DEVELOPED CAVITATION- CAVITY DYNAMICS

David R. Stinebring
Michael L. Billet
Jules W. Lindau
Robert F. Kunz
Applied Research Laboratory
P.O. Box 30
State College, PA 16804
USA

σ > σi
INTRODUCTION
V & P1 Noncavitating Flow DECREASING
PRESSURE
When we discuss high-speed supercavitating BUBBLES

bodies, it primarily involves the fully developed σ ≤ σi

supercavitating phase. For many cases the body V & P2 Cavitation Inception or
Limited Cavitation
must transition from fully wetted, through the CAVITY

partially cavitating regime before reaching σ << σi

supercavitation. Therefore, an understanding of V & P3
Developed Cavitation
all phases of cavitation is necessary. CAVITY

σ <<< σi
FLOW REGIMES AND BASIC
V & P4
DEFINITIONS
Supercavitation

Cavitating flows are commonly described by the Figure 1: Schematic of cavitation flow regimes
cavitation number, s , expressed as
Limited cavitation occurs at an intermediate value
P− P of the cavitation number where amount of a given
σ ≡ v (1)
1 2 type of cavitation is minimized. Limited
ρV
2 cavitation can be vaporous or gaseous. The
former type of cavitation is caused by the
where, P and V are the reference pressure and explosive growth of bubbles due to the rapid
velocity, respectively, Pv is the vapor pressure at conversion of liquid to vapor at the bubble wall
the bulk temperature of the liquid, and ? is the whereas the latter is due to a slower mode of
mass density of the liquid. If a cavitation bubble growth caused by the transport of
experiment is conducted by holding the velocity noncondensible gas into the bubble. The
constant and varying the reference pressure, condition at which cavitation initially appears is
various amounts of cavitation can be observed as called cavitation inception.
shown in Figure (1). Noncavitating flows occur at
sufficiently high pressures where there is no Of particular importance in the study of cavitation
evidence of bubbles. Supercavitation occurs at is the minimum pressure coefficient, C p min given
very low pressures where a very long vapor cavity
by
exists and in many cases the cavity wall appears
glassy and stable except near the end of the cavity. P −P
Between these flow regimes is limited cavitation C = min
p 1
and developed cavitation.
min ρV 2 (2)
2

where, Pmin is the minimum pressure in the liquid.

The "classical theory for scaling vaporous limited
cavitation" states that, s = −C p min and C p min =
− Cp min Cpmin

constant. This implies that when scaling from one

1
flow state to another, the characteristics of the Also, it is very important to determine the
flow field and its boundaries remain geometrically definition of limited cavitation when comparing
and kinematically similar. However, real flows experimental results. In general, these studies into
often do not obey the classical theory because of scale effects can be divided into two general types
so-called "scale effects" which arise from changes as follows:
in velocity, size, fluid properties and microbubble
distribution. Experimental results clearly show Viscous Effects: Scale effects that act on the flow
that in many cases, the limited cavitation number outside the cavitation bubble which influence
can be greater or less than the minimum pressure the local pressure in the liquid flow.
coefficient and in some cases the minimum
pressure coefficient is not constant, such as in 1. Flow field changes due to variations in
vortical flows. One example of scale effects can Reynolds number, Froude number, and
be noted from the "standard" cavitation tests Mach number including steady and
conducted at many facilities for the ITTC. Figure turbulent pressure fluctuations.
(2) shows some of the cavitation data and the
varied appearance of the cavitation data can be 2. Departures from exact geometric
noted in Figure (3). similarity such as those due to roughness
and manufacturing.

Bubble Dynamic Effects: Scale effects that act on

the bubble growth process which cause the
liquid pressure at the cavitation bubbles to
depart from the equilibrium vapor pressure
corresponding to the bulk temperature of the
liquid.

1. Time effects.
2. Heat transfer effects.
3. Surface tension effects
4. Transport of noncondensible gas
5. Liquid tension, i.e., microbubbles.
Figure 2: ITTC cavitation data
The various factors which causes scale effects and
the influence on the cavitation number can be
ascertained by employing the Rayleigh-Plesset
equation to describe the growth of a "typical"
cavitation bubble. This equation can be written as
 3  2 S 4 µR& 
− C (t ) + σ + C (t )  ρ V 2
1
ρ  RR&& + R& 2  = P ( t ) − −
 2  G R R  p T  2 (3)

where,

P( t ) − P
C (t ) =
p 1
ρV 2
2

and,

Figure 3: Photographs of cavitation appearance for Pv (T ) − Pv (Tr ( t ))

ITTC data C (t ) = .
T 1 2
ρV
Extensive research has been conducted into scale 2
effects, such as in References (1), (2), and (3).

2
 Thus Cp(t) is the time varying pressure coefficient and in Figure (5) for a 1/8-caliber ogive. Figure
which describes the variation of the liquid (6) shows the effect of boundary layer transition
pressure outside of the bubble. The on limited cavitation.
thermodynamic coefficient (CT) describes the
effect of heat transfer on the vapor pressure in the
bubble. The term PG is the partial pressure of gas
2S
inside the bubble and the term is the “liquid”
R
tension of the bubble. Multiplying Equation (3)
by a time interval (dt), integrating over a time
interval (Tc) which is typical of a cavitation
process, and solving for s yields

2S
PG R
σ = −Cp + − − φ − CT (4)
1 1
ρ V∞2
ρV∞
2
2 2
where, φ is a bubble dynamic parameter. The
φ

bars denote averages over the interval TC. In the

absence of significant dynamic effects, Equation Figure 4: Comparison of CIT and ARL Penn State
(4) reduces to an equilibrium equation inception data for 50.8-mm (2.0-inch) Schiebe
noses
2S
P −
G R −C .
σ = −C + (5)
p 1 T
ρV 2
2

Referring to Equation (5), it is noted that the terms

2S
PG , , and CT all cause bubble dynamic scale
R
effects and the liquid tension effect and
thermodynamic coefficient will reduce the
cavitation number at inception and the gas
(3) pressure in the bubble will increase the cavitation
number at inception. Viscous scale effects are
contained with the average pressure
coefficient Cp . This can be written as Figure 5: Desinent cavitation number versus
Reynolds number for 1/8-caliber ogives
∆P T ∆P R ∆P U
−Cp = −Cp + + +
1 1 1
s
ρ V∞2 ρ V∞2 ρ V∞2
2 2 2

where CPR is the average local pressure in the

( )
absence of surface roughness ∆P R , turbulence,

(∆P ) and flow unsteadiness (∆P ) .

