Cavitation of A Water Jet in Water

Brigham Young University
BYU ScholarsArchive
Theses and Dissertations
2012-04-18
Cavitation of a Water Jet in Water

Michael Marshall Wright
Brigham Young University - Provo
Follow this and additional works at: https://scholarsarchive.byu.edu/etd
Part of the Mechanical Engineering Commons
BYU ScholarsArchive Citation

Wright, Michael Marshall, "Cavitation of a Water Jet in Water" (2012). Theses and Dissertations. 3175.
https://scholarsarchive.byu.edu/etd/3175
This Thesis is brought to you for free and open access by BYU ScholarsArchive. It has been accepted for inclusion
in Theses and Dissertations by an authorized administrator of BYU ScholarsArchive. For more information, please
contact scholarsarchive@byu.edu, ellen_amatangelo@byu.edu.
Michael M. Wright
A thesis submitted to the faculty of

in partial fulfillment of the requirements for the degree of
Master of Science
Tadd T. Truscott, Chair

Steven E. Gorrell
R. Daniel Maynes
Department of Mechanical Engineering

June 2012
Copyright © 2012 Michael M. Wright

All Rights Reserved
ABSTRACT
Michael M. Wright
Department of Mechanical Engineering, BYU
Master of Science
Cavitation is a phenomenon that occurs in liquids when the pressure drops below the vapor
pressure of the liquid. Previous research has verified that cavitation bubble collapse is a dynamic
and destructive process. An understanding of the behavior of cavitation is necessary to implement
this destructive mechanism from an axisymmetric jet for underwater material removal. This work
investigates the influence of jet pressure and nozzle diameter on the behavior of a cloud of cav-
itation bubbles generated by a submerged high-pressure water jet. First, this investigation is put
into context with a condensed historical background of cavitation research. Second, a description
of the cavitation-generating apparatus is given. Next, the experimental methods used to explore
the behavior of the cavitation clouds are explained. Finally, the results of the investigation, includ-
ing propagation distance, cloud width and area, pulsation frequency, and cloud front velocity are
presented. Among the results is a discussion of the significant experimental factors affecting the
behavior of the cavitation clouds. It is shown that the Reynolds number, specifically the diameter
of the nozzle, has a significant effect on the measurements. In some cases the jet pressure, and sub-
sequent jet velocity, had a less significant effect than was expected. Overall, this research describes
the cavitation cloud formed when a submerged high-speed water jet discharges.
Keywords: cavitation, submerged water jet, cavitation cloud, propagation distance, nozzle diame-
ter, pulsation frequency, Reynolds number, front velocity
ACKNOWLEDGMENTS
I must first acknowledge Dr. Tadd Truscott for accepting me as one of his first graduate
students. He has been incredibly supportive and helpful throughout my graduate studies. I am
appreciative of my graduate committee for their support and direction. Kevin Cole and Ken Forster
were instrumental in helping with data acquisition and part fabrication, respectively.
I am grateful for the Naval Research Enterprise Internship Program (NREIP) and Dr. David
Beal at the Naval Undersea Warfare Center (NUWC) for the opportunity to spend an incredible
summer working on this research in Rhode Island. I also thank Amanda Dropkin at the Naval
Undersea Warfare Center for her collaboration on this project.
Finally, I express special thanks to my loving wife, Emily, and my son, Elliott, for their
continual support.
TABLE OF CONTENTS
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
NOMENCLATURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Chapter 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.2 Mechanics of Cavitation . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.3 Cavitation Erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Theoretical Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Chapter 2 Experimental Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.1 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.1 Experimental Configuration . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.2 Water Jet Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.3 Experimental Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Instrumentation and Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4 High-speed Photography and Image Processing . . . . . . . . . . . . . . . . . . . 17
2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Chapter 3 Experimental Results and Discussion . . . . . . . . . . . . . . . . . . . . . 22

3.1 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.1.1 Duration of Cavitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1.2 Propagation Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.1.3 Cavitation Cloud Width . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.1.4 Cavitation Cloud Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.1.5 Pulsation Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.1.6 Cloud Front Velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.1.7 Cavitation Erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.2 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Chapter 4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
v
Appendix A Uncertainty and MATLAB® Scripts . . . . . . . . . . . . . . . . . . . . . 51
A.1 Uncertainty Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
A.2 MATLAB® Scripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
A.2.1 Propagation Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
A.2.2 Length Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
A.2.3 Cloud Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
A.2.4 Pulsation Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
A.2.5 Front Velocity Measurement . . . . . . . . . . . . . . . . . . . . . . . . . 65
A.2.6 Duration Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
vi
LIST OF TABLES
2.1 Experimental Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.2 Reynolds Numbers of Experimental Parameters . . . . . . . . . . . . . . . . . . . 14
2.3 Image Calibration Conversion Factors . . . . . . . . . . . . . . . . . . . . . . . . 18
vii
LIST OF FIGURES
2.1 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.2 Comparison between cavitating water jet and a jet of pure gas . . . . . . . . . . . . 11
2.3 Water jet valve configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.4 Nozzle geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.5 Data acquisition arrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.6 LabVIEW™ block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.7 Cavitation cloud area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.8 Single pixel column image intensity . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.9 Raw image intensity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.10 One set of frequency spectrum data . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.1 Image sequence: 0.45m, 1.0 mm nozzle, 20.7 MPa . . . . . . . . . . . . . . . . . 23

3.2 Pressure vs. Time for 1.0 mm nozzle at 20.7 MPa . . . . . . . . . . . . . . . . . . 25
3.5 Pressure vs. Time for 1.0 mm nozzle at 24.1 MPa . . . . . . . . . . . . . . . . . . 29
3.9 Cavitation from a 1.0 mm nozzle at four pressures: 17.2, 20.7, 24.1, 27.6 MPa . . . 33
3.10 Cavitation at 20.7 MPa from three nozzles: 0.7 mm, 1.0 mm, 2.0 mm . . . . . . . . 34
3.11 Time to discharge a column of water vs. Reynolds number . . . . . . . . . . . . . 35
3.12 Maximum propagation distance . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.13 Comparison of propagation distance vs. Reynolds number . . . . . . . . . . . . . 38
3.14 Non-dimensionalized propagation distance vs. Reynolds number . . . . . . . . . . 39
3.15 Cavitation cloud width vs. Reynolds number . . . . . . . . . . . . . . . . . . . . . 40
3.16 Cavitation cloud area vs. Reynolds number . . . . . . . . . . . . . . . . . . . . . 41
3.17 Cavitation cloud pulsation frequencies vs. Reynolds number . . . . . . . . . . . . 43
3.18 Front velocity of cloud pulses vs. Reynolds number . . . . . . . . . . . . . . . . . 44
viii
NOMENCLATURE
σ Cavitation number
Po Flow Pressure
Pv Vapor Pressure
ρ Density
Vj Jet velocity
Ainlet Pipe area
Aexit Nozzle area
P Pressure
z Height
g Gravity
f Friction factor
L Pipe length
D Pipe diameter
K Minor losses
℘ Power
Vf Front velocity
CD Coefficient of drag
Aequiv Area of equivalent cavitator
Aj Jet area
λ Slenderness ratio
Dc Cavity diameter
dj Jet diameter
Lcavity Cavity length
Re Reynolds number
ix
CHAPTER 1. INTRODUCTION
The principle of cavitation has been a topic of research for well over a century. Damage to
solid materials resulting from cavitation was first observed in the nineteenth century, but it was not
until the twentieth century that a proper investigation was able to definitively claim that cavitation
eroded solid materials in certain flow situations [1]. Research on cavitation continues today in a
few principal areas. These areas include, but are not limited to: the composition and behavior of
cavitation, methods to prevent or reduce the damage caused by cavitation, and techniques to take
advantage of the cavitation phenomenon for mining, cleaning, etc. The aim of the work presented
here is to quantify the observed behavior of cavitation from a water jet.
1.1 Objectives
This research examines the behavior of a cavitation cloud created when a high-pressure
axisymmetric water jet is submerged in quiescent water. In spite of abundant research on various
subtopics of cavitation, little research has been done in this particular area. Much of the available
research surrounding cavitation deals with the interaction of cavities and solid surfaces, specifi-
cally, bubble formation and collapse near a solid surface, cavitation erosion, and methods to either
incorporate or mitigate cavitation in real-world applications. Research shows that a surface can
be damaged if cavitation occurs on or near it. Measurements of the geometry and performance
of cavitation bubbles being transported downstream from an axisymmetric water jet are presented
here.
When a high-pressure column of water is discharged through a small-diameter straight
nozzle into quiescent water the accompanying drop in pressure from the accelerated column of
water can result in the formation of cavitation bubbles. These vapor bubbles conglomerate into a
cavitation cloud whose momentum carries the cloud away from the nozzle.
1
The objective of this work is to determine what happens to the cavitation cloud after it is
emitted from the nozzle of the jet. Various experimental parameters including nozzle diameter and
jet pressure are modified and the resulting effects on the cavitation cloud are recorded. High-speed
photography and pressure instrumentation were the means of capturing the dynamic behavior of
these events. From this data the relationship between nozzle diameter and jet pressure, specifi-
cally Reynolds number (Re), in regards to propagation distance, cloud width and area, pulsation
frequency, and cloud front velocity was discovered.
The remainder of Chapter 1 describes some background information regarding cavitation.
A few historical details are given on the discovery of cavitation. A brief description of the fluid
mechanics of cavitation follows. Next is an outline of some pertinent contemporary research on
cavitation. A section of theoretical calculations and expectations concludes the chapter. After the
background is laid out, Chapter 2 introduces the experimental methods. Included here are descrip-
tions of the design of the experimental apparatus, the instrumentation and data acquisition that was
required, as well as information on the high-speed photography setup and image processing. The
experimental results and discussion follow in Chapter 3. These results detail the behavior of the
cavitation cloud for the various experimental parameters. Finally, Chapter 4 is a summary of the
present and future work.
1.2 Background
Research on cavitation is plentiful. Prominent scientists from the past played major roles
in the discovery of cavitation and our modern understanding of this phenomenon. This section
introduces the first investigations into cavitation as well as descriptions of the fluid mechanics that
govern it. Relevant contemporary research is then presented to introduce the study found in this
thesis.
1.2.1 History
The term cavitation was first used to describe voids in liquid flows beginning in the late
nineteenth century. The renowned Osbourne Reynolds [2], famous for the non-dimensional Reynolds
number, performed “experiments showing the boiling of water in an open tube at ordinary temper-
2
atures” in 1894. In his experiments he forced water through a converging–diverging glass tube.
In the throat of which, he noticed the formation of a small white cloud, and heard sounds similar
to those heard when water in a kettle boils. The cloud was composed of vaporous voids resulting
from the drop in pressure as the water accelerated through the throat of the nozzle.
Around the turn of the twentieth century hydromechanical systems—pumps, valves, pro-
pellers, etc.—were becoming more powerful and began showing curious signs of damage. The
intensity of the damage was particularly unnerving on new, fast transatlantic steamers [3]. The ro-
tational velocity of their propellers was greatly increased as these ships were designed with more
powerful engines. The result being that the propellers were experiencing erosion problems almost
immediately upon setting sail. In 1915, the Board of Invention and Research in London, formed a
commission to determine the cause of this propeller damage [1].
This commission investigated five possible causes of the damage: metal surface structure,
working stress, water impingement, cavitation, and water hammer. The first three possible causes
were shown to not produce any pitting damage. The other two causes, cavitation and water ham-
mer, showed the propensity to erode the propellers. The commission came to the conclusion that
“cavitation will only produce erosion when accompanied by conditions which cause
the cavities to collapse in such a way and in such a position that the energy of collapse
is concentrated on a small portion of the propeller surface” [1].
It would be desirable to be able to predict and control the conditions that lead to erosion.
The research herein presented tries to address this. One of the members of the commission, who
was not a member of the Royal Society, S. S. Cook [4], was allowed to publish his contributions
to the commission’s study in 1928. He concluded that water droplets eroded the turbine blades
by the water hammer effect; which can be described as high-pressure waves in a flow as a result
of the flow being abruptly stopped by a solid or other obstruction. He also postulated that the
velocity of the surface of a vapor cavity accelerates as the cavity collapses, which is related to
the internal pressure of the cavity [4]. Rayleigh improved on Cook’s theory by accounting for the
compressibility of the vapor during the collapse of the cavity. These two results provide the basis
for calculating the pressure developed by a collapsing cavitation bubble, and have been used by
3
many subsequent researchers to help explain how a simple vaporous cavity can transmit enough
energy to erode strong materials like steel and rock.
1.2.2 Mechanics of Cavitation
Cavitation is a phenomenon that occurs in liquids, and depends on the non-dimensional

