakil Ahmmed on 12 March 2018.
The user has requested enhancement of the downloaded file.

THEORITICAL STUDY ON NONCAVITATING &
CAVITATING FLOW OVER A 2-D HYDROFOIL
By
MOHAMMAD SHAKIL AHMMED
BACHELOR OF SCIENCE
BANGLADESH UNIVERSITY OF ENGINEERING AND

TECHNOLOGY, DHAKA.
October, 2009
THEORITICAL STUDY ON NONCAVITATING & CAVITATING
FLOW OVER A 2-D HYDROFOIL
Submitted by
MOHAMMAD SHAKIL AHMMED
Approved as to style & contents
Bangladesh University of Engineering & Technology
ii
DECLARATION
It is hereby declared that this thesis or any part of it has not been submitted elsewhere for
the award of any degree or diploma.
iii
DEDICATION
Dedicated
TO
Our Respectable Parents
iv
ACKNOWLEDGEMENT
Writing at this paragraph has been deleted intentionally
v
ABSTRACT
Two-dimensional finite volume method (FVM) based on Reynolds averaged Navier-Stokes

equations (RANS) has been used to simulate incompressible flow around hydrofoil in both
non-cavitating and cavitating conditions. The NACA0012 hydrofoil section is used as test
case. A cavitation model based on bubble dynamics equations is included to investigate the
unsteady behavior of cavitation at different cavitation numbers. Turbulence model
Renormalization Group (RNG) k-ε model with enhanced wall treatment is used to capture
boundary layer around hydrofoil surface and their effectiveness are evaluated. It is observed
that by k-ε turbulence model shows better performance. For this reason, k- ε model
associated with multiphase model is used only in case of cavitating flow. To implement all
of these, a commercial CFD software package, FLUENT 6.3.26 is used for numerical
solutions of the governing equations.
The cavitating study first presents an unsteady behavior of the partial cavity and then super
cavity attached to the foil. The computed result in terms of pressure coefficient, lift
coefficient and drag coefficient at different cavitation numbers have been shown
graphically and in the tabular form. The flow pattern and hydrodynamic characteristics are
also studied at different angles of attack. Unsteady turbulent cavitation flows around
NACA0012 hydrofoil is simulated by two dimensional computations of viscous
compressible turbulent flow model. Flows are calculated with SIMPLE type finite volume
scheme & to simulate turbulent flows RNG k- ε model is used with enhanced wall
treatment. The results of NACA0012 foil show quasi-periodic vortex cavitation
phenomena. Satisfactory agreement with the published numerical results is obtained in both
cavitating and noncavitating conditions.
vi
CONTENTS
DECLARATION iii
DEDICATION iv
ACKNOWLEDEMENT v
ABSTRACT vi
NOMENCLATURE xi
LIST OF TABLES xiv
LIST OF FIGURES xv
CHAPTER ONE 1
1. INTRODUCTION 1
1.1 Basics of Cavitation 1

1.2 Types of Cavitation 2
1.3 Partial Cavitation 3
1.3.1 Fixed Cavities 6
1.3.2 Cavity Detachment 6
1.4 Super Cavitation 7
1.4.1 Cavity Pressure 8
1.4.2 Cavity detachment 9
1.4.3 Cavity length 11
1.5 The physical phenomena 12
1.5.1 The main effects of cavitation of hydraulics 12
1.5.2 The main forms of vapor cavities 13
1.6 Cavitation in real fluid flows 13
vii
1.61 Cavitation regimes 13
1.6.2 Liquid vapor interfaces 14
1.7 Dimensionless numbers 16
1.7.1 Cavitation Number, σ 16
1.7.2 Void Fraction, α 16

1.7.3 Reynolds Number, Re 17
1.7.4 Strouhal Number, St 17
1.7.5 Pressure Coefficient Cp, Drag and Lift Coefficient 17

1.8 Literature Review 18
1.9 Objective with Specific Aims 21
CHAPTER TWO 22
2. THEORITICAL FORMULATION 22
2.1 Introduction 22
2.2 Governing Equation 23
2.3 The Dynamics of Spherical Bubble 24
2.3.1 Assumption 24
2.3.2 Boundary and Initial Condition 24
2.3.3 Rayleigh-Plasset Equation 25
2.3.4 Interpretation of Rayleigh-Plasset Equation in Terms of Energy Balance 27
2.3.5 The Collapse of Vapor Bubble 28
2.3.5.1 Assumptions 28
2.3.5.2 The Interface Velocity 28
2.3.5.3 The Pressure Field 30
2.4 Turbulence Modeling 31
2.4.1 Two Equation (k-ε) Turbulence Modeling 32
2.4.2 Wall function 33
2.5 Cavitation Modeling 35
2.5.1 Multiphase Modeling 36
viii
2.5.2 Transport Equation-Based Empirical Cavitation model 36
2.5.2.1 Model-1 37
2.5.2.2 Model-2 37
2.5.2.3 Model-3 37
2.6 Numerical Methodology 38
2.6.1 Pressure-Based Algorithm for Steady-State Computations of Cavitating Flows 41
2.6.2 Pressure-Based Algorithm for Time dependent Computations of Cavitating 45
Flow
CHAPTER THREE 49
3. NUMERICAL SIMULATION 49
3.1 Simulation Using Fluent 6.3.26. Based on Finite Volume Method 49

3.2 Steady Flow Over a Hydrofoil 50
3.2.1 Geometry of CAV2003 Hydrofoil 50
3.2.2 Computational Domain 50
3.2.3 Grid Generation 51
3.2.4 Grid Study 53
3.3 Boundary Condition 54
3.4 Solver Initialization and Flow Solution 55
CHAPTER FOUR 57
4. RESULTS AND DISCUSSION 57
4.1 Definitions 57
4.2 Noncavitating Analysis 58
4.2.1 Influence of Mesh 58
4.2.2 Influence of Turbulence Model 58
4.2.3 Comparison of Other Researchers 60
4.3 Cavitating Analysis 60
4.3.1 Numerical Simulation 61
ix
4.3.2 Analysis of Results 61
4.4 Computation of Unsteady Turbulent Cavitations Flow over NACA0012 72
CHAPTER FIVE 84
5. CONCLUSIONS AND RECOMMENDATION 84
5.1 Conclusion 84
5.2 Recommendation 85
REFERENCES 86
APPENDIX 86
Appendix A: 87
x
NOMENCLATURE

A surface area vector
CD drag coefficient
Cf frictional coefficient
Cp pressure coefficient
C1ε,C2ε turbulence constants
CL lift coefficient
DV drag force
Ce empirical constant
Cc empirical constant
fg gas much fraction
fv vapor much fraction
p pressure
pv vapor pressure
p∞ system pressure
Re vapor generation source term
Rc vapor condensation source term
f frequency of the vortex shedding
fi flux in the ith direction
F convective flux
Gk generation of turbulent kinetic energy
Gb generation of turbulent kinetic energy due to buoyancy
~ generation of turbulent kinetic energy due to mean velocity gradients
Gk
Gω generation of ω
G production of turbulent viscosity
I turbulence intensity
xi
k turbulent kinetic energy
l length scale (chord of the foil)
L characteristic length
n unit vector
N faces number of faces enclosing cell
Re Reynolds number
R bubble radius
S reference area
S source of  per unit volume
St Strouhal number
Sk, Sε user-defined source terms (for k-ε)
Sk , Sω user-defined source terms (for k-ω)

u flux
u/ Fluctuating velocity
V cell volume
Uavg mean flow velocity
U ref reference velocity
U free stream velocity

Y k, Y ω dissipation of k and ω due to turbulence
y distance from wall surface
y+ dimensionless parameter representing a local Reynold’s number
u+ nondimensional velocity
u* friction velocity
Greek symbols
r radial coordinate
ρ density
ρl liquid density
xii
ρm mixture density
ρv vapor density
σ cavitation number
ν kinematic viscosity
ω specific dissipation
 molecular kinematic viscosity
ε turbulent dissipation rate
 ij
momentum transport
τw wall shear
, turbulent kinematic viscosity
σk,σε turbulent Prandtl numbers for k and ε respectively
τw shear stress at the wall,
Г diffusion coefficient for 
α volume fraction
δij Kronecker delta function
γ surface tension
τij shear stress
∞ free stream
μt turbulent viscosity
xiii
LIST OF TABLES
Table No. Caption Page No.
4.1a Variation of lift coefficient at different angles (α) of attack 57
4.1b Drag coefficient (CD) at different angles of attack 57
4.2 Comparison of lift coefficient CL for non cavitating case. 58
4.3 Comparisons of the lift coefficient due to cavitation number 60
4.4 Lift & drag coefficient at various cavitation numbers at unsteady 61

condition
xiv
LIST OF FIGURES
Figure Caption Page No.

1.1 Cavitation patterns (a) Traveling bubble cavitation (b) Attached or 3
Sheet Cavitation (c) Tip vortex cavitation (d) Shear cavitation. Taken
from Franc [6].
1.2 Observed Cavitation Patterns on a 2D NACA 16012 Hydrofoil as a 5
function of the angle of attack α and the Cavitation number σ taken
from Franc & Michel [2], note that σ is the Cavitation number defined
as in Equation (1.1).
1.3 Partial Cavitation 6
1.4 Supercavity behind a Two-dimensional NACA 16012 Hydrofoil 8
(Reynolds number 106, cavitation parameter 0.07, angle of attack 17
deg.)
1.5 Supercavity 9
1.6 Cavity detachment (a) Non-tangential detachment (b) Tangential 10
detachment
1.7 Local effect of surface Tension on Cavity detachment 10
1.8 Liquid Vapor Interface 1.5
2.1 Bubble formation 24
2.2 Schematic of a Bubble in an infinite domain of Liquid 26
2.3 Evolution of Rt  and R t  during Bubble collapse 29
3.1 Schematic Diagram of the Flow Field around NACA0012 Hydrofoil 51

with Boundary condition
3.2 Grid Lines in Mesh: (a) Overall View and (b) Close-up View Near the 52
xv
Hydrofoil.
4.1 Comparison of lift Coefficient at various angle of attack 62

4.2 Drag Coefficient under noncavitating condition 62
4.3 Comparison of the lift coefficient due to variation of cavitation 63
number at 10 degree angle of attack
4.4 Variation of drag coefficient due to the variation of cavitation number 63
at 10 degree angle of attack
4.5 Comparison of lift & drag coefficient with variation of cavitation 64
number at 10 degree angle of attack with other researchers
4.6 Lift & drag coefficient with variation of cavitation number in unsteady 65
condition.
.
4.7 Time history of lift coefficient at 4 degree angle of attack 66
4.8 Time history of drag coefficient at 4 degree angle of attack 66
4.9 Contours of pressure coefficient (left) and the vaor volume fraction 67
(right) at σ=0.2 at steady state condition
4.10 Contours of pressure coefficient (left) and the vaor volume fraction
(right) at σ=0.4 & σ=0.6 respectively at unsteady state condition 67
(right) at σ=0.8 at unsteady state condition 68
(right) at σ=1.0 & σ=1.2 respectively at unsteady state condition 68
4.13 Pressure distribution over the hydrofoil at σ= 0.2 69
xvi
Pressure distribution over the hydrofoil at σ= 0.2 in unsteady
4.18 71
condition
4.19 Pressure distribution over the hydrofoil at σ= 0.4 in unsteady 72

condition
condition
4.21 73
condition
4.22 condition 73
4.23 condition 74
condition
Periodic forming & collapsing of vortex cavitation at σ=0.6,
4.4.1 78
attack =6˚
4.4.2 Experimental vortex cavity 80

4.4.3 Numerial pressure distribution by Kubota et al. 80
4.4.4 Pressure contour obtained by the numerical calculation by Chen Ying, 81
LU Chaun-jing, WU Lei 81
Pressure contour obtained at present
4.4.5
4.4.6 82
Velocity & pressure distribution of vortex cavitation
4.4.7 Velocity & pressure distribution of vortex cavitation obtained by Chen 82

4.4.8 Ying, LU Chuan-Jing, WU Lei.
83
Period of lift coefficient of NACA0012
Period of lift coefficient of NACA0012 obtained by Chen Ying, LU 83

4.4.9
Chuan-jing, WU Lei.
.
xvii
xviii
CHAPTER ONE
INTRODUCTION
1.1 Basics of Cavitation

Cavitation means cavity formation when the local pressure falls below the vapor
pressure at a given temperature. The appearance of vapor cavities inside an initially
homogeneous liquid medium occurs in very different situations. According to the flow
configuration and the physical properties of the liquid, these show different
characteristics.
Cavitation can be defined as the breakdown of a liquid medium under very low
pressures. This makes cavitation relevant to the field of continuum mechanics and it
applies to cases in which the liquid is either static or in motion.
According to Franc and Michel [1], “Cavitation can be defined as the breakdown of a
liquid medium under very low pressures”.
Hydrodynamic cavitation is associated with fluid flow and occurs when part of the fluid
changes from the liquid phase to the vapor phase. This change occurs approximately at
constant temperature. The vapor bubbles were considered to be void space, a cavity, and
the regions of the fluid in which vapor exists are called cavities. Cavitation is a
widespread phenomenon in liquids. As long as high fluid velocities are occurring, there
is the risk of cavitation. It often occurs in many hydraulic devices such as turbines,
pumps, propellers, free jets, river dams, pipe systems, valves, engines and reactor
cooling systems. In marine applications the occurrence of cavitation is common. Not
only does it appear in propulsion systems like propellers and water-jet impellers, but
also on fins, domes, rudders, struts, hydrofoils and torpedoes. In all these cases,
cavitation is generally an undesirable phenomenon which preferably should be
eliminated or well controlled. There are, however, some application areas where
cavitation is advantageous, as for instance in medical applications, ultrasonic’s materials
cleaning, rock cutting, sonoluminescence, super-cavitating projectiles and torpedoes.
Chapter One: Introduction
This study is particularly concerned with hydrodynamic cavitation, i.e. cavitation in

flowing liquids. This includes flow around hydrofoil, wings or propeller blades.
1.2. Types of Cavitation

When the phase change occurs in flowing liquids, e.g. a decrease of the pressure below
the saturation pressure due to an expansion of fluid, hydrodynamic cavitationcomes to
light. On the other hand, acoustic cavitation may occur in a quiescent or nearly
quiescent liquid. When an oscillating pressure field is enforced on a liquid medium,
cavitation bubbles may appear within the liquid when the oscillation amplitude is large
enough. Naturally, hydrodynamics cavitation and acoustic cavitation may occur at the
same time.
Cavitation can take different form as it develops from its inception. In case the pressure
is mostly above the saturation pressure, cavitation is strongly dependent on the basic
non-cavitating or fully wetted flow .As cavitation develops, the vapor structures disturb
and modify the flow and a new often-unsteady flow pattern evolves. Cavitation pattern
can be divided into different groups [2]:
 Bubble or “travelling” cavitation. Bubbles may appear in regions of low
pressure and low pressure gradients as a result of rapid growth of small
air nuclei present in the liquid. The bubbles are carried along by the
flow and disappear when they enter a region with higher pressure [3].
 Attached or sheet cavitation. When a low pressure region is formed near
the leading edge of a streamlined object in the flow, the liquid flow
separates from the surface and a pocket of vapor is formed[4,5].
 Cloud cavitation. When a vapor sheet detaches from the surface and is
advected with the flow, a region with a large number of vapor
structures is formed. This region is usually called cloud cavitation,
although it consists of a vertical flow region with many vapor bubbles.
This type is usually erosive when collapsing near a surface.
 Vortex cavitation. In the low- pressure core of vortices the pressure may
be low enough for cavitation to occur. This type of cavitation is often
found at the tip of lifting surface and is therefore also denoted by tip
vortex cavitation.
2
 Shear cavitation. In region with high shear vorticity is produced. As a

result coherent rational structures are formed and pressure level drops
in the core of the vortices, which become potential sites for cavitation.
Flow situation with shear cavitation can be found in wakes, submerged
jets at high Reynolds number and separated flow regions which
developed with foils at large angle of attack
(a) (b)
(c) (d)
Figure 1.1: Cavitation patterns (a) Traveling bubble cavitation (b) Attached or
sheet cavitation (c) Tip vortex cavitation (d) Shear cavitation Taken
from Franc[6].
1.3 Partial Cavitation