T (∆PU )

Scale effects are very important in understanding

the physics of cavitation. They are important as
can be noted in Figure (4) for a Schiebe headform

3
 noted by the effect of liquid temperature variation
on the cavitation number. It is the same as for
limited cavitation and it will become more
significant as the temperature is increased,
resulting in the cavity temperature being lower
than the bulk temperature of the fluid. One of the
first theoretical and experimental investigations
into the variation of cavitation number with
temperature has been conducted by Stahl and
Stepanoff(4). A detailed investigation into this
effect was conducted by Holl, Billet, and others at
Penn State that developed the entrainment
theory(5),(6). This investigation provides insite into
the characteristics of developed cavitation.
Figure 6: Effect of boundary layer transition on
limited cavitation Most of this thermodynamic scale effect
investigation used a zero-caliber-ogive body
Developed Cavitation shown in Figure (7). At small cavity lengths (L/D
< 2.0), the cavity is very cyclic and the re-entrant
The subject of a fully developed cavity and the jet formed at the cavity closure strikes the leading
various similarity laws which predict the cavity edge of the cavity. The photograph of the cavity
parameters have been studied by many researchers on the zero-caliber ogive was taken with
and some of these results are discussed in other continuous lighting so the cavity appears like seen
lectures. However, this research has largely been to the naked eye. Figures (8 to 11) show cavity
divided into two types of cavity flows: (1) behavior seen with high-speed movies and very
ventilated cavities, and (2) vaporous cavities. short-duration strobe flash lighting. The unsteady
These cavities have been shown to have one nature of the cavity is apparent. Observing the
similar characteristic in that the cavity shape is a cavitation with continuous lighting provides a
single-valued function of the cavitation number time average view of the cavity, which also can
based on cavity pressure. Thus, the cavitation lead to problems in investigating developed
number for developed cavitation is based on cavitation. The cavity length is a measure of the
cavity pressure (Pc) and can be expressed as cavitation number. However, determination of the
cavity length visually is subjective, since an
P− P observer is seeing a time average of the cavity,
σ= c
1 2 (6) and pressure measurements are affected by the
ρV
2 ∞ transient nature of the cavity.

where Pc ≈ PV + PG − ∆PV for vaporous cavity

and Pc would be PVE for the ventilated cavity. The
rear of these cavities is a complex and often
highly turbulent region of the flow where the
contents of the cavity are carried or entrained
away. However, in the case of vaporous
cavitation where the cavity contains mostly vapor,
the loss of vapor is immediately made up by the
evaporation from the walls of the cavity. On the
other hand, a continual supply of gas is needed to
form a ventilated cavity.

Fluid properties of the liquid inherently affect the

vaporization process along the cavity wall, which Figure 7: Photograph of cavity on a zero-caliber
supplies vapor to the cavity in a similar manner as ogive illuminated with continuous lighting
is the case of limited cavitation. This scale effect
is given the term "thermodynamic effect," and is

4
 Figure 10: Photograph of cavity on zero-caliber
ogive illuminated with a 3-µ sec flash

Figure 8: Sequence from a high-speed movie

showing a complete cavity cycle- 5000 frames per
second- 6.35-mm (0.25-inch) diameter zero-
caliber ogive, Velocity = 25 m/sec

Figure 11: Sequence from high-speed movie

showing the reentrant jet moving through the
cavity and striking the cavity leading edge
(between frames 11 and 12)- 5000 frames per
second- 6.35-mm (0.25-inch) diameter zero-
caliber ogive, Velocity = 15 m/sec

The high-speed movie sequences show that the

unsteady nature of the cavities is due to the
Figure 9: Detail of cavity after reentrant jet strikes
reentrant jet moving through the cavity and
leading edge, first image in sequence is below
striking the cavity leading edge. This causes the
cavity to detach from the cavitator leading edge
and break into helical vortices connected by
vortex filaments, Figures 9 and 10.

5
The cavity on the zero-caliber ogive is similar to model diameter as shown in Figure (15). This is
the bubble shape of the wake associated with the in agreement with several other investigators such
separated flow from the model nose. At the end as Wade(9).
of this bubble there is a re-entrant jet and the
velocity of this jet of liquid can be expressed as
1
Vj = V (1 + σ ) 2
(7)

which can be derived from Bernoulli's equation

which assumes that the frictionless fluid. This
condition is similar to the vortical flow near the
end of the wake bubble. At appropriate values of
Reynolds numbers a regular vortex sheet is
observed in the wake region aft of bubble closure. Figure 12: Maximum temperature depression
At higher Reynolds numbers a turbulent wake versus cavity tap position for Freon 113;
with irregular mixing results. For the case of the Temperature= 175° F, L= 0.50, 0.875, and 1.25-
cavity closure on the body, the re-entrant travels inch, V=64 ft/sec
along the body surface to the cavity separation
location

As the re-entrant jet affects the cavity dynamics,

the body experiences a time dependent drag.
Gilbarg and Serrin(7) show that this change of drag
can be related to the momentum of the re-entrant
jet and hence a characteristic dimension of this jet.