Cavitation number σ reaching a flow-dependent critical value [5]. The critical value is reached
when a liquid flow significantly increases velocity, or the pressure of the flowing fluid drops to the
vapor pressure as described by:
Po − Pv
σ= 1 2
; (1.1)
2 ρV j
where Po defines the pressure of the flow, Pv is the vapor pressure, ρ is the fluid density, and V j
is the velocity of the flow [6]. The propensity for a flow to cavitate increases as σ nears zero.
The type of flow, i.e. internal pipe flow, axisymmetric external flow, flow over a propeller, etc.,
determines how low σ must be before cavitation begins, there is no single value for σ below
which cavitation is guaranteed and above which it is not possible. When σ nears the flow-specific
critical value vapor cavities begin to emerge from nucleation sites, expand, and can coalesce into
a cavitation cloud. As the velocity of the flow decreases, the pressure rises back above the vapor
pressure and the cavities begin to collapse. From experiments performed by Wu and Chahine [7],
the interior composition of coalesced cavities is primarily water vapor. The compressibility of the
internal vapor results in high internal gas pressure that is inversely proportional to the radii of the
cavities [4, 8]. Research presented by Ooi [9] demonstrates that the Cavitation number becomes
less dependent on the Reynolds number of the flow as the flow velocity is increased. For his
experiments this was confirmed in flows greater than 20 m/s. The water jets used in the current
investigation are much faster than 20 m/s. Gopalan et al. [10] discuss the existence of vorticity and
strain in the shear layer of an axisymmetric jet, which contributes to the inception of cavitation by
lowering the local pressure. They, as well as O’Hern [11], show that cavitation inception occurs
in the cores of streamwise vortices. Soyama et al. [12] show that a submerged water jet behaves
significantly different than a water jet in air. The jet’s shear layer exhibits considerable cavitation
potential resulting in high erosion rates. For the current study an axisymmetric jet was used to
4
generate cavitation. The cavities appeared to form in the shear layer initially, until the rest of the
jet would cavitate.
Once formed, cavitation bubbles are unstable and can break into smaller parts then reform
back to their original shape [3]. The life of a gas bubble consists of many collapse and rebound
episodes [13]. This result will be shown in subsequent sections. The frequency of an oscillating
bubble is not affected by the shape of the bubble; however, the oscillation frequency can be re-
duced if the bubble is near a solid boundary [14]. Hutli and Nedeljkovic [15] have shown that
the cavitation emitted from a submerged jet is periodic in nature. An axisymmetric cavitating jet
pulses from the nozzle because of a reentrant jet at the tail end of the cavitation cloud [16]. Sato
et al. [17] also show from high-speed photography that a chain-reaction of bubble collapses results
from local collapses in the cavitation cloud that send a pressure wave through the flow. According
to Hansson and Mørch [18], the cloud collapse begins at the periphery and moves inward. This
work quantifies some of these behaviors.
1.2.3 Cavitation Erosion
Cavitation bubble collapse is a dynamic event [5] for which extensive research is avail-
able regarding the potential for cavitation bubbles to erode surfaces. It should be noted that this
section describes the effects of single bubble interactions more than the interactions of clouds of
bubbles. Cavitation erosion, or pitting, is the removal of surface material as a result of the remark-
ably destructive interaction of shockwaves and micro-jets. According to Reisman et al., “Bubbly
shockwaves play a critical role in generating pressure pulses, which lead to cavitation noise and
damage” [19]. The dynamic nature of the cavity collapse generates shockwaves, and if the col-
lapse occurs near a solid boundary an extremely high-pressure micro-jet can impinge on the solid
surface [20]. Considerable research shows that these shockwaves and micro-jets are strong enough
to cause yielding and failure of solid materials [4, 20–24].
Coleman and Scott [21] present research showing submerged cavitating jets that exhibited
varying erosion results for five metals, mostly aluminum alloys typically used in environments
where cavitation can occur. They claim impinging jets deform the metal surfaces, which leads to
ductile flow, tearing, and finally removal of the torn metal fragments. Bourne and Field [22] add
that collapsing bubbles generate high-speed jets that impinge on the contact surface. The mecha-
5
nism of these impinging jets was further explained by Jayaprakash et al. [24] and Brujan et al. [23].
For a single bubble at a standoff distance between 0.9 and 1.1 (the ratio of distance from wall to
maximum bubble radius) the spherical bubble becomes toroidal as a jet penetrates from the far side
to the bubble-wall boundary. This torus subsequently breaks up into several spherical bubbles that
also impinge on the boundary. These smaller bubbles have increasingly higher pressures [4, 8] that
result in even faster jet impingement. Annular erosion patterns are common in single cavitation
bubble applications and this torus explanation clarifies why this pattern occurs [20, 21, 23]. On a
macro-scale, annular patterns are also common, but this is from the fact that most of the cavitation
forms in the shear layer of the jet [10–12].
Philipp and Lauterborn [20] show that three effects of single bubble collapse lead to cav-
itation erosion. First, a shock wave generated upon bubble collapse increases the local pressure.
Second, the impinging liquid jet exerts high pressures on the surface. Finally, during collapse the
bubble is drawn closer to the surface increasing the potential damage from either the first or sec-
ond effect. One practical application of these cavitation principles was shown by Lu et al. [25],
where laser-drilled holes in metal plates underwater were cleaner than laser-drilled holes in plates
in air. Micro-jets from the cavitation bubble collapse that were formed by a laser heating the water
impacted their metal specimens with much higher forces than laser ablation alone.
1.3 Theoretical Approach
Before any experiments were performed it was necessary to make a few theoretical predic-
tions of what would occur. One application using high-speed jets in a submerged environment that
has been rigorously studied is the use of a jet to form a supercavity around a submerged torpedo.
Unclassified defense documents outline some of this work, which provided a starting place for
predictions of the behavior of the cavitation in this study.
Without being able to measure the nozzle exit velocity a theoretical prediction based on
standard fluid flow in a pipe was made [6]. A predicted jet velocity is necessary to predict σ ,
which is used in predicting the cavity geometry. The theoretical jet velocity is:
6
v
Ainlet u 2 (Pinlet − Pexit ) + ρg (zinlet − zexit )
Vj = ; (1.2)
u
Aexit t
u 2 2
Ainlet fL Ainlet
ρ Aexit + D + Aexit ΣK − 1
where Ainlet is the area of the pipe, Aexit is the area of the nozzle exit, Pinlet is the reservoir pressure,
Pexit is the quiescent water pressure, g is gravity, zinlet and zexit are the relative heights of the
reservoir and nozzle, f is the friction factor for stainless steel pipe, L is the length of the pipe, D
is the pipe diameter, and ΣK is the sum of all minor losses, such as bends and constrictions, in the
system. This is the velocity used in Equation 1.1 to predict whether or not the flow will cavitate.
The formation of cavitation from a high-speed jet in quiescent water can be mathematically
modeled as cavitation formed by a disk cavitator in a moving flow [26]. A comparison between the
power of the cavity front (℘out ) to the power of the jet (℘in ) is made to determine the front velocity
of the cavity generated by the high-speed jet. These are given by:
1
℘out = ρV f2 AequivCDV f ; (1.3)
2
1
℘in = ρ(V j −V f )2 A j (V j −V f ); (1.4)
2
where ρ is the density of the water, V f is the front velocity, V j is the jet velocity, Aequiv is the
equivalent area of the cavitator, and A j is the area of the jet. CD = 0.82, which is the coefficient of
drag for a disk cavitator as given by Savchenko [27]. Equating Equations 1.3 and 1.4 and solving
for the front velocity gives:
Vj
Vf = . (1.5)
AequivCD 1/3

1+ Aj
The geometry of the cavity modeled as if generated by a disk cavitator can be described as
a function of σ given by May [26]:
L
λ= = 1.067σ −0.658 − 0.52σ 0.465 (1.6)
d
7
where λ is the slenderness ratio, L is the length of the cavity, d is the diameter of the widest part of
the cavity. The cavity diameter can be determined from the jet diameter, d j , according to May [26]
by:
r
dj CD
Dc = . (1.7)
0.205 σ
The length of the cavity is found from λ and Dc , as:
Lcavity = λ Dc . (1.8)
Using the experimental parameters discussed later in Chapter 2, theoretical values were
calculated for Equations 1.1–1.8. The jet velocity was predicted to range from 100 m/s to about
130 m/s. From these velocities the range of σ was from 0.0201 at low jet velocity down to 0.0124
at the highest jet velocity. The predicted cloud front velocities were only between 27 m/s and 35
m/s, and theoretical propagation lengths extend from 0.30 m as far as 1.5 m downstream. These
theoretical calculations will be compared to experimental values in Chapter 3.
1.4 Conclusion
This chapter introduced cavitation by giving a short history of its discovery, describing
the fluid mechanics that govern the formation of cavitation, and presenting previous studies of
cavitation behavior. The erosive capabilities of cavitation have clearly been established. The final
section outlined a theoretical approach to determining cavitation behavior as a function of the size
and velocity of the initial water jet. The research to be presented expands on the behavior of
cavitation clouds already described. Some interesting relationships between jet size and pressure
will be described, and comparisons between predicted and measured cavitation behavior will be
given.
8
CHAPTER 2. EXPERIMENTAL METHODS
This chapter describes the experimental setup and parameters used in the study of cavitation
cloud behavior. The design and operation of the water jet apparatus is explained first. Next is a
description of the instrumentation and data acquisition used in the experiments. Lastly, a section
is devoted to the high-speed photography and image processing techniques used in this study.
2.1 Experimental Setup
This section illustrates the design of the water jet. The jet device and the full test setup are
described in detail. The peculiarities of the design are clarified and justified. Security measures
to prevent injury were in place during all testing. The extremely high-pressure gas used in the
experiment required all components to be rated for such pressures. The entire experiment was
done in a lockable room, and warning signs were used to prevent accidents due to tampering.
Proper safety equipment, including eye protection, was required of all persons in the room during
the experiments.
2.1.1 Experimental Configuration
In the literature on cavitation there have been various methods of cavitation generation. The
method adapted to this experiment used high-pressure gas to accelerate a column of water through
a nozzle with sufficient velocity that the resulting jet cavitated. This method created a single shot
or discharge of the jet that lasted only as long as it took to discharge the fixed volume of water;
it was not a continuous stream. A layout of the experimental setup is shown in Figure 2.1. The
main components of the setup were the jet, a rigid frame onto which the jet was mounted, and
a tank of quiescent water. In addition to these components there was electronic instrumentation,
including power supplies, a high-speed camera, pressure transducers, and a computer to which the
electronics were connected.
9
Gas reservoir
Release valve
Water inlet
Static pressure
Dynamic pressure transducer
transducer
gth
Bleed valve e len Gas inlet
Pip
Frame
Nozzle
Water tank
Figure 2.1: Experimental test setup showing the jet apparatus mounted on a frame to be discharged
into a water tank. The water inlet hose, compressed gas inlet hose, and pressure transducers are
also visible.
The water tank was made of glass and measured 1.83m×0.61m×0.61m. A rigid framework
of 2.54 cm square extruded aluminum surrounded the tank to which the jet was affixed. The
framework was fixed to the floor to prevent movement when the jet discharged. The nature of
the jet and tank configuration required that the jet be mounted at an angle below horizontal. Full
adjustability of the framework allowed the jet to be aligned at various angles. The decline also
ensured that the water column was fully discharged through the nozzle without the pressurized
gas simply passing over the top of the water in the pipe. Figure 2.2 shows the difference between
the cavitating jet and pure gas exiting the nozzle at the end of the jet discharge. It is unlikely
that this particular jet could be feasibly used in a horizontal position due to the possibility for the
pressurized gas to escape without discharging the water. Additionally, to eliminate the possibility
of air being entrained in the jet from the surface it was requisite that the nozzle be at a sufficient
depth in the tank of water. This depth was approximately 20 cm, and throughout the experiments
no air entrainment was observed.
10
(a) Cavitating water jet: 1.0 mm nozzle, 20.7 MPa
(b) Gas exiting nozzle after jet discharge
Figure 2.2: Comparison between cavitating water jet and gas following the jet discharge. (a)
Cavitation bubbles move away from the nozzle without expanding much. (b) The pressurized gas
expands rapidly upon exit after discharging all of the water.
2.1.2 Water Jet Design
The jet apparatus consisted of a series of valves and pipes. These valves separated a reser-
voir of ultrahigh-pressure nitrogen (nitrogen was chosen for its inert properties and availability at
ultrahigh pressure) from a length of pipe filled with tap water at the end of which was a small-
diameter orifice nozzle. Figure 2.3 shows the valve arrangement. A tank of ultrahigh-pressure
nitrogen at 40 MPa was regulated into a smaller storage reservoir at the desired test pressure. The
reservoir was a 1.15-liter carbon-fiber-wrapped paintball cylinder rated for 31 MPa. Test pressures
11
C
B
D
A
E
F
Figure 2.3: Water jet valve configuration. High-pressure nitrogen (A) fills reservoir (C) through
valve (B) where a transducer (D) measures the static pressure. Check valve (F) separates the gas
from water filled through valve (H) by hose (G). Opening valve (E) releases the gas to discharge
the water.
never exceeded 28 MPa. Tap water filled a length of 6.35 mm diameter pipe through a separate
port. A high-pressure ball valve and check valve separated the pressurized nitrogen from the pipe
filled with water. The check valve was instrumental in preventing water from contaminating the
gas reservoir. One of three standard zero-degree-spray-angle pressure washer nozzles was attached
to the end of the length of pipe to create the water jet. Figure 2.4 illustrates the geometry of the
nozzles. The length of the throat increased with increased diameter from 3.0 mm in the 0.7 mm di-
ameter nozzle up to 3.4 mm and 4.5 mm in the 1.0 mm and 2.0 mm diameter nozzles, respectively.
All of the components that would be in contact with water were made of stainless steel to
prevent corrosion. The mating connector between the small reservoir and the jet had to be custom
fabricated out of high-strength carbon steel because the thread size of the small reservoir was not
standard. The entire system was designed to withstand pressures up to 31 MPa, slightly greater
than the maximum pressure to be tested. The weakest component had a factor of safety of 1.13.
12
6.0 6.5 7.0
D* 39°
L* 1.5* 2.5 2.4 7.5 8.0

*All dimensions in millimeters
D = 0.7, 1.0, 2.0 mm
L = 3.0, 3.4, 4.5 mm
Figure 2.4: Geometry of the zero-degree-spray-angle nozzles used in the experiments. The length
of the throat of the nozzle increased with increased diameter. For 0.7 mm diameter, length was
3.0mm. For 1.0 mm diameter, length was 3.4 mm. For 2.0 mm diameter, length was 4.5 mm.
2.1.3 Experimental Variables
Experiments were performed for a series of nozzle diameters, reservoir pressures, and water
column lengths as a way to compare the effect each of these parameter had on the behavior of the
cavitation. Table 2.1 outlines these experimental parameters. Various constraints on the pressurized
gas and material sizes led to the particular parameters shown in Table 2.1. Nitrogen was chosen
as the pressurized gas because of its inert properties. It is commercially available at a maximum
of 41.4 MPa, which constrained the experimental pressures to be less than this maximum. The
maximum test pressure of 27.6 MPa was chosen to provide a factor of safety greater than 1.0 for
the entire device as previously described. The three nozzle sizes were the three smallest diameter
nozzles that were easily compatible with the apparatus design and not cost-prohibitive. Preliminary
calculations, as described in Chapter 1, predicted σ = 0.0124–0.020, which provided support that
the jet using the chosen parameters would lead to cavitation. The Reynolds numbers of the jet,
Re = ρV D/µ, with the given parameters are outlined in Table 2.2.
The primary data set of 36 tests (three tests for each nozzle at each of four pressures) was
collected for the 0.45 m pipe. Initial experiments exhibited pulsation, which was believed to be
13
dependent on the pipe length. It was assumed that the longer pipes would affect the pulsation of the
cloud; the results of which are explained in the next chapter. The other two pipe lengths provided
comparisons to the primary data.
Table 2.1: Experimental Parameters
Pipe Length Nozzle Diameter Reservoir Pressure