Cavitation or vaporization of a fluid is a phase changed observed in high speed flows
wherein the local absolute pressure in liquid reaches the vicinity of the vapor pressure at
the ambient temperature. The collapse of vapor bubbles can cause damage (pitting or
erosion) to metal surface and have an adverse effect on the performance of lifting
surface with extensive cavitation.
3
To predict inception of cavitation it is desirable to have indices to provide quantitative

measures of the dynamic flow condition under which the local pressure drops to the
vapor pressure. The vapor pressure of a fluid is a fundamental property which depends
on the temperature. However, a parameter is needed that would assume a unique value
for each set of dynamically similar cavitating conditions. Furthermore, the parameter
should be able to describe the flow condition relative to those for which cavitation is
absent, incipient or various stages of development.
In flowing liquids, the tendency of a flow to cavitate is indicated by the so-called

cavitation number which is expressed by
p   p v T 
 (1.1)
1
U  2
2
where
 p  is defined as the absolute ambient pressure
 pv is the vapor pressure
1
 U  2 is the dynamic stagnation pressure.
2
 U  is the reference velocity.
4
Figure1.2: Observed Cavitation patterns on a 2D NACA 16012 Hydrofoil as a

function of the angle of attack α and the Cavitation number σ. Taken
from Franc & Michel [2], note that σ is the Cavitation number defined
as in Equation (1.1).
Thus the vapor cavitation number is the ratio of the net pressure inhibiting cavitation to
the relevant dynamic pressure, i.e. it is a measure of tendency of the moving fluid to
cavitate. When σ is large, cavitation is unlikely; when σ is small the likelihood is great.
The cavitation number’s basic importance stems from the fact that it is an index of
dynamic similarity of flow condition under which cavitation occurs.
The present work addresses the problems of cavitating hydrofoil in two dimensions. As
a prelude to the theoretical treatment of this problems, we briefly review certain
physical aspects concerning the nature of cavitation that need to be kept in mind in
order to correctly describe the cavitating flow around a two dimensional hydrofoil.
5
1.3.1 Fixed Cavities

Fixed cavitation is defined as the cavitating condition in which a statistically fixed
cavity is formed by the guiding surface of the liquid, i.e. the rigid boundary of an
immerged hydrofoil, and the free surface of the liquid flow. Figure 1.3 illustrates a case
on partial cavitation, a fixed cavity which does not envelop the rear end of the hydrofoil,
but closes of the hydrofoil surface. At sufficiently low value of the cavitation number,
the sheet cavitation on the back of a hydrofoil will extend to beyond the trailing edge,
which a case of super cavity (Figure1.5) will be formed.
U∞ Partial cavity
Figure 1.3: Partial Cavitation
1.3.2 Cavity Detachment

Experiments have revealed that the location of the cavity detachment point can have a
significant effect on the cavity extent and the cavity volume. In the case of a sharp
leading edge, the cavity will develop from this point. If however, the leading edge is a
smooth curve, the cavity detaches from the point downstream of the laminar boundary
layer separation point. The position of the separation and the detachment point and the
correlation between them has been studied extensively by Arakeri [7]. He developed a
semi-empirical method to predict the position of viscous laminar cavitation separation
on an arbitrary body. However, the method only applies in the Reynolds range for
which the cavitating body posses boundary layer separation under noncavitating
condition.
From a numerical point of view, the position of the cavity detachment point has an
important influence on the quantitative aspects of the cavity shape .For a given
cavitation number, σ, when the cavity detachment point moves downstream, the cavity
become fatter and shorter. However, the physical validity of the result will also be
affected .As for example if the cavity detachment point is set too far forward on a
6
hydrofoil section, the cavity streamline may be intersect the foil, thereby resulting in the
non-physical solution.
The smooth detachment condition, also known as the Brillounin-Villat condition,

requires that the slope and the curvature of the cavity and the detachment point must
equal that of the foil surface at the same position. This condition ensures that the cavity
dose not intersects the foil at the leading edge and that the pressures on the wetted foil,
upstream on the cavity, are larger than the cavity pressure.
The smooth detachment point (SDP) is thus found by requiring the slope of the pressure
distribution with respect to the foil arc-length at the cavity leading edge to be equal to
zero. Detaching the cavity upstream of the SDP will result in a cavity that intersects the
foil surface, while detaching the cavity downstream of the SDP will produce pressure
upstream of the cavity which is smaller than the cavity pressure.
However in reality, the cavity will detach well downstream of the SDP. Furthermore,
pressures smaller than the cavity pressure have been measured on the wetted flow
upstream of the cavity detachment point. The criterion formulated by Arakeri [7] has
since then been extended in the case of cavitating hydrofoils by Franc & Michel, [8].
The result obtained by Franc & Michel [8] confirm that the cavity detachment take place
just downstream on the boundary- layer separation, but not at the minimum pressure
point. They formulated a cavity detachment criterion by which the actual detachment
point is determined as the one among all the theoretically possible detachment points,
for which the computation predicts a laminar separation just upstream.
In practice, the application of the criterion formulated by Franc & Michel, [8] consists
of the moving the detachment point on the foil until a boundary layer calculation shows
laminar separation just ahead of the cavity leading edge. Thus it is apparent that an
important factor governing the location of the cavity detachment is viscosity through the
behavior of the boundary layer.
1.4 Supercavitation
As the cavitation parameter is decreased, a small cavity attached to a hydrofoil will
extend and grow longer. It becomes a supercavity as soon as it ceases to close on the
cavitator wall but inside the liquid, downstream of the cavitator. Simultaneously, the lift
7
of the hydrofoil decreases while its drag increases. For very high relative velocities
between the liquid and the body, it is practically impossible to use non-cavitating foils,
such as the conventional ones used in aerodynamics. In such cases, different types of
supercavitating foils have been designed for better efficiency, such as truncated foils
with a base cavity or supercavitating foils with non-wetted upper sides.
.
Figure 1.4: Supercavity behind a Two-dimensional NACA 16012 Hydrofoil

(Reynolds number 106, cavitation parameter 0.07, angle of attack 17
deg.)
1.4.1 Cavity Pressure

As an example, consider a supercavity attached to a two-dimensional foil, as
schematically shown in Figure 1.5. The cavity is made of a mixture of vapor and non-
condensable gas, so that the pressure inside is:
pc  p v  p g (1.2)
8
In this equation, as usual, pv stands for the vapor pressure and p g , for the partial gas
pressure. The cavity pressure pc is generally considered as constant in time and uniform
throughout the cavity.
The presence of gas in the cavity is due to diffusion through the interface of gases
dissolved in the liquid. If the concentration of gas at saturation in the liquid is large, the
partial pressure of gas in the cavity will be large. On the contrary, with deaerated water
at room temperature, which is usually the case in hydrodynamic tunnels, the value of
p g is small with respect to pv . Another case which deserves attention is ventilated
cavities for which the gas pressure can be high, depending mainly on the ambient
pressure and the injection flow rate.
U∞ Supercavity
Figure 1.5: Supercavity
The tangential shear stress on the cavity interface is due to the friction between the
vapor-gas layers of the cavity and the external flow. Its value is generally small since
the density of the fluid inside the cavity is much smaller than that of the liquid. Hence,
the shear stress on the cavity is usually negligible.
1.4.2 Cavity Detachment

If no special element, such as a step or a sharp edge, is designed to fix the detachment
point of the cavity, the position of detachment is unknown a priori and a detachment
criterion is necessary to predict the location of the cavity detachment. Two main criteria
are available.
Villat-Armstrong criterion
Villat and Armstrong criterion was established on the basis of inviscid two-dimensional
flow theory [1]. Within the framework of this theory, the cavity must detach tangentially
9
to the solid wall (Figure 1.6). If not, the velocity on the free streamline would tend to
zero at detachment, which is incompatible with the condition of constant pressure or
constant velocity along the whole free streamline.
Cavity
V=0 at detachment
(a) (b)
Figure 1.6: Cavity detachment (a) Non-tangential detachment,(b)
Tangential detachment
The Villat-Armstrong criterion assumes that the cavity is a zone of minimum pressure
for the whole flow field. According to this criterion, the detachment point is then the
point of minimum pressure on the wall. Its determination is often calculated via an
iterative procedure. Initially, it can be chosen at the minimum pressure point taken from
non-cavitating conditions. However, this initial guess should be corrected to take into
account the change in the pressure distribution due to the development of cavitation and
to make sure that the detachment point is actually a point of minimum pressure of the
supercavity flow.
Laminar separation criterion
From the examination of cavitation inception on spheres, circular cylinders and ogives,
Arakeri [7] proved that viscous effects are predominant in cavitation inception. They
showed that laminar separation on a wall provides a site for inception and that the
elimination of an existing laminar separation by stimulating the boundary layer with a
trip has a strong effect on inception. From those experiments, Arakeri [7] proposed a
criterion to predict the position of detachment of a cavity. This criterion also takes into
account the local effects of surface tension which obliges the free streamline to be non-
tangential to the body at detachment (Fig. 2.6).
Cavity detachment Cavity detachment

without surface tension with surface tension
Figure 1.7: Local effect of surface tension on Cavity detachment
10
Franc and Michel [8] confirmed this viscous effect in the case of hydrofoils and proved
that it is relevant not only for cavitation inception but also for supercavitation. They
showed that a well developed cavity always detaches downstream of laminar separation
of the boundary layer. The existence of separation, which generates a relatively dead
zone, is the only opportunity for a cavity to remain attached to the wall and to be
sheltered from the incoming flow. If the boundary layer does not separate, the cavity is
swept away by the flow and cannot attach to the smooth wall.
In conclusion, the laminar separation criterion assumes that a supercavity detaches at the
laminar separation point of the boundary layer. The position of laminar separation and
even its existence in some cases may strongly differ from the case of the fully wetted
flow because of the changes in pressure distribution due to developed cavitation.
1.4.3 Cavity Length

The length of a supercavity is one of the most important parameters of the cavity flow. It
is measured from detachment to closure and may be affected, experimentally, by large
uncertainties due to the instability of the closure region.
The length of a supercavity increases when the relative cavity under pressure σc
decreases. This is easily understandable since; in that case, the pressure difference
between the reference point and the cavity decreases, resulting in smaller pressure
gradients in the whole flow field except in the vicinity of cavity closure. Then the
streamlines tend to have a smaller curvature and to become closer to straight lines
parallel to the upstream velocity.
In many cases, it is possible to model the experimental dependence of the cavity length
with the cavity under pressure for low values of that parameter by a power law:
l
 A  n (1.3)
c
where c stands for a characteristic size of the body. Throughout this study, we use σ to
define the cavity under pressure rather than σc. The exponent n is found equal to 2 if the
body is located in an infinite medium. The value of A depends on the body shape and
position.
11
1.5 The Physical Phenomena
1.5.1 The Main Effects of the Cavitation in Hydraulics

If a hydraulic system is designed to operate with a homogeneous liquid, additional vapor
structures due to cavitation can be interpreted, by analogy with the case of mechanical
systems, as mechanical clearances. The vapor structures are often unstable, and when
they reach a region of increased pressure, they often violently collapse since the internal
pressure hardly varies and remains close to the vapor pressure. The collapse can be
considered analogous to shocks in mechanical systems by which clearances between
neighboring pieces disappear. Following this, a number of consequences can be
expected:
 alteration of the performance of the system (reduction in lift and increase in drag
of a foil, fall in turbomachinery efficiency, reduced capacity to evacuate water in
spillways, energy dissipation, etc.);
 the appearance of additional forces on the solid structures;
 production of noise and vibrations;
 wall erosion, in the case of developed cavitation if the velocity difference
between the liquid;
 and the solid wall is high enough.
Thus, at first glance, cavitation appears as a harmful phenomenon that must be avoided.
In many cases, the free cavitation condition is the most severe condition with which the
designer is faced. To avoid the excessive financial charges that would be associated with
this, a certain degree of cavitation development may be allowed. Of course, this can be
done only if the effects of developed cavitation are controlled.
The negative effects of cavitation are often stressed. However, cavitation is also used in
some industrial processes to concentrate energy on small surfaces and produce high
pressure peaks. For this purpose, ultrasonic devices are often used. Examples of such
positive applications include:
 the cleaning of surfaces by ultrasonic or with cavitating jets,
 the dispersion of particles in a liquid medium,
 the production of emulsions,
12
 electrolytic deposition (the ion layers that cover electrodes are broken down by
cavitation, accelerating the deposition process),
 therapeutic massage and bacteria destruction in the field of medical engineering,
 the limitation of flow rates in confined flows due to the development of
supercavities.
1.5.2 The Main Forms of Vapor Cavities

Cavitation can take different forms as it develops from inception. Initially, it is strongly
dependent on the basic non-cavitating flow structure. However, as it develops, the vapor
structures tend to disturb and modify the basic flow. Cavitation patterns can be divided
into three groups. These are:
 Transient isolated bubbles. These appear in the region of low pressure as a result
of the rapid growth of very small air nuclei present in the liquid .They are carried
along by the main flow and subsequently disappear when they enter areas of
high enough pressure.
 Attached or sheet cavities. Such cavities are often attached to the leading edge of
a body, e.g. on the low-pressure side of blades and foils.
 Cavitating vortices. Cavitation can appear in the low-pressure core of vortices in
turbulent wakes or, as a regular pattern in tip vortices of 3-D wings or propeller
blades.
Some patterns do not fall easily in these classes. For example, on the low-pressure
surfaces of foils or propeller blades, vapor structures with a very short lifetime can
appear. They have the form of attached cavities but are transported similarly to traveling
bubbles.
1.6 Cavitation in Real Fluid Flows
1.6.1 Cavitation Regimes

For practical purposes, it is useful to consider two distinct steps in cavitation
development:
 cavitation inception, i.e., the limiting regime between the non-cavitating and the
cavitating flow;
 developed cavitation, which implies a certain permanency and extent of the
cavitation or a significant fall in performance of machines.
13
The distinction is important in the context of acceptance or otherwise of cavitation in

industrial situations. In the case of undeveloped cavitation, inception or desinence
thresholds are of interest. For developed cavitation, the manufacturer must focus on the
consequences of cavitation on the operation of the hydraulic system.
In the case of attached cavities, a further distinction may be useful: partial cavities,
which close on the wall, and supercavities, which close away from the boundary
(typically a foil).
1.6.2 Liquid Vapor Interfaces

Cavitating flows, like other two-phase liquid-gas flows, are characterized by the
presence of numerous interfaces. However, their response to external perturbations, for
example a pressure rise, can be very different from the case of liquid-gas flows.
Two-phase flows containing gas bubbles are not usually subject to rapid changes in
mean density (except in the case of shock waves). This is because the non- condensable
nature of the gas confers a kind of global stability to the flow.
In cavitating flows, however, the interfaces are subjected on one side to a constant
pressure, practically equal to the vapor pressure. Thus, they cannot sustain an increase
or decrease in external pressure without rapidly evolving in both shape and size. They
are extremely unstable.
It is almost impossible to use intrusive probes to take measurements within a cavitating

flow due mainly to the cavitation the probe itself generates. However, if the liquid is
transparent, it is possible to visualize the interfaces, as they reflect light very effectively.
Interfaces can generally be considered as material surfaces and their observation from
one-shot photographs (short flash durations of the order of a microsecond) or from high-
speed photography or video (at a typical rate of ten thousand frames per second) gives
an idea of the flow dynamics.
Concerning the exchange of liquid and vapor across an interface, the mass flow rate m
(per unit surface area) across the interface is proportional to the normal velocities of
14
either the liquid or the vapor relative to the interface. Mass conservation across the
interface gives
 dn   dn 
m   l Vln     v Vvn   (1.4)
 dt   dt 
In this equation, the indices ℓ and v refer to the liquid and vapor phases respectively,
dn
and the index n to the normal component of the velocities. The symbol is the
dt
normal velocity of the interface.
LIQUID
VAPOR
Vln
dn Vvn
dt
Figure 1.8: Liquid Vapor Interface
If the flow rate through the interface is negligible, i.e. m  0 (which is actually
assumed in most cases), the three normal velocities are equal and the interface is a
material surface, i.e. a surface made up of the same fluid particles at different instants.
Two cases are of particular interest:

 For a spherical bubble whose radius R(t) is a function of time, the normal
velocity of the interface is dR . In the case of negligible flow rate:

dt
Vln  Vvn  dR .
dt
 For a steady cavity attached to a wall surrounded by a flowing liquid, dR is

dt
zero. If it is assumed that the mass flow rate through the interface is negligible,
the normal velocities of the liquid and the vapor at the interface are also zero.
15
Thus the liquid velocity of the outside flow at the interface is tangential to the
cavity wall.
1.7 Dimensionless Numbers

Relevant dimensionless numbers used in this research are as follows:
1.7.1 Cavitation Number, σ

The dimensionless cavitation number, σ was introduced by Thoma [9]. The number is a
measure for the sensitivity of the flow for cavitation to occur and is useful to facilitate
the comparison of result of experiments and numerical simulation. The cavitation
number σ is defined as
p   p v (T )
 (1.5)
1
 U 2 
2
where p [Pas],   [ kgm 3 ] and U 2  [ ms 1 ] are the free stream pressure, free stream
density and free-stream velocity, respectively, and where p v T  is the vapor pressure of
water at temperature T [K] . Note that a higher cavitation number indicates that the
pressure in the flow much decrease more before cavitation occurs. A small cavitation
number indicates that a smaller decrease in pressure causes cavitation. Thus, low a
cavitation number corresponding to high susceptibility for cavitation.
1.7.2 Void Fraction, α

The void fraction α within a volume V [m3] of a fluid follows from the fluid density
   v , sat T   1    l , sat T  as
Vv    l , sat T 
  (1.6)
V  v , sat T    l , sat T 
where Vv [m3] is the volume of vapor within the volume V of the fluid and where
 v, sat T  [kgm-3] and  l , sat T  [kgm-3] are the saturated vapor and liquid density at
temperature at T, respectively.
Experimentally, it is very difficult to determine the void fraction to any location in the
flow. Numerically the void fraction is used for visualization and analysis purpose.
16
Employing the equilibrium cavitation model the determination of the void fraction is
just a post- processing step evaluating Equation (1.6).
1.7.3 Reynolds Number, Re

The Reynolds number is the ratio of the inertia forces to viscous forces and thus it
quantifies the relative importance of these two types of forces given the flow conditions.
The Reynolds number Re is defined as:
 U  L UL
Re   (1.7)
 
where   is the density of the fluid, U  a characteristic velocity of the flow, L a
characteristic length scale [m],   the dynamic fluid viscosity [Pas], and        is
the kinematic fluid viscosity [m2 s-1 ]. The flow about a hydrofoil of chord length
c  0.1m of pure water at saturation pressure and at a velocity of U   6ms 1 has a
Reynolds number of Re  5.9  10 5
1.7.4 Strouhal Number, St

The Strouhal number St is employed to quantify the oscillating frequency in unsteady
flows. For cavitation flows the Strouhal number St is defined by:
fl
St  , (1.8)
U
where f [Hz]is the cavity shading frequency , l is the mean cavity length [m] and U  is
the free stream velocity . Often, it is difficult to accurately obtain a mean cavity length
for unsteady cavitation. So, for convenience we define a different Strouhal number
St c based on the chord length c of the foil instead of on the mean cavity length:
fc
St c  , (1.9)
U
1.7.5 Pressure Coefficient C p , Lift and Drag Coefficients

The dimensionless pressure coefficient C p is defined as:
p  p
Cp  (1.10)
1
 U 2 
2
17
where p is the local pressure in the flow field, and where p ,   and U  are the free-
stream pressure, the free-stream density and free-stream velocity, respectively. In the
following we usually employ the pressure coefficient.
Neglecting skin friction, the drag and lift forces can be obtained from
 
F   pn ds . (1.11)
S

where S surface of the object, p the pressure on the surface of the object and n the unit
normal pointing into the object, i.e. out of the computational domain. In 2D we will use
lower-case symbols, i.e.
 
f   pn dC (1.12)
C
with C the closed curve of the object. For two-dimensional flow about a 2D geometry

the lift force l per unit length in span-wise direction is equal to the component of f in
the direction normal to the free-stream, which in our case is f y . For three dimensional
flow the lift force L is equal to Fz . The drag forces d per unit length in span-wise
direction and the drag force D are equal to f x and Fx for two- dimensional or three
dimensional flows, respectively.
1.8 Literature Review

Research on cavitation data back to the days of Euler who observed the occurrence of
cavitation in high speed of water flow during his studies on rotating flow machinery.
The word cavitation has been introduced by Froude who described the voids filled with
vapor as cavities [10]. In 1895 Person was amongst the first to observe the negative
effects of cavitation on the performance of a ship propeller [11]. He was the first to
build the cavitation tunnel to investigate the problems due to cavitation experienced on
p   p v (T )
the propeller on the ship Turbinia. The cavitation number   was
1
 U 2 
2
introduced by Thoma [1, 9] in the context of the experimental investigation on water
turbine and pumps.
18
In order to study the physical aspect of cavitation many experiments have been carried
out throughout the years. Theoretical and numerical approach followed soon with two
main areas of research [1]: bubble dynamics and developed – or super civitities.
A large body of work has been published on bubble dynamics. We mansion , amongst of
many others, Rayleigh [12] and Plesset [12], after whom the Rayleigh–Plesset equation
is named which describes the temporal evolution of the radius of a vapor bubble in an
incompressible, viscous liquid. The evaluation is driven by effects of pressure variations
and surface tension.
The field of developed cavities started more than century ago, e.g. Helmholz [9, 13] and
Kirchhoff [14], with the work on free-streamline theory or wake theory by using
conformal mapping techniques or the non-linear hodograph technique. Birkhoff &
Zarantello [15] describe the hodograph technique in detail. Wu [16] points out that this
theory can only be used for cavitation flow around simples geometries like bluff bodies
and flat plates, but cannot be used for cavitating flow around arbitrary bodies like
hydrofoil and propeller blades. In 1953 Tulin [17] applied linearization procedure to the
problem of the flow about a super cavitating symmetric profile at zero angle of attack
and zero cavitation number since then many researchers have extended the linear theory
to flows around arbitrary bodies at any cavitation number.
The introduction of computers 1970s brought about a large number of numerical

methods of based on linear theory, which has been extended to three-dimensinal flow
problem by the use of lifting surface theory. Most lifting surface theory methods deal
with sheet cavitation by imposing a transpiration type of (linearized) boundary condition
on the solid surface below the sheet cavity. The advantages of lifting surface methods
are their short computation times, enabling fast assessment and improvement of designs.
The drawback of linear theory is that the partial cavity flows around hydrofoils it
predicts that the length and volume of cavity will increase when the thickness of the
hydrofoil of increased, which contradicts in experimental observation. Also for unsteady
sheet cavitation the dynamic motion of sheet cavity is not predicted and linearized
theory has a limit ability to describe complex flows with enough accuracy.
A different approach to simulate cavitating flows emerged in the 1990s. Method using
the Euler or Navier-Stokes equation were developed together with the transport equation
19
for the void fraction, with two- phase flow equations or with other cavitation closure
model equations. As classified by the 22nd ITTC special committee in 1999 [18] these
approaches can be grouped into number of categories. 1) Interface tracking methods 2)
Volume of fluid methods 3) Discrete bubble methods 4) Two phase flow methods. It
must be noted that the distinction between some of these group is not always completely
clear and that combinations of the categories are used by different authors.
In the last decade various methods for numerical simulation of cavitating flow were
developed. Most of the studies treat the two phase flow as a single vapor–liquid phase
mixture flow. The evaporation and condensation can be modeled with different source
terms that are usually derived from the Rayleigh–Plesset bubble dynamics equation.
This approach was first made by Kubota et al. [19] who used the linear part in the
Rayleigh–Plesset to equation to describe the evolution of bubble radius as a function of
surrounding pressure. Different other cavitation models that included more complex
relations between pressure and bubble radius were derived from the Rayleigh– Plesset
equation – for example Schnerr and Sauer [20] and Frobenius et al. [21], but they all
included some quantities (like bubble number density and initial bubble diameter) that
are impossible or very hard to determine. For example the recommended (estimated)
value for bubble number density that has to be included in the mentioned models is 104
m−3 according to Kubota et al. [19], 108 m−3 according to Schnerr and Sauer [22] or to
Frobenius et al. [21] and even 1012 m−3 according to Alajbegovic et al. [22].
Recently different authors proposed to consider a transport equation model for the void
ratio, with vaporization/condensation source terms to control the mass transfer between
the two phases (Singhal et al. [23], Kunz et al. [24], Merkle et al. [25], Senocak and
Shvy [26] and Owis and Nayfeh [27]). This method has the advantage that it can take
into account the time influence on the mass transfer phenomena through empirical laws
for the source term. It also avoids using, hard to determine, quantities like bubble
number density and initial bubble diameter. The other way to model cavitation process
is by the so called barotropic state law that links the density of vapour–liquid mixture to
the local static pressure. The model was proposed by Delannoy and Kueny [28] and later
widely used by other Coutier-Delgosha et al. [29], Hofmann et al. [30], Lohrberg et al.
[33] and Song 524 M. and He [32]). The results obtained with the barotropic cavitation
model show very good correlation to the experiments but the past simulations lacked in
20
robustness of the numerical algorithms, which resulted in numerical instability and

sometimes, poor convergence. A cavitation model, based on bubble dynamics equations,
described in Singhal et al. [23] is included in the code and it was used to describe the
unsteady behavior of cavitation including the shedding of vapor structures.
1.9 Objectives with Specific Aims

The main objective of this study is to simulate the cavitating flow around hydrofoil
using Finite Volume Method (FVM) [33-34]. For the numerical simulation of cavitating
flow a bubble dynamics multiphase cavitation model is used to describe the generation
and evaporation of vapor phase. The investigation is carried out to simulate cavitating
and noncavitating flow around the NACA0012 hydrofoil. The specific aims of this
study are:
 To investigate noncavitating and cavitating condition for different cavitation

number.
 To study cavitation inception on the hydrofoil surface at different angles of attack.
 To study the shape and general behavior of the cavitation.
 To obtain lift and drag force.
 To compute velocity and pressure distribution on the foil.
 To study with the unsteady turbulent cavitations flows.
 To investigate the vortex phenomena on the hydrofoil surface
 To validate predicted results by comparing with published numerical results
21
CHAPTER TWO
THEORETICAL FORMULATION
2.1 Introduction
In this chapter the governing equations, dynamic evolution of a spherical bubble and the
computational methodology are presented. First, the Navier-Stokes equations governing
the flow of incompressible single-phase fluids are presented. Secondly, the dynamic
evolution of a spherical bubble with a fixed center, which undergoes uniform pressure
variations at infinity is considered. This simple model demonstrates the main features of
many practical cases such as bubble collapse, bubble formation from a nucleus, bubble
oscillations, etc. Experience shows that more complicated situations, involving the
motion of the bubble center for example, can be approximately dealt with this model.
From an historical viewpoint, the liquid motion induced by a spherical cavity in an

infinite medium under uniform pressure at infinity seems to have been first considered
by Besant in 1859. It was solved for a non-viscous liquid by Rayleigh [12] to interpret
the phenomenon of cavitation erosion. In 1948, Cole used the model of a spherical
bubble containing a non-condensable gas and applied it to sub- marine explosions.
Plesset [12] considered the general case of bubble evolution for a viscous and non-
compressible liquid.
The Reynolds Navier-Stokes equations and the turbulence modeling concept are
provided. Following these, cavitation modeling is discussed, and a transport equation-
based interfacial dynamics cavitation model is developed.
Finally, the chapter is concluded with the presentation of numerical methodology and
the pressure-based algorithms for steady and time-dependent computations. The
governing equations are presented in Cartesian coordinates for the ease of discussion.
Issues related to the implementation of the models are also discussed
Chapter Two: Theoretical Formulation
2.2 Governing Equation

The conservative or divergence form of the system of equations which governs the time
dependent two dimensional flow of an incompressible Newtonian fluid is [34]:

divu   0 (2.1)
 u  p
  div uu     divgrad u   S x (2.2)
 t  x
 v  p
  div vu     divgrad v   S x (2.3)
 t  y

where,  = density, p = pressure, t = time, u  ui  vj = velocity vector, =
viscosity, and S = source term. The above equations are known as conservation
equations as these equations obey the conservation principles of mass, momentum and
energy. It is clear that there are significant commonalities among the above mentioned
various equations. So these equations can be written in a general form and that general
equation can be solved numerically instead of solving each equation individually. By
introducing a general variable  the conservative form of all fluid flow equations can be
written as:
 u  p
  div u     div grad    S x (2.4)
 t  x
In words
Rate of increase Net rate of flow of Rate of increase Rate of increase
of  of fluid +  out of fluid = of  due to + of  due to
element element diffusion source (Source
(Unsteady) (Convective term) (Diffusive term) term)
Equation (2.4) is the so-called transport equation for property  . In order to bring out
the common features we have, of course, had to hide the terms that are not shared
between the equations in the source terms. By setting  equal to 1, u, v and selecting
appropriate values for the diffusive coefficient  and source term we may obtain
Equations (2.1) to (2.3). Equation (2.4) is used as the starting point for computational
procedures in the finite volume method.
23
2.3 The Dynamics of the Spherical Bubbles

2.3.1. Assumptions
The main assumptions are as follows:
 the liquid is incompressible and either Newtonian or inviscid;
 gravity is neglected;
 the air content of the bubble is constant, its inertia is neglected as in any
exchange of heat with the surroundings. This adiabatic assumption is valid when
considering rather large bubbles;
 the bubble is saturated with vapor whose partial pressure is the vapor pressure at
the liquid bulk temperature.
The functions to be determined, in the liquid domain r  Rt  , are the velocity u r , t  and
the pressure p r , t  induced by the evolution of the bubble (Figure 2.1).
u(r, t)
r
p(r, t) R(t)
Figure 2.1: Bubble formation
2.3.2. Boundary and Initial Condition

The mass transfer through the interface is neglected, so that the liquid velocity at the
interface u r , t  is equal to the interface velocity R  dR dt
In the case of a viscous fluid of kinematic viscosity  , the normal stress at the interface
is:
u
t rr R, t    p R, t   2  (2.5)
r rR
The balance of normal forces is given by:

2S
 t rr R, t   p v  p g t   (2.6)
R
24
where p g stands for the partial pressure of the gas inside the bubble. Assuming an
adiabatic transformation of the gas, the instantaneous gas pressure is related to the initial
pressure p g 0 by the following expression:
3
 R 
p g t   p g 0  0  (2.7)
 Rt  
Where  is the ratio of heat gas capacities c pg and cvg thus, the pressure on the cavity
interface is given by:

3
R  2S u
p R , t   p g 0  0    2 (2.8)
 R  R r r R
Far from the bubble, the liquid is assumed at rest so that u , t   0 and the pressure
p, t  also denoted p t  is assumed given.
The equilibrium of the interface of bubble requires the following condition to be

satisfied
2S
p  p g  pv  (2.9)
R
For the initial conditions (subscript zero), the bubble is assumed to be in equilibrium,
i.e., R 0   0 , so that Equation (2.9) is satisfied:
2S
p 0  p g 0  pv  (2.10)
R0
2.3.3. Rayleigh-Plesset Equation

Rayleigh has presented a mathematical analysis of the growth and collapse of a
spherical cavity. The equation in its most general form is known as the Rayleigh-Plesset
equation [3]. A spherical bubble of radius R(t) in an infinite domain of liquid is shown
in Figure 2.2. The pressure in the liquid P∞(t) is assumed to be known and the
temperature of the liquid is assumed to be constant. The pressure Pg t  and temperature
Tg t  inside the bubble are homogeneous and uniform.
25
Liquid P t , T t  u(r,t)
Bubble
R(t)
Figure 2.2: Schematic of a Bubble in an infinite domain of Liquid.
Due to spherical symmetry, the flow is of source (or sink) type and irrotational. The

mass conservation equation for an incompressible fluid, i.e. divV  0 gives
R2
u (r, t )  R (2.11)
r2
In this very particular case, the viscous term of the Navier-Stokes equation is zero. Thus,
for both viscous and non-viscous fluid, the momentum equation is:
u u 1 p
u  (2.12)
t r  r
from which one infers, taking Equation (2.11) into account:
R2  2  R  R    1 p
 4
R  2 R  2  (2.13)
r2 r r5   r
Integrating with respect to r and considering the conditions at infinity, one obtains:
p r , t   p  t   R  2 R 2  R  R 
2 4
R   (2.14)
 r r 4r 4 
This equation is equivalent to the Bernoulli equation for a variable unsteady flow of
non-viscous liquid. On the interface r  R Equation (2.14) gives:
26
pR, t   p  t   3 R 2
 RR (2.15)
 2
Finally, with expression (2.8) for the pressure at the interface, and noting that:
u 2 R
 (2.16)
r r R R
Equation (2.15) becomes:

3
 3  R  2S R
  RR  R 2   p v  p  t   p g 0  0    4 (2.17)
 2   R R R
This equation, known as the Rayleigh-Plesset equation, allows us to determine the

temporal evolution of the radius R and consequently the pressure field in the liquid
when the law p t  is given. For a non-viscous liquid, the last term on the right hand
side vanishes. The corresponding equation is known as the Rayleigh equation. Both
equations are differential and highly non-linear, due to the inertial terms. This results in
various specific features are presented in this chapter.
In the two following sections, the Rayleigh equation will be used to solve the problem
of bubble collapse and bubble explosion. In most cases, the inertial forces are dominant
and viscosity does not play a significant role. The role of surface tension is often
secondary in the case of bubble collapse.
2.3.4. Interpretation of the Rayleigh-Plesset Equation

Noting that:
 
RR
3 2
2
R 

1 d 2 3
2 RR 2 dt
R R   (2.18)
the Rayleigh-Plesset equation (2.17) can be written as follows:

3
  R0  
d
dt
  
2 R R   PV  p g 0 
2 3
  P t  4 R 2 R  8 SR R  16  R R 2 (2.19)
  R  
The term on the left-hand side represents the variation in kinetic energy of the liquid
body. The first term on the right-hand side represents pressure forces acting on the
liquid, while the surface tension forces are represented by the second term. The
27
dissipation rate due to viscosity is expressed as  2e e d , where

ij ij eij stands for the
deformation rate tensor and the integral is taken over the entire liquid volume.
2.3.5 The Collapse of Vapor Bubble

2.3.5.1. Assumptions
In the present section, the effects of viscosity, non-condensable gas and surface tension
are all ignored.
Before the initial time, the bubble is supposed to be in equilibrium under pressure P 0 ,
which is equal to Pv , according to Equation (2.10). From the instant t  0 , a constant
pressure P , higher than Pv , is applied to the liquid. It results in the collapse of the
bubble in a characteristic time  called the Rayleigh time.
This simple model allows us to describe the global features of the first bubble collapse
for an almost inviscid liquid such as water. However, it does not provide an account of
the successive rebounds and collapses actually observed in various physical situations. It
should be noted that, if surface tension were not ignored, the collapse would be only
slightly accelerated.
2.3.5.2. The Interface Velocity

With the previous assumptions, the Rayleigh-Plasset equation (2.17) can be integrated
using relation (2.18) to give:
R 2 R 3  
2
3

 p   p v  R 3  R0 3  (2.20)
As R is negative during collapse, one obtains:
2 p   p v  R0 
3
dR
  3  1 (2.21)
dt 3   R 
The radius tends to zero and the radial inwards motion accelerates without limit. The
numerical integration of this equation allows the calculation of the radius R(t) as a
function of time. The characteristic collapse time or Rayleigh time is:
28
R0
3  dR 

2 P  P0 
0
3
R0
 0.915R0
P  P0
(2.22)
1
R3
 5 6 
The constant 0.915 is the approximate value of where  is the factorial
6 4 3
gamma function.
The value of  is in good agreement with the experimental values for a large range of
initial values of the bubble diameter from about one micrometer to one meter. As an
example, in the case of water, a bubble with an initial radius of 1 cm collapses in about
one millisecond under an external pressure of 1 bar.
R
1 R / R0
0
1 t/τ
Figure 2.3: Evolution of Rt  and Rt  during Bubble collapse

The behavior of Rt  and R t  are shown in Figure 2.3. While the mean value of the
collapse velocity is R0  , R tends to infinity at the end of collapse. For R approaching

0, the interface velocity has the following strong singularity:
3 3
2 p   p v  R0  2
R R  2
R   0.747 0  0  (2.23)
3   R   R
At the end of the collapse, the radius evolves according to the law:
2
R0   t  5
 1.87   (2.24)
R   
29
With the previous numerical values, it is found that R  720 m s for R R0  1 20 .
Such high values of velocity, of the order of half of the velocity of sound in water, lead
us to believe that liquid compressibility must be taken into account in the final stages of
collapse.
It must be kept in mind that some other physical aspects, such as the presence of non-
condensable gas or the finite rate of vapor condensation, will modify bubble behavior.
However, the Rayleigh model exhibits the main features of bubble collapse, particularly
its short duration and the rapid change in its time scale.
2.3.5.3. The Pressure field

The pressure field pr , t  can be determined from Equation (2.14) in which R is known
 can be deduced by derivation, which gives:
from Equation (2.21) and R
3
   p   p v R0
R (2.25)
 R4
The result by the calculation is :
p r , t   p  R  R0  R4  R0 3 
3
r , t     3  4  4  3  1 (2.26)
p  p v 3r  R  3r  R 
The behavior of non-dimensional pressure  at several instants is shown in Figure 2.3.

It exhibits a maximum within the liquid as soon as the bubble radius becomes smaller

than 1 3

4 R0  0.63R0 . The maximum pressure is:
4
 R0 3  3
 3  1
p max  p   4 R 
 max   1
(2.27)
p  pv  R0 3
 3
 3  1
R 
and it occurs at distance rmax from the bubble center given by:
1
 R0 3  3
  1 
rmax 3
 R3  (2.28)
R  R0 
 4 R 3  1
When R R0 becomes small, the two previous relations give approximately:
30
3 3
1 R  R 
 max  4  0   0.157  0  R
4 3R R (2.29)
rmax 3
 4  1.59 (2.30)
R
Very high pressures close to the bubble interface are reached. For example, for
R R0  1 20 , p max  1,260 p bars if p   p v is one bar.
Attention must be paid to the kind of pressure wave that appears in Figure 2.3 during the
collapse of the bubble. As only pressure and inertia forces are taken into account in the
present model, this pressure wave propagating inward must be considered as the effect
of inertia forces only. More complicated models exhibit a similar behavior. From a
physical viewpoint, the violent behavior of bubble collapse results from two main facts:
— the pressure inside the bubble is constant and does not offer any resistance to liquid
motion;
— the conservation of the liquid volume, through spherical symmetry (Eq. 2.11),tends
to concentrate liquid motion to a smaller and smaller region.
2.4 Turbulence Modeling

Turbulence modeling is the key issue in most CFD simulation. The appropriate
treatment of turbulent model will be crucial to the success of CFD. Turbulence could be
thought of as instability of laminar flow that occurs at high Reynolds numbers.
Whenever turbulent is present in a certain flow it appears to be the dominant over all
other flow phenomena. Turbulent flow contains small fluctuation. This resolution of
such small motion requires fine grids and time steps. Such that a direct simulation
becomes unfeasible for high Reynolds numbers. That is why successful modeling of
turbulence greatly increases the quality of numerical simulation
A review for the assessment of turbulence modeling is given by Spalart (2000).

Commonly known models can be classified as follows:
• Zero and one equation models; Spalart-Allmaras model
• Two equation model; k-ε model, k-ω model
31
• Reynolds stress model
• Algebraic stress model
• Detached eddy simulation (DES)model
• Larges eddy simulation(LES) model
2.4.1 Two Equation (k-ε) Turbulence Model

Among the mentioned models, k-ε model Jones and Launder [35] has become the
popular one with well-known deficiencies. It is computationally tractable and robust. A
wide variety of flow structures have been computed with this model. In this model, two
partial differential equations, one for the turbulent kinetic energy (k) and a second for
the rate of dissipation of the turbulent kinetic energy (ε) are solved. The second
transported variable (turbulent dissipation ε) determines the scale of the turbulence,
where as the first variable k, the energy in the turbulence [38].
In this study, Renormalization-group (RNG) k-ε model is used to simulate the

cavitating flow over hydrofoil. The RNG–based k-ε turbulence model is derived from
the instantaneous Navier-Stokes equations, using a mathematical technique called
“Renormalization Group” (RNG) methods. The analytical derivation results in a model
with constant different from those in the standard k-ε model, and additional terms and
functions in the transport equations for k and ε.
The transport equation for the RNG k-ε model for turbulence kinetic energy, k, and its
rate of dissipation ε are below:
 

k    ku i     k  eff k   Gk  Gb    YM  S k (2.31)
t xi x j  x j 
       2
   u i       eff  C1 Gk  C 3 Gb   C 2   R  S  (2.32)
t xi x j  x j  k k
In this equation, G k represents the generation of turbulence kinetic energy due to the
mean velocity gradients. Gb is the generation of turbulence kinetic energy due to the
32
buoyancy. YM represents the contribution of the fluctuation dilatation in compressible
turbulence to the overall dissipation rate. The quantities  k and   are the inverse effect
Prandtl numbers for k and ε respectively and C 1 and C 2 are empirical constants, S k
and S  are user define source terms.
The turbulence kinetic energy k is given by

3
k U avg I 2 (2.33)
2
where U avg is the mean flow velocity.
The turbulence intensity I and the turbulence length l can be found from the flowing
equations[38,39]:
1
l  0.07  L I  0.16 1
(2.34)
Re  8
Also, the turbulence dissipation rate ε defined as
3
3
k2
  C 4
(2.35)
l
The constants in the RNG k-ε model are considered as:

C1  1.42 , C 2  1.68 , C   0.09 ,     k  1.393
2.4.2 Wall Function

The implementation of k-ε turbulence model requires a special attention in the vicinity
of a solid wall. Physically, the flow near the no-slip boundaries is different than the free
stream flow. Viscous effects dominate in the near-wall region. Because of this
difference in the flow structure, a treatment is needed along with the k-ε turbulence
model. A common approach to resolve the near-wall turbulence is the so-called wall
functions [35].
The nondimensional normal distance from the wall, y+ (local Reynolds number), and the
nondimensional velocity, u+, can be used to correlate the profiles in the near-wall region
of turbulent flows
33
ut  wall y u
u  u  y  (2.36)
u  
where u* is the friction velocity and ν is the kinematic viscosity. The above definition of
y+ is referred to as “original definition”. The wall function treatment for fully developed
turbulent flow in the vicinity of a solid wall, assuming that flow has two-layer structure
(viscous sub layer followed by the log layer), is as follows [36]:
y y   11.63

u*   1 (2.37)
 log Ey


  y   11.63
where E and κ are constants. The above definitions for the nondimensional quantities
cause a singularity when there is separation in the flow. To eliminate the computational
singularity, one can use the concept of local equilibrium between production and
dissipation of turbulent kinetic energy. The following can be written [37]:
 wall
 C k (2.38)

This formula relates the wall shear stress to the turbulent kinetic energy. Since k is not
zero at separation, the above form does not cause computational singularities. Using the
flow equilibrium assumption, one can define the nondimensional normal distance,
referred to as “equilibrium definition”, as follows:
1
C  4 k yp

y  (2.39)

where yP is the normal distance of the first grid point away from the wall.
It is recommended that, for standard wall function in k-ε model or when transitional
flows option is not active in k-ω model the y-plus value should be 30 < y+ < 300. (A y+
value close to the lower bound y+ = 30 is most desirable.)
On the other hand, for enhanced wall treatment in k-ε model or when transitional flows
option is enabled in SST k-ω model the y+ at the wall-adjacent cell should be on the
order of y+ = 1. However, a higher y+ is acceptable as long as it is well inside the
viscous sublayer (y+ < 4 or 5).
34
It should be noted here that the y+ value from Equation (2.39) is based on a turbulent
boundary layer on a flat plate. Therefore, it is used only as an estimate in the present
case as the geometry is not actually a flat plate. The y+ values are also solution-
dependant. The actual value of y+ for the hydrofoil form is obtained with the viscous
flow solution. Furthermore, the real y+ is not a constant but varies over the wall surface
according to the flow in the boundary layer.
2.5. Cavitation Modeling

A common approach in cavitation modeling is to use the homogeneous flow theory. In
this theory, the mixture density concept is introduced and a single set of mass and
momentum equations are solved. Some of the existing studies solve the energy equation
and determine the density through suitable equations of state. Since most cavitating
flows are isothermal, arbitrary barotropic equations have been proposed to supplement
the energy consideration [28].
Various methods for numerical simulation of cavitating flow were developed in the past.
Most of the studies are two phase flow as a single liquid–vapor phase mixture flow. The
evaporation and condensation can be modeled with different source terms that are
usually derived from the Rayleigh–Plesset bubble dynamics equation. This approach
was first made by Kubota et al. [19] who used the linear part in the Rayleigh–Plesset to
equation to describe the evolution of bubble radius as a function of surrounding
pressure. At presents different authors proposed to consider a transport equation model
for the void ratio, with vaporization/condensation source terms to control the mass
transfer between the two phases. The cavitation model implemented here is based on the
so called “full cavitation model”, developed by Singhal et al.[23]. It accounts for all first
order effect i.e., phase change, bubble dynamics, turbulent pressure fluctuation and non-
condensable gases. However, the original approach assuming single-phase, isothermal,
variable fluid density flows, the cavitation model in FLUENT is under the frame work
of multiphase flows. It has the capability to account for (N-phase) flows or flows with
multiphase species transport, the effect of slip velocities between the liquid and gaseous
phases, and the thermal effects and compressibility of both liquid and gas phase. The
cavitation model can be used with the mixture multiphase model
35
2.5.1 Multiphase Model

A single fluid (mixture model) approach is used [40]. The basic approach consists of
using standard (Navier–Stokes) viscous flow equation and conventional turbulent model
(RNG k-ε modal). The mass and momentum conservation equations (Eq.2.42&2.43)
together with the transport equation (Eq.2.44) and the equation of the turbulence model
from the set of equations from which fluid density (which is the function of the vapor
mass fraction f v ) is computed. The  m  f v (mixture density-vapor mass fraction)
relation is:
1 fv 1  fv
  (2.40)
m v l
The volume fraction of the vapor phase  v  is related to of the mass fraction of the
vapor phase with:
m
v  fv (2.41)
v
The mass conservation equation for the mixture is:

 m   . m vm   0 (2.42)
t
The momentum conservation equation for the mixture is

t
   
 m vm   . m vm vm   p  .  m vm  vm T  mg  F (2.43)
And the transport equation for the vapor is:


 m f v   . m vm f v   Re  Rc (2.44)
t
2.5.2 Transport Equation-Based Empirical Cavitation Models

Different forms of the transport equation-based cavitation models have been proposed in
literature. The main differences are due to different source terms that are needed for
cavity generation and destruction. All the models have introduced empirical factors to
regulate the mass transfer. These empirical factors have been tuned based on numerical
experimentation. So far, satisfactory results have been produced with these models,
provided that the computational algorithm addresses the numerical issues specific to
cavitating flows. However, the derivation of these models has not been clarified in the
literature and no satisfactory physical arguments have been made regarding the source
of the empirical factors. In what follows, these models are presented and discussed.
36
2.5.2.1 Model-1 [27]

Several researchers have adopted this model [25]. Both volume fraction and mass
fraction forms have been adopted in this model. Evaporation and condensation terms are
both functions of pressure. The liquid volume fraction form is considered.
 L  C MIN PL  PV ,0  L C prod MAX PL  PV ,0 1   L 

    L u   dest  (2.45)
t 
V 0.50  LU 2  t   
0.50  LU 2  t 
1
The empirical factors adopted the following values, Cdest =1.0, Cprod=8.0x10 . These
parameters are determined through numerical experimentation and have been non-
dimensionalized with free stream values to have the correct dimensional form as the
convective terms. The same parameters can be used for different geometries and flow
conditions provided that they are non-dimensionalized with the free stream values. The
discussion holds for the other cavitation models also.
2.5.1.2 Model-2 [24]

The liquid volume fraction is chosen as the dependent variable in the transport equation.
The evaporation term is a function of pressure whereas the condensation is a function of
the volume fraction.
 C MIN PL  PV ,0  L C prod  L 1   L 