Measurements made inside of a developed cavity

also indicate the physics of the vaporization
process. As shown in Figure (12), the maximum
temperature depression from inside the cavity
indicating that most of the vaporization process is
occurring near the beginning of the cavity where Figure 13: Cavity temperature depression versus
the curvature is greatest. This maximum temperature for Freon 113, V=19.5 m/sec, L=3.18
temperature depression is affected by the air cm (1.25-inch), two different air
content of the liquid as shown in Figure (13). contents
Consequent measurement of the cavity pressure
indicates that this is due to the presence of
noncondensible gas (partial pressure above vapor
pressure). The magnitude of the pressure
unsteadiness that occurs in the cavity as shown in
Figure (14). This is due to the re-entrant jet.
The geometric characteristics of developed
cavities were determined from a series of
experiments conducted by Billet and Weir(5),(6).
From measurements of the cavity profiles, the
maximum cavity diameter (Dm), cavity half-length
(A), and cavity surface area were determined for Figure 14: Cavity tap pressures versus time for
both natural and ventilated cavities. Data shows water; temperature= 66° C, L= 2.2 cm (0.875-
that for a given model geometry and cavitation inch), V=19.5 m/sec
number, ventilated and natural cavities have the
same profile, within experimental error. This was
found to be independent of reference velocity, and
cavity dimension are directly proportional to the

6
 coefficient regions which are determined by the
cavity closure conditions.

An extensive analysis of the various factors which

influence air entrainment for ventilated cavities in
the twin-vortex regime is presented by Cox and
Clayden(11). They show theoretically that high
rates of air entrainment are associated with the
formation of the twin-vortex system. They also
present experimental data for ventilated cavities
on 2.54-, 1.91-, and 1.27-cm-diameter disks which
show that for a given rate of air entrainment, the
cavity length to disk diameter ratio increases as
Figure 15: Ratio of maximum cavity to model the Froude number based on disk diameter
diameter versus corrected cavitation number increases.

Waid(12) also measured flow coefficients for

For ventilated cavities, the flow coefficient is ventilated cavities on a series of three-dimensional
defined as hydrofoils. The results obtained for both the re-
(8)
Q entrant jet and two-vortex regimes show the same
C ≡
Q
VD2 trends as the data obtained by Swanson and
O'Neill(10).
where Q is the ventilation rate and D is the
reference dimension. This is a dimensional Gas diffusion was found to have a significance on
representation of the volume flow rate of gas the cavity pressure (above vapor pressure) and on
needed to sustain a cavity. This volume flow rate the measured thermodynamic effect. Thus, the
is the same regardless if it is a ventilated or vapor entrainment rate needed to sustain a natural cavity
cavity. This assumption is supported by has contributions not only from the vapor but also
Reichardt, who demonstrates that the drag and from gas diffusion. Gas diffusion was also found
geometry for similar vaporous and ventilated to have a significant effect on the gas needed to
cavities are the same. sustain a ventilated cavity. Billet and Weir(13)
conducted an experimental investigation into this
One of the first experimental investigations of air effect. An analysis of this effect was conducted
entrainment was carried out at the California based on a mathematical model of diffusion
Institute of Technology in 1951 by Swanson and developed by Brennen(14).
O'Neill(10). In their experiments an air filled
cavity was maintained behind a 2.54-cm disk over Gas diffusion will occur when there exists a
a cavitation number range of 0.25 to 0.08 and a concentration gradient between the freestream and
corresponding Froude number range of 10 to 15. the cavity. This concentration gradient can be
Their results show that as the supply of air to the calculated from the partial pressures of the gas.
ventilated cavity was decreased from a maximum The maximum partial pressure of the gas in the
flow coefficient of 1.6, the cavity size decreased freestream can be determined from a knowledge
only slightly over a large range of air supply rates of the air content and Henry's Law
until a certain minimum rate of air supply was
reached. Thereafter, a slight decrease in the air
supply led to a rapid reduction in the size of the PFS = α ⋅ β (9)
cavity for small values of flow coefficients.
where, a is the air content in ppm by moles and ß
In the region where the cavity size varies rapidly is Henry's Law constant, which depends on
with the flow coefficient, the cavity has a re- temperature. The mass flow rate of diffused gas
entrant jet closure. A further increase in flow into the cavity is proportional to the concentration
coefficient causes the re-entrant jet to break down gradient and can be expressed in terms of the
and be replaced by the twin-vortex regime. This difference in partial pressures as
demonstrates the existence of two distinct flow

7
 Supercavitation
 P −P 
M& = f (transport )  ρ FS G

DG  L β  (10) When a cavity length is much greater than the
body dimensions, this flow regime is described as
supercavitation. The cavity stays attached to the
where, f(transport) is a dimensional function of body and the cavity closure is far downstream.
the parameters influencing the diffusion The length of the cavity does not vary
mechanism and PG is the partial pressure of significantly even though considerable oscillations
noncondensible gas in the cavity. This transport can occur at its closure. However, the cavity acts
function can be determined from a knowledge of as if it were an extension of the body. In this case,
the free shear layer over the cavity surface. A the same flow field would exist around a solid
schematic of this process is shown in Figure (16) body having a shape comprising of the wetted
from Brennen(14). nose plus the free-cavity profile. Unlike
developed cavities, the gross features of
supercavitation on a headform are not only
functions of the cavitation number but also the
Froude number-gravity becomes important.

Supercavities have many characteristics of

classical free streamline flows. The cavity interior
is essentially at constant pressure, and the cavity
walls are essentially free-stream surfaces of
constant velocity. Cavity pressures approach
vapor pressure. These assumptions are the basis
of many theoretical treatments of cavity geometry
Figure 16: Schematic for an undersaturated cavity and forces. Many of these theories will be
(Brennen [14]) presented during this course; however, only cavity
dynamics will be discussed in this section.
The results of this analysis are shown in Figure
(17). In this figure, the effect on flow coefficient There exists a transition between the re-entrant jet
is given for both an undersaturated and closure condition and the twin-vortex closure
oversaturated condition. These data were then condition. This has been discussed in many
corrected for gas diffusion effects and compare previous investigations. However, a study of
favorably to the data for no gas diffusion. water-entry modeling by Stinebring and Holl(15)
describes this transition of flow regimes in detail
and will be discussed briefly.