(m) (mm) (MPa)
0.45 0.7, 1.0, 2.0 17.2, 20.7, 24.1, 27.6
0.61 1.0, 2.0 17.2, 20.7, 24.1, 27.6
0.91 1.0, 2.0 17.2, 20.7, 24.1, 27.6
Table 2.2: Reynolds Numbers of Experimental Parameters
Reynolds number
Nozzle Diameter Pipe Length
17.2 MPa 20.7 MPa 24.1 MPa 27.6 MPa
0.7 mm 0.45 m 63,000 70,000 75,000 80,000
0.45 m
1.0 mm 0.61 m 90,000 99,000 107,000 114,000
0.91 m
0.45 m
2.0 mm 0.61 m 181,000 198,000 214,000 229,000
0.91 m
2.2 Operation
This section describes the operating procedure of the jet to generate cavitation (refer to
Figure 2.3). The procedure for discharging the water jet is as follows:
1. With all valves closed, open valve (B) separating the nitrogen tank from the reservoir (C)
2. Release nitrogen through the regulator to fill the reservoir, then close the separating valve
3. Fill the pipe with water through valve (H), and bleed off excess gas through the bleed valve
near the nozzle (not shown in the figure)
14
4. With the data acquisition programs running and the camera ready, trigger the camera
5. Immediately after triggering the camera open the main release valve (E)
6. Close the release valve as soon as the jet fully discharges
This procedure was performed for each nozzle diameter at each of the four experimental
pressures. The nozzle could be easily changed by unscrewing it and screwing in the next size
without moving the apparatus, which was firmly fixed to the aluminum frame. After the initial 36
experiments the length of pipe for the water column was replaced by a longer pipe. In order to
exchange this pipe the entire jet was removed from the frame and the pipe replaced. The jet angle
changed slightly for the longer pipes, but this did not affect the experiments.
2.3 Instrumentation and Data Acquisition
Pressure instrumentation inside the jet allowed for precise pressure measurements to be
made throughout each run. An Omega PX309-7.5KG5V static pressure transducer was connected
to the pressurized reservoir to determine the test pressure. A pressure snubber, a connector con-
sisting of a network of tiny passageways, separated the reservoir from the transducer to prevent
pressure spikes that could damage the sensitive transducer diaphragm.
Knowing the pressure at the nozzle exit was necessary to compare experimental data with
theoretical values. An Omega DPX101-5K dynamic pressure transducer with a resonant frequency
of 500 kHz was used to measure this pressure. The transducer was connected just upstream of the
nozzle by an extension of pipe. The length of pipe extended the transducer out of the tank of water
since it was not waterproof (refer to Figure 2.1).
The pressure transducers each output a voltage signal between 0–5 volts. They were cal-
ibrated to produce 97 mV/MPa and 168 mV/MPa for the static and dynamic transducers, respec-
tively. The uncertainty of the static transducer measurements was approximately ±1.1 MPa, and
the uncertainty for the dynamic transducer was approximately ±0.5 MPa [28]. A schematic show-
ing how the transducers and camera are connected to the computer is shown in Figure 2.5. The
static and dynamic pressure transducers were powered separately by power supplies. The camera
output two signals, one for the trigger and one for the shutter. The trigger signal was a positive
15
DAQ PC
PS1
PS2 Trigger
Camera
Dynamic
Static
Figure 2.5: Data acquisition from transducers and camera to the computer. Both the static and
dynamic transducer were powered by individual power supplies. Their output signal was routed
into a data acquisition bus with two signals from the high-speed camera. The combined signals
were read into a computer for further analysis.
Figure 2.6: LabVIEW™ program block diagram showing how the four signals are read and stored
in the computer for further analysis.
16
five volts beginning at the moment the trigger was depressed, and continued for the duration of
the filming. The second camera signal was a pulse of five volts each time an image frame was
recorded. This allowed for images to be correlated exactly to the pressure data. The four voltage
signals, two from the transducers and two from the camera, were converted to digital signals to be
analyzed in the computer.
The analog voltage signals from the pressure transducers and camera were converted to
digital values by a NI USB-6211 DAQ and read into a LabVIEW™ program. See Figure 2.6. Data
acquisition was set at 21 kHz in this program. This sampling rate reduced the likelihood of aliasing
the five volt pulsed signal.
2.4 High-speed Photography and Image Processing
High-speed photography was the primary method of analyzing the cavitation behavior. For
this investigation a single high-speed Photron SA-3 camera, whose frame rate was fixed at 5000
fps with shutter speed of 1/10000 s, provided clear flow visualization. There was still slight motion
blur with these camera settings near the nozzle exit where the velocity was greatest. At this frame
rate the recorded images had a maximum resolution of 768×128 pixels. To ensure that the entire
cavitation cloud would be captured in each image frame the camera was rotated to the same angle
as the jet. All images presented in figures here are horizontal, but it should be kept in mind that the
jet was actually angled downward.
MATLAB® was used to process the high-speed images from which the overall propagation
distance, cloud width and area, cloud frontal velocity, and pulsation frequency were computed. All
of the scripts used are included in Appendix A.2. To be able to measure the geometry of the
cavitation in the images, image calibration was necessary to convert measurements from pixels to
centimeters. With the camera in position an image of a ruler with a centimeter scale in the water
tank near the nozzle provided this relationship. Many pairs of points in the calibration image were
chosen with the two points in each pair separated by one centimeter on the imaged ruler. The
average number of pixels between each pair of points became the pixel-to-centimeter conversion
factor. Table 2.3 summarizes the conversions used for the various configurations. The camera
position had to be changed throughout the experiments as the cavitation cloud propagated out of
17
Table 2.3: Image Calibration Conversion Factors
Configuration Conversion (pixels/cm)

0.45m (17.2-20.7 MPa) all nozzles 25.02
0.45m (24.1-27.6 MPa) 2.0 mm nozzle 20.7
0.61m all pressures and nozzles 19.5
0.91m all pressures and nozzles 19.25
the image frame. Each new camera position required a new pixel-to-centimeter conversion factor
that was determined by the method described.
The propagation distance was measured from the nozzle exit to the location where the
cavitation cloud collapsed. The pulses of each run collapsed at a parameter-specific common
location. This process converted the distance between two manually chosen points in an image
from pixels to centimeters. The chosen points were the stationary nozzle and the point of general
collapse. The same procedure was used for the cloud width, except the points were measured at
the upper and lower cloud edges.
To find the area of the cavities a more rigorous image processing approach was necessary.
In order to find the area the images were converted into binary images. To accurately convert the
many shades of gray into either black or white each image went through multiple processes. The
nozzle was first eliminated from each image. Next, the background of each image was turned
white using a disk structuring element five pixels in size. At this point the image was converted to
a black jet on a white background. After inverting the black and white to get a white jet on a black
background, the bwarea function was used to measure the area of each cavitation cloud. Figure 2.7
shows a raw image and the processed image from which the area was calculated. The white area
bordered in red is the area that was measured.
To compute the cloud front velocity many sets of consecutive images were used. Three sets
of multiple consecutive images were chosen from each run and two points were manually chosen in
each image—one at the nozzle and the other at the cloud front just before the cloud collapsed. The
change in front position between consecutive images gave the distance traveled between frames.
The velocity was then calculated by dividing this distance by the known frame-separation time
(0.0002 s).
18
Figure 2.7: Cavitation cloud area measured from an image. Only the darkest parts of the image
corresponding to definite cavities are measured. The red outline shows what area was measured.
Figure 2.8: Representative image of how the intensity of a single column of pixels (white line) at
a fixed distance downstream from the nozzle is used to measure the frequency of pulsation.
The frequency at which each jet pulsed was determined using each image of the individual
runs. The average color intensity of one column of pixels was measured at a fixed distance from
the nozzle (see Figure 2.8). This intensity was measured at the same position for each image
over the entire discharge of the jet. The intensity of the pixel column fluctuated according to
the cloud pulsation (see Figure 2.9). A passing pulse was wider and darker which resulted in an
average intensity closer to the value for total black. When the intensity values over the life of the
cavitation were plotted the result was what appeared to be a noisy signal. Each discharge resulted
in approximately 700 data points for the 0.7 mm diameter jet down to only approximately 250 data
points for the 2.0 mm diameter jet. A Fast Fourier Transform (FFT) was used to find the component
frequencies of the noisy signal. To give better results the data for each run was concatenated with
itself three times (see Figure 2.9) before being run through the FFT. MATLAB’s fft command
was used to find the frequency spectrum of the intensity variation at this vertical segment of each
discharge.
The most prominent frequencies of each run were plotted with those of the other runs, see
Chapter 3. Figure 2.10 shows the resulting frequency spectrum of the data from Figure 2.9. There
is an obvious peak between 500–1000 Hz. From this example, the frequencies of the three highest
19
140
135
130
125
Intensity
120
115
110
105
100
95
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Time [s]
Figure 2.9: Concatenated intensity values for each image over the life of a jet at 20.7 MPa from
the 0.45 m pipe and 1.0 mm diameter nozzle (Re = 99,000).
2000
1500
Magnitude
1000
500
0
0 500 1000 1500 2000 2500
Frequency [Hz]
Figure 2.10: Example of the frequency spectrum of the measured intensity for a jet at 20.7 MPa
from the 0.45 m pipe and 1.0 mm diameter nozzle (Re = 99,000).
20
peaks would be recorded. After doing this process for every experimental run the many data are
plotted to visualize any frequency trends. In the next chapter, all of the cavitation measurements
are presented with their associated uncertainties as calculated in Appendices A.1 and A.2.
2.5 Conclusion
This chapter described the experimental methods used to investigate the behavior of cavi-
tation clouds from an axisymmetric water jet. The experimental configuration and water jet design
were explained, as well as the operating procedure. The experimental variables—nozzle diame-
ter, jet pressure, and pipe length—were given, including the reasoning of why they were chosen.
Explanations of the pressure instrumentation, data acquisition, and high-speed photography were
given to clarify how the experimental results were obtained. The remainder of this thesis presents
the findings of the investigation.
21
CHAPTER 3. EXPERIMENTAL RESULTS AND DISCUSSION
This chapter presents a discussion of cavitation cloud behavior. The results are divided into
sections—cavitation duration, propagation distance, cloud width, cloud area, pulsation frequency,
and cloud front velocity. The principal case against which other cases are compared was for the
0.45 m pipe, 1.0 mm nozzle, at 20.7 MPa. Though all of the tested cases are included in the plotted
data, only the most significant are illustrated in figures.
3.1 Experimental Results
A representative display of the cavitating jet is shown in Figure 3.1 for cavitation from the
0.45 m pipe length, 1.0 mm nozzle, at 20.7 MPa. The images reveal that the event was a cloud
of small bubbles rather than a single supercavity. If the cavitation had been a single cavity the
imaging technique of backlighting would have shown the translucence of the cavity. The apparent
turbulent nature of the cavitation cloud and its opacity tend to suggest that the cavitation is a cloud
of small bubbles. There is the possibility that a larger cavity exists inside the bubble cloud, but
that is impossible to determine from these images. Savchenko [27] reported that in tests where
a continuous water jet was used to generate a single supercavity supplementary, or artificial, gas
injection was necessary for velocities less than 100 m/s. This means that at lower velocities no
single cavity was formed without adding gas to the cavitation vapor pockets. The current tests all
had predicted jet velocities greater than 100 m/s, but the streams were not steady. This unsteadiness
could explain why no supercavities were formed. For all of the tests the cavitation cloud formed at
the nozzle, propagated downstream, and collapsed.
The first column of Figure 3.1 shows the initial formation of a cavitation cloud. The first,
middle, and last frame of each column shows the time from when the first cavitation bubbles were
visible. The time step between each frame is ∆t = 0.0004 s. The other two columns of the figure are
sequences of the same jet from later times. Throughout the discharge every run exhibited pulsation.
22
−3
x 10
0.0032 s 0.0238 s 0.1152 s
0.024
4 0.116
0.025
5 0.117
0.026
6 0.118
0.027 Pulse Begins
7 0.119
0.0072 s 0.0278 s 0.1192 s
0.028
[s]
[s]
[s]
8 0.12
0.029
9 0.121
Pulse Ends
0.03
10 0.122
0.031
0.0112 s [MPa]
0.0318 s 0.1232 s
11
[MPa] 0.123
[MPa]
0 5
19.5 20 19.219.319.4
Figure 3.1: Image sequence with a ∆t=0.0004 s of three time periods of a cavitating water jet from
a 0.45 m long pipe with a 1.0 mm nozzle at 20.7 MPa (Re = 99,000). Time stamps show the time
from cavitation inception. The cavitation is formed at the nozzle, moves downstream as a cloud,
and collapses. The nozzle exit pressure as a function of time plotted to the right of each sequence
shows jet pulsation (Frame rate: 5000; shutter: 1/10000 s; resolution: 768×128).
23
The sequence shows how a pulse of cavitation bubbles traveled from the nozzle downstream and
collapsed at a certain distance. See Figure 3.1, column 2, frames 1-9. In column 2, frame 1, near
the nozzle a pulse begins. It is seen to grow longitudinally as well as axially. The pulse continues to
grow until the cloud separates from the nozzle and a new pulse begins. The fifth frame of column
2 shows another pulse beginning before the previous pulse fully collapses. In frame 9 the pulse
that began in frame 1 finally collapses.
Along the vertical axis of each column of images, a pressure plot from the nozzle exit
shows that the pressure inside the pipe was unsteady. This unsteady effect caused the jet pulsation
from the nozzle. The collapse and rebound of the bubbles in the free stream was the result of a
different phenomenon related to gas bubbles in liquid (see [13, 14]). The mass collapse occurred
as described in Chapter 1, where individual bubbles collapse on the periphery of the cloud sending
pressure waves through the cloud. These waves cause more individual bubbles to collapse until
the entire cloud collapses. See [17, 18] for further reading on this. It is sufficient for this research
that the cloud of bubbles collapsed at a common location based on the experimental parameters.
Pulses from the image sequences can be correlated with the pressure plots. For example, in the
third column of Figure 3.1 a new pulse is seen in the eleventh frame from the top (t = 0.1192 s)
and a pressure increase begins just before this time (about 0.1186 s), as marked in the pressure
plot. Since the new pulse is already far away from the nozzle in frame 11, it can be assumed
that this pulse began around 0.1186 s because of the pressure surge, before it is seen at 0.1192
s. Additionally, it appears that a decrease in pressure registered on the plots corresponded to the
cavitation cloud moving away from the nozzle just before a new pulse begins. Before the pulse
that began around 0.1186 s moves downstream and collapses (see arrows in Figure 3.1), there are
two additional pulses. The first new pulse occurs around 0.12 s, and is visible in frame 13. The
second new pulse begins around 0.121 s and is visible in frame 16 just as the original pulse that
began in frame 11 finally collapses.
Figure 3.2 plots the pressure of the full cavitation sequence for the 1.0 mm diameter nozzle
at 20.7 MPa (Re = 99,000) from Figure 3.1. The dashed lines outline which portions of the full test
are shown in the image sequences. The leftmost pair of lines is for the first column of Figure 3.1,
and it can be seen that the pressure is ramping up. The center pair of lines is for the center column
just as the pressure reaches a plateau. The rightmost pair of lines separates the third column near
24
20
15
Pressure [MPa]
10
0
0 0.05 0.1 0.15
Time [s]
Figure 3.2: Pressure–time plot for Figure 3.1. The leftmost pair of lines delineates the pressure
corresponding to the first column of Figure 3.1. The center pair of lines is for the center column of
Figure 3.1. The rightmost pair of lines if for the third column of Figure 3.1 near the end of the life
of the jet.
the end of the life of the jet. The pressure oscillated during the discharge of each jet as seen for
one run in Figure 3.2.
Figures 3.3, 3.4, 3.6, 3.7, and 3.8 also show the formation and general tendencies of cavi-
tation from the 0.45 m pipe with varied pressure and nozzle diameter. Figure 3.3 shows cavitation
from the 1.0 mm nozzle at 17.2 MPa (Re = 90,000). Figure 3.4 shows cavitation from the same
nozzle at 24.1 MPa (Re = 107,000). Figure 3.6 shows cavitation again from the same nozzle at
27.6 MPa (Re = 114,000). Figure 3.7 shows cavitation from the 0.7 mm nozzle at 20.7 MPa (Re =
70,000). Figure 3.8 shows cavitation from the 2.0 mm nozzle at 20.7 MPa (Re = 198,000). These
image sequences illustrate the cavitation cloud behavior for varied pressure and varied jet diameter.
Figure 3.3 illustrates the initialization of a cloud from the 1.0 mm nozzle and 0.45 m pipe
at 17.2 MPa. The frame separation time for this and the following two sequences (Figures 3.4
and 3.6) is ∆t=0.0002 s. The first column shows that a small pulse exits the nozzle and quickly
collapses. The pressure plot on the vertical axis shows how the pressure increases from near zero
25
to 5 MPa in a very short time (0.0042 s). Once the full pressure stabilizes somewhat, columns
2 and 3 show that only minor pressure fluctuations are present. These fluctuations are still large
enough to cause the jet to pulsate though. The image sequence shows the growth and collapse of
the cavitation cloud.
It can be seen in Figure 3.4 that for the same size nozzle, 1.0 mm, at higher pressure, 24.1
MPa, (Re = 107,000) the cloud is slightly larger and travels a little farther. Once again the first col-
umn shows the ramp up of the pressure, and the other columns show the growth and collapse of the
cavitation cloud. The erratic pressure spikes at times: 0.0502, 0.0509, ... , 0.1174, 0.1181, 0.1187,
0.1196, 0.1197, 0.1204, are baffling. A plot of the pressure for the full discharge, Figure 3.5, shows
the pressure for the entire cavitation event, and highlights these spikes in pressure. It is assumed
that these spikes are due to gas bubbles trapped next to the transducer diaphragm that compressed
and expanded during the discharge of the jet. The experimental setup allowed for excess gas to be
released from the system to prevent compressibility from affecting the dynamic transducer. Mul-
tiple additional runs were taken after ensuring all gas was released from the system, but some of
these runs occasionally resulted in pressure spikes similar to Figure 3.5. Without being able to
visually determine whether or not there was any trapped gas it is difficult to definitively make any
conclusions about the obvious pressure spikes.
The pressure spikes are again seen in Figure 3.6; a discharge from the 1.0 mm nozzle at
27.6 MPa (Re = 114,000). The first twelve frames of the third column show the formation of the
cavitation cloud pulse. Around 0.0993 s the pressure spikes appear to affect the uniformity of the
pulsation, as the cloud does not appear to form in the same way.
Comparisons to Figure 3.1 between jet diameter are shown in Figures 3.7 and 3.8. The
small 0.7 mm diameter nozzle results in less intense cavitation, and the cloud is shorter and nar-
rower (Fig. 3.7). The larger diameter 2.0 mm jet exhibited intense cavitation that propagated far
downstream (Fig. 3.8). In order to fit the image sequence on the page, columns 1 and 2 are cropped
to allow column 3 to show how much farther downstream the cloud propagated. The cloud col-
lapsed outside of the frame. Measurements of the propagation lengths given in Section 3.1.2
required that the frame had a wider field of view to capture the entire jet.
The image sequences of Figure 3.9 and Figure 3.10 are representative comparisons between
nozzle diameter and reservoir pressure. Figure 3.9 shows the cavitating jet from the 1.0 mm nozzle
26
−3
x 10
0
0.0000 s 0.0268 s 0.0868 s
0.027 0.087
0.5
0.0275 0.0875
0.028 0.088
1.5
0.0285 0.0885
[s]
[s] [s]
0.029 0.089
2.5
0.0295 0.0895
0.03 0.09
3.5
0.0305 0.0905
4
0.0042 s 0.0310 s [MPa]
0.0910 s [MPa]
[MPa]
0.031 0.091
2 4 17.217.417.617.8 16.8 17 17.2
27
0.015 0.047 0.117
0.0150 s 0.0470 s 0.1170 s
0.0155 0.0475 0.1175
0.016 0.048 0.118
0.0165 0.0485 0.1185
0.017 0.049 0.119
[s] [s] [s]
0.0175 0.0495 0.1195
0.018 0.05 0.12
0.0185 0.0505 0.1205
0.019 0.051 0.121
[MPa] [MPa] [MPa]