2
 L
    L u   dest  (2.46)
t  
0.50  LU 2   L t   Lt
5
The empirical factors adopted in this study have the following values, C prod =9.0x10 ,
4
Cprod=3.0x10 .
2.5.1.3 Model-3 [23]

The working fluid is assumed to be a mixture of liquid, liquid vapor and noncondensible
gas. The vapor mass fraction is the dependent variable in the transport equation. Source
terms Re and Rc that are included in the transport equation define vapor generation
(liquid evaporation) and vapor condensation, respectively. Source terms are the function
of the local flow condition (Static pressure, velocity) and fluid properties (liquid and
vapor phase densities, saturation pressure and liquid vapor surface tension). The source
terms are derived from the Rayleigh-Plesset equation where high order terms and
viscosity term have been left out. The source terms can be expressed as [23].
37

 m f v   . mvm f v   Re  Rc (2.47)
t
k 2 pv  p
Re  C e l v 1  f v  f g  , when p  pv (2.48)
 3 l
k 2 pv  p
Rc  C e l l fv when p  pv (2.49)
 3 l
Where C e and C c are empirical constants, and k is the local kinetic energy,  surface
tension, f v vapor mass fraction and f g mass fraction of noncondensible (dissolve)
gases. C e and C c were determined by comparing experimental and numerical results at

different combination of initial conditions and geometries [23]; their values are taken as
0.02 and 0.01 respectively.
2.6 Numerical Methodology

The baseline Navier-Stokes solver [37,41], employs a pressure-based algorithm and a
finite volume approach to solve the fluid flow and energy equations, written in body
fitted curvilinear coordinates, on collocated, multi-block grids in 2D and 3D domains.
For the present study, the transport equation-based cavitation models, described earlier,
are implemented into the solver and related modifications, regarding the convection
schemes and the pressure-based algorithm, have been made for both steady and time-
dependent computations. To help describe the underlying algorithm, the steady-state
generic transport equation is adopted in vector form as

  u         q (2.50)
where  is the generalized dependent variable, Γ is the diffusion coefficient, and the
second term on the right hand side represents the source term for the transported
quantity  . The above equation is transformed to an integral form, suitable for finite
volume discretization, using the divergence theorem,
 
 u     nds   q dV
s v
(2.51)
which upon integration yields the following equation
38
F i , j  F 1 F1 F 1 F 1  bi , j (2.52)

i , j i ,j i, j  i, j
2 2 2 2
Where bi is the integrated form of the source term, and F represents the flux of  at
each control volume face and is composed of a convective and a diffusive part as
follows:
 
 u      n S cf
conv diff
Fcf  Fcf  Fcf (2.53)
The diffusive flux is discretized using the second order central difference scheme,
whereas the choice of discretization scheme for the convective flux often depends on the
flow conditions and fluid physics [37]. In this study two schemes are considered,
namely the first order upwind (FOU) and the second order controlled variation scheme
(CVS). The first order upwind scheme is employed mainly to probe the sensitivity of the
solution with respect to the choice of the convection operators. The CVS scheme is
employed in all cases for experimental validation. Both convection schemes are
illustrated considering the one-dimensional transport equation of the variable  . In the
first order upwind (FOU) scheme the value of the dependent variable is estimated using
the upwind neighbor value. If one lets fi+1/2 be the first order flux at a control volume
face, determined through first order extrapolations of two immediate neighboring cells,
then the scheme can simply be written as
n 1   n 1  
f 1  i MAX   1 u n 1 ,0   i MAX    1 u n 1 ,0  (2.54)
i
2  i  2 i 2   i 2 i 2 
where u is the velocity component in the corresponding direction.
Higher order spatial accuracy can be obtained by employing more grid points for
extrapolation. However, it is known that second or higher order accurate schemes for
convection terms produce oscillations around discontinuities. Later, in the results
section, it will be shown that shock-like discontinuities in density profiles do appear in
cavitating flows. If a linear second order upwind scheme is used, oscillations in the
vicinity of discontinuities can interfere with the mass transfer cavitation model and lead
to unrealistic solution or even divergence. For this reason, the numerical methodology
must introduce a convection scheme that does not generate nonphysical oscillations in
the vicinity of sharp gradients and discontinuities.
39
The second order accurate CVS scheme is based on the total variation-diminishing
(TVD) concept and is suitable for the present problem. In the CVS scheme, the
convective flux is estimated using the second-order TVD scheme to improve the formal
order of accuracy and the local characteristic speeds are assigned according to the value
of the local convective speed. The second order net flux term for the linearized implicit
version of the CVS is presented as follows:
n
1    1     
  1 b 1  Q 1  1    r 1    i 1  i 
( 2) ( 2) n 1
f 1 f 1 (2.55)
i i 2 i  2  i  2 i
2  2  2   
i
2 2 
n
1    1     
  1   b 1  Q 1  1    r 1    i  i 1 
n 1
2 2  2
i i i
2  2  2   
i

n
1       
  3   r 3   b 3  Q 3   i  2  i 1 
4 i  2   i  2   i  2 i
2 
n
1       
  3   r 3    b 3  Q 3  i 1   i  2 
4 i  2   i  2   i  2 i
2 
The superscript (2) stands for second order accuracy, (n+1) represents the value at the
current iteration, and subscript (i) indicates cell center location. The function ψ(r) and Q
are the Minmod flux limiter and the dissipation function, respectively, and they are
defined as follows:
f i2  f 3 f i 1  f 1
i i
 1, r   max0, min 1, r , r 
1  2
r 
1  2
(2.56)
i
2
f i 1  f 1
i
2
fi  f 1
i i
2 2
1  b2 
     , if b  
Q 1  Q b 1  2    (2.57)
i
2  i  2 
b, if b  
The parameter δ is used to regulate the numerical viscosity; in this study it is assigned as
zero, which results in the clipping of the fluxes. Local characteristic speed b is assigned
according to the value of the cell face velocity.
40
2.6.1 Pressure-Based Algorithm for Steady-State Computations of

Cavitating Flows
The pressure-based algorithm, adopted for steady-state computations, follows the spirit
of the well-established SIMPLE algorithm [42, 43], with substantial extension to treat
issues associated with curvilinear coordinates and multiblock interface. Basically, the
momentum equations are discretized as

u 
u nb  VP  d P P  bPu ,
 
Aup u p   Anb (2.58)
 
where A up and Anb
u
are the coefficients of the cell center and neighboring nodes,

respectively, due to contributions from convection and diffusion terms. V P and bPu
represent the volume of the cell and the source term, respectively. Note that the
 d operator is the discrete form of the gradient operator. For the present study, there is
no source term in the momentum equations; hence, it does not appear in the following
formulation. The above form can also be written in a more convenient way.
 
u p  H u P  D P  d P P (2.59)
where DP is written as
VP / APu 0 0 
 v 
DP   0 VP / A P 0  (2.60)
 0 0 V P / APw 

For collocated grids, DP reduces to a scalar since APu = APv = APw . H is a linear operator
resulting from the discretization of convection and diffusion terms.
The solution procedure is based on the predictor-corrector approach, where the
discretized momentum equations is cast as

 

u P  H u  P 
 DP  d P n 1 P
(2.61)
indicating that the velocity field at any given location is updated based on the existing
values of the neighboring velocity and pressure. The superscript (n-1) stands for the
41
values from the previous iteration. Subtraction of Eq. (2.61) from Eq. (2.59) leads to the
following correction term
 
u p  H u P  D P  d P P (2.62)
The steady-state continuity equation is written in discretized form as
 

 u  n S cf  0 (2.63)
In order to derive corrections for the pressure field, the continuity equation, Eq. (2.71),
is converted to a pressure correction equation by substituting the corrected velocities:
 
 D d P P  n S cf 
P

  U  P ,  (2.64)
where
 
U  u  n S cf (2.65)
Clearly, the above pressure correction equation has a diffusive nature. Note that the

H u  corrections are neglected to have a simpler form of the pressure-correction
equation.
A variety of pressure-based algorithms have been proposed in years following the

SIMPLE algorithm[37]. The differences among these algorithms are due to different

approximations related to the H u  term. Moukalled and Darwish [44] have reviewed
the various pressure-based algorithms and present them in a unified formulation.
In the pressure-based algorithm, the pressure-correction equation has been revised to

achieve successful solutions for highly compressible flows. This formulation will be
described in the context of noncavitating flows with compressibility effects to motivate
the present cavitating flow method. For highly compressible flows, density needs to be
corrected to account for the strong pressure-density dependency. For such a formulation
the flux terms in the continuity equation is
u      u   u     u    u    u    u 
      
(2.67)

where u term is the mass flux entering the control volume. Starred variables represent
42
the predicted value and primed variables represent the correction terms. The inclusion of
the above equation, along with a relation that couples the density to pressure, leads to
the following pressure correction equation:
 P  C  PP (2.68)
 
  D  d P   n S cf 
P

  C  PPU  
P

   U  
P  
  C P PP D d P   n S cf 
P
(2.69)
By comparing Eq. (2.64) and (2.69), one can see that the characteristic of the pressure-
correction equation is altered from a pure diffusive nature to a mixed convective-
diffusive nature in regions where density is a function of pressure. As discussed by Shyy
[37], the relative importance of the first and second terms in Eq. (2.69) depends on the
local Mach number; for low Mach number flows, only the first term prevails, while for
high Mach number the second term becomes important. The Mach number dependency
can be shown through the equation of state. The fourth term is a nonlinear second-order
correction term and it can either be neglected or included in the source term to stabilize
the computation in early iterations.
In the cavitation models described earlier, a convection equation with pressure

dependent source terms is solved to determine the density field. Because of this
coupling between pressure and density, the pressure-correction equation needs to be
reformulated even though the Mach number effect is not explicitly addressed in the
model. Once the cavitation model is implemented into a pressure-based algorithm, the
pressure-correction equation exhibits a convective-diffusive nature in cavitating regions
and purely diffusive nature in the liquid phase. In the present algorithm, the following
relation between density correction and pressure correction is introduced to establish the
pressure-density coupling
 P  C 1   l P PP (2.70)
where C is an arbitrary constant. It should be emphasized that the choice of this constant
does not affect the final converged solution because of the nature of the pressure-
correction equation. Different values of C simply lead to different paths for reaching the
converged solution. However, the convective-diffusive nature of the pressure correction
equation is directly affected by the choice. It can easily be shown that the ratio between
43
the convective strength, the second term in Eq.(2.69), and diffusive strength, first term
in Eq. (2.69), is directly related to C (1-αL). In the cavity region, the liquid mass fraction
decreases, and the pressure-correction equation is of a clear convective-diffusive nature.
On the other hand, in the liquid region, the pressure-correction equation returns to a
purely diffusive type. Furthermore, it is found that a very large value for C can
destabilize the computation in early stage of the iteration process. For this reasons, it is
suggest that C=0(1) be used. It should be noted that the above discussion is based on
normalized variables, with density of the liquid is assigned to 1.
The present pressure-velocity-density coupling scheme is along a path responsive to the
cavitation dynamics. It is found that the proposed scheme mimics the  variation
P
adequately, especially near αL=0, where the variation is very steep. From the point of
view of convergence rate, the scheme performs satisfactorily compared to the non-
cavitating flow computations.
Due to the convective-diffusive nature of the pressure-correction equation for cavitating

flows, the coefficient matrix is noticeably asymmetric. Hence, iterative matrix solvers,
designed especially for the fast efficient solution of the symmetric problems need new
insight into the precondition treatment. In the present study, the conventional relaxation
technique is employed.
Next, the practical implementation of the pressure-velocity-density coupling for

cavitating flows is presented. For simplicity, the following one-dimensional case is
considered to illustrate the nature of the revised pressure correction equation. Note that a
first order upwind scheme is used for the second term in Eq. (2.69):
 
1  C 1   l i 1 max[u 1 ,0]  Pi 1
a i Pi   aiinc   (2.71)
i
 2 
 
1  C 1   l i 1 max[u 1 ,0]  Pi 1  bi
  a iinc  
i
 2 
   inc 
1  C 1   l i 1 max[u 1 ,0]    a i 1  C 1   l i max[ u 1 ,0] 
a i   aiinc    (2.72)
i i
 2   2 
44
In this equation, aiinc inc

1 and a i 1 are the coefficients stemming from an incompressible
formulation, bi is the source term, and  L is the liquid volume fraction. Subscripts (i+1)
and (i-1) stand for the neighboring grid nodes in the east and west direction,
respectively. The above form is a combined incompressible-compressible formulation
that preserves the incompressible nature in the liquid phase. In the cavitating region, it
accounts for the pressure-density dependency in a nonlinear fashion, in accordance with
the local value of αL. This modification is key to a stable computation in which the
uniform vapor pressure is recovered in the final converged solution.
Another aspect is that, similar to compressible flow computations, the density at the cell
face is up winded both in the discretized momentum and pressure correction equations.
The criterion for upwinding is based on the value of liquid volume fraction; that is,
wherever αL is less than 1.0, the cell-faced density value is estimated based on an
upwinded formula. In regions of sharp density gradients, a single point upwinded
extrapolation for density, instead of a two-point interpolation can significantly improve
the convergence level. It should also be emphasized that Eq. (2.68) is not limited to the
cavitation model employed in this study; it can easily be adopted for other cavitation
models.
Finally the major steps of the algorithm are summarized.
1. Compute the discretized momentum equation, Eq. (2.61).

2. Compute the convective-diffusive pressure-correction equation, Eq. (2.69).
3. Correct the velocity and pressure field.
4. Compute the discretized scalar equations for αL, k and ε, resulting from cavitation and
turbulence models, using the corrected velocity and pressure field.
5. Compute the density field and turbulent viscosity field
6. Go to Step 1 and iterate until convergence.
2.6.2 Pressure-Based Algorithm forTime-Dependent Computation of

Cavitating Flows
In SIMPLE type of pressure-based methods, the equations are solved successively by
employing iterations. The advantage gained by implicit differencing of the equations is
coarsened during iterative process and time dependent computations can be highly
45
expensive due to need for iterations. Issa [26] have developed a pressure-based
algorithm called the Pressure-implicit with Splitting of Operators (PISO) for the
solution of unsteady flows. For steady flows this algorithm can be interpreted as an
extension of the SIMPLE algorithm. PISO algorithm employs the splitting of operations
in the solution of the implicitly discretized momentum and pressure or pressure-
correction equations. This makes the solution procedure sequential in time domain and
enables the exact solution of discretized equations at each time step with a formal order
of accuracy. Operator splitting method eliminates the need for under relaxation adopted
in SIMPLE type of algorithms.
Issa [26] has presented the compressible formulation of the algorithm. In this
formulation the pressure-density coupling is introduced only through the time dependent
term of the continuity equation. Bressloff [45] has extended the PISO algorithm for
high-speed flows by adopting the pressure-density coupling procedure in all-speed
SIMPLE type of methods.
In the case of cavitating flow computations, the typical relaxation factors used in the
iterative solution process are smaller compared to the ones used in computations of
single-phase flows, and in addition smaller time steps are needed to study the cavitation
dynamics. Hence, time-dependent computations with an iterative SIMPLE type
pressure-based method become impractical. To eliminate this difficulty, the PISO
algorithm [46] for curvilinear coordinates with all-speed treatment is modified and
extended for unsteady cavitating flow computations. The formulation of the present
algorithm shares similar features of the formulation presented in Moukalled and
Darwish [44].
In the predictor step the discretized momentum equations are solved implicitly using the
old time pressure to obtain an intermediate velocity field. A backward Euler scheme is
used for the discretization of the time derivative term.

   ut
n 1

 

u P  H u  P 
 D P  d P n 1 P
(2.73)
The intermediate velocity field does not satisfy continuity and needs to be corrected
using the continuity equation as a constraint. In the first corrector step, a new velocity
46

field u P , and a new pressure field, P  are sought. The discretized momentum equation
at this step is written as
 u nP1
 
 
u P  H u  P 
 DP  d P n 1
P

t
(2.74)
Eq. (2.73) is subtracted from Eq. (2.74), leading to velocity correction terms
 
u P  u P  D P  d P P (2.75)
If the pressure field depends on the density field, such as in high-speed flows or in
cavitating flows, density field needs to be corrected and the following is written, as
previously discussed in the previous section
 P   Pn 1   P  P  C  PP (2.76)
The discretized continuity equation written for the new velocity field and density field
reads the following
 P   Pn 1
VP    u   n S cf P  0 .