It has been observed that for a ventilated cavity

there exists a condition when the cavity length for
a given flow condition suddenly grows to four or
five times its original length with only a slight
increase in the ventilation air flow rate. This
instability in the cavity length may be attributed to
the transition between the re-entrant jet and twin-
vortex flow regimes.

Previous discussions on developed cavitation are

for flows where gravity effects are not important
and the re-entrant jet is the gas entrainment
Figure 17: Flow coefficient data with diffusion for mechanism. However, at low velocities or longer
6.35-mm diameter zero-caliber ogive, V= 18.3- cavity lengths and gravity is normal to the flow,
m/sec the cavity will move upward to the body
centerline as a result of buoyancy. The flow
velocities of the top surface of the cavity will be

8
different than on the lower surface, which will
result in a net circulation. As a result, a new CAVITATOR HYDRODYNAMICS
cavity equilibrium system is established where
entrainment now occurs through the vortex cores For supercavitating flow the cavitator is located at
of the twin-vortex structure at cavity closure. the forward most location on the body, and the
cavity downstream of the cavitator covers the
A photograph of the cavity in the re-entrant jet body. The shape of the cavity is defined by the
regime is shown in Figure (18). The conical nose cavitation number based on cavity pressure. The
is supported by support struts so the aft region the simplest form of cavitator is a disk where the drag
cavity can be investigated with minimum coefficient is defined as,
disturbance. The opaque appearance of the cavity C d = 0.82 (1 + σ ) . (11)
is primarily due to the violent mixing caused by
the re-entrant jet. As transition to the twin-vortex There have been numerous expressions derived
regime takes place, the cavity becomes clear at the for cavity dimensions as a function of cavitation
leading edge but with some mixing due to the re- number. Reichardt (16) showed that the cavity
entrant jet. The photograph in Figure (19) shows length divided by the cavity diameter is
the cavity in the twin vortex regime. This independent of the shape of the cavitator and is
transitional region between flow regimes was also only a function of cavitation number,
observed for models which have afterbodies (sting
mounted) although this effect was not as
L σ + 0.008
pronounced. The full growth of the cavity could = . (12)
not be realized because of interference at the d m σ ( 0.066 + 1.7 σ )
downstream end of the cavity. Reichardt also developed an expression for the
cavity diameter divided by the cavitator diameter,

0.5
d m  Cd 
 ,
=
d  8  (13)
 σ − 0.132 σ 
7

which is only a function of cavitation number and

cavitator drag coefficient, C d .
By substituting the definition of drag coefficient
into equation (13) and rearranging, the drag force
required to create an axisymmetric cavity of a
given diameter is:

Figure 18: Photograph of cavity for a 45° conical π

Dcavity = ρ V∞ d m ( σ − 0.132 σ
2 2 8/ 7
nose illuminated with continuous lighting ) (14)
8

where, ρ is the mass density of the fluid. For a

given velocity and cavity diameter, the drag is
only a function of the cavitation number. It is not
a function of the geometry of the cavitator.

The shape of the cavity is approximately elliptical.

A number of researchers have developed formulas
for the cavity, most notably, Logvinovich (17).
The cavity radius is given as,

2
 R1 2   t  χ
R = Rk 1 − 1 − 2  1 −  (14)
 R 
k   
Figure 19: Photograph of twin vortex cavity  tk
downstream of a strut-mounted 45° cone

9
 specific physics problems, so it is especially
appropriate for graduate student work.
where, Rk , is the maximum cavity radius, t k , is
Three facilities are specially utilized for
the time of formation to the cavity midpoint, χ , is
supercavitation studies and are: (1) 48-inch
a correction factor, χ = 0.85, and R1 and t are the Diameter Water Tunnel, (2) 12-inch Diameter
radius and time of formation at the matching Water Tunnel, and (3) Ultra-High-Speed
station, x1 , where, Cavitation Tunnel. The 12-inch Diameter Water
x − x1 Tunnel was built in 1951 and the Ultra-High-
t= Speed Cavitation tunnel was built in 1962. These
Vk three facilities will be discussed briefly.
and where, x, is the streamwise distance, and Vk ,
is the cavitator velocity. 48-inch (1.22 m) Diameter Water Tunnel
(Garfield Thomas Water Tunnel)
Since there is a gas filled cavity over the This large water tunnel is a variable –speed,
variable-pressure tunnel primarily intended for
body, there is a loss of buoyancy so the body
propulsion studies of body-propulsor systems. It
must be supported by contact with the cavity is a closed circuit, closed-jet water tunnel. A
wall. The cavitator must be at an angle of detailed description of the facility is given in
attack to provide lift to support the forward Figure (20).
section of the body, and the aft section can be
supported by the afterbody planing or control The tunnel has two honeycombs in the 3.66 m
surfaces or a combination of both. The cavity diameter "settling section" to not only straighten
shape will be influenced by foreshortening the flow, but also reduce the test section
due to the cavitator at angle of attack, and turbulence level below 0.1 percent. The
perturbations due to downwash from lift and cylindrical working section is four feet in
bouyancy effects. These effects upon the diameter and 4.27 m long. The water velocity in
the working section is variable up to 24.4 m/sec.
cavity shape are beyond the scope of this
And is controlled by a four-bladed, adjustable
paper and are covered in Logvinovich. pitch impeller which is 2.41 m in diameter and is
driven by a 1491 kW electric motor.
EXPERIMENTAL FACILITIES AND TEST
PROGRAMS The pressure in the working section may be
reduced to 20.7 kPa, which permits a wide range
These next sections will briefly cover some of the of cavitation numbers for testing. The air content
facilities and test programs at ARL Penn State of the water in the tunnel is controlled through the
related to cavity dynamics, cavity ventilation use of a degasser that permits removal of air from
requirements, cavitator testing, and control the water. Thus, the water in the working section
surface testing. They are intended to show the can be less than air-saturated regardless of the
hardware and techniques that have been used to working section.
investigate the physics of developed cavitation.
The degassing system is located in the tunnel by-
Testing Facilities at ARL Penn State pass system, which also has filters and the tunnel
pressure control systems. This by-pass can handle
The Garfield Thomas Water Tunnel of the from 500 gpm to 3000 gpm from the tunnel and
Applied Research Laboratory Penn State is a can be operated during cavitation testing to
complex of hydrodydynamic and hydroacoustic remove air and bubbles. Gas removal from the
test facilities that are registered with the water is accomplished with a Cochrane Cold-
International Towing Tank Conference, an Water Degasifier. This degasser consists of an
organization of member counties that design and upright cylindrical tank where the upper portion is
test ships and other marine structures in tanks and filled with plastic saddles. Water taken from the
tunnels. Since the 48-inch Diameter Water tunnel is sprayed into the top of the tank, the
Tunnel began operation in 1949, many additional interior being maintained at a high vacuum. The
facilities and capabilities were added. Many of water spray falling over the saddles exposes a thin
the smaller water tunnels are utilized to research