0.0192 s 0.0512 s 0.1212 s
10 15 23 24 25 22 23 24
28
30
25
Pressure [MPa] 20
15
10
0
0 0.05 0.1 0.15
Time [s]
Figure 3.5: Pressure versus time plot for Figure 3.4. The leftmost pair of dashed lines is for the
first column of Figure 3.1. The center pair of lines is for the center column of Figure 3.1. The
rightmost pair of lines if for the third column of Figure 3.1 near the end of the life of the jet.
at the four test pressures (17.2 MPa, 20.7 MPa, 24.1 MPa, 27.6 MPa) with Re = 90,000–114,000 at
approximately the same time (0.1000 s) after cavitation inception. This time, 0.1000 s, was chosen
to avoid effects from the pressure ramp up and effects near the end of the jet discharge. Each of the
sequences, (a)–(d), displays a cloud pulse and its propagation distance. The maximum propagation
distance is manifested as the location beyond which the dark cavitation cloud fails to pass. It is
here that the main collection of bubbles collapses. In all of the image sequences there are still
what appear to be tiny bubbles traveling downstream after the general collapse of the cloud. This
is the lighter gray shading in the images which lengthens and spreads over time. These seem to be
remnant cavitation bubbles, but are extremely small and dispersed. It is unlikely that the lighter
gray shading is a collection of gas bubbles from the pressurized gas or water supply, as this was
not seen when the water jet was discharged at pressures less than one MPa. These smaller bubbles
(gray areas) are not considered in any of the forthcoming measurements. The maximum distance
the clouds propagate is shorter for the lowest pressure in sequence (a), and appears to increase with
increasing pressure to sequence (d) where the maximum distance is the farthest.
29
0.0174 s 0.0574 s 0.0974 s
0.0175 0.0575 0.0975
0.018 0.058 0.098
0.0185 0.0585 0.0985
0.019 0.059 0.099
[s] [s] [s]
0.0195 0.0595 0.0995
0.02 0.06 0.1
0.0205 0.0605 0.1005
0.021 0.061 0.101
0.0216 s 0.0616 s 0.1016 s

[MPa] 0.0215 [MPa] 0.0615 [MPa] 0.1015
16 18 20 22 24 25 26 27 24 25 26
30
0.0154 s 0.0654 s 0.1254 s
0.0155 0.0655 0.1255
0.016 0.066 0.126
0.0165 0.0665 0.1265
0.017 0.067 0.127
[s] [s] [s]
0.0175 0.0675 0.1275
0.018 0.068 0.128
0.0185 0.0685 0.1285
0.019 0.069 0.129
0.0196 s 0.0696 s 0.1296 s

[MPa] 0.0195 [MPa] 0.0695 [MPa] 0.1295
8 10 12 19 20 21 19 19.5 20
31
−3
x 10
0.0014 s 0.0154 s 0.0354 s
0.0155 0.0355
1.5
0.016 0.036
2
0.0165 0.0365
2.5
3 0.017 0.037
[s] [s] [s]
3.5 0.0175 0.0375
4
0.018 0.038
4.5
0.0185 0.0385
5
0.019 0.039
[MPa] 5.5
0.0014 s 0.0196 s 0.0396 s
[MPa] 0.0195 [MPa] 0.0395
1 2 3
12 14 18.8 19
32
0.1046 s 0.1004 s
Max. Distance
New Pulse
New Pulse Max. Distance
0.1058 s 0.1016 s
(a) (b)
0.0996 s 0.1030 s
New Pulse
Max. Distance
New Pulse
Max. Distance
0.1008 s 0.1042 s
(c) (d)
Figure 3.9: Comparison of cavitation (∆t=0.0002 s) from a 1.0 mm nozzle, 0.45-meter pipe at four
pressures: (a) 17.2 MPa (Re = 90,000), (b) 20.7 MPa (Re = 99,000), (c) 24.1 MPa (Re = 107,000),
(d) 27.6 MPa (Re = 114,000). Each sequence shows at least one new pulse and the maximum
distance at which the cavitation cloud collapses. The pulsation and propagation distance are clearly
marked for each pressure.
33
0.1026 s 0.1004 s 0.0350 s
0.1044 s 0.1022 s 0.0368 s
(a) (b) (c)
Figure 3.10: Cavitation cloud propagation is dependent on nozzle orifice diameter from the 0.45-
meter pipe at 20.7 MPa for (a) 0.7 mm (Re = 70,000), (b) 1.0 mm (Re = 99,000), (c) 2.0 mm
(Re = 198,000) diameter nozzle (∆t=0.0002 s). The cavitation cloud is much bigger as the nozzle
diameter increases. Note that the cavitation cloud from the 2.0 mm diameter nozzle has a faster
development and shorter total duration. It lasted less than 0.1000 seconds, so (c) shows 0.0350 s
to 0.0368 s.
The relationship between cavitation cloud behavior for the three nozzles (0.7 mm, 1.0 mm,
2.0 mm) at 20.7 MPa is portrayed in Figure 3.10. Again the sequences begin around 0.1000 s,
except sequence (c). The duration of cavitation from the 2.0 mm diameter nozzle was less than
0.1000 s. For this reason the 2.0 mm sequence, (c), begins at 0.0350 s instead. Each of the
sequences portrays the jet in its fully-developed state; meaning that the pressure is at its steady
value after ramping up from zero, as seen in Figures 3.2 and 3.5. The intensity of the cavitation
is discernible as a darker cloud. There is a distinct separation between cavitation pulses evidenced
by the lighter areas in the images, especially in (c). It is clear from these image sequences that
larger exit orifices lead to more cavitation bubbles and farther propagation distances. Over the
same length of time (0.0018 s) the 2.0 mm jet produced much more cavitation than the smaller 0.7
and 1.0 mm nozzles. All of these values are quantified in subsequent sections.
34
3.1.1 Duration of Cavitation
The length of time required for the column of water to be expelled through the nozzle de-
pended on three variables: reservoir pressure, nozzle diameter, and pipe length. Without additional
measurement techniques, the only way to measure the duration of the experiments was visually.
This section presents measurements of the duration of each cavitation event, timed from the first
visible cavitation bubbles until the nitrogen gas exited the nozzle.
Figure 3.11 shows the duration of the cavitation cloud as a function of the Reynolds num-
ber (Re = V d/ν). The data in this figure, and all subsequent plots, are differentiated by the test
parameters. The length of pipe from which the jet discharged is shown by marker shape; triangle,
square, and circle for the 0.45, 0.61, and 0.91 m pipes, respectively. Nozzle diameter is shown by
Pipe − Diameter
0.25 0.45m−0.7mm
0.45m−1mm
0.45m−2mm
Cavitation Duration [s]
0.2 0.61m−1mm
0.61m−2mm
0.91m−1mm
0.15 0.91m−2mm
0.1
0.05
0
0.5 1 1.5 2 2.5
Re 5
x 10
Figure 3.11: Time to discharge a column of water as a function of Reynolds number. Notice that
higher Re resulted in shorter times. The larger 2.0 mm diameter nozzle discharged in a shorter
amount of time than the smaller 1.0 mm diameter nozzle. Also, the discharge was faster for the
shorter pipes, entirely based on volume flow rate.
35
marker color; black, blue, and red for the 0.7, 1.0, 2.0 mm diameter nozzles, respectively. The four
test pressures are manifested by the conglomeration of the data in groups of four.
As expected from mass conservation; the duration of each discharge decreased as Re in-
creased; the larger column of water took longer to be fully expelled from the nozzle. For example,
visible cavitation issuing from the 0.91 m pipe with the 1.0 mm nozzle at 20.7 MPa (Re = 99,000)
lasted 0.2438 s, but only lasted 0.1415 s from the 0.45 m pipe with the same nozzle and pressure
(Re = 99,000). The columns of water also discharged faster through the larger 2.0 mm diameter
nozzle than the smaller 1.0 mm nozzle. The discharge time from the 0.45 m pipe at 20.7 MPa
was 0.1415 s from the 1.0 mm nozzle (Re = 99,000), and 0.0446 s for the 2.0 mm nozzle (Re =
198,000). Finally, it appears that higher pressure slightly decreased the discharge time. The dura-
tion of visible cavitation shortened by only 0.0275 s from 0.1521 s to 0.1246 s when the pressure
was increased from 17.2 MPa to 27.6 MPa from the 1.0 mm nozzle and 0.45 m pipe (Re = 90,000
and 114,000, respectively).
3.1.2 Propagation Distance
Multiple measurements of the propagation distance were taken for each experimental run
(see Table 2.1). For the primary pipe length, 0.45 m, a total of nine length measurements were
made for each pair of parameters—nozzle diameter and jet pressure. There were a total of eight
measurements taken for each run for the 0.61 m pipe, and only four measurements for each run for
the 0.91 m pipe. Although fewer experiments were run for the longer pipes, pipe length did not
seem to have a significant effect on the measurements, so more data were unnecessary.
The predicted propagation distances at the four test pressures, according to Equation 1.8,
ranged from 0.3 m to 1.5 m downstream from the nozzle. The actual propagation distance was
measured manually, as described in Section 2.4. Figure 3.12 visually shows that for a jet from the
0.45 m pipe at 20.7 MPa the nozzle diameter was the primary differentiator between the propa-
gation distances. Measurements from the three images show that the cavitation cloud collapsed
farther downstream at 27.8 cm for the 2.0 mm diameter nozzle (Re = 114,000) than the 0.7 mm di-
ameter nozzle (Re = 70,000), which collapsed at only 11.7 cm, and the 1.0 mm diameter nozzle (Re
= 99,000), which collapsed at 15.4 cm. In fact, this trend was the same for all experimental param-
36
(a) 0.7 mm diameter nozzle; 11.7 cm
(b) 1.0 mm diameter nozzle; 15.4 cm
(c) 2.0 mm diameter nozzle; 27.8 cm
Figure 3.12: Maximum propagation distance marked by the arrow for each nozzle diameter, (a)
0.7 mm (Re = 70,000), (b) 1.0 mm (Re = 99,000), (c) 2.0 mm (Re = 198,000), from the 0.45-meter
pipe at 20.7 MPa.
eter configurations. It is assumed that this occurred mainly because of the increased momentum of
the column of water from the larger diameter nozzle.
The measured distances are much shorter than predicted. The theory predicts the length of
a cavity behind a disk cavitator in a constant velocity flow. However, in the water jet experiments
the velocity decreases significantly as the flow moves downstream. Energy is lost, thus the actual
cavity cannot grow as large as the predicted cavity. The lost energy is likely due to viscosity, which
was neglected in the equations resulting in the significant discrepancy.
Figure 3.13 illustrates the linear dependence of the propagation distance on the Reynolds
number. The average collapse distance from the nozzle of the multiple measurements for each
run is the value plotted in Figure 3.13. Scatter in the multiple data points and the method of cal-
culating the lengths are accounted for in the uncertainty measurements as random and systematic
uncertainty, respectively (see uncertainty in Appendix A.1 and MATLAB® scripts in Appendix
A.2).
Nozzle diameter had the greatest effect on the cavitation cloud propagation distance, dis-
tinguished by color in the plots. The propagation distance for the smallest diameter nozzle (0.7
mm) was the shortest, ranging from 10±1.45 cm to 13±1.6 cm. The 1.0 mm nozzle resulted in
37
     