(2.77)
t
Inclusion of the corrected velocity field, Eq. (2.75) and density field, Eq. (2.76),
transforms the above discretized continuity equation into a pressure-correction equation
to be solved in the first corrector step
C  PP
t
 
V P    n 1 D d P   n S cf  P

  C P U  
P
 
   n 1U  P . (2.78)
When Eq. (2.73) is subtracted from Eq. (2.74), the terms containing operator H is lost.
Hence a second corrector step is needed to satisfy the mass conservation. In the second

corrector step a new velocity field, u P , and a new pressure field, P  , are sought. The
discretized momentum equation at this step is written as

  ut
n 1
 
 
u P  H u  P 
 DP  d P  P
P
(2.79)
47
Eq. (2.74) is subtracted from Eq. (2.79), leading to the following correction terms
  
 
u P  u   H u   u   P  DP  d P P (2.80)
The density field needs to be corrected at this step also as follows:
 P   P      P   Pn 1   P   P . (2.81)
Inclusion of the corrected velocity field, Eq. (2.80) and density field, Eq. (2.81),
transforms the discretized continuity equation into a pressure-correction equation to be
solved in the second corrector step
C  PP
t
 
V P     D d P   n S cf 
P

  C  P U  P

   U  
P  
  

    H u   u   n S cf 
P
(2.82)
Issa [28] has shown that addition of each corrector step increases the accuracy by order
of one in time. In this study, it is found that, instead of adding an additional corrector
step, repeating the first corrector and second corrector step one more time gives the
correct time-dependent behavior, such as the vortex shedding behind a cylinder.
Additionally, since the splitting error depends on time step size, smaller time steps are
needed for time accurate computations.
Oliveria and Issa have discussed the coupling of temperature equation to the PISO
algorithm for buoyancy-driven flows. Their study has shown that the PISO algorithm is
amenable to different arrangements in the operator splitting procedure. An improved
method is developed to handle the strong coupling of velocity and temperature in
buoyancy-driven flows. For turbulent cavitating flow computations, special attention is
also needed to couple the cavitation model and the turbulence model equations in the
operator splitting procedure.
As explained in the previous section, the density at the cell face is upwinded based on a
single point extrapolation, both in the discretized momentum and pressure-correction
equations, to enhance mass and momentum conservation in regions of sharp density
gradients.
48
CHAPTER THREE
NUMERICAL SIMULATION
The numerical simulation is conducted in two stages. Firstly, computational models are
created and simulations are run with 2D turbulence models on a hydrofoil. The
computed results are compared to other numerical results to validate the computational
models. Secondly, the main concentration is given on the modeling & computation of
unsteady turbulent cavitation flows and to investigate the vortex phenomena. All of the
investigations are carried out using very efficient commercial software FLUENT 6.3.
26.
3.1 Simulation Using FLUENT 6.3.26

FLUENT 6.3.26 uses a finite volume-based algorithm to transform the governing
physical equations to algebraic equations that can be solved numerically. In such an
approach, the computational domain is subdivided into individual, discrete control
volumes, or cells. The governing equations about each cell are then integrated, yielding
discrete equations that conserve each quantity on a control-volume basis. Consider the
following steady-state conservation equation for transport of a scalar quantity  written
in integral form for an arbitrary control volume, V [38]:
   
   . dA     . dA  S dV (3.1)
V
where ρ = density

 = velocity vector (= ui + vj in 2D)

A = surface area vector
 = diffusion coefficient for 

 = gradient of 
S = source of  per unit volume

Chapter Three: Numerical Simulation
This equation is applied to each cell in the computational domain. FLUENT discretizes
this integral equation as:
N faces  N faces 
 
 f  f A f 
f
 f
     A f  S  V
 n
(3.2)
Where, Nfaces = number of faces enclosing cell
 = value of  convected through face f


f = mass flux through the face
Af = area of face f, A
   
   = magnitude of   normal to face f
 n
V = cell volume
The equations solved by FLUENT that lead to a full description of the flow field around
a given object take the same form as the discretized equation above.
3.2 Steady Flow over a Hydrofoil
3.2.1 Geometry of CAV2003 Hydrofoil

The section of the hydrofoil is presented in Figure 3.1 which shows a schematic view of
the NACA0012 hydrofoil geometry. The nominally two–dimensional configuration is
mounted at a zero degree angle-of-attack. The equation for the upper surface of the
symmetric foil geometry of CAV2003 hydrofoil is provided as:
2 3 4
y x  x  x x x
 a0  a1    a 2    a3    a 4   , (3.3)
c c c c c c
y  a 0 x  a1 x   a 2  x   a3  x   a 4  x 
2 3 4
(3.4)
with a 0  0.11858 , a1  0.02972 , a 2  0.00593 , a 3  0.07272 , a 4  0.02207 .
y  y c and x  x c is the dimensionless coordinate along the chord line ranging from
-0.05 at the leading edge to 0.05 at the trailing edge. Using same kind of calculation
geometry for NACA0012 can be obtained.
3.2.2 Computational Domain
50
The flow field around the hydrofoil is modeled in two dimensions. The flow from left to
right with the hydrofoil of chord length c 1 m submersed in an incompressible fluid is
considered. The computational domain is sketched in Figure 3.1. The hydrofoil is
located almost at the middle of a channel of length at left & right side of the foil is 25c
& 30c respectively and height 20c.
Figure 3.1 shows the total 2D computational domain and boundary conditions. The inlet
boundary condition is specified velocity inlet with a constant velocity profile. Upper and
lower boundaries are velocity inlet as the inlet. The outlet uses is a constant pressure
boundary condition. The foil itself is a no-slip wall, i.e., u = 0, v = 0 at the foil surface.
In the physical domain the flow is not confined. Nevertheless, a fictitious external
rectangular boundary is needed at a large distance from the hydrofoil in order to solve
the governing equations numerically. Also the flow at exit is treated as a pressure outlet.
The problem setup together with the important dimensions is shown in Figure 3.1.
Figure 3.1: Schematic Diagram of the Flow Field around NACA0012

Hydrofoil with Boundary condition
3.2.3 Grid Generation

GAMBIT, the preprocessor of FLUENT is used to generate the two dimensional grid
around a NACA0012 hydrofoil in this study. The quality and efficiency of the numerical
51
simulation is highly dependent on the construction of the grid used in the computational
model. Several factors must be considered when generating a grid to ensure that the best
possible numerical results are obtained with a particular solution algorithm. Grid point
placement can have a substantial effect on the stability and convergence of the
numerical solver. In order for a computational fluid dynamics code to provide a
complete flow field description for a particular problem, the user must specify a grid
that tells the flow solver at what locations in the problem domain the solution is to be
computed. The specification of the grid’s construction can have a major influence on the
fidelity of the solution and can, in fact determine whether a solution is even attainable.
A typical computational mesh is used for simulation as shown in Figure 3.2. This
particular mesh has approximately 17080 nodes, 33880 faces and 16800 quadrilateral
cells with considerable mesh concentration both around the hydrofoil and in the wake.
To facilitate meshing, we employed a C-type grid system around the hydrofoil.
(a)
52
(b)
Figure 3.2: Grid Lines in Mesh: a) Overall View b) Close-up View Near the
Hydrofoil
3.2.4 Grid Study

A detailed grid study is carried out for the non-cavitating flow past the NACA0012
hydrofoil using the commercial software FLUENT 6.3.26. The flow geometry, the inlet
and outlet conditions are specified in Figure 3.1. The C-grid topology is adopted, and
the boundary condition used in the simulations is shown also in that figure.
Grid is used to investigate the influence of grid parameters such as the number of
control volumes, the minimum grid spacing around the hydrofoil on lift force and drag
force. A typical grid is shown in Figure 3.2. Most of the cells are located around the foil
and a contraction of the grid is applied in its upstream part to obtain an especially fine
discretization of the areas where cavitation is expected.
53
Table 3.1: Parameter of grid study

Grid No of No of No of Lift Drag
no. Cells faces nodes force force
1 12555 25374 12819 1.4347 0.2876
2 17100 34565 17465 0.8648 0.4085
3 17100 34565 17465 0.8091 0.3032
4 17100 34565 17465 0.6019 0.0161
5 16800 33880 17080 0.5017 0.0383
First of all to minimize the CPU time, we tried to create a mesh topology that would
allow a coarse grid in the region near the hydrofoil. The first configuration is rejected
due to lack of required minimum spacing near the hydrofoil. So we increase the number
of grids around the hydrofoil specially at the leading edge to achieve accuracy
comparable to the result of Nobuhide Takasugui, Hajime Yamaguchi, Hiroharu Kato,
Masatsugu Maeda, H. Ghassemi , Chen Ying, LU Chuan-jing, WU Lei. At GRID 2,3 &
4 we set the distance between hydrofoil & the boundary less, as a result the value of CL
is has greater value than GRID5. After trying with these schemes we moved to the
GRID5 scheme & found a better result than the other remaining scheme. In GRID 2,3 &
4 we have tried with different state such as cavitating, noncavitating or steady and
unsteady analysis maintaining the same geometry to find out a result which is similar
with the values of other researchers, but the result shows quite difference from the
predicted one. Than we have increased the boundary distance from the hydrofoil to get a
better result as we have to decrease the lift coefficient from the previous one and lift
coefficient has a great dependency on the distance between the hydrofoil & the
boundary. The GRID 5 shows the value of lift & drag coefficient similar to the value
computed by Chen Ying. So we have choose the Grid 5 for our further analysis.
3.3. Boundary condition

Four boundary conditions of the respective domain are used in this study. The upstream
and the downstream boundary of the domain are modeled as velocity inlet and pressure
outlet respectively and both the upper side and lower side of the domain have velocity
inlet boundary conditions. The surface of the foil is modeled as a wall.
54
Velocity inlet boundary conditions are used to define the flow velocity, along with all
relevant scalar properties of the flow, at flow inlets. The total (or stagnation) properties
of the flow are not fixed, so they will rise to whatever value is necessary to provide the
prescribed velocity distribution.
This boundary condition is intended for incompressible flows where the magnitude and
direction of the inlet velocity is known. The velocities are set to given values, and the
first derivative of the pressure with respect to the axial direction is taken equal to zero.
p p
u, v   u, v given ,  0
n x
The pressure outlet boundary condition is used to model flow exit where the details of
the flow and gauge pressure are not known. Pressure outlet boundary conditions require
the specification of a static (gauge) pressure at the outlet boundary. The user has no
need to define any condition at the exits: Fluent extrapolate the required information
from the interior of the domain. As long as the flow at the exits is expected to be well-
developed and incompressible (in this study, given the geometry of the domain and the
free stream velocity), application of the pressure outlet boundary condition to the exits
boundary is a reasonable choice.
To set the static pressure at the pressure outlet boundary, entering the appropriate value
for Gauge pressure in the pressure outlet panel has great impact in simulate the solution
and the Gauge pressure is strongly related to the operating pressure set.
Velocity inlet boundary condition at upper & lower portion of the foil is used for the
physical geometry of interest, and the expected pattern of the flow solution. They can
also be used to model zero-shear slip walls in viscous flows.
Finally, the wall boundary condition is used to separate fluid and solid regions. In
viscous flows (such as in this study), the no-slip or zero tangential velocity boundary
condition is enforced when the wall boundary condition is imposed; the shear stress and
associated friction drag is computed based on the flow details in the local flow field.
The boundary condition of the hydrofoil is shown in Figure 3.1
3.4 Solver Initialization and Flow Solution

After the grids are constructed, the next step is to import them into FLUENT 6.3.26 the
numerical solver. Since each grid is exported from GAMBIT in FLUENT 6.3.26’s
55
native format, the import process is straightforward. After the grids are imported, the
solver is initialized. This procedure involved several steps, such as:
-Selecting the solver formulation
-Defining physical models
-Specifying fluid properties
-Specifying boundary conditions
-Adjusting residual
-Initializing the flow field
-Iterating
For all of the grids generated in this study, the pressure based solver formulation is used.
This approach solves the continuity, momentum, and energy equations sequentially as
opposed to simultaneously. Because the pressure based solver is traditionally used for
incompressible and mildly compressible flows, given the flow conditions being
investigated, this study is well-tailored for its use. Also, the solution controls (under-
relaxation parameters, etc) are left at the default settings for all cases. Next, the flow
around the geometries modeled in this study is approximated as steady-state. Although
this choice precluded the ability to capture vortex shedding and other time-dependent
effects, the majority of pertinent flow features and their associated drag effects (like
separation point, pressure drag, and skin friction drag) could still be accurately modeled.
One reason for choosing steady-state simulations is because of the reduced
computational load they placed on the computer – given the total number of simulations
needed, run time is a major limiting factor. The other, more compelling reason for
choosing a steady-state modeling approach is that the properties of interest in this study
is steady state values. Although, in reality, separation point, lift, pressure, and drag all
vary with respect to time on a microscopic scale, their net values and net effects can
usually be considered steady properties and can accurately be modeled as such. The user
interface updates based upon whether the steady or unsteady solver is selected. The time
step size, the number of iterations per time step, the total number of time steps, and the
convergence limit for each time step needs be specified when the unsteady solver is
used. On the other hand, the total number of iterations and the convergence limit needs
to be specified for steady solver.
56
For the geometries modeled in this research, definition of the physical models simply
involves specifying turbulent simulation is desired in the solution computation. A
turbulent model of hydrofoil is also simulated to compare with experimental results. In
this stage, The RNG (Renormalization group) k-ε model with Enhanced Wall Treatment
is used in all of the cases. For velocity- pressure coupling the SIMPLE algorithm is used
for unsteady case. The second-order discretization scheme for the pressure equation and
quick discretization for momentum equation are used throughout this study.
The mixture model is a simplified multiphase model that can be used to model
multiphase flows where the phase moves at different velocities. In this study we
consider two phase only, water- liquid as a primary phase and vapor as secondary phase.
Specifying the fluid properties and the boundary conditions are straightforward. For
every computational model in the study, the default fluid properties for water-liquid
(water-liquid at standard conditions) and water vapor are used. Also, the boundary
condition types are all specified during the grid generation step, so the only addition
needed is to specify the inlet velocity.
For all flow cases, the flow field is initialized from the inlet boundary condition. This
process is necessary to provide a starting point for the evolution of the iterative solution
process. In every case, after the flow is successfully initialized, the solution is iterated
until one of the following three conditions is attained: convergence, divergence, or non-
decaying oscillation of the residuals. Convergence is declared if the x-velocity, y-
velocity, and continuity residuals all dropped below 0.001.
57
CHAPTER FOUR
RESULTS AND DISCUSSION
4.1 Definitions
The major benefit of CFD analysis is its ability to compute the values of every flow
parameter at each grid point in the domain studied, giving a very descriptive picture of
the entire flow field. The present study can be classified into two main parts: Firstly, the
non cavitating flow and cavitating flow over the proposed NACA0012 hydrofoil &
secondly, modeling & computations of unsteady turbulent cavitations flow. In this
study, different hydrodynamic parameters are computed for a wide range of cavitation
number (0.2-1.4) which is calculated as:
p   p v (T )
 (4.1)
1
 U 2 
2
where p [Pas],   [ kgm 3 ] and U 2  [ ms 1 ] are the free stream pressure, free stream
density and free-stream velocity, respectively, and where p v T  is the vapor pressure of
water at temperature T [K] . Note that a higher cavitation number indicates that the
pressure in the flow much decreases greatly before cavitation occurs. A small cavitation
number indicates that a smaller decrease in pressure causes cavitation. Thus, a low
cavitation number corresponds to high susceptibility for cavitation.
The drag coefficient is computed as
D L
CD  CL  (4.2)
1 1
 ref U ref 2 S  U 2  S
2 2
where D = drag force, L= Lift force and S = appropriate reference area
As explained earlier, the drag force D consists of two parts: the pressure drag and the
skin friction drag. Pressure drag is simply the component of the pressure force acting in
the axial direction. Skin friction drag is a function of the fluid viscosity and the velocity
profile at the surface of the body. FLUENT 6.3.26 directly computes both pressure drag
and skin friction drag by numerically integrating the incremental value of each at every
nodal point along the body surface.
Chapter Four: Results and Discussion
The pressure coefficient is calculated using the formula:
p  p
Cp  (4.3)
1
 U 2 
2
where p  local pressure
p  = free stream pressure
 ref = reference density
U   free stream velocity
In the following we employ the - C p coefficient.
4.2 Non-Cavitating Analysis