10
water film to a vacuum, which permits the rapid Published description: ARL Penn State Report NORD
removal of air. The water collects at the bottom 16597-56, Lehman, 1959.
of the degasser and is returned to the bypass Applied Research Laboratory, Fluid Dynamics
system. Department, The Pennsylvania State University, USA

Figure 20: Description of the ARL Penn State 48-

inch diameter water tunnel

12-inch (0.305 m) Diameter Water Tunnel

The facility is typically used for flow studies that
do not require wake-operating propellers. This is
a closed-circuit tunnel, which has the similar
working section velocity and pressure range as the
Garfield Thomas Water Tunnel. As is the case of
other facilities, the pressure is independent of the
working section velocity and details are given in
Description of Facility: Closed Circuit, Closed Jet
Figure (21),
Type of Drive System:4-Blade Adjustable Pitch
Impeller
One of the unusual features of this facility is the
Total Motor Power:2000 HP Variable Speed
use of two interchangeable working sections for
(1491 kW)
different test purposes. A circular section is 0.305
Working Section Max. Velocity: 18.29 m/s
m in diameter, and a rectangular one is 0.51 m
Max. & Min. Abs. Pressures:413.7 to 20.7 kPa
high and 11.4 cm wide; each is 0.76 m long. The
Cavitation number Range:>0.1 dependent on
circular working section is used to study the
velocity and/or J-range
bodies employed in three-dimensional flow
problems. The rectangular one is to study
Instrumentation: Propeller dynamometers, five-
hydrofoils and slots and other two-dimensional
hole probes, pitot probes, lasers, pressure sensors,
flow problems.
hydroplanes, planar motion mechanism, force
balances.
The tunnel drive consists of a mixed-flow pump
that forms a part of the lower leg and one turn of
Type and Location of Torque & Thrust
the tunnel circuit. The pump is driven by a 150 hp
Dynamometers: Model internally mounted, 150 hp
electric induction motor through a variable-speed
limit (111.85 kW).
fluid coupling and reducing gear. The method of
cooling the tunnel water is by a heat exchanger,
Propeller or Model Size Range: Model size from
which is an integral part of the circuit.
76.2 mm to 635.0 mm id.
The tunnel is also connected to the by-pass circuit
Tests Performed: Forces, flow field, and pressure
of the 48-inch diameter water tunnel so that the air
distributions on bodies of revolution, hydrofoils,
content and free air can be controlled.
propellers, etc. Cavitation performance and noise
measurements of propellers, foils, hydrodynamic
shapes, etc. Steady state and time-dependent
force and torque measurements on powered
models. Flow visualization.

Other Remarks: Tunnel turbulence level is 0.1

percent in test section. Air content can be
controlled as low as 1 ppm per mole.
Measurement can be made of hydrodynamic
functions for stability and control of submerged
vehicles. Directional hydrophone system for
relative acoustic measurements. Description of Facility: Closed Circuit, Closed Jet
Test Sections:(1) Circular: 304.8 mm dia x 762.0
mm long

11
(2) Rectangular: 508.0 mm x 114.3 mm x 762.0 would be encountered in a system employing
mm long hydrocarbon lubricants. As the fluid leaves the
Type of Drive: Mixed Flow Peerless Pump bearing, it enters an accumulator tank and from
Total Motor Power: 150 hp (11.8 kW) there is pumped by means of a triplex pump back
Working Section Max. Velocity:24.38 m/s to the tunnel. Pressure control is obtained by by-
Max. & Min. Abs. Pressures: 413.7 to 20.7 kPa passing a portion of the fluid from the discharge
Cavitation Number Range:>0.1 dependent on of the triplex pump to the accumulator tank and/or
velocity bleeding some fluid from the main tunnel circuit.
Instrumentation:Lasers, pressure sensors, A surge tank is located in the pressure control
hydrophones circuit in order to remove fluctuations caused by
Model Size Range:50.8 mm max. dia. the triplex pump. The liquid in the tunnel by-pass
is filtered down to remove particles above two
Tests Performed: Steady and time-dependent microns.
force and pressure measurements on unpowered
models. Noise measurements on cavitating Also, a 14 kw heater is located in the by-pass
models. Three-dimensional flow problems circuit. It has the capability to heat the liquid to
(circular section). Two-dimensional flow temperatures in excess of 176°C. Water, Freon
problems (rectangular section). Axial-flow pump 113, and alcohol have been used as working fluids
tests. in this facility.

Other Remarks: Independent gas control of air

content. Water filtration with 25-micrometer
filters. Intermittent operation with drag-reducing
additive injection. Partial neutralization of
additive downstream of test section.
Published Description: ARL Penn State Report NORD
16597-56, Lehman, 1959.