50
52 52 64 75 87 108 129 150
40
Propagation Distance [cm]
Pipe ï Diameter
0.45mï0.7mm
30 0.45mï1mm
0.45mï2mm
0.61mï1mm
0.61mï2mm
20 0.91mï1mm
0.91mï2mm
Theory
10
0
0.5 1 1.5 2 2.5
Re 5
x 10
Figure 3.13: Comparison of propagation distance versus Reynolds number. Nozzle diameter has
the greatest effect on cavitation cloud propagation distance. The distance increases with higher Re,
but notice that the length of the water column had relatively no effect on the propagation distance.
Theoretical distances are plotted for the given Re.
typical propagation lengths of 15±2.4 cm to 20±2.95 cm. The jet from the largest diameter nozzle
(2.0 mm) had a wider range of possible distances traveled at 20±2.96 cm to 32±4.81 cm. Fig-
ure 3.13 shows that the pipe length, distinguished by marker shape, had seemingly no effect on the
propagation distance, except for the 0.45 m pipe and 2.0 mm diameter nozzle. Pressure affected
the distance somewhat, but in some cases this effect was less significant; compare the maximum
propagation distances in Figure 3.9 where the four jets appear to all collapse in a similar location.
Based on this data, it can be concluded that for the tested parameters larger diameter jets
are capable of longer distances. This trend is significant in that it outlines how the jet parameters
can be scaled. A change in the diameter of the jet affects both Re and the propagation distance.
Future work calls for a larger-scale experiment, and the relationship between propagation distance
and Reynolds number will be essential to scaling up the jet. Future work is described further in
Section 4.2.
38
0.5
Pipe − Diameter
0.45m−0.7mm
0.45 0.45m−1.0mm
0.45m−2mm
0.4 0.61m−1mm
0.61m−2mm
0.91m−1mm
theory
0.35
0.91m−2mm
L/L
0.3
0.25
0.2
0.15
0.5 1 1.5 2 2.5
Re 5
x 10
Figure 3.14: Non-dimensionalized propagation distance versus Reynolds number. Clear separation
along the horizontal axis is an effect of the jet diameter. Vertical scatter is from the different lengths
of pipe.
Figure 3.14 compares the non-dimensional propagation distance to Reynolds number. The
length was non-dimensionalized by the predicted length of the cavity from Section 1.3. The data
are grouped in sets of four by the test pressure, and in color by the nozzle diameter. This plot
illustrates the large discrepancy between the theoretical and measured propagation distances. The
discrepancy increases with jet pressure.
3.1.3 Cavitation Cloud Width
One of the measured traits of the cavitation clouds was the typical width of the cloud. Once
cavitation began the cloud expanded radially, as well as longitudinally to a certain size before fully
collapsing. After collapsing, smaller bubbles continued to spread, but the exact makeup of these is
uncertain and their expansion was neglected. Only the width of the main cloud was measured. The
previous image sequences illustrate how the main cavitation cloud had limited width expansion.
This characteristic allowed for measurements of the typical cloud width to be made.
39
8
Jet Diameter
0.7 mm
7
1.0 mm
2.0 mm
6 Theory
Cloud Width [cm]
0.5 1 1.5 2 2.5