This section will present the methodology that has been applied to select the mesh and
the model of turbulence used for the cavitating computations. The comparisons will be
based on convergence criteria and predicted values of lift and drag.
4.2.1 Influence of the Mesh

First of all, an analysis of the influence of the mesh on the non cavitating predictions is
performed. In the x-y plane C-grid topology is employed as illustrated in Figure 3.1. A
total of 16800 cells are employed in the 2D model. A typical grid is selected for
simulation of the flow.
4.2.2 Influence of Turbulence model

To simulate non cavitating flow there exists a great influence of different turbulence
models. Here, we simulate the noncavitating flow with the turbulence models of RNG k-
ε with enhanced wall Treatment. The values of lift & drag coefficient are calculated at
different angles of attack, in which the viscous and pressure parts are analyzed
separately. The results show a good agreement on the predictions.
Finally, we compared the k-ε model with the result of Nobuhide Takasugi, Hajime
Yamaguchi, Hiroharu & Masatsugu.
56
Table 4.1a: Variation of lift coefficient at different angles (α) of attack
Angle of attack (α) Lift coefficient (CL)
0 -0.0004
1 0.0650
2 0.1200
3 0.1920
4 0.2540
5 0.3180
6 0.3820
Table 4.1b: Drag coefficient (CD) at different angles of attack
Angle of attack (α) Drag coefficient (CD)
0 0.0187
1 0.0184
2 0.0162
3 0.01469
4 0.0117
5 0.0065
6 0.000937
57
4.2.3 Comparison with other Researchers

Computed lift and drag coefficient for the non cavitating flow with turbulence model k-ε
is compared with the results of Nobuhide Takasugi, Hajime Yamaguchi, Hiroharu &
Masatsugu.
Table 4.2: Comparison of lift coefficient CL for non cavitating case
Angle of attack (α), Present value CL Experimental value from

degree Takasugi,Hajime,Kato&
Maeda
0 0 0
1 0.065 0.05
2 0.120 0.100
3 0.192 0.160
4 0.254 0.210
5 0.318 0.310
6 0.382 0.36
The present result shows a good agreement on the prediction of the total lift. We have
done our further work with the k-ε model with enhanced wall treatment for cavitating
calculations.
4.3 Cavitating Analysis

The steady & unsteady phenomena in cavitating condition is analyzed in this section.
This section presents results computed for the typical cavitation numbers such as σ=0.2,
0.4, 0.6, 0.8, 1.0, 1.2 & 1.4. It also presents hydrodynamic parameters computed for
different cavitation numbers and the effect of angle of attack on cavitating flow.
Predicted results are compared with the numerical & experimental results of other
researchers.
58
4.3.1. Numerical Simulation

In the simulation of the cavitating flow, the commercial software package FLUENT
6.3.26 is used. It is a two/three dimensional structure mesh code that solves a set of time
dependent Reynolds-averaged Navier-Stockes equations (URANS) in a conservative
form. The numerical model uses an implicit finite volume method, based on a SIMPLE
algorithm of Patankar [41], associated with multiphase and cavitation model. For
numerical simulation of cavitating flow, a bubble dynamics cavitation model is used to
describe the cavity formation. The RNG k-ε turbulence model with enhance wall
treatment is used as a turbulence model. The corresponding y+, the distance from the
wall to the first grid points in viscous unit is on the order of unity. A second order
central scheme is used for discretization for space except for the convective terms. The
convective term in the momentum equation is discretized by the QUICK scheme for non
cavitating flow and second order implicit scheme is used for cavitating problem.
Pressure based solver SIMPLE is used as the velocity pressure –coupling algorithm.
The convergence criterion was determined by observing the evaluation of different flow
parameters (velocity magnitude at inlet, static pressure behind the hydrofoil) in the
computational domain. For computation, residual is taken be 10-4. An important
parameter in FLUENT 6.3.26 cavitation model is the mass fraction of the
noncondensable gas (N.C.G). We found that the result is quite sensitive to the value of
N.C.G. Based on the preliminary test computation the value 1× 10-6 was found to give
reasonable results and used for computation.
Choosing time step has great influence on simulation of cavitating flow. Different time
step values were tested, eventually the time step for unsteady computation was set to
5×10-5 and approximately 30 iteration per time step were needed to obtain a converged
solution. The value  l  998.2 kg/m3,  v  0.5542 kg/m3,  l  10 3 pas,
 l  1.34  10 5 pas,   0.0717 N/M for liquid and vapor density, liquid and vapor
dynamical viscosity and surface tension respectively, were used for simulation.
4.3.2. Analysis of the result

To predict the behavior of the cavitating flow for different values of cavitation number
such as σ = 0.2, 0.4, 0.6, 0.8, 1.0, 1.2 & 1.4, we first present comparisons of the
computed lift and drag coefficient for cavitating flow with
N.Takasugi,H.Yamaguchi,H.Kato,M.Maeda & H.Ghassemi. Table 4.3 shows that the
lift coefficient and the drag coefficient are in good agreement with published results.
59
Table 4.3: Comparisons of the lift coefficient due to cavitation number(α=10 deg.)
Cavitation Present value Experimental Values of CL Experimental

number(σ) of CL values of CL from values of CL
from H.Ghassemi
Takasugi,
Yamaguchi,
Kato, Maeda
0.2 0.270 0.200 0.170 ----
0.4 0.372 0.300 0.390 0.370
0.8 0.544 0.600 0.610 0.540
1.0 0.583 0.600 0.600 0.600
1.2 0.605 0.600 0.600 0.610
1.4 0.615 0.600 0.600 0.600
The time average value of lift coefficient calculated by present method for the cavitation
number σ = 1.2 and σ = 1.4 are very close to the numerical result of Nobuhide
Takasugi,Yamaguchi, H. Kato, M. Maeda & H.Ghassemi. The total lift is also due to
two effects: the pressure field and the viscous force calculated by Equation (4.2). The
drag force is the result of the convective motion of the hydrofoil through the fluid.
Because of no-slip condition of the wall, a tangential velocity gradient is created in the
direction normal to the wall.
The drag & lift coefficient also calculated in unsteady condition. For calculating at
unsteady condition the RNG k-ε model is used with enhanced wall treatment. The
values of lift & drag coefficient are given below:
60
Table 4.4: Lift & drag coefficient at various cavitation number at unsteady
condition
Cavitation number(σ) Lift coefficient(CL) Drag coefficient(CD)
0.2 0.270 0.049
0.4 0.440 0.053
0.6 0.623 0.049
0.8 0.789 0.034
1.0 0.858 0.012
1.2 0.863 -0.010
1.4 0.850 -0.030
61
Figure 4.1:Comparison of the lift coefficient at various angle of attack
Figure 4.2: Drag coefficient under non cavitating condition
62
Figure 4.3:Comparison of the lift coefficient due to variation of cavitation number at 10

degree angle of attack
63
Figure 4.4: Variation of drag coefficient due to the variation of cavitation number
at 10 degree angle of attack
Cavitation number (σ)
Figure 4.5: Comparison of lift & drag coefficient due to the variation of cavitation
number at 10 degree angle of attack with other researchers
64
Figure 4.6: Lift & drag coefficient with variation of cavitation number in unsteady
condition.
65
Figure 4.7: Time history of lift coefficient at 4 degree angle of attack
Figure 4.8: Time history of drag coefficient at 4 degree angle of attack
66
σ=0.2
Figure 4.9 : Contours of pressure coefficient (left) and the vaor volume fraction
(right) at σ=0.2 at steady state condition
σ=0.4
σ=0.6
Figure 4.10 : Contours of pressure coefficient (left) and the vaor volume fraction
(right) at σ=0.4 & σ=0.6 respectively at unsteady state condition
67
σ=0.8
Figure 4.11: Contours of pressure coefficient (left) and the vaor volume fraction
(right) at σ=0.8 at unsteady state condition
σ=1.0
σ=1.2
Figure 4.12: Contours of pressure coefficient (left) and the vaor volume fraction
(right) at σ=1.0 & σ=1.2 respectively at unsteady state condition
68
Pressure distribution curve of NACA0012 at steady condition (cavitating):
Figure 4.13: Pressure distribution over the hydrofoil at σ= 0.2
69
70
Pressure distribution curve of NACA0012 at unsteady condition:
Figure 4.18: Pressure distribution over the hydrofoil at σ= 0.2 in unsteady condition
71
72
73
74
4.4. Computations of unsteady turbulent cavitations flow over

NACA0012 hydrofoil
Computations was carried out under the condition of σ=0.6, with attack of 6º,
Re=1000000. After computation proceeds for enough time steps, it approaches a period
of phenomenon shown in figure 4.4.1, which means, the cavity lengthen & shorten
periodically, together with vortex cavitation near the trailing edge the foil upside. When
cavity length increasing, the pressure & density gradient between cavity & trailing edge
part increases, that makes a vortex to form, until cavity length & vortex scale reaches
maximum. Then the velocity field, the vortex cavitation was flowed downstream &
collapses the enhanced pressure pushes the cavity shortens a minimum length, and then
starts another period.
In this case, the cavity length oscillates between about 1/4 to 1/3 chord length, in
accordance with the trailing edge vortex cavitation growth & collapsing.
Trailing edge vortex cavity together with leading edge attached cavity have been
observed in experiment by Kubota et al [17] in Fig 4.4.2 & simulated by numerical
method further, which is shown in Fig 4.4.3., also pressure contour obtained by
numerical method by Chen Ying, LU Chaun-Jing, WU Lei shown in figure 4.4.4.And
after that the pressure contour found by the Fluent 6.3.26 through analysis of flow over
the hydrofoil NACA0012 has shown in figure 4.4.5 for comparison with the other
researchers. The near portion of the cavity oscillates cyclically. Not only does the cavity
length change, but the cavity also rises up at its rear part. The unsteady characteristics
computed here agree well with the experimental observations of attached cavitation.
The detailed pressure distribution & vortex velocity field computed here by us in Fig
4.4.5. and Fig 4.4.6. The pressure distribution & vortex velocity field computed by Chen
Ying, LU Chaun-Jing, and WU Lei is given in Fig 4.4.7 for the comparison with the
present case. It is obvious to see the similarity of our simulation results those ones
ahead.The foil surface pressure where cavity exists is almost constant & equal to vapor
pressure. Figure 4.4.6 shows the close-ups of velocity vectors around the cavity &
Figure 4.4.7 shows the velocity vectors around the cavity found by Chen Ying, LU
Chaun-Jing & WU Lei. The present value is quite similar with the value found by the
researchers. The velocity vectors around the cavity suggest that the mechanism of the
cavitation cloud shedding.The density change of fluid due to cavitation has a strong
effect on vorticity dynamics. Figure 4.4.8 presents lift coefficient in two typical
periods.It has been proved by the experiments that the Strouhal number of cavity
75
shedding & collapsing is nearly constant at such cavitation number by Le et al. And
obvious period can be seen in lift coefficient figure. Lift force oscillates between 0.33-
0.57 in one period , reaches its maximum value almost at the moment when leading
cavity approaches its most length and the vortex cavitation grows up to its extremity.
This case is mainly because that these moments, the total pressure inside leading cavity
& vortex cavity reached the maximum value. Figure 4.4.9 shows the lift force found by
the Chen Ying , LU Chuan-Jing, WU Lei & the figure shows the lift force oscillates
between 0.33-0.58 in one period which is quite similar to the present value. But the
curve shown by the researchers is slight different in a sense that it had shown more ups
and downs than the present because in present case time period is taken less than the
published results, because it needs more time to deal with the larger time period. Again
any jet is not used in present case as it was used by the researcher, that is why re-entrant
flow is not simulated. The partial cavity surrounded body is in high unstable force
environment, which may lead to unsteadiness of movement of underwater vehicle or
rotation of turbo-machinery.
76
(1) (2)
(3) (4)
(5) (6)
77
(7) (8)
(9) (10)
Figure 4.4.1(a): Formation of vortex on the upside of the foil NACA0012
78
Figure 4.4.1(b): Periodic forming & collapsing of vortex cavitation at σ=0.6,
attack =6˚
79
Figure 4.4.2: Experimental vortex cavity
Figure 4.4.3: Numerial pressure distribution by Kubota et al.
80
Figure 4.4.4: Pressure contour obtained by the numerical calculation by Chen

Ying, LU Chaun-jing, WU Lei
Figure 4.4.5: Pressure contour obtained at present
81
Figure 4.4.6: Velocity & pressure distribution of vortex cavitation
Figure 4.4.7: Velocity & pressure distribution of vortex cavitation obtained by

Chen Ying, LU Chuan-Jing, WU Lei.
82
Figure 4.4.8: Period of lift coefficient of NACA0012
Figure 4.4.8: Period of lift coefficient of NACA0012 obtained by Chen Ying, LU

Chuan-jing, WU Lei.
83
84
85
CHAPTER FIVE
CONCLUSION AND RECOMMENDATION
5.1 Conclusion
Two-dimensional finite volume method has been applied to simulate incompressible flow
around NACA0012 hydrofoil using CFD software FLUENT 6.3.26. From the above study,
following conclusion can be drawn:
1) Finite volume method seems very prospective for simulation of flow around
hydrofoil in both non-cavitating and cavitating conditions.
2) To capture boundary layer, turbulence model k-ε Renormalization-group (RNG)

with enhanced wall treatment is used for simulation of steady flow around hydrofoil
in non-cavitating condition. It is observed that the RNG k-ε model with enhanced
wall treatment compute the lift coefficient almost accurately. However, only RNG
k-ε model with enhanced wall treatment is used for simulation of cavitating flow
because of its better performance.
3) Unsteady turbulent cavitation flows around a NACA0012 hydrofoil is simulated by

two dimensional computations of viscous compressible turbulent model. The flows
are calculated with finite volume scheme associated with a barotropic vapor/ liquid
state law which strongly links density & pressure variation. To simulate turbulent
flow RNG k-ε model with enhanced wall treatment is used. The results of the
NACA0012 foil shows quasi-periodic vortex cavitation phenomenon.
4) Cavitating flow around NACA0012 hydrofoil at different cavitation numbers is
analyzed by two dimensional multiphase cavitation models with k-ε turbulence
model.
5) Pressure based 2D solver with RNG k-ε turbulent model is used for simulation of
cavitating flow at different angle of attack. Presence of hydrofoil with a positive
angle of attack causes the sharp pressure and velocity gradient near the nose. The
lift force and drag force increase with the increase of angle of attack.
Chapter Five : Conclusion and Recommendation
In general, it can be concluded that, the RNG k- ε model with enhanced wall treatment
shows better performance for noncavitating and cavitating condition.
5.2 Recommendation for Further Study

Following topics can be investigated further from cavitating flows:
(i) In this study, two dimensional models is used, however three dimensional model
also could have been used although the computational cost could increase in that
case.
(ii) This research is restricted only to sheet cavitation. However, the formation and
collapse of cloud cavity and related generation and radiation of shocks waves can be
investigated.
(iii) The unsteady turbulent flow at the wake behind a cavitating hydrofoil can be
studied.
(iv) In this study , we have not consider the jet for generating re-entrant flow in unsteady
turbulent cavitation flow analysis. This effect of jet on the foil can be studied.
85
APPENDIX
Appendix A: Hydrodynamic Parameters
Figure A.1: Two dimensional definition of l max , l t max , t max and l d .
87
REFRENCE
1. Franc, J.P., Michel, J.M., (2004): Fundamentals of Cavitation. Kluwer Academic

Publishers.
2. Franc, J.P.,Michel, J.M., (1988): Unsteady Attached Cavitation on an Oscillating
Hydrofoil. Journal of Fluid Mechanics, 193:171–189.
3. Brennen, C. E., (1995): Cavitation and Bubble Dynamics, Oxford University Press,
Oxford.
4. Arndt, R.E.A, (2002): Cavitation in Vortical Flows. Annual Review in Fluid
Mechanics, 34: 143–175.
5. Franc, J.P., (2001): Partial Cavity Instabilities and Re-Entrant Jet. In CAV2001:
Fourth International Symposium on Cavitation. California Institute of Technology,
Pasadena, USA.
6. Franc. J.P. (2005): Cavitation. In Fluid Dynamics of Cavitation and Cavitating
urbopumps. CISM, International Centre For Mechanical Sciences.
7. Arakeri, V. (1975): Viscous Effects on the Position of the Cavitation Separation of
the Smooth Bodies, Journal of Fluid Mechanics, 68(4): pp, 779-77
8. Franc and Michel (1985): Attached Cavitation and Boundary Layer: Experimental
Investigation and Numerical Treatment, Journal of fluid Mechanics, 154: pp, 63-90.
9. Knapp, R.T., Daily, J.W., Hammitt,F.G., (1970): Cavitation. McGraw-Hill Book

Company.
10. Thornycroft, J., Barnaby, S.W. (1895): Torpedo Boat Destroyers. Minutes of
Proceedings of the Institution of Civil Engineers, 122:51–103.
11 Trevena, D.H., (1987): Cavitation and Tension in Liquids. Adam Hilger.
12. Plesset, M.S.,(1949): The Dynamics of Cavitation Bubbles. Journal of Applied
Mechanics, 16:277.
13. Helmholtz, H.,(1868): On Discontinuous Movments of Fluid . Philosophical
Magazine, 36:337-346.
14. Kirchhoff, G.,(1869): Zur Theorie freier Fl¨ussigkeitsstrahlen. J. reine angew. Math.,
70:289–298.
15. Birkhoff, G., Zarantonello, E.H., (1957): Jets, Wakes, and Cavities. Academic Press
Inc.
16. Wu.T.Y-T., (1972): Cavity and Wake Flows. Annual Reviews of Fluid Mechanics ,
4:243-284.
17. Tulin, M.P., (1953): Steady Two-Dimensional Cavity Flows about Slender Bodies.
David Taylor Model Basin Report 834, 1953.
86
18. ITTC 1999, (1999): Final report of the Specialist Committee on Computational
Method for Propeller Cavitation. In Proceedings of 22nd International Towing Tank
Conference, volume 3.
19. Kubota, A., Hiroharu, K., Yamaguchi,H., (1992): A New Modeling of cavitating
flows, a Numerical Study of Uunsteady Ccavitation on a Hydrofoil section, J. Fluid
Mech. 240 , 59–96.
20. Schnerr, G.H., Sauer, J.,(2001): Physical and Numerical Modeling of Unsteady
Cavitation Ddynamics, in: 4th International Conference on Multiphase Flow, ICMF-
2001, New Orleans, USA.
21. Frobenius, M., Schilling, R., Bachert, R., Stoffel,B., (2003): Three-Dimensional,
Unsteady Ccavitation effects on a single Hydrofoil and in a radial Pump –
Measurements and Numerical Simulations, Part two: Numerical simulation, in:
Proceedings of the Fifth International Symposium on Cavitation, Osaka, Japan.
22. Alajbegovic, A., Groger, H.A., Philipp, H., (1999): Calculation of Transient
Cavitation in Nozzle using the Two-Fluid model, in: 12th Annual Conference on
Liquid Atomization and Spray Systems, Indianapolis, IN, USA.
23. Singhal, A. K., Athavale, M. M., Li, H.Y. (2002 ): Mathematical Basis and
Validation of the Full Cavitation Model, Journal Fluid Engineering, Vol. 124, No. 3,
pp. 617–624.
24. Kunz, R., Boger, D., Chyczewski,T., Stinebring, D., Gibeling, H., (1999): Multi-
phase CFD Analysis of Natural and Ventilated Cavitation about Submerged Bodies,
ASME FEDSM99-7364 ASME FEDSM99-7364, SAN Francisco.
25. Merkle, C.L., Feng, J., Buelow, P.E.O., (1998): Computational Modeling of the
Dynamics of Sheet Cavitation, in: 3rd Int. Symp. on Cavitation, Grenoble, France.
26. Senocak, I., Shvy, I.W., (2002): A Pressure-based method for Turbulent Cavitating
flow Simulations, J. Comput. Phys. 176 , 363–383.
27. Owis, F.M., Nayfeh, A.H., (2004): Numerical Simulation of 3-D Incompressible,
Mmulti-phase Flows over Cavitating Projectiles, Eur. J. Mech. B Fluids 23, 339–
351.
28. Delannoy, Y., Kueny, J.L., (1990): Two Phase Flow Approach in Unsteady
Cavitation Modeling, in: Cavitation and Multiphase Flow Forum, ASMEFED,vol.
98, pp. 153–158.
29. Coutier-Delgosha, O., Fortes-Patella, R., Reboud, J.L., (2003): Evaluation of
Turbulence Model Influence on the Numerical Simulations on Unsteady Cavitation,
J. Fluids Engrg. 125, 38–45.
30. Hofmann, M., Lohrberg, H., Ludwig, G., Stoffel, B. B., Reboud, J.L., Fortes-
Patella, R., (1999): Nume rical and Experimental Investigations on the Self –
Oscillating behaviour of Cloud Cavitation – Part 1: Visualisation, in: Proceedings of
the 3rd ASME/JSME Joint Fluids Engineering Conference, San Francisco, CA.
31. Lohrberg, H., Stoffel, B., Fortes-Patella, R., Reboud, J.L.,(2002): Numerical and
Experimental Investigations on the Cavitation flow in Ccascade of Hydrofoils, Exp.
Fluids 33, 578–586.
87
32. S Song, C. C., He, J., (1998): Numerical simulation of cavitating flows by single-
phase flow approach, in: 3rd International Symposium on Cavitation, Grenoble,
France, pp. 295–300.
33. Anderson, J. (2005): Computational Fluid Dynamics: The Basics with Applications,
McGraw-Hill, Inc., New York.
34. Versteeg, H.K & Malalasekera, W. (1995): An Introduction to Computational Fluid
Dynamics, Longman Scientific & Technical, England
35. Launder, B.E., Spalding, D. B., (1974): The Numerical Computation of Turbulent
Flows, Compute. Method. Appl. M., 3, pp. 269-289.
36. White, F.M., (1974): Viscous Fluid Flow, McGraw-Hill, New York.
37. Shyy, W., (1997): Computational Modeling for Fluid Flow and Interfacial
Transport, Elsevier, Amsterdam, The Netherlands.
38. Fluent Inc., (2005): FLUENT, User's Guide, Lebanon, NH 03766.
39. Fluent Inc. (2005): GAMBIT 2.2.30: Training notes, Fluent 6.3.26.
40. Dular, M., Bacher, R., Stoffel, B., Sirok, B. (2005): Experimantal Evaluation of
Numerical Simulation of Cavitating Flow Around Hydrofoil, Europian Journal of
Machanics B/Fluids 24, 522-538.
41. Shyy, W., Thakur, S.S., Ouyang, H., Liu, J. and Blosch, E., (1997): Computational
Techniques for Complex Transport Phenomena, Cambridge University Press, New
York.
42. Patanker, S.V., Spalding, D.B., (1972): A Calculation Procedure for Heat Mass and
Momentum Transfer in Three Dimensional Parabolic Flows, Int. J. Heat Mass Tran,
15,pp1787-1806.
43. Patankar, S.V., (1980): Numerical Heat Transfer and Fluid Flow, Hemisphere,
Washington DC.
44. Moukalled, F., and Darwish, M., 2000, “A Unified Formulation of the Segregated
Class of Algorithms for Fluid Flow at All Speeds,” Numer. Heat Tr. B-Fund, 37,
pp.103-139
45. Bressloff, N.W., 2001, “A Parallel Pressure Implicit Splitting of Operators
Algorithm Applied to Flows at All Speeds,” Int. J. Numer. Meth. Fl., 36, pp. 497-
518.
46. Thakur, S.S., and Wright, J.F., 2002, “An Operator-Splitting Algorithm for
Unsteady Flows at All Speeds in Complex Geometries,” submitted for publication.
47. Pouffary, B., Fortes-Patela, R., Reboud, J.L., (2003): Numerical Simulation of
Cavitating Flow Around a 2D Hydrofoil: A Barotropic Approach, Fifth
International Symposium on Cavitation(Cav2003), Osaka, Japan, November 1-4.
48. Coutier-Delgosha, O., Jacques Andre Astolfi, (2003) : Numerical Reduction of
Cavitating Flow on a Two –Dimensional Symetrical Hydrofoil with a Single Fluid
Model, Fifth International Symposium on Cavitation(Cav2003), Osaka, Japan,
November 1-4.
49. Yoshinori SAITO, Ichiro NAKAMORI, Tosshiaki IKOHAGI (2003): Numerical
Analysis of Unsteady Vaporous Cavitating Flow Around a Hydrofoil, Fifth
International Symposium on Cavitation(Cav2003), Osaka, Japan, November 1-4.
88
50. Kawamura, T., Sakuda, M. (2003): Comparison of Bubble and Sheet Cavitation
Models for Simulation of Cavitation Flow Over a Hydrofoil, Fifth International
Symposium on Cavitation(Cav2003), Osaka, Japan, November 1-4.
51. Coutier-Delgosha, O., Reboud , JL. & Delannoy, 2003a: Numerical Simulation in
Unsteady Cavitation Flows. Int. J. for Numerical Methods in Fluids 42(5), 527-548.
52. Yuan W & Schnerr G.H. (2002): Optimization of Two -Phase Flow in injection
Nozzles Interaction of Cavitation and external jet formation. Proc Fluid Eng.
Summer Meeting.
53. Chein, K.Y., (1982): Prediction of Channal and Boundary-Layer Flows with Low
Reynolds numberTurbulence Model. AIAA., vol20, No.1
54. Chen Ying, LU Chuan-jing, WU Lei: Modeling and Computation of Unsteady
Turbulent Cavitation Flows
55. DELANNOY Y, KUNEY J.L. Two phase flow approach in unsteady cavitation
modeling[C]. Cavitation and Multiphase Flow Forum, ASME, New York ,1990,98:
153-158
56. Song C., HE J. Numerical simulation of cavitating flows by single-phase flow
approach[C]. 3rd Int. Symp. on Cavitation. Grenoble , France,1998,295-300.
57. Merkle C.L, Feng J. and Buelow P.E.O. Computational modeling of the dynamics of
sheet cavitation[C], 3rd Int. Symp. on Cavitation, Grenoble , France,1998,307-313.
58. Chen Y. et. al. Modeling hydrodynamic non equilibrium in cavitating flows[J]. J.
Fluids Engineering, 1996, 118(1):172-178.
59. Kunz R, Boger D., Chyczewski T.et.al. Multiphase CFD analysis of natural &
ventilated cavitation about submerged bodies[C]. 3 rd ASME/JSME Joint Fluid
Engineering Conference. San Francisco, CA, USA, 1999
60. Beyong R.S., Satoru Y, Xin Y., Application of Preconditioning method to gas liquid
two phase flow computations[J]. J. Fluids Engineering, 2004, 126(4): 605-612.
61. Song C.C.S., HE J. Numerical simulation of cavitating flows by single-phase flow
approach [C]. 3rd Int. Symp. on Cavitation. Grenoble , France,1998,295-300.
62. Arndt R.E.A, Song C.C.S., Kjeldsen M.et.al. Instability of partial cavitation: a
numerical / experimental approach[C]. 23rd Symp. on Naval Hydrodynamics. Rouen,
France, 2000.
63. Qiao Q. Numerical modeling of natural and ventilated cavitating flows[D]. Ph.D.
Thesis, Mineapolis, USA: University of Minnesota, 2004.
64. Spalding D. Numerical computations of multiphase flows[C]. VKI for Fluid
Dynamics. Brussels, Belgium, 1981.
65. Markatos N. Modeling of two phase transient flow and combustion of granular
propellants [J]. Int. J. Multiphase Flow. 1986. 12: 913-933.
66. WU Lei, LU Chuan-jing, XUE Lei-ping. An approach in modeling two-dimensional
partially cavitating flow [J]. Journal of Hydrodynamics, Ser. B, 2002, 14(1): 45-51.
67. WU Lei, LU Chuan-jing, LI Jie et al. Numerical simulations of 2-D Periodic
unsteady cavitating flows [J]. Journal of Hydrodynamics, Ser. B, 2006, 18(3): 341-
344.
89
68. FU Hui-ping, LU Chuan-jing, WU Lei. Research on characteristics of flow around
cavitating body of revolution [J]. Journal of Hydrodynamics, Ser. A, 2005, 20(1):
84-89. (in Chinese)
69. Stutz B, Reboud J. L. Measurements within unsteady cavitation [J]. Exp. Fluids,
2000, 29: 545-552
70. H. Ghassemi. Boundary Element method applied to the cavitating hydrofoil & marine
propeller. Scientia Iranica , Vol.10, No.2,pp: 142-152
90
View publication stats

Ahmmed - 2009 - Numerical Study of Cavitating and Noncavitating Flow Around 2D Hydrofoil

Uploaded by

Copyright:

Available Formats

Ahmmed - 2009 - Numerical Study of Cavitating and Noncavitating Flow Around 2D Hydrofoil

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ahmmed - 2009 - Numerical Study of Cavitating and Noncavitating Flow Around 2D Hydrofoil

Uploaded by

Copyright:

Available Formats

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

Numerical Study of Cavitating and Noncavitating Flow Around 2D Hydrofoil

Thesis · January 2009

Mohammad Shakil Ahmmed

CFD Modelling of Anaerobic Biogas Digestor View project

The user has requested enhancement of the downloaded file.

MOHAMMAD SHAKIL AHMMED

BANGLADESH UNIVERSITY OF ENGINEERING AND

MOHAMMAD SHAKIL AHMMED

Approved as to style & contents

Bangladesh University of Engineering & Technology

Our Respectable Parents

Writing at this paragraph has been deleted intentionally

Two-dimensional finite volume method (FVM) based on Reynolds averaged Navier-Stokes

LIST OF TABLES xiv

1.1 Basics of Cavitation 1

1.7.2 Void Fraction, α 16

1.7.5 Pressure Coefficient Cp, Drag and Lift Coefficient 17

3.1 Simulation Using Fluent 6.3.26. Based on Finite Volume Method 49

4. RESULTS AND DISCUSSION 57

U free stream velocity

4.1a Variation of lift coefficient at different angles (α) of attack 57

4.1b Drag coefficient (CD) at different angles of attack 57

4.2 Comparison of lift coefficient CL for non cavitating case. 58

4.3 Comparisons of the lift coefficient due to cavitation number 60

4.4 Lift & drag coefficient at various cavitation numbers at unsteady 61

Figure Caption Page No.

3.1 Schematic Diagram of the Flow Field around NACA0012 Hydrofoil 51

4.1 Comparison of lift Coefficient at various angle of attack 62

4.19 Pressure distribution over the hydrofoil at σ= 0.4 in unsteady 72

4.4.2 Experimental vortex cavity 80

4.4.7 Velocity & pressure distribution of vortex cavitation obtained by Chen 82

Period of lift coefficient of NACA0012 obtained by Chen Ying, LU 83

1.1 Basics of Cavitation

This study is particularly concerned with hydrodynamic cavitation, i.e. cavitation in

1.2. Types of Cavitation

 Shear cavitation. In region with high shear vorticity is produced. As a

1.3 Partial Cavitation

To predict inception of cavitation it is desirable to have indices to provide quantitative

In flowing liquids, the tendency of a flow to cavitate is indicated by the so-called

 pv is the vapor pressure

 U  is the reference velocity.

Figure1.2: Observed Cavitation patterns on a 2D NACA 16012 Hydrofoil as a

1.3.1 Fixed Cavities

Figure 1.3: Partial Cavitation

1.3.2 Cavity Detachment

The smooth detachment condition, also known as the Brillounin-Villat condition,

Figure 1.4: Supercavity behind a Two-dimensional NACA 16012 Hydrofoil

1.4.1 Cavity Pressure

Figure 1.5: Supercavity

1.4.2 Cavity Detachment

Cavity detachment Cavity detachment

Figure 1.7: Local effect of surface tension on Cavity detachment

1.4.3 Cavity Length

1.5 The Physical Phenomena

1.5.1 The Main Effects of the Cavitation in Hydraulics

1.5.2 The Main Forms of Vapor Cavities

1.6 Cavitation in Real Fluid Flows

1.6.1 Cavitation Regimes