Figure 21: Description of the ARL Penn State 12-

inch diameter water tunnel

Ultra-High-Speed Cavitation Tunnel

Much of the basic studies on cavitation have been Description of Facility:Closed Circuit, Closed Jet
conducted in the Ultra-High-Speed Cavitation Type of Drive System:Centrifugal Variable Speed
Tunnel. The characteristics of this tunnel are Drive
given in Figure (22). Total Motor Power: 75 hp (55.9 kW)
Working Section Max. Velocity: 83.8 m/s
This tunnel is a high-speed, high-pressure system Max. & Min. Abs. Pressures:8274.0 to 41.4 kPa
which is designed so that the velocity in the test Cavitation Number Range:>0.01 dependent on
section can be varied from 13.7 to 100 m/sec. The velocity and velocity
tunnel pressure can be varied to a maximum of 8.3 Instrumentation:Pressure and temperature sensors,
MPa. The inside diameter of the test section of lasers
3.81 cm. Temperature Range:16°C to 176°C
Model Size Range:12.7 mm max. dia.
The entire tunnel circuit is made of stainless steel Test Medium:Water, Freon 113, Alcohol
except for the lucite test section windows and the
bronze centrifugal pump. The small amount of Tests Performed: Incipient and desinent
corrosion in the system coupled with a bypass cavitation studies. Development cavittion studies.
filtering circuit allows for a very clean system. Cavitation damage.
The bearing for the centrifugal pump shaft is Other Remarks: Stainless steel tunnel. Bronze
lubricated by drawing water from the tunnel pump. Three filter banks for removal of water,
circuit through the bearing and thus this method of acides, solid particles (10 micrometers) depend on
lubrication avoids possible contamination which fluid media.

12
Published Description: ARL Penn State TM 75-188,
Weir, Billet & Holl, 1975. The cavity cycling frequency Strouhal number
based on body diameter is defined as
Figure 22: Description of the ARL Penn State 1.5-
inch diameter water tunnel fD
S=
Measurements of Cavity Dynamics in the V∞
Reentrant Flow Regime where, f is the cycling frequency, and D is the
body diameter. A plot of cycling frequency
As stated previously, if the product of the Strouhal number as a function of cavitation
cavitation number and Froude number is above a number is presented in Figure (24). The data of
critical value, then the cavity will be in the Knapp (18) for hemispherical nosed models and
reentrant flow regime. The oscillation of the from Stinebring (8) for zero caliber ogives are
cavity in the reentrant regime affects the also shown in the figure. The trend of lower
ventilation gas entrainment and the body cycling frequency for lower cavitation numbers,
dynamics. Therefore, understanding the dynamics i.e., larger cavities, is the same for all models for
of reentrant jet cavities is important. vaporous and ventilated cavities. The Strouhal
number can also be defined such that the
A 45° conical-nosed model having a 2.54-cm characteristic length scale is the ratio of cavity
(1.0- inch) diameter afterbody was instrumented length to model diameter, L/D. A plot of Strouhal
with an internally mounted piezoelectric number as a function of L/D is presented in Figure
transducer, Figure 23. It was expected that the (25). The plot shows that the cavity length is the
reentrant jet striking the cavity would cause a characteristic dimension for cavity cycling and
significant pressure rise that could be measured that the Strouhal number based on cavity length is
with the transducer. Spectral analyses were a constant, approximately, S= 0.16.
performed on the transducer output to determine
the frequency of the cavity cycling. Tests were
conducted with both vaporous and ventilated
cavities over a range of cavitation numbers and
velocities of 9.1, 13.7 and 15.2 m/sec (30, 45, and
50 ft/sec). High-speed movies of the cavitation
were taken and compared to the spectrum
analyzer results. To determine if the boundary
layer thickness affects the cycling frequency, a
number of tests were conducted with distributed
roughness applied to the conical nose. Lastly, the
results were compared with the findings of
previous investigations in water tunnels covering
a wide range of model sizes and geometries.
Figure 24: Cavity cycling Strouhal number, based
on body diameter, as a function of cavitation
number for vaporous and ventilated cavities for a
number of investigations having different test
geometries

Figure 23: Schematic of model for measuring

cavity cycling showing the location of the
pressure transducer (inset)

13
 Figure 26: Photograph of the models used to
examine the effect of afterbody on gas
Figure 25: Cavity cycling Strouhal number, based entrainment
on cavity length divided by body diameter, as a
function of cavity length divided by body The results of the study for a tunnel velocity of
diameter for vaporous and ventilated cavities for a 13.7- m/sec (45-ft/sec) are displayed in Figure
number of investigations having different test (27). The model requiring the lowest flow rate
geometries for a given cavitation number was the one with no
afterbody. The model having the afterbody one-
Effect of Afterbody Shape Upon Gas half of the cone diameter required the largest
Entrainment ventilation flow rate (except at the highest
cavitation numbers), and the model with the
Since most of the gas entrainment takes place in afterbody equal to the cone diameter was between
the aft region of a body, the afterbody shape the two. It would seem that the model with no
should affect the required ventilation rate to afterbody should require the largest flow rate,
generate a cavity. Tests were performed in the because of the greater volume of gas inside the
ARL Penn State 0.305-m (12-inch) diameter cavity, but this was not the case. The reentrant jet
water tunnel to examine afterbody effects upon behavior could possibly account for this effect.
gas entrainment. The models used for this study Observations show that gravity causes the
are shown in Figure (26). Each model had a reentrant jet to move along the bottom of the
25.4-mm (1.0-inch) diameter 45° apex angle cavity for the model with no afterbody. An
conical nose. One model had a 25.4-mm (1.0- afterbody causes a “guiding effect” upon the
inch) afterbody, the second had a 12.7-mm (0.5- reentrant jet that results in more mixing at the top
inch) afterbody, and the third model was of the cavity. It is not known why the smaller
supported by three struts and had no afterbody. afterbody required the largest ventilation. The
All models had six holes for introduction of data does show that afterbody design is important
ventilation gas and a pressure tap for measuring to minimize ventilation requirements.
cavity pressure. Tests were conducted at
velocities of 9.1, 13.7, and 15.2-m/sec (30, 45,
and 50 ft/sec). To minimize gaseous diffusion
across the cavity wall, the test conditions were
selected so the cavity pressure was close to the
partial pressure of gas in the surrounding water.