Re 5
x 10
Figure 3.15: Cavitation cloud width versus Reynolds number. The 2.0 mm jet exhibited wider cav-
itation clouds than the 1.0 mm or 0.7 mm jet, but all measured values became increasingly different
from predicted widths. The pulsatory nature of the jets dramatically influences the experimental
uncertainty.
Multiple width measurements were taken for each run of the three nozzles and four pres-
sures. The measurements were taken at a point downstream where the main cavitation cloud
reached its maximum width, which was not dependent on the length of pipe used. This point
was closer to the nozzle for the shorter propagating clouds, and farther from the nozzle for the
longer propagating clouds.
Figure 3.15 shows the average cloud width for cavitation for each of the nozzle diameters
and pressures tested. The predicted widths from Equation 1.7 were between 2.2 cm for cavitation
from the 0.7 mm nozzle at the lowest pressure and 7.9 cm for cavitation from the 2.0 mm nozzle at
the highest pressure. The discrepancy between the measured and predicted widths in Figure 3.15
increases with Re. The same reasoning as that used for the propagation length discrepancies is
again used here–viscosity and constant flow velocity were not considered in the predictions. The
measured widths for all three nozzle diameters are close together, ranging from 0.95 cm to 1.6
40
cm. As can be seen in the figure, significant measurement uncertainty exists; which is primarily
due to the rapid diameter changes as the cavity pulsated. The width visibly increased with nozzle
diameter, but these experiments emphasize that the cloud width is not as dependent on jet velocity
as the theory predicts.
3.1.4 Cavitation Cloud Area
In addition to the length and width measurements of the cavitation cloud, the longitudinal
cross-sectional area was also measured, as described in Section 2.4. It should be noted that this
area is not simply the product of the length and the width of the cloud, but is measured by counting
the number of pixels. Figure 3.16 illustrates the dependence of the cloud area on the Reynolds
number. The relationship is distinctly linear. The linear relationship between cloud area and Re is
similar to that illustrated in Figure 3.13, meaning the area is more dependent on the propagation
50
45
40
35
Cloud Area [cm2]
30
25
20
15
Jet Diameter
10 0.7 mm
5 1.0 mm
2.0 mm
0.5 1 1.5 2 2.5
Re 5
x 10
Figure 3.16: Comparison of the cavitation cloud area versus Reynolds number. The 2.0 mm jet
generates a larger cavitation cloud than the smaller diameter jets.
41
length than the cloud width. The nozzle diameter, specifically, had a major impact on the size of
the cavitation cloud. The points plotted in Figure 3.16 show the average area over the life of the
cavitation for each experimental run. Data for the three pipe lengths are combined into a single
point for each Reynolds number.
Over the range of test pressures the cloud from the 2.0 mm nozzle covered a larger area (30
and 45 cm2 ) than the cloud from the 1.0 mm (10-17 cm2 ) and 0.7 mm nozzles (5-9 cm2 ). These
results are distinctly visible in the image sequence of Figure 3.10, which compares cavitation
between the three nozzle diameters.
3.1.5 Pulsation Frequency
As can be seen in the sequences of Figures 3.1, 3.3, 3.4, 3.6–3.10 the jet exhibited an inter-
esting pulsation as the water was forced through the nozzle. Previous researchers have described
this pulsation and some associated trends, but failed to quantify the frequencies of these pulsations
(see [13–16]). The high-speed camera, as well as the dynamic pressure transducer, recorded this
pulsation. A trend for the pulsation rate was determined from the images according to the variation
of image intensity for a single column of pixels, by the method previously described in Section 2.4.
In Figure 3.17 the most prominent two or three frequencies of the pulsing cavitation cloud
for each run are plotted versus the Reynolds number. All of the frequency data are plotted to illus-
trate trends in the data. These trends would not be noticeable if only a single average frequency had
been plotted for each experimental run. The plotted frequencies are noticeably clustered according
to nozzle diameter. The longer pipes were included in the study initially under the suspicion that
pipe length would greatly influence the pulsation frequency. This proved to not be the case, as
manifested in the frequency plot.
Just as with the propagation distance the pulsation frequency of each jet depended more
on nozzle diameter than reservoir pressure or pipe length. The cavitation from each diameter jet
appears to be grouped by frequency as well. There appears to be a slight decrease in frequency
with increasing Reynolds number, which matches the trend described by Hutli [15], where the
frequency decreased as the jet pressure increased. Jets from the largest diameter nozzle (2.0 mm)
exhibited generally lower frequency content ranging from 250 Hz to 600 Hz. Jets from the 1.0
mm nozzle had pulsations ranging between 450 Hz and 800 Hz. The smallest diameter jets (0.7
42
1200
Pipe − Diameter
0.45m−0.7mm
0.45m−1mm
1000 0.45m−2mm
0.61m−1mm
0.61m−2mm
Frequency [Hz]
800 0.91m−1mm
0.91m−2mm
600
400
200
0.5 1 1.5 2 2.5
Re 5
x 10
Figure 3.17: Cavitation cloud pulsation frequencies versus Reynolds number. The pulsation fre-
quency was dependent on nozzle diameter, as seen by the color groupings, and jet pressure. The
primary pulsation frequency of the cavitating jet for the 2.0 mm diameter nozzle ranged between
250–600 Hz. The range for the jet with the 1.0 mm diameter nozzle was 450–800 Hz. The 0.7 mm
diameter nozzle gave the jet a pulsation frequency greater than 700 Hz.
mm) pulsed at higher frequencies between 700 Hz and 1000 Hz, generally. Frequency analysis
of the images and pressure data included signals around 100–130 Hz, which are not included in
Figure 3.17. Analysis of a group of images without cavitation resulted in only frequencies in this
range, and it was determined to be caused by the flickering of the backlighting. The pressure
data exhibited frequencies similar to those of the images, even the 100–130 Hz oscillations. It is
assumed that this is because the computer was plugged into the AC wall power.
3.1.6 Cloud Front Velocity
Each cavitation cloud grew longitudinally and spread radially as it moved away from the
nozzle. This action resulted in decreased velocity and energy downstream. It was expected that
higher reservoir pressures resulted in faster downstream velocities. Theoretically, for the experi-
43
35
30
Pulse Velocity [m/s]
25
20
15
0.7 mm nozzle
1.0 mm nozzle
2.0 mm nozzle
10
0.5 1 1.5 2 2.5
Re 5
x 10
Figure 3.18: Front velocity of cloud pulses versus Reynolds number. The measured values are
closer to the predicted velocities for the 0.7 mm diameter nozzle than the measured velocities for
the 1.0 mm and 2.0 mm diameter nozzles. The same relationship to jet pressure shown by the
theoretical front velocity is not seen in the measured data.
mental parameters in this study the exit velocity of the jet ranged from 100 m/s to 127 m/s. These
values were pressure-dependent and nozzle diameter-independent. The velocity of the front of the
cavitation cloud just before it collapsed was measured and compared to the predicted front veloc-
ity from Section 1.3. This measurement was sensitive to the location at which it was taken. The
cloud front velocity decreased dramatically from the nozzle exit to the point of collapse because
of mainly viscous effects.
Figure 3.18 shows the relationship between cloud front velocity and the Reynolds number
of the jet. The plotted data are the average front velocity of multiple sets of images for each
run. Pipe length had minimal effect on the front velocity, so all of the data from the three pipes
are averaged into the single value plotted for each set of parameters. Error bars are included to
show that the measurement uncertainty was typically only ±1–2 m/s. The nozzle diameter has
little effect on the predicted front velocity. From Equation 1.5 the front velocity is dependent
44
on the jet exit velocity, which at each test pressure was constant for the three nozzle diameters.
The predicted velocities increased with pressure for each nozzle diameter, as seen in Figure 3.18.
The experimental data do not follow the trend of increased front velocity for increased pressure.
There is a weak correlation of the opposite trend where the experimental front velocity decreases
with increasing Re. This decrease in front velocity from low Re to higher Re is common for the
measurements of all three nozzle sizes.
The sensitivity of the measured velocity on the location downstream could be the reason
that the data do not follow the predicted trends. As previously stated, the predicted values from
Equation 1.5 are based on the initial jet velocity and the area drag due to the diameter of the cavity.
The distance downstream at which the velocity measurement is made is not taken into account in
these predictions. Additionally, the initial jet velocity is independent of the nozzle diameter, but
the propagation distance is not. Once the cavitation bubbles are formed the higher Re jets (larger
nozzle diameter and higher pressures) result in farther propagation distances; assumedly because
the larger diameter jets have greater momentum, and higher pressure gives higher initial velocity.
The longer distance traveled results in more energy being lost to effects like viscosity. The lower
energy then causes the front velocity to be reduced, as was measured. Preliminary measurements of
the front velocity at relatively the same location downstream from the nozzle for each run showed
that the velocity would more likely follow a trend of increasing with Re.
3.1.7 Cavitation Erosion
An attempt was made to erode solid material with this experimental water jet. Due to the
setup of the jet this proved difficult. Cavitation is known to erode even the strongest materials.
To test whether this jet would erode solid material placed downstream from the nozzle multiple
discharges were fired at a piece of clear polystyrene. The solid polystyrene was placed 22 cm
away from the nozzle, a few centimeters short of the maximum average propagation distance for
the nozzle and pressure used. The nozzle diameter was 2.0 mm and the pressure used was 20.7
MPa, giving a Re = 198,000. It was assumed that even a few small eroded pits would be visible
in the clear solid since tiny scratches are visible. The jet was aligned with the solid specimen and
discharged 30 times at the clear polystyrene. Each discharge was monitored visually to ensure
that it hit the solid surface. When the solid was removed from the water, there were no visible
45
markings. No further testing was completed due to the length of time necessary and inconvenience
of running each test. In the literature it can take hundreds of jet pulses to affect any measurable
erosion. Future work is needed to adapt this mechanism to be able to be easily discharged many
times to erode a solid downstream.
3.2 Conclusion
This chapter presented the experimental results from a cavitating submerged water jet. The
behavior of the cavitation cloud was outlined with respect to the experimental parameters: nozzle
diameter, jet pressure, and water column length. Image sequences showed how the cavitation
cloud developed and propagated downstream. Longer pipe lengths lengthened the time it took
to discharge the jet, but affected little else. The jet lasted approximately 0.22 s longer when the
pipe length was doubled from 0.45 m to 0.91 m. The propagation length was shown to be highly
dependent on the nozzle diameter, and less so on the jet velocity. The measured lengths were
shorter than what was predicted, only ranging between 10–35 cm. Other geometric factors, such
as width and longitudinal area, were also measured, and show that increasing the jet diameter
resulted in greater cavitation. Each experimental case exhibited an interesting pulsation that was
quantified according to Reynolds number. The ultimate differentiator turned out to be the jet
diameter. Cavitation from jets with larger diameters pulsed at slightly lower frequencies, generally.
The frequency ranges for the three nozzle diameters, 0.7 mm, 1.0 mm, 2.0 mm, were >700 Hz,
450–800 Hz, and 250–600 Hz, respectively. Finally, the velocity of the front of the cavitation
cloud was measured and described in spite of some measurement challenges. An attempt was
made to erode a solid material with the cavitation cloud. The inconvenience of the experimental
setup prevented any measurable damage from being done. This is where significant future work is
necessary.
46
CHAPTER 4. CONCLUSIONS
The research presented in this thesis described the behavior of cavitation clouds from a
submerged water jet. Before cavitation clouds could be studied, a means of generating cavitation
had to be developed. A column of water discharged by ultrahigh-pressure nitrogen through a
small-diameter nozzle was the method used. There are various ways of generating cavitation and
this method was suited to the investigation of the behavior of cavitation clouds.
4.1 Summary
An axisymmetric water jet will cavitate when submerged in water if its velocity is suffi-
ciently high. For a column of water discharged into quiescent water by pressurized gas at pressures
from 17.2 to 27.6 MPa this cavitation appeared as a cloud of small bubbles. The cavitation cloud
exhibited a pulsation behavior that has not been fully quantified in the literature. Here it was shown
that the frequency at which the cloud pulsed was dependent on Reynolds number of the water jet
(see Figure 3.17). The frequency of pulsation decreased as the jet velocity and diameter increased.
It was found that low Re jets resulted in higher frequencies. For Re < 80,000 the range of com-
mon frequencies was 500–1000 Hz. For higher Re = 80,000–120,000, the range was approximately
400–800 Hz. The highest Re = 180,000–230,000, exhibited slightly lower frequencies, 250–600
Hz.
Each pulse of cavitation traveled away from the nozzle until it collapsed. The distance at
which this collapse occurred was also shown to be linearly dependent on Re, but more specifically
on the nozzle diameter. The distance increased as the diameter of the jet increased. For Re =
60,000–80,000 the distance traveled until collapse was approximately 10–14 cm. For Re = 90,000–
120,000 the distance was 15–20 cm. For Re = 180,000–230,000 the distance was 17–32 cm.
The longitudinal area and average width were measured to show how the cavitation ex-
panded from a thin water jet. Again, these were dependent on Re. The cloud area increased lin-
47
early with Re, whereas the width of the cloud followed a less distinct trend. The average maximum
cloud width remained relatively constant, around 1.2 cm, for all cases tested.
Lastly, it was determined that the relationship between the experimental parameters and
the velocity of the front of the cavitation cloud was somewhat less clear. The location at which
the velocity measurements were made played a significant role in what the data showed. The front
velocity decreases with Re, which was not what the theory predicted. As with some of the other
measurements, the fact that viscosity was neglected greatly affected how the experimental and
theoretical data were related.
4.2 Future Work
There have been myriad studies related to cavitation. This research focused on one partic-
ular aspect related to clouds of cavitation bubbles. The hope is that this work forms a foundation
for future research using a cavitating water jet to erode solid materials at a distance in underwater
environments.
At this time, a proposal has been written to a federal agency, the Strategic Environmental
Research and Development Program, to use cavitation from a water jet as a way to disassemble
unexploded ordnance (UXO) in submerged environments. Currently, there are no environmentally
friendly ways to eliminate UXO underwater. A device to cut the material into pieces from which
the explosives could be easily removed without contaminating the environment has been proposed.
In order for the research presented here to be of use for this device further research is
necessary. A numerical model is being developed at the Naval Undersea Warfare Center to allow
the adjustment of design parameters. This model needs to be validated and tested. With this model
a scaled-up cavitating jet could be tested to determine the maximum propagation distance, and
thus, the distance away from a solid at which the solid could be effectively cut. Once the model
is used to determine what size of a jet is necessary, a full-scale jet would be built and tested in a
submerged environment.
The previous research presented in Chapter 1 shows that cavitation erosion has the potential
to be an effective cutting mechanism. If a cavitation device that can remotely cut submerged solids
without polluting its surroundings is developed many underwater environments will become safer
for commercial and private use.
48
REFERENCES
[1] Parsons, C. A., and Cook, S. S., 1919. “Investigations into the causes of corrosion or erosion
of propellers.” Journal of the American Society of Naval Engineers, 31(2), pp. 536–541. 1, 3
[2] Reynolds, O., 1901. “Experiments showing the boiling of water in an open tube at ordinary
temperatures.” Papers on Mechanical and Physical Subjects, 2, pp. 578–587. 2
[3] Kornfeld, M., and Suvorov, L., 1944. “On the destructive action of cavitation.” Journal of
Applied Physics, 15, pp. 495–506. 3, 5
[4] Cook, S. S., 1928. “Erosion by water hammer.” Proceedings of the Royal Society of London,
119, pp. 481–488. 3, 4, 5, 6
[5] Knapp, R., Daily, J., and Hammitt, F., 1970. Cavitation. McGraw-Hill Book Company, New
York. 4, 5
[6] Munson, B. R., Young, D. F., and Okiishi, T. H., 2006. Fundamentals of Fluid Mechanics.,
fifth ed. John Wiley and Sons, Inc. 4, 6
[7] Wu, X., and Chahine, G., 2007. “Characterization of the content of the cavity behind a
high-speed supercavitating body.” Journal of Fluids Engineering, 129, pp. 136–145. 4
[8] Rayleigh, L., 1917. “On the pressure developed in a liquid during the collapse of a spherical
cavity.” Philosophical Magazine, 34, pp. 94–98. 4, 6
[9] Ooi, K. K., 1985. “Scale effects on cavitation inception in submerged water jets: a new look.”
Journal of Fluid Mechanics, 151, pp. 367–390. 4
[10] Gopalan, S., Katz, J., and Knio, O., 1999. “The flow structure in the near field of jets and its
effect on cavitation inception.” Journal of Fluid Mechanics, 398, pp. 1–43. 4, 6
[11] O’Hern, T. J., 1990. “An experimental investigation of turbulent shear flow cavitation.” Jour-
nal of Fluid Mechanics, 215, pp. 365–391. 4, 6
[12] Soyama, H., Yanauchi, Y., Sato, K., Ikohagi, T., Oba, R., and Oshima, R., 1996. “High-
speed observation of ultrahigh-speed submerged water jets.” Experimental Thermal and Fluid
Science, 12, pp. 411–416. 4, 6
[13] Trilling, L., 1952. “The collapse and rebound of a gas bubble.” Journal of Applied Physics,
23(1), pp. 14–17. 24, 42
[14] Strasberg, M., 1953. “The pulsation frequency of nonspherical gas bubbles in liquids.” Jour-
nal of the Acoustical Society of America, 25(3), pp. 536–537. 24, 42
49
[15] Hutli, E., and Nedeljkovic, M., 2008. “Frequency in shedding/discharging cavitation clouds
determined by visualization of a submerged cavitating jet.” Journal of Fluids Engineering,
130. 5, 42
[16] Sato, K., Sugimoto, Y., and Ohjimi, S., 2009. “Structure of periodic cavitation clouds in
submerged impinging water-jet issued from horn-type nozzle.” In Ninth Pacific Rim Interna-
tional Conference on Water Jetting Technology. 5, 42
[17] Sato, K., Sugimoto, Y., and Ohjimi, S., 2009. “Pressure-wave formation and collapses of cav-
itation clouds impinging on a solid wall in a submerged water jet.” In Seventh International
Symposium on Cavitation CAV2009. 5, 24
[18] Hansson, I., and Morch, K. A., 1980. “The dynamics of cavity clusters in ultrasonic (vibra-
tory) cavitation erosion.” Journal of Applied Physics, 51(9), pp. 4651–4658. 5, 24
[19] Reisman, G. E., Wang, Y. C., and Brennen, C. E., 1998. “Observations of shock waves in
cloud cavitation.” Journal of Fluid Mechanics, 355, pp. 255–283. 5
[20] Philipp, A., and Lauterborn, W., 1998. “Cavitation erosion by single laser-produced bubbles.”
Journal of Fluid Mechanics, 361, pp. 75–116. 5, 6
[21] Coleman, S. L., and Scott, V. D., 1995. “Comparison of tunnel and jet methods for cavitation
erosion testing.” Wear, 184, pp. 73–81. 5, 6
[22] Bourne, N. K., and Field, J. E., 1995. “A high-speed photographic study of cavitation dam-
age.” Journal of Applied Physics, 78(7), p. 4423. 5
[23] Brujan, E. A., Keen, G. S., Vogel, A., and Blake, J. R., 2002. “The final stage of the collapse
of a cavitation bubble close to a rigid boundary.” Physics of Fluids, 14(1), pp. 85–92. 5, 6
[24] Jayaprakash, A., Chahine, G., and Hsiao, C. T., 2010. “Numerical and experimental study of
the interaction of a spark-generated bubble and a vertical wall.” In ASME 2010 International
Mechanical Engineering Congress and Exposition. 5, 6
[25] Lu, J., Xu, R. Q., Chen, X., Shen, Z. H., Ni, X. W., Zhang, S. Y., and Gao, C. M., 2004.
“Mechanisms of laser drilling of metal plates underwater.” Journal of Applied Physics, 95(8),
pp. 3890–3894. 6
[26] May, A., 1975. Water entry and cavity-running behavior of missiles 75-2
Tech. rep., NAVSEA Hydroballistics Advisory Committee, http://www.dtic.mil/cgi-
bin/GetTRDoc?AD=ADA020429&Location=U2&doc=GetTRDoc.pdf. 7, 8
[27] Savchenko, Y. N., 2001. “Supercavitation–problems and perspectives.” In CAV 2001: Fourth
International Symposium on Cavitation. 7, 22
[28] Figliola, R., and Beasley, D., 2006. Theory and Design for Mechanical Measurements.,
fourth ed. John Wiley and Sons, Inc. 15, 51
50
APPENDIX A. UNCERTAINTY AND MATLAB® SCRIPTS
A.1 Uncertainty Calculations
This section shows the calculation steps to determine the experimental uncertainty of the
various length measurements, and other measurements. The process is according to Figliola [28].
That text is where the student-t values were found as well. The equations are listed in this section,
but the calculations of the values are done in the MATLAB® codes listed in Section A.2.
The static and dynamic pressure measurements had errors of ±1.1 MPa and ±0.5 MPa,
respectively. These were acceptible, as pressure did not greatly affect any of the test results.
The random error resulting from multiple measurements is found from the equation:
q
P= Σ (Pi )2 (A.1)
Si
Pi = √ , (A.2)
Ni
where Si is the standard deviation of the measured values, and Ni is the number of values. The
systematic error from the measurement calculations is given by:
s 2
∂L
B= Ucalibration . (A.3)
∂ calibration
The calibration factors are listed in 2.3. Ucalibration is the error in picking pixels set to 5 pixels.
L is the equation of what is being measured; this is either length, velocity, area, etc. The total
uncertainty is found from the following equation:
q
UTotal = B2 + (tαν P)2 , (A.4)
where tαν is the student-t value for the number of measurements. The probability I used was 95%.
51
A.2 MATLAB® Scripts
A.2.1 Propagation Distance
The propagation distance was measured with this code and each value was loaded into the
next code:
% Cavitation Measurements
% Michael Wright
% October 28, 2011
clc
clear all
% Create [NxMxj] 3D array where each image is [NxMx1]
% Create list of files into an array
fiLes = dir(’/Volumes/WRIGHTDRIVE/Supercav_BYU/JET/...
Images/Images18/07mm/2500_1/*.tif’);
newpath = ’/Volumes/WRIGHTDRIVE/Supercav_BYU/JET/...
Images/Images18/07mm/2500_1/’;
%--------------------------------------------------------------------------
addpath(newpath)
x = 2320;
y = x+100;
for j = x:20:y
imshow(fiLes(j).name);
iMage(:,:,j) = imread(fiLes(j).name); % Read in each image to a 3D array
imshow(iMage(:,:,j)) % Show image on the screen
n = 1; % Number of distances desired
M(:,:,j) = ginput(2*n); % [X,Y,image #] of chosen points
% Count the number of pixels between chosen points
% In piXels each row contains the measurements from a single image
for i = 1:n
52
piXels(j,i) = sqrt((M(2*i,1,j)-M(2*i-1,1,j))^2 +...
(M(2*i,2,j)-M(2*i-1,2,j))^2);
% Conversion factor for each run: pixels/cm
cFc = 25.02;
cFf = 20.7;
cF24 = 19.5;
cF36 = 19.25;
end
end
close all
piXels;
Length = piXels./cFf; % Length of jet in cm
L = Length(Length~=0)
A.2.2 Length Uncertainty
The length measurements taken from the previous code are plotted here, and include the
measurement uncertainty.
clc;
clear all
% Length output from processing (Row is pressure)
L187 = [10.6 10.4 8.48 11.993 6.742 9.353 11.867 9.233 9.033;
14.232 11.713 9.322 11.35 10.793 9.952 10.472 11.112 10.834;
12.79 10.796 10.75 13.67 13.15 10.23 13.315 12.912 12.432
12.842 12.232 13.11 13.631 12.632 11.592 12.994 14.39 11.232];
L181 = [15.072 14.029 14.988 14.75 16.747 12.151 12.31 17.549 15.348;
17.626 13.71 15.428 14.11 14.588 15.788 14.19 15.188 13.43;
17.506 17.715 17.467 16.187 17.747 16.428 17.388 15.629 17.146;
15.108 18.466 20.744 21.103 17.786 19.824 19.864 17.186 16.667];
53
L182a =[28.382 28.738 25.38 27.578 26.542 24.145 22.91 21.903 22.539;
27.818 27.099 28.937 29.618 29.658 28.898 26.14 26.661 28.577];
L182b =[25.845 30.486 32.611 35.508 35.413 34.058 30.097 26.042 29.76;
31.161 34.832 34.252 32.609 30.93 31.014 27.006 34.941 31.014];
L241 = [17.231 14.052 15.846 14.564 15.847 13.847 16.564 11.436;

17.385 15.949 19.181 16.923 17.128 16.615 15.897 16.821;
17.951 18.104 17.646 17.181 18.616 17.283 18.308 15.181;
16.667 20.821 19.744 19.898 19.949 20.564 20.667 17.693];
L242 = [20.872 21.18 21.026 23.847;

21.181 19.541 24.308 23.078;
23.488 27.18 28.053 27.231;
28.821 28.615 25.694 27.643];
L361 = [13.507 15.065 14.234 15.325;

14.597 15.065 17.663 15.896;
19.169 18.39 17.664 17.351;
18.702 16.364 20.052 16.052];
L362 = [16.832 22.082 17.87 20.68;

23.273 23.898 24.884 23.534;
25.925 22.857 26.241 28.675;
24.001 33.768 27.689 24.99];
%-------------------------------------
% t-distribution values
t18 = 2.306; % for 9 values
t24a = 2.365; % for 8 values
54
t24b = 3.182; % for 4 values
t36 = 3.182; % for 4 values
%-------------------------------------
% P_i values
for j = 1:2
P182a(j) = std(L182a(j,:))/sqrt(length(L182a(j,:)));
P182b(j) = std(L182b(j,:))/sqrt(length(L182b(j,:)));
end
P182 = [P182a P182b];
clear P182a P182b
for i = 1:4
P187(i) = std(L187(i,:))/sqrt(length(L187(i,:)));
end
%---------------------------------------------------
% uncertainty of measured scatter
S187 = sqrt(P187(1)^2+P187(2)^2+P187(3)^2+P187(4)^2);
S181 = sqrt(P181(1)^2+P181(2)^2+P181(3)^2+P181(4)^2);
S182 = sqrt(P182(1)^2+P182(2)^2+P182(3)^2+P182(4)^2);
S241 = sqrt(P241(1)^2+P241(2)^2+P241(3)^2+P241(4)^2);
S242 = sqrt(P242(1)^2+P242(2)^2+P242(3)^2+P242(4)^2);
S361 = sqrt(P361(1)^2+P361(2)^2+P361(3)^2+P361(4)^2);
S362 = sqrt(P362(1)^2+P362(2)^2+P362(3)^2+P362(4)^2);
%---------------------------------------------------------------
% differential Pixel lengths
x187 = L187./25.02;
55
x181 = L181./25.02;
x182a = L182a./25.02;
x182b = L182b./20.7;
x182 = [x182a;x182b];
x241 = L241./19.5;
x242 = L242./19.5;
x361 = L361./19.25;
x362 = L362./19.25;
% uncertainty of pixel choice
for i = 1:4
B187(i) = sqrt(x187(i,1)^2+x187(i,2)^2+x187(i,3)^2+...
x187(i,4)^2+x187(i,5)^2+x187(i,6)^2+x187(i,7)^2+...
x187(i,8)^2+x187(i,9)^2);
B181(i) = sqrt(x181(i,1)^2+x181(i,2)^2+x181(i,3)^2+...
x181(i,4)^2+x181(i,5)^2+x181(i,6)^2+x181(i,7)^2+...
x181(i,8)^2+x181(i,9)^2);
B182(i) = sqrt(x182(i,1)^2+x182(i,2)^2+x182(i,3)^2+...
x182(i,4)^2+x182(i,5)^2+x182(i,6)^2+x182(i,7)^2+...
x182(i,8)^2+x182(i,9)^2);
B241(i) = sqrt(x241(i,1)^2+x241(i,2)^2+x241(i,3)^2+...
x241(i,4)^2+x241(i,5)^2+x241(i,6)^2+x241(i,7)^2+x241(i,8)^2);
B242(i) = sqrt(x242(i,1)^2+x242(i,2)^2+x242(i,3)^2+x242(i,4)^2);
B361(i) = sqrt(x361(i,1)^2+x361(i,2)^2+x361(i,3)^2+x361(i,4)^2);
B362(i) = sqrt(x362(i,1)^2+x362(i,2)^2+x362(i,3)^2+x362(i,4)^2);
end
%-----------------------------------------------------------------
% average lengths
m182a = mean(L182a,2);
m182b = mean(L182b,2);
56
m187 = mean(L187,2);
m181 = mean(L181,2);
m182 = [m182a; m182b];
m241 = mean(L241,2);
m242 = mean(L242,2);
m361 = mean(L361,2);
m362 = mean(L362,2);
% TOTAL uncertainty of each set of parameters

for k = 1:4
M(1,:) = m187;
M(2,:) = m181;
M(3,:) = m182;
M(4,:) = m241;
M(5,:) = m242;
M(6,:) = m361;
M(7,:) = m362;
U(1,k) = sqrt(B187(k)^2 + t18*P187(k)^2);

U(2,k) = sqrt(B181(k)^2 + t18*P181(k)^2);
U(3,k) = sqrt(B182(k)^2 + t18*P182(k)^2);
U(4,k) = sqrt(B241(k)^2 + t24a*P241(k)^2);
U(5,k) = sqrt(B242(k)^2 + t24b*P242(k)^2);
U(6,k) = sqrt(B361(k)^2 + t36*P361(k)^2);
U(7,k) = sqrt(B362(k)^2 + t36*P362(k)^2);
end
%------------------------------------------------
% Create a figure
X = [17.2 20.7 24.1 27.6];
57
figure
hold on
errorbar(X,M(1,:),U(1,:),’-.^k’,’MarkerFaceColor’,’k’,’MarkerSize’,8)
errorbar(X,M(2,:),U(2,:),’-^k’,’MarkerFaceColor’,’c’,...
’MarkerSize’,8,’MarkerEdgeColor’,’k’)
errorbar(X,M(3,:),U(3,:),’--^r’,’MarkerFaceColor’,’r’,...
errorbar(X,M(4,:),U(4,:),’-sk’,’MarkerFaceColor’,’c’,...
errorbar(X,M(5,:),U(5,:),’--sr’,’MarkerFaceColor’,’r’,...
errorbar(X,M(6,:),U(6,:),’-ok’,’MarkerFaceColor’,’c’,...
errorbar(X,M(7,:),U(7,:),’--or’,’MarkerFaceColor’,’r’,...
hold off
xlabel(’Pressure [MPa]’)
ylabel(’Propagation Distance [cm]’)
legend(’0.46m-0.7mm’,’0.46m-1mm’,’0.46m-2mm’,...
’0.61m-1mm’,’0.61m-2mm’,’0.91m-1mm’,’0.91m-2mm’,’Location’,’EastOutside’)
print(’-depsc’,’-tiff’,’-r1200’,’Length_Feb17’)
A.2.3 Cloud Area
Significant image processing was required to convert the images into binary images without
losing important data. This code measures the area of the cavitation cloud.
% Michael Wright
% Cavitating Submerged Water Jet
% 10/19/2011
% This program outlines the cavitation cloud in an image of a submerged
58
% cavitating water jet.
tic
clc;
clear all;
% Add the new directory to the path being searched
addpath(newpath)
nozzle = imread(fiLes(1).name); % Read in first image to remove nozzle

background = imopen(nozzle,strel(’disk’,5));
nozzle2 = nozzle + background;
level = graythresh(nozzle2);%-0.5;
Bnoz = im2bw(nozzle2,level);
Bnoz = bwareaopen(Bnoz,30);
Wnoz = imcomplement(Bnoz);
X = ones(7);
Nozzle = imdilate(Wnoz,X);
%%
i = 1780;
k = 1780;%numel(fiLes)-20;
for j = i:k%numel(fiLes)
iMage = imread(fiLes(j).name); % Read each image to a 3D array iMage
% Open image and use structuring element to turn the background blank
background = imopen(iMage,strel(’disk’,5));
59
iMage2 = iMage + background;
% Convert the image to binary--black jet on white background

level = graythresh(iMage2)-0.35;
Bjet = im2bw(iMage2,level);
Bjet = bwareaopen(Bjet,30);
% Switch black/white of binary image

Wjet = imcomplement(Bjet);
% use first image to eliminate nozzle from subsequent images

WJn = Wjet-Nozzle; % White jet without nozzle
for p = 1:128
for q = 200:768
if WJn(p,q) < 0
WJn(p,q) = 0;
end
end
end
figure % Show jet without the nozzle
imshow(WJn,’InitialMagnification’,200);
pause(0.1)
% Plot the boundaries around each cloud

[B,L,N,A] = bwboundaries(WJn);
hold on
for m = 1:length(B)
boundary = B{m};
plot(boundary(:,2),boundary(:,1),’r’,’LineWidth’,2)
pause(0.1)
60
end
hold off
% Computes the area of the jet in pixels
Area(j) = bwarea(WJn);
end
%Find the average cloud area in square centimeters
Cloud = 0.0016.*Area(1,i:k);
C = [min(Cloud) max(Cloud) mean(Cloud) std(Cloud)];
A.2.4 Pulsation Frequency
The following code measures the image color intensity of a column of pixels at a fixed
distance downstream from the nozzle. These intensity values are run through the fft function to
find the frequency spectrum of the jets.
tic
clc;
clear all;
fiLes = dir(’/Volumes/WRIGHTDRIVE/Supercav_BYU/JET/Images/...
Images18/1mm/3000_1/*.tif’);
% Add the new directory to the path being searched
newpath = ’/Volumes/WRIGHTDRIVE/Supercav_BYU/JET/Images/...
Images18/1mm/3000_1/’;
addpath(newpath)
%%
i = 1;
clear Inten
for j = 10+50:numel(fiLes)-90 %change beginning point for each set
iMage = imread(fiLes(j).name);
61
iCol = iMage(:,620);
Inten(i) = sum(iCol)/length(iCol);
i = i+1;
end
Fs = 5000;
y = [Inten, Inten, Inten];
Nsamps = length(y);
t = (1/Fs)*(1:Nsamps);
y_fft = abs(fft(y));
y_fft(1) = [];
y_fft = y_fft(1:Nsamps/2);
f = Fs*(0:Nsamps/2-1)/Nsamps;
figure(2)
plot(t,y)
xlabel(’Time [s]’)
ylabel(’Intensity’)
% print(’-depsc’,’-tiff’,’-r1200’,’Intensity’)
figure(3)
plot(f,y_fft)
xlabel(’Frequency [Hz]’)
ylabel(’Magnitude’)
% print(’-depsc’,’-tiff’,’-r1200’,’FreqSpectrum’)
%==========================================================
% frequency values from "Frequency.m" amplitude spectrum
clc
clear all
% all 18 inch data
62
s18_07 = [10 39 120 33 116 15 21 41 116,...
113 686 790 31 77 113 750 851 882 954 1103 24...
80 118 760 915, 16 114 781 114 750 56 427 678...
754 830, 17 115 582 663 721 888 113 664 713...
835 924 105 713 922];
t18_07 = 0.0068977573.*[2500 2500 2500 2500 2500...
2500 2500 2500 2500, 3000 3000 3000 3000...
3000 3000 3000 3000 3000 3000 3000 3000...
3000 3000 3000 3000, 3500 3500 3500 3500...
3500 3500 3500 3500 3500 3500, 4000 4000...
4000 4000 4000 4000 4000 4000 4000 4000...
4000 4000 4000 4000];
s18_1 = [115 627 663 692 907 116 616 637 665 918 115 536 692 821,...
110 561 627 112 544 581 678 114 491 582, 111 577 112 533...
589 740 32 112 583 640 1078, 42 109 233 543 643 810 467...
519 570 778 115 460 550 788];
t18_1 = 0.0068977573.*[2500 2500 2500 2500 2500 2500 2500 2500...
2500 2500 2500 2500 2500 2500, 3000 3000 3000 3000 3000 3000...
3000 3000 3000 3000, 3500 3500 3500 3500 3500 3500 3500 3500...
3500 3500 3500, 4000 4000 4000 4000 4000 4000 4000 4000 4000...
4000 4000 4000 4000 4000];
s18_2 = [111 269 491 313 411 609 370 440 511 660, 94 280 317...
392 332 410 723 18 308 453 490 634, 26 230 536 24 500 23 711,...
40 402 522 582 51 278 353 429 41 226 288 350];
t18_2 = 0.0068977573.*[2500 2500 2500 2500 2500 2500 2500 2500...
2500 2500, 3000 3000 3000 3000 3000 3000 3000 3000 3000...
3000 3000 3000, 3500 3500 3500 3500 3500 3500 3500,...
4000 4000 4000 4000 4000 4000 4000 4000 4000 4000 4000 4000];
63
% all 24 inch data
s24_1 = [35 120 580 689 629 706 788, 120 550 598 638 118 580...
630 654 720, 118 475 580 685 118 464 542 613, 520 606...
120 404 566 627];
t24_1 = 0.0068977573.*[2500 2500 2500 2500 2500 2500 2500,...
3000 3000 3000 3000 3000 3000 3000 3000 3000, 3500 3500...
3500 3500 3500 3500 3500 3500, 4000 4000 4000 4000 4000 4000];
s24_2 = [115 398 620, 473 510 883, 312 393, 380 512];
t24_2 = 0.0068977573.*[2500 2500 2500,...
3000 3000 3000, 3500 3500, 4000 4000];
% all 36 inch data

s36_1 = [118 469 610 649 747, 117 530 649,...
118 470 508 618 803, 119 560 797];
t36_1 = 0.0068977573.*[2500 2500 2500 2500 2500,...
3000 3000 3000, 3500 3500 3500 3500 3500,...
4000 4000 4000];
s36_2 = [392 427 479, 346 483 562, 282 365, 325 421 111];
t36_2 = 0.0068977573.*[2500 2500 2500, 3000 3000 3000,...
3500 3500, 4000 4000 4000];
% shape denotes pipe length, color denotes nozzle

figure(1)
hold on
plot(t18_07,s18_07,’k^’,’MarkerSize’,8,’MarkerFaceColor’,’k’,...
’MarkerEdgeColor’,’k’)
plot(t18_1,s18_1,’k^’,’MarkerSize’,8,’MarkerFaceColor’,’c’,...
64
plot(t18_2,s18_2,’r^’,’MarkerSize’,8,’MarkerFaceColor’,’r’,...
plot(t24_1,s24_1,’ks’,’MarkerSize’,8,’MarkerFaceColor’,’c’,...
plot(t24_2,s24_2,’rs’,’MarkerSize’,8,’MarkerFaceColor’,’r’,...
plot(t36_1,s36_1,’ko’,’MarkerSize’,8,’MarkerFaceColor’,’c’,...
plot(t36_2,s36_2,’ro’,’MarkerSize’,8,’MarkerFaceColor’,’r’,...
axis([16.1 28.6 0 1000])
ylabel(’Frequency [Hz]’)
legend(’0.45m-0.7mm’,’0.45m-1mm’,’0.45m-2mm’,’0.61m-1mm’,...
’0.61m-2mm’,’0.91m-1mm’,’0.91m-2mm’,’Location’,’EastOutside’)
hold off
% print(’-depsc’,’-tiff’,’-r1200’,’FrequencyPress’)
A.2.5 Front Velocity Measurement
This code is similar to the length measurement, but it divides distances by the time between
frames to find the velocity of the cloud front.
% Supercavitating Jet Measurements

% Michael Wright
% October 28, 2011
%
clc
clear all
65
%--------------------------------------------------------------------------
addpath(newpath)
z = 3574
for j = 3398:3403
imshow(fiLes(j).name);
iMage(:,:,j) = imread(fiLes(j).name); % Read each image to a 3D array
imshow(iMage(:,:,j)) % Show image on the screen
n = 1; % Number of distances desired
M(:,:,j) = ginput(2*n); % [X,Y,image #] of chosen points
% Count the number of pixels between chosen points
% In piXels each row contains the measurements from a single image
for i = 1:n
piXels(j,i) = sqrt((M(2*i,1,j)-M(2*i-1,1,j))^2 +...
(M(2*i,2,j)-M(2*i-1,2,j))^2);
% Conversion factor for each run: pixels/m
cFc = 2502;
cFf = 2070;
cF24 = 1950;
cF36 = 1925;
end
end
close all
66
piXels;
Length = piXels./cF24; % Length of jet in cm
% Time between each frame
dT = 1/5000;
for m = 1:numel(Length)-1
Velocity(m,1) = (Length(m+1) - Length(m))/dT; % Velocity in m/s
end
Velocity
%=========================================================
This code uses the values obtained from the previous to plot the velocities.
clc
clear all
%==========================================================================
% Velocity raw data where row is pressure
va725 = [43.89 57.911 32.15 57.945 58.078 57.935 53.762...
73.938 44.007 47.993 33.965 32.011 33.828 34.089 47.97...
1 37.968 33.993 85.875 62.126 49.831 30.011 63.983 41.997...
43.944 57.968];
va730 = [41.955 69.933 39.995 27.973 83.924 79.991 47.956 65.908...
50.015 75.898 61.807 57.944 65.975 71.855 79.87 69.965 75.927...
85.924 79.918 72.009 80.079 43.884 49.9 64.203 53.955 50.059...
83.933 78.033 59.877 71.991];
va735 = [87.981 61.977 77.86 87.895 73.956 77.92 72.037 75.92...
64.024 74.001 76.042 83.875 76 74.07 51.938 71.958 47.945...
63.956 65.994 49.906 50.052 79.932 55.954 49.977 67.927 66.01...
71.925 75.958 87.901];
va740 = [84.084 59.695 54.075 43.843 51.964 72.059 73.922 57.909...
63.924 53.975 51.991 71.916 65.962 77.986 75.911 61.976 40.114...
67
79.958 49.948 71.878 76.015 85.949 71.782 64.016 49.953 79.909 67.886];
va125 = [97.873 85.937 87.938 65.97 79.927 77.932 81.87 86.001...