14
 Figure 29: Photograph of cavitator mounted on
force cell balance
Figure 27: The effects of afterbody arrangement
on ventilation requirements Photographs of cavities produced by conical and
disk cavitators are shown in Figures (30 to 32).
Cavitator Testing The differences in cavity appearance, by using
continuous and stroboscopic lighting, are shown
A photograph of a model for cavitator testing in in Figures (30 and 31). Short exposure times,
the ARL Penn State 12-inch diameter water tunnel such as with stroboscopic lighting, should be used
is shown in Figure (28). The models can when investigating unsteady cavity dynamics.
incorporate systems for introducing ventilation Strobe lights can be syncronized with the
gas, cavitator force balances, pressure scanning rate of video systems to provide
measurement systems, and other instrumentation. exposure times of a few microseconds for each
Just downstream of the cavitator in the figure is a video frame. The video record of the cavity can
gas deflector that redirects the gas flow to then be compared with transient measurements,
minimize disturbances to the cavity. The strut- such as forces, by use of a master clock. The
mounted model is attached to the large dark plate, photograph in Figure (32) shows a stable cavity
which is the section that mounts to the water and the effect of the strut mounting on the cavity.
tunnel test section. A typical mounting, for a When investigating the downstream end of a
cavitator at angle of attack, to a force balance, is cavity, a sting mount is used, or as shown in
shown in Figure (29). Care must be taken to Figure (19) the cavitator can be supported with
ensure that the design for the introduction of careful design of the struts.
ventilation gas does not have a significant effect
upon the force measurements.

Figure 30: Photograph of cavity formed by a

conical nosed cavitator- continuous lighting

Figure 28: Photograph of model for testing

cavitators

15
 (35). The wall cavity is approximately 25.4-mm
(1.0-inch) thick at the leading edge of the
hydrofoil. Reference lines are drawn along the
hydrofoil, 25.4-and 38.1-mm (1.0 and 1.5-inches)
from the base. The photograph in Figure (35)
shows the pressure side view with the hydrofoil at
high angle of attack. As shown in the figure, the
tunnel wall cavity strikes the hydrofoil leading
edge at the 25.4-mm (1.0-inch) reference line.
The cavity along the hydrofoil surface is deflected
down due to the constant pressure cavity and the
pressure gradient along the surface. This
deflection of the cavity changes the wetted area
Figure 31: Photograph of cavity formed by a and affects the lift and drag forces. The hydrofoil
conical nosed cavitator- stroboscopic lighting cavity can be seen attached to the trailing edge in
the figure.

Figure 32: Photograph of cavity formed by a disk

at angle of attack

Cavity Piercing Hydrofoil/Control Surface

Testing

Cavity piercing supercavitating hydrofoil/control

surfaces have been tested in the ARL Penn State
12-inch diameter water tunnel. Some hydrofoils
having wedge shaped cross sections are shown in
Figure (33). The hydrofoil is mounted to a force
balance and test fixture, as shown in Figure (34).
The test fixture mounts through the water tunnel
test section wall, and is designed so that there is a
small gap between the base of the hydrofoil and Figure 33: Cavity piercing hydrofoils having
the test section wall. The hydrofoil angle of wedge shaped cross sections
attack can be adjusted by rotating the test fixture.
A two-dimensional cavity is created along the test
section wall upstream of the hydrofoil. This is
done by ventilating downstream of a wedge
mounted to the tunnel wall. The hydrofoil then
pierces the tunnel wall cavity as shown in Figure

16
 Figure 34: Cavity piercing hydrofoil showing Figure 36: Photograph of suction surface of cavity
force balance and tunnel mounting fixture piercing hydrofoil that shows cavity disturbances
due to imperfections along leading edge.

Figure 35: Photograph of pressure side of cavity

piercing hydrofoil

The suction side view of a hydrofoil at high angle

of attack is shown in Figure (36). This view Figure 37: Photograph of cavity piercing hydrofoil
shows disturbances in the cavity wall due to at low angle of attack, with base cavitation
imperfections along the hydrofoil leading edge.
For a hydrofoil with a very sharp leading edge and
at large angles of attack, the cavity is transparent
and stable near the hydrofoil. At low angles of
attack the cavity is attached at the base as shown
in Figure (37), and is termed base cavitation. At
intermediate angles of attack between base
cavitation and supercavitation there can be what is
termed partial cavitation. With partial cavitation
there may be intermittent cavitation along the
suction face, a small, attached cavity or a vortex
cavity. Figure (38) shows the stable sheet cavity
and supercavitating tip vortex formed by a higher
aspect ratio hydrofoil. The forces associated with
the different forms of hydrofoil cavitation will be Figure 38: Photograph of cavity from higher
discussed next. aspect ratio cavity piercing hydrofoil

17
Typical lift and drag coefficient curves as a
function of angle of attack for a cavity piercing
hydrofoil, are shown in Figure (39). The
reference area for reducing forces to coefficient
form was the wetted area if the tunnel wall cavity
was not affected by the hydrofoil surface pressure
gradient. The force coefficients have also been
calculated based upon the actual wetted area, as
determined from videotapes of the tests, see
Figure (35). The complex shape of the force
coefficient curves is due in part to the transitions
between base, partial, and supercavitating flow
regimes. Figure 39: Lift and drag coefficients for a cavity-
piercing hydrofoil as a function of angle of attack
While the figure shows the averaged force
coefficients, the time dependent forces due to
cavity fluctuations can also be very important.
Time dependent forces for base cavitation are
shown in Figures (40 and 41), partial cavitation in
Figures (42 and 43), and supercavitating
conditions in Figure (44 and 45). For all figures,
at time zero the hydrofoil is fully wetted. At
approximately 1.25 seconds the tunnel wall cavity
is established and the hydrofoil pierces the cavity.
After establishment of the tunnel wall cavity, both
lift and drag decrease for all conditions, as
expected. The primary difference due to the
changes in forms of cavitation can be seen in the
lift plots. The lift for base cavitation and
supercavitation show relatively little fluctuation.
However, with partial cavitation there are large Figure 40: Drag force as a function of time, base
changes in lift. All tests in the tunnel were cavitation
documented with video cameras using
stroboscopic lighting synchronized with the video
scanning. These changes in lift were associated
with intermittent bursts of cavitation on the
hydrofoil suction surface. In some cases the lift
fluctuations were up to 50%.

For these fins with wedge shaped cross sections,

the intermittent cavitation occurred when the
relative flow direction near the leading edge was
close to the wedge half-angle in the direction of
flow. This relative flow direction is affected by
the perturbation in the tunnel wall cavity due to
the pressure side pressure gradient. The “effective
half-angle” is then also affected by the cavity Figure 41: Lift force as a function of time, base
perturbation. Practical designs must take into cavitation
consideration these factors that affect the unsteady
cavity behavior.

18
Figure 42: Drag force as a function of time, partial Figure 45: Lift force as a function of time,
cavitation supercavitation

SUMMARY
This report presents a summary of the cavitation
flow regimes and experimental techniques used at
ARL Penn State for investigating developed
cavitation physics.

References:

1. Arakeri, V. H. and A. J. Acosta, "Viscous

Effects in Inception of Cavitation on
Axisymmetric Bodies," J. of Fluids Engr.,
ASME, December 1973, pp. 519-526.

2. Gates, E. M. and M. L. Billet, "Cavitation

Figure 43: Lift force as a function of time, partial Nuclei and Inception," IAHR Symposium,
cavitation Tikyo, Japan, 1980, pp. 3-25.

3. Holl, J. W., R. E. A. Arndt, and M. L. Billet,

"Limited Cavitation and the Related Scale
Effects Problem, The Second international
JSME Symposium on Fluid Machinery and
Fluidics, Tokyo, Japan, 1972, pp. 303-314.

4. Stahl, H. A. and A. J. Stephanoff,

"Thermodynamic Aspects of Cavitation in
Centrifugal Pumps," J. of Basic Engr.,
ASME, Vol 78, 1956, pp. 1691-1693.

5. Holl, J. W., M. L. Billet, and D. S. Weir,

"Thermodynamic Effects on Developed
Cavitation," J. of Fluids Engr., Vol. 97, No.
Figure 44: Drag force as a function of time, 4, Dec. 1975, pp. 507-514.
supercavitation
6. Billet, M. L., J. W. Holl, and D. S. Weir,
"Correlations of Thermodynamic Effects for
Developed Cavitation," J. of Fluids Engr.,
Vol. 103, Dec. 1981, pp. 534-542.

19
7. Gilbarag, D., and J. Serrin, "Free Boundaries 16. May, A., “Water Entry and the Cavity-
and Jets in the Theory of Cavitation," J. of Running Behavior of Missiles,” SEAHAC
Math and Physics, Vol 29, April 1950, pp. 1- Technical
12. Report 75-2, Naval Surface Weapons
Center, White Oak Laboratory, Silver
8. Stinebring, D. R., M. L. Billet, and J. W. Holl, Spring, MD, 1975.
"An Experimental Study of Cavity Cycling
for Ventilated and Vaporous Cavities," 17. Logvinovich, G., “Hydrodynamics of Free-
International Symposium on Jets and Boundary Flows,” Trudy TsAGI, 1980.
Cavities, ASME, Miami Beach, FL, Nov.
1985, pp. 1-4. 18. Knapp, R. T., J. W. Daily, and F. G.
Hammitt, “Cavitation,” McGraw-Hill, 1970.
9. Ward, R. L., "Cavity Shapes for Circular
Disks at Angles of Attack," CA Inst. of
Tech. Hydrodynamic Lab Report, E-73.4,
1957.

10. Swanson, W. M., and J. P. O'Neill, "The

Stability of an Air-Maintained Cavity
Behind a Stationary Object in Flowing
Water," Hydrodynamic laboratory, CA Inst.
of tech, M24.3, Sept 1951.

11. Cox, R. N. and W. A. Clayden, "Air

Entrainment of the Rear of a Steady
Cavity," Symposium on Cavitation in
Hydrodynamics, National Physical
Laboratory, Sept. 1955.

12. Ward, R. L., "Experimental Investigation of

Forced Ventilated Cavities on Hydrofoils of
Various Chord Lenths," Report No.
LMSC/D243664, Lockheed Missiles and
Space Corp, Sunnyvale, CA, Dec. 1968.

13. Billet, M. L. and D. S. Weir, "The Effect of

Gas Diffusion on the Flow Coefficient for a
Ventilated Cavity," J. of Fluids Engr.,
ASME, Vol. 97, Dec. 1975, pp. 501-506.

14. Brennen, C., "The Dynamic Balances of

Dissolved Air and Heat in Natural Cavity
Flows," J. of Fluid Mechanics, Vol. 37, Part
1, 1969, pp. 115-127.

15. Stinebring, D. R., and J. W. Holl, "Water

tunnel Simulation Study of the Later Stages
of Water Entry of Conical Head Bodies:
Phase II – Effect of the Afterbody on Steady
State Ventilated Cavities," ARL Penn State
Report No. TM-79-206, Dec. 1979.

20