85.891 65.97 67.915 75.956 75.896 83.956 81.96 45.943...
55.969 47.928 62.093 61.956];
va130 = [75.952 63.931 81.989 43.923 56.003 77.882 59.991...
59.947 69.96 83.931 57.972 43.962 83.91 73.945 95.913 81.934...
65.965 79.928 83.933 67.936 76.083 69.844 81.951 79.931 93.91...
93.931 92.04 71.859 73.961];
va135 = [89.879 77.8 79.935 69.854 85.982 101.87 63.949 81.939...
85.927 85.882 79.935 59.952 51.958 91.916 109.93 55.95 99.92...
83.999 77.931 99.92 89.942 53.956 67.94 79.951 66.022 83.956 81.887];
va140 = [101.97 51.881 55.951 61.954 95.939 95.899 99.924 93.941...
85.909 79.994 111.82 51.994 65.96 79.918 77.941 83.933 101.92...
93.96 83.907 103.92 107.92 73.985 97.889 93.921];
va225 = [64.005 56.033 53.862 50.032 60.075 63.881 73.807 61.947...

61.906 59.947 67.901 71.902 59.951 77.933 55.959 77.962 89.808 67.945];
va230 = [65.945 45.956 49.966 81.978 71.91 69.925 61.95 83.936...
73.937 71.959 75.775 73.939 94.12 81.779 93.977 89.849 74.03...
81.856 81.928 61.942 71.845 93.924 125.9];
va235 = [57.964 55.577 89.371 79.709 108.8 79.639 96.188 77.375...
99.033 94.236 94.179 101.49 57.917 67.64 31.4 53.139 43.478...
55.572 94.184 65.227 82.146 101.57 67.491 74.947 65.149 48.308...
91.833 94.155 79.71 79.706 72.478 74.868 79.719 58.065 77.274...
55.51 86.957 65.248 55.552 96.63 79.708];
va240 = [62.898 55.442 55.564 65.216 86.978 103.86 87.015 74.837...
113.6 96.597 128.02 135.25 96.618 94.227 77.278 103.85 53.139...
65.216 60.386 123.19 94.203 106.26 106.32 106.26];
68
vb125 = [66.655 48.751 94.839 76.872 71.793 69.261 53.801 38.47...
53.838 74.384 61.503 43.597 51.281 61.535 61.536 43.58 41.033...
76.934 74.357 46.176];
vb130 = [64.101 74.409 46.096 71.793 74.351 71.818 69.199 53.846...
41.031 64.102 79.483 79.511 61.511 105.12 76.921 87.178...
69.229 61.557];
vb135 = [117.91 74.358 79.48 56.458 82.04 87.179 64.109 97.529 84.481...
74.377 102.59 102.53 76.915 82.043 76.929 99.999 53.854...
71.794 71.794 94.871];
vb140 = [46.134 64.102 41.025 135.85 97.436 89.749 115.38 79.414...
61.635 51.191 110.28 87.156 102.56 64.273 74.155 105.12 84.601...
82.051 59.088 107.63];
vb225 = [56.431 56.387 56.285 76.897 20.457 56.345 41.239 82.06...

64.094 59 84.612 105.11];
vb230 = [84.553 58.974 56.415 69.266 79.475 56.35 79.534 61.58...
61.496 46.221 77.051 58.911];
vb235 = [56.329 53.838 56.41 84.607 84.621 102.58 117.97 84.611...
66.649 43.628];
vb240 = [84.578 79.544 105.15 94.837 64.101 76.921 76.979 69.224...
84.57 87.175 79.549];
vc125 = [70.119 60.999 70.159 84.262 71.437 49.378 88.325 ...

72.714 92.406 64.752];
vc130 = [120.75 76.629 77.956 53.326 57.107 68.831 97.407 58.441];
vc135 = [89.61 66.236 67.556 97.388 80.528 93.498 90.994 75.254 92.196];
vc140 = [87.008 85.712 78.004 97.396 100.02 74.016 50.649 58.498 66.181];
vc225 = [72.698 73.772 69.113 68.659 58.416 71.431];

vc230 = [63.635 67.551 59.751 67.529 51.803 80.375 74.004 87.009...
69
97.47 64.899 59.714];
vc235 = [67.504 76.629 74.019 97.499 89.733 76.878 46.764];
vc240 = [54.482 67.544 66.221 79.731 86.364 77.97 76.643 68.828 62.415];
%==========================================================================
P = [17.2 20.7 24.1 27.6];
M(1,:) = [mean(va725) mean(va730) mean(va735) mean(va740)];
M(4,:) = [mean(vb125) mean(vb130) mean(vb135) mean(vb140)];
M(5,:) = [mean(vb225) mean(vb230) mean(vb235) mean(vb240)];
M(6,:) = [mean(vc125) mean(vc130) mean(vc135) mean(vc140)];
M(7,:) = [mean(vc225) mean(vc230) mean(vc235) mean(vc240)];
pa7 = [2.06*std(va725)/length(va725) 2.04*std(va730)/length(va730)...

2.04*std(va735)/length(va735) 2.06*std(va740)/length(va740)];
pb1 = [2.093*std(vb125)/length(vb125) 2.11*std(vb130)/length(vb130)...
2.093*std(vb135)/length(vb135) 2.093*std(vb140)/length(vb140)];
pb2 = [2.201*std(vb225)/length(vb225) 2.201*std(vb230)/length(vb230)...
2.262*std(vb235)/length(vb235) 2.228*std(vb240)/length(vb240)];
pc1 = [2.262*std(vc125)/length(vc125) 2.365*std(vc130)/length(vc130)...
2.306*std(vc135)/length(vc135) 2.306*std(vc140)/length(vc140)];
pc2 = [2.571*std(vc225)/length(vc225) 2.228*std(vc230)/length(vc230)...
2.447*std(vc235)/length(vc235) 2.306*std(vc240)/length(vc240)];
b = [0.506143459 0.670229957 0.559484344 ...
70
0.71395341 0.591797935 0.603702947 0.440279194;
0.726983035 0.807992835 0.756995585 0.772059922...
0.593313756 0.58239626 0.615818856;
0.760424639 0.84383483 1.197936383 0.964080439...
0.635963204 0.657790376 0.529531155;
0.688421238 0.854167381 1.097486477 0.99269112...
0.703865448 0.616505281 0.559002256];
B = b’;
U(1,:) = [sqrt(B(1,1)^2 + pa7(1)^2) sqrt(B(1,2)^2 + pa7(2)^2)...

sqrt(B(1,3)^2 + pa7(3)^2) sqrt(B(1,4)^2 + pa7(4)^2)];
sqrt(B(2,3)^2 + pa1(3)^2) sqrt(B(2,4)^2 + pa1(4)^2)];
sqrt(B(3,3)^2 + pa2(3)^2) sqrt(B(3,4)^2 + pa2(4)^2)];
U(4,:) = [sqrt(B(4,1)^2 + pb1(1)^2) sqrt(B(4,2)^2 + pb1(2)^2)...
sqrt(B(4,3)^2 + pb1(3)^2) sqrt(B(4,4)^2 + pb1(4)^2)];
U(5,:) = [sqrt(B(5,1)^2 + pb2(1)^2) sqrt(B(5,2)^2 + pb2(2)^2)...
sqrt(B(5,3)^2 + pb2(3)^2) sqrt(B(5,4)^2 + pb2(4)^2)];
U(6,:) = [sqrt(B(6,1)^2 + pc1(1)^2) sqrt(B(6,2)^2 + pc1(2)^2)...
sqrt(B(6,3)^2 + pc1(3)^2) sqrt(B(6,4)^2 + pc1(4)^2)];
U(7,:) = [sqrt(B(7,1)^2 + pc2(1)^2) sqrt(B(7,2)^2 + pc2(2)^2)...
sqrt(B(7,3)^2 + pc2(3)^2) sqrt(B(7,4)^2 + pc2(4)^2)];
%=================================================================
figure(1)
hold on
errorbar(P,M(1,:),U(1,:),’-.^k’,’MarkerFaceColor’,’k’,’MarkerSize’,8)
errorbar(P,M(2,:),U(2,:),’-^k’,’MarkerFaceColor’,’c’,...
errorbar(P,M(3,:),U(3,:),’--^r’,’MarkerFaceColor’,’r’,...
71
errorbar(P,M(4,:),U(4,:),’-sk’,’MarkerFaceColor’,’c’,...
errorbar(P,M(5,:),U(5,:),’--sr’,’MarkerFaceColor’,’r’,...
errorbar(P,M(6,:),U(6,:),’-ok’,’MarkerFaceColor’,’c’,...
errorbar(P,M(7,:),U(7,:),’--or’,’MarkerFaceColor’,’r’,...
hold off
ylabel(’Pulse Velocity [m/s]’)
legend(’0.46m-0.7mm’,’0.46m-1mm’,’0.46m-2mm’,’0.61m-1mm’,...
’0.61m-2mm’,’0.91m-1mm’,’0.91m-2mm’,’Location’,’Best’)
% print(’-depsc’,’-tiff’,’-r1200’,’Velocity_ALL’)
A.2.6 Duration Measurements
This code plots the duration of the cavitation events.
clc
clear all
% Time chart for thesis
P = [17.2,20.7,24.1,27.6];
% row is nozzle, column is pressure, z is pipe
T(:,:,1) = [0.2197,0.2015,0.1930,0.1793;
0.1521,0.1415,0.1330,0.1246;
0.0471,0.0446,0.0435,0.0389];
T(:,:,2) = [0,0,0,0;
0.1895,0.1627,0.1640,0.1557;
0.0574,0.0536,0.0506,0.0490];
72
T(:,:,3) = [0,0,0,0;
0.2586,0.2438,0.2334,0.2240;
0.0824,0.0752,0.0734,0.0688];
figure(1)
hold on
%plot(P,T(1,:,1),’^k’,’MarkerFaceColor’,’k’,’MarkerSize’,8)
plot(P,T(2,:,1),’-^k’,’MarkerFaceColor’,’c’,’MarkerSize’,8,...
plot(P,T(3,:,1),’--^r’,’MarkerFaceColor’,’r’,’MarkerSize’,8,...
plot(P,T(2,:,2),’-sk’,’MarkerFaceColor’,’c’,’MarkerSize’,8,...
plot(P,T(3,:,2),’--sr’,’MarkerFaceColor’,’r’,’MarkerSize’,8,...
plot(P,T(2,:,3),’-ok’,’MarkerFaceColor’,’c’,’MarkerSize’,8,...
plot(P,T(3,:,3),’--or’,’MarkerFaceColor’,’r’,’MarkerSize’,8,...
hold off
axis([16 28 0 0.28])
ylabel(’Time of Jet [s]’)
legend(’0.46m-1mm’,’0.46m-2mm’,’0.61m-1mm’,’0.61m-2mm’,...
’0.91m-1mm’,’0.91m-2mm’,’Location’,’EastOutside’)
%print(’-depsc’,’-tiff’,’-r1200’,’Time’)
73

Cavitation of A Water Jet in Water

Uploaded by

Copyright:

Available Formats

Cavitation of A Water Jet in Water

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

Cavitation of A Water Jet in Water

Uploaded by

Copyright:

Available Formats

Brigham Young University

Theses and Dissertations

Cavitation of a Water Jet in Water

Follow this and additional works at: https://scholarsarchive.byu.edu/etd

Part of the Mechanical Engineering Commons

BYU ScholarsArchive Citation

A thesis submitted to the faculty of

Tadd T. Truscott, Chair

Department of Mechanical Engineering

Copyright © 2012 Michael M. Wright

Cavitation of a Water Jet in Water

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii

Chapter 2 Experimental Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Chapter 3 Experimental Results and Discussion . . . . . . . . . . . . . . . . . . . . . 22

2.1 Experimental Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.1 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3.1 Image sequence: 0.45m, 1.0 mm nozzle, 20.7 MPa . . . . . . . . . . . . . . . . . 23

1.2.2 Mechanics of Cavitation

Cavitation is a phenomenon that occurs in liquids, and depends on the non-dimensional

1.2.3 Cavitation Erosion

1.3 Theoretical Approach

2.1 Experimental Setup

2.1.1 Experimental Configuration

(b) Gas exiting nozzle after jet discharge

2.1.2 Water Jet Design

L* 1.5* 2.5 2.4 7.5 8.0

2.1.3 Experimental Variables

Table 2.1: Experimental Parameters

Pipe Length Nozzle Diameter Reservoir Pressure

Table 2.2: Reynolds Numbers of Experimental Parameters

6. Close the release valve as soon as the jet fully discharges

2.3 Instrumentation and Data Acquisition

2.4 High-speed Photography and Image Processing

Configuration Conversion (pixels/cm)

3.1 Experimental Results

0.027 Pulse Begins

0.0155 0.0475 0.1175

0.016 0.048 0.118

0.0165 0.0485 0.1185

0.017 0.049 0.119

[s] [s] [s]

0.0175 0.0495 0.1195

0.018 0.05 0.12

0.0185 0.0505 0.1205

0.019 0.051 0.121

[MPa] [MPa] [MPa]

0.018 0.058 0.098

0.0185 0.0585 0.0985

0.019 0.059 0.099

[s] [s] [s]

0.0195 0.0595 0.0995

0.02 0.06 0.1

0.0205 0.0605 0.1005

0.021 0.061 0.101

0.0216 s 0.0616 s 0.1016 s

0.016 0.066 0.126

0.0165 0.0665 0.1265

0.017 0.067 0.127

pa7 = [2.06std(va725)/length(va725) 2.04std(va730)/length(va730